EVALUATION  OF FUGITIVE
   DUST EMISSION  DATA
E. H. PECHAN & ASSOCIATES, INC.
      3514 University Drive
      Durham, NC 27707
       (919) 493-3144

-------
                                     October  11,  1991


            DRAFT  REPORT
         EVALUATION OF FUGITIVE
           DUST EMISSION  DATA
          Contract  No.  68-D9-0168
          Work Assignment No.  25
               Prepared by:

            William R. Barnard
               T.  Allan  Dean
            Patricia M.  Carlson
       E.H.  Pechan & Associates,  Inc.
           3514 University Drive
             Durham, NC  27707
               Prepared  for:

             Charles C.  Masser
Air & Energy  Engineering Research Laboratory
    U.S.  Environmental Protection Agency
     Research Triangle Park, NC   2771 1

-------
                                    NOTICE
This document is a preliminary draft.  It has not been formally released by
the, U.S. Environmental Protection Agency and should not at this stage be
construed to represent Agency policy.  It is being circulated for comments on
its technical merit and policy implications.

-------
                        TABLE OF CONTENTS
INTRODUCTION
EMISSIONS DATA ASSESSMENT .................   2-1
     VEHICLE TRAVEL ON UNPAVED ROADS  ...........   2-1
          Introduction  ..................   2-1
          Historical Development  .............   2-1
          Organization  ..................   2-3
          Studies of Primary Importance ..........   2-4
               Study 1   ..................   2-4
               Study 2  ..................   2-6
               Study 3  .................    2-11
               Study 4  .................    2-12
               Study 5  .................    2-19
               Study 6  .................    2-22
               Study 7  .................    2-26
               Study 8  .................    2-30
               Study 9  .................    2-33
               Study 10 .................    2-39
               Study 11 .................    2-43
               Study 12 .................    2-49
          Studies of Secondary Importance ........    2-63
               Study 13 .................    2-63
               Study 14 .................    2-65
               Study 15 .................    2-66
               Study 16 .................    2-67
               Study 17 .................    2-68
               Study 18 .................    2-69
     VEHICLE TRAVEL ON PAVED ROADS  ...........    2-71
          Introduction  .................    2-71
          Studies of Primary Importance .........    2-72
               Study 1  .................    2-72
               Study 2  .................    2-75
               Study 3  .................    2-79
               Study 4  .................    2-82
               Study 5  .................    2-84
               Study 6  .................    2-88
               Study 7  .................    2-92
          Studies of Secondary Importance ........   2-102
               Study 9  .................   2-102
               Study 10 .................   2-103
               Study 11 . .  ' ...............   2-104
     MINING .......................   2-105
          Introduction  .................   2-105
          Studies of Primary Importance .........   2-106
               Study 1  .................   2-106
               Study 2  .................   2-117
               Study 3  .................   2-122
               Study 4  .................   2-124

                               iii

-------
          Studies of Secondary Importance .....•••   2  134
               Study 5  .	   2-134
               Study 6	•  •   •   2-35
     STORAGE PILES	...-•••   <*-\ 3 '
          Introduction		   ^07
          Studies of Primary Importance ......-••   ~^~ JL
               Study 1	   2-37
               Study 2	  .   -   2~1^
               Study 3	   ;]„
               Study 4	   2-1 5b
          Studies of Secondary Importance 	   2-158
               Study 5	   2-158
     CONSTRUCTION ACTIVITIES  ..............   2-159
          Introduction  ..... 	  .....   2-159
               Study 1	   2-159
               Study 2	   .   2-162
               Study 3	   2-165
          Introduction	-   2-170
          Studies of Primary Importance 	   .   2-170
               Study 1	   2-170
               Study 2	   2-174
     LANDFILLS	   2-180
     UNPAVED PARKING LOTS	   2-184
     WIND EROSION	   2-185
          Introduction	   2-185
          Studies of Primary Importance 	 ...   2-186
               Study 1   	   2-186
               Study 2	   2-188
               Study 3	   2-190
               Study 4	   2-194
               Study 5	   2-197
               Study 6	   2-199
               Study 7	   2-203
               Study 8	   2-206
          Studies of Secondary Importance ...  	   2-208

ACTIVITY DATA AND EMISSION FACTORS USED FOR INVENTORY
     PURPOSES	    3-1
     INTRODUCTION ........ 	    3-1
     GENERAL INFORMATION  	    3-1
     AGRICULTURAL TILLING ..... 	    3-3
          Non EPA Federal, State, and Local Agencies  .  .  .    3-4
               Silt content	    3-4
               Number of tillings and acres of land planted    3-4
          EPA regional office or state or local air
               pollution agencies 	    3-4
          Emission inventory documents  	    3-4
               Silt content	    3-4
               Number of tillings and acres of cropland  .  .    3-5
     WIND EROSION	    3-5
          Emission factors  .... 	  .....    3-5
          State  Agencies	    3-7

                                iv

-------
     CONSTRUCTION 	    3~8
          Non EPA Federal, State, and Local Agencies   .  .  .    3-8
          EPA regional office or state and local air
               pollution agencies 	    3-9
          Emission inventory documents   	    3-9
     MINING AND QUARRYING	     3-10
          Non EPA Federal, State, and Local Agencies   .  .     3-10
          EPA regional office or state and local air
               pollution agencies 	     3-11
          Emission inventory documents   	     3-11
     UNPAVED ROADS  	     3-11
          Non EPA Federal, State, and Local Agencies   .  .     3-12
               Vehicle Speeds 	     3-12
               Vehicle weight, wheels and distribution   .     3-12
               Vehicle miles travelled (VMT) and road
                    mileage	     3-12
          EPA regional office or state and local air
               pollution agencies 	     3-13
               Silt content	     3-13
               VMT	     3-13
          Emission inventory documents   	     3-13
               Silt content	     3-13
               VMT and mileage	     3-13
     PAVED ROADS	     3-14
          Emission factors  	     3-14
          Non EPA Federal, State, and Local Agencies   .  .     3-15
               VMT	     3-15
               Other	     3-15
          EPA regional office or state and local air
               pollution agencies 	     3-16
               Silt	     3-16
               Other	     3-16
          Emission inventory documents   	     3-16
     STORAGE PILES  	     3-16

EVALUATION AND RECOMMENDATIONS	    4-1
     EVALUATION	    4-1
          Methodology	    4-1
          Sampling Equipment  	    4-5
          Quality Assurance 	    4-6
          Emission Factor Development 	    4-8
          Documentation 	     4-10
          Activity Data Requirements  	     4-11
     RECOMMENDATIONS  	     4-12

REFERENCES	    5_-|

APPENDIX A	    A_-|
     NOTES AND INFORMATION CONCERNING ACTIVITY DATA SOURCES    A-1
                                v

-------
                          LIST  OF  TABLES
            PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
            EXPOSURE DATA AND EMISSION FACTORS  	
            EMISSION FACTORS FOR FIRST PHASE  	
            TEST PARAMETERS AND EMISSION FACTORS FOR SECOND
            EMISSION FACTORS FOR THIRD PHASE   	
            PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
            CONCENTRATION MEASUREMENTS AND ASSOCIATED FIELD
TABLE 2.1.
TABLE 2.2.
TABLE 2.3.
TABLE 2.4.
     PHASE
TABLE 2.5.
TABLE 2.6.
TABLE 2.7.
     DATA ........................
TABLE 2.8.  PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
TABLE 2.9. EXPOSURE DATA AND EMISSION FACTORS  ......
TABLE 2.10.  PARAMETERS MEASURED AND CORRESPONDING
     EQUIPMENT   .....................
TABLE 2.11.  EXPOSURE DATA AND EMISSION FACTORS  .....
TABLE 2.12.  MEASURED EXPOSURES AND CALCULATED EMISSION
     FACTORS  ......................
TABLE 2.13.  PARAMETERS MEASURED AND CORRESPONDING
     EQUIPMENT   .....................
TABLE 2.14.  EXTENT OF SAMPLING FOR VARIOUS INDUSTRIES   .
TABLE 2.15.  MEASURED DUST CONCENTRATIONS FOR TOTAL
     PARTICULATES AND PM10   ...............
TABLE 2.16.  EMISSION FACTORS FOR TOTAL PARTICULATES AND
TABLE 2.17.  PARAMETERS MEASURED AND CORRESPONDING
     EQUIPMENT   .....................
TABLE 2.18.  EXPOSURE DATA AND EMISSION FACTORS  .....
TABLE 2.19.  PARAMETERS MEASURED AND EQUIPMENT EMPLOYED
TABLE 2.20.  MEASURED EMISSION FACTORS  .........
TABLE 2.21.  EQUATIONS FOR PREDICTING MEDIAN EMISSION
     FACTORS3 ......................
TABLE 2.22  EEM'S EXPOSURE DATA AND CALCULATED EMISSION
     FACTORS  ......................
TABLE 2.23.  MRI'S EXPOSURE DATA AND CALCULATED EMISSION
     FACTORS
TABLE 2.24.  PEI'S EXPOSURE DATA AND CALCULATED EMISSION
     FACTORS
TABLE 2.25.  TRC'S EXPOSURE DATA AND CALCULATED EMISSION
     FACTORS
TABLE 2.26.  USS'S EXPOSURE DATA AND CALCULATED EMISSION
     FACTORS   	
TABLE 2.27.  EMISSION FACTORS FOR VARIOUS SPEEDS   .  .  '.
TABLE 2.28.  MEASURED PARAMETERS AND CORRESPONDING
     EQUIPMENT   	
TABLE 2.29.  CALCULATED EMISSION FACTORS  	
TABLE 2.30.  PARTICULATE CONCENTRATIONS AND EMISSION
     FACTORS   	
TABLE 2.31.  PARAMETERS MEASURED AND CORRESPONDING
     EQUIPMENT   	
2-27
2-29

2-32

2-35
2-35

2-37

2-38

2-39
2-42
2-44
2-48

2-48

2-54

2-56

2-58

2-59

2-61
2-64

2-76
2-78

2-81

2-82
                                VI

-------
TABLE 2.32.  EXPOSURE DATA AND EMISSION FACTORS  .....     2-84
TABLE 2.33.  PARAMETERS MEASURED AND  CORRESPONDING
     EQUIPMENT   	 .........     2-85
TABLE 2.34.  EXPOSURE DATA AND EMISSION FACTORS  	     2-87
TABLE 2.35.  PARAMETERS MEASURED AND  CORRESPONDING
     EQUIPMENT	     2~°9
TABLE 2.36.  EXTENT OF SAMPLING FOR VARIOUS  INDUSTRIES  .     2-90
TABLE 2.37.  MEASURED DUST CONCENTRATIONS FOR  TOTAL
     PARTICULATES AND PM10                  -  -	     2~91
TABLE 2.38.  EMISSION FACTORS FOR TOTAL PARTICULATES  AND
     PM10	     2~92
TABLE 2.39.  PARAMETERS MEASURED AND  CORRESPONDING
     EQUIPMENT	     2-93
TABLE 2.40.  EXPOSURE DATA AND EMISSION FACTORS  	     2-95
TABLE 2.41.  PARAMETERS MEASURED AND  CORRESPONDING
     EQUIPMENT	     2-98
TABLE 2.42.  CALCULATED EMISSION FACTORS   	  .  -    2-100
TABLE 2.43.  PARAMETERS MEASURED AND  CORRESPONDING
     EQUIPMENT   .....  	    2-107
TABLE 2.44.  EXTENT OF SAMPLING AT VARIOUS MINING
     ACTIVITIES  	  ..........    2-107
TABLE 2.45.  DISPERSION MODELLING DATA  	    2-109
TABLE 2.46.  AVERAGE EMISSION RATES BY OPERATION AND  MINEa   2-117
TABLE 2.47.  DISPERSION MODELING DATA 	    2-121
TABLE 2.48.  DISPERSION MODELING PARAMETERS  	    2-123
TABLE 2.49-  SOURCE CHARACTERIZATION  AND METEOROLOGICAL
     PARAMETERS  ......  	  ....    2-125
TABLE 2.50.  SAMPLING RUNS FOR EACH OPERATION	    2-128
TABLE 2.51.  EMISSION FACTORS DERIVED USING  EXPOSURE
     PROFILING:    .	    2-130
TABLE 2.52.  EMISSION FACTORS DERIVED FROM QUASI-STACK
     TESTING:  DRILLING	  .    2-131
TABLE 2.53.  EMISSION FACTOR DERIVED  USING EXPOSURE
     PROFILING WITH TETHERED BALLOON:  BLASTING  .  . .  .  .    2-131
TABLE 2.54.  EMISSION FACTORS DERIVED FROM UPWIND DOWNWIND
     MODELING:   COAL LOADING	    2-132
TABLE 2.55.  EMISSION FACTOR DERIVED  FROM  UPWIND-DOWNWIND
     METHOD:  DOZER, DRAGLINE, AND SCRAPER  	    2-133
TABLE 2.56.  PREDICTIVE EQUATIONS DEVELOPED  THROUGH
     REGRESSION  ANALYSIS OF  FIELD DATA3  	    2-134
TABLE 2.57.  CONCENTRATION MEASUREMENTS AND  SAMPLE  RUN
     TIMES	    2-141
TABLE 2.58.  MEASURED EXPOSURES AND CALCULATED EMISSION
     FACTORS	    2-144
TABLE 2.59.  MEASURED EXPOSURES AND CALCULATED EMISSION
     FACTORS FOR MATERIAL LOAD-OUT   	    2-149
TABLE 2.60.  MEASURED EXPOSURES AND CALCULATED EMISSION
     FACTORS FOR CONVEYOR STACKING   	    2-150
TABLE 2.61.  MEASURED EXPOSURES AND CALCULATED EMISSION
     FACTORS FOR CONVEYOR TRANSFER   	    2-152
TABLE 2.62.  DISPERSION MODELING DATA FOR  UNLOADING TRUCK   2-154
                                VII

-------
TABLE 2.63  EXPOSURE MEASUREMENTS AND CALCULATED  EMISSION
     FACTORS	    2-157
TABLE 2.64.  DISPERSION MODELING DATA AND CALCULATED  SOURCE
     STRENGTHS	    2-161
TABLE 2.65.  SOURCE STRENGTHS CALCULATED FOR VARIOUS  WIND   2-164
TABLE 2.66.  PRIMARY EQUIPMENT FOR EACH OPERATION OF  ROAD
     CONSTRUCTION PROJECT  	    2-166
TABLE 2.67.  DISPERSION MODELING PARAMETERS AND RESULTS  .    2-168
TABLE 2.68.  FINDINGS OF REGRESSION ANALYSIS   	    2-169
TABLE 2.69.  PARAMETERS MEASURED AND CORRESPONDING
     EQUIPMENT	    2-171
TABLE 2.70.  EXPOSURE DATA AND EMISSION FACTORS	    2-173
TABLE 2.71.  PARAMETERS MEASURED AND CORRESPONDING
     EQUIPMENT	    2-175
TABLE 2.72.  EMISSION MEASUREMENTS AND CALCULATED EMISSION
     FACTORS	    2-177
TABLE 2.73.  PREDICTIVE EMISSION FACTOR EQUATIONS FOR SOIL
       PREPARATION AND MAINTENANCE OPERATIONS  	    2-179
TABLE 2.74.  CONCENTRATIONS AND CORRESPONDING  SOURCE
     STRENGTHS	    2-183
TABLE 2.75.  DISPERSION PARAMETERS - AGRICULTURAL SITES  .    2-190
TABLE 2.76.  CONCENTRATION MEASUREMENTS AND SAMPLE RUN
     TIMES	    2-194
TABLE 2.77.  PARAMETERS MEASURED AND CORRESPONDING
     EQUIPMENT	    2-195
TABLE 2.78.  MEASURED CONCENTRATIONS DOWNWIND  OF  EXPOSED
     AREAS	    2-196
TABLE 2.79-  WIND EROSION SAMPLING PARAMETERS	    2-199
TABLE 2.80.  WIND EROSION PARAMETERS	    2-202
TABLE 2.81.  WIND TESTING DATA	    2-205
TABLE 4.1.  EVALUATION OF COMPUTATIONS BY JUTZE AND AXETELL,
     1974	    4-4
TABLE 4.2.  EVALUATION OF COMPUTATIONS BY LEMON ET AL.,
     1974	    4-4
TABLE 4.3.  EVALUATION OF COMPUTATIONS BY COWHERD ET  AL..
     1974	    4_8
                               Vlll

-------
                            SECTION 1

                          INTRODUCTION


     The Joint Emissions Inventory Oversight Group (JEIOG) was
formed to advise and coordinate the research plans and needs of
the Air and Energy Engineering Research Laboratory (AEERL) and
the Office of Air Quality Planning and Standards (OAQPS) related
to the development of emissions inventory methodologies.  The
work detailed in this report is part of that coordination and
needs effort.

     One of the areas of high priority for JEIOG is area source
emissions inventory methodologies for fugitive dust sources.
Emission inventory methodologies for fugitive dust sources of
PM10 have received little attention with regard to the
availability of data sources, accuracy of emission factors used
to develop inventories, frequency of updates to either emission
factors or activity data, or an evaluation of alternative methods
of determining emissions from these sources.  As more local,
state and regional agencies are required to submit regulatory
plans for attainment of the PM10 ambient air quality standard,
methods for accurately determining the emissions from fugitive
dust sources will be crucial.  As a consequence, EPA needs to
have a definitive understanding of the current status of
methodologies used to inventory these sources, so that guidance
can be developed for the preparation of emissions inventories.
In addition, EPA also needs to identify and prioritize
appropriate research and development goals for the development of
new, and enhancement of existing, fugitive dust emission
estimation techniques.  In order to help assess the current
status of fugitive PM10 emissions estimation methodologies and to
develop and prioritize the research and development goals for
these sources, EPA issued E.H. Pechan and Associates, Inc. a work
assignment to evaluate current fugitive PM10 emissions
estimations methodologies and to provide an assessment of the
research and development goals required for the development of
new or the enhancement of existing methodologies.

     The work carried out as part of this assignment was
implemented in two broad tasks.  These tasks were:

1 .   An evaluation of existing sources of field-gathered
     emissions data, for all 'fugitive dust area sources.  This
     evaluation included assembly and documentation of currently
     available sources of data.  Evaluation and discussion of the
     data included, but was not limited to, extent of data
     collection, type and extent of quality assurance utilized,
     documentation of raw data, and peer review procedure
     utilized.  In addition, a specific subtask was to include a
     determination of the process necessary to fully document the

                               1-1

-------
     fugitive dust emissions data gathered during the NAPAP
     program.  In particular, the NAPAP data evaluated was that
     concerning emissions from unpaved roads.

2.   A subset of the fugitive dust sources examined as part of #1
     above were examined to determine and document the
     methodologies currently used to estimate emissions.  This
     subset included wind erosion, unpaved roads, agricultural
     tilling, construction/demolition activities, mining and
     quarrying, paved roads, and storage pile activities.
     Documentation of the current methodologies used to determine
     emissions included evaluation and discussion of the two
     principal inputs to an emissions estimation methodology,
     namely emission factors and source activity data.  As part
     of this task, documentation of the currently used emission
     factor included the identification of which data summarized
     as part of #1 above was used to develop these emission
     factors.  The evaluation of the source activity data
     included an examination of the completeness or extent of
     activity input, magnitude of the inventory area to which the
     activity data is applicable or available (i.e., local,
     regional, or national), accuracy of the data (if possible),
     and where appropriate, frequency and cost of updates
     required.  Again, a specific subtask was the determination
     of the process necessary to document the source activity
     data gathered during the NAPAP program, in particular, that
     data that applies to the unpaved road emissions estimates

     The results of this evaluation are presented below.  Over 80
studies were evaluated to determine whether or not they were
suitable for inclusion in this report.  The studies presented in
the review were selected based upon whether or not actual
emissions measurements were made for that source.  Normally, only
emissions from uncontrolled sources were considered, however, a
few sources were reviewed if we felt the results warranted
inclusion from either a data quality perspective or from a
historical development perspective.  Some studies that are
reviewed included emissions measurements from both controlled and
uncontrolled sources.

     The report is divided into four sections.  Section one is
this introduction.  Section two details the emissions data,
methodologies, equipment configurations, sampling sites,
parameters measured, quality assurance aspects and findings of
the studies reviewed.  Section three examines the sources of
activity data for the subset of sources considered in #2 above.
Section four presents our evaluation of the reviewed studies and
presents our comments and recommendations for future research and
development activities.

     The determination of the requirements for documenting the
information from the 1985 NAPAP emissions measurements and

                               1-2

-------
activity data used to determine the emissions estimates from
unpaved roads for the 1985 NAPAP emissions inventory are not
included here.  That information was provided to the Work
Assignment Manager separately.
                               -3

-------
                            SECTION 2

                    EMISSIONS DATA ASSESSMENT
VEHICLE TRAVEL ON UNPAVED ROADS

Introduction

     Unpaved roads are a , significant source of particulate
emissions in the United States.  This category has received more
attention by researchers than any other fugitive particulate
source.

     The majority of studies reviewed in this report used one of
two approaches to estimate emissions:  dispersion modeling or
exposure profiling.  The first technique generally involves
measurement of dust concentrations upwind and downwind from the
road, followed by solution of a generalized dispersion model to
determine the source strength in units of mass of dust per unit
of road length per unit time.  This is in turn converted into an
emission factor  (mass per vehicle-mile or per vehicle-kilometer)
by dividing by the number of vehicle passes.

     Exposure profiling has been used to measure fugitive
particulate emissions from several source categories.  As it is
applied in the case of roads  (paved and unpaved),  it involves
measurement of the horizontal mass flux of dust downwind from the
road by isokinetically sampling the air at several points over
the height of the plume simultaneously.  Background levels of
dust are subtracted from the exposure calculated at each downwind
sampling height.  Exposure is then integrated with respect to
height to obtain the mass emitted per unit of road length.
Dividing this by the number of vehicle passes during the test
gives the emission rate in mass per vehicle-mile (or per vehicle-
kilometer) .

Historical Development

     According to Turner (1970), atmospheric dispersion modeling
in general had its beginnings in the 1930's.  Cowherd et al.
(1974) found that the first application of dispersion modeling
for determining particulate emissions from unpaved roads was by
undergraduate students at the School of Engineering at the
University of New Mexico in 1.971 .   (No documentation for this
study has been found).  High volume  (hi-vol) samplers were used
then, and the most recent dispersion studies have also relied
primarily on hi-vols to measure particulate concentrations.  Of
the studies which determined emission factors using dispersion
modeling, only Jutze and Axetell's 1974 work used other equipment
to measure concentrations.  They used a beta gauge developed by
GCA Corporation.  The primary difference among  the dispersion

                               2-1

-------
modeling studies has been in the number and placement of hi-vols
used to estimate concentration.  The most recent documented study
which employed dispersion modeling to measure particulate
emissions from vehicle travel on unpaved roads was performed in
1977 by McCaldin.

     Exposure profiling is a much more recent development in the
field of air pollution measurement than dispersion modeling.  It
was developed by the Midwest Research Institute (MRI) under a
contract for EPA in 1973-74, specifically for the purpose of
measuring fugitive particulate emissions.  The first system
consisted of several high-volume sampling "heads" mounted
vertically on a profiler tower.  The sampling heads were all
attached to a single vacuum blower by flexible hosing.  The
orientation and intake velocity of the sampling heads were not
adjusted during individual tests:  the investigators noted that
meteorological changes during individual tests were
insignificant.  Dust deposition between the road and the profile
tower was considered part of vehicle emissions and, therefore,
dustfall buckets were used to measure it.

     This system has undergone several changes since its original
development.  For example, dust deposited between the road and
the profiler is no longer measured or included in the emission
factor.  Suction at each sampling head is provided by independent
blowers, facilitating variation in sampling speed between heads.
Another refinement of the methodology is the use of warm-wire
anemometers at each sampling intake to continuously monitor wind
speed and adjust the sampling velocity to match it in order to
maintain isokinetic conditions.

     Several organizations have applied this methodology, each
with its own variations in specific procedures or equipment
configurations.  MRI positions its profiler filters horizontally
such that air is drawn in through a settling chamber (to catch
particles larger than about 50 ^im, since these are generally not
suspended) before it passes up through the filter.  Other
organizations use vertically mounted filters in the profiler.
Background dust levels are measured using a profiling tower in
some cases and a standard hi-vol in others.  Determination of
particle size distribution has been performed in several ways:
cascade impaction, scanning electron microscopy, and stacked
filters.  Between 1984-86 the Southern Research Institute (SoRI)
evaluated this methodology by conducting a side-by-side
comparison of five teams using the method but with variations in
the specific implementation.' This is the most recently
documented application of exposure profiling in measuring
particulate emissions from unpaved roads.

     Several relevant studies did not employ either profiling or
dispersion modeling methodologies.  Rather than set up a plume
sampler beside the road, Roberts (1973) towed it behind the car

                               2-2

-------
to measure the concentration of the plume and multiplied this by
the estimated volume of the plume after traveling one mile to
obtain the mass emitted per vehicle-mile.  Handy et al.  (1975)
measured dust deposition near unpaved roads on a kilogram per
month basis.  Pinnick et al. (1985) measured particle
concentrations and size distributions using a variety of optical
particle measuring devices.

Organization

     Reports that are of primary importance to the development of
emission factors for vehicle travel on unpaved roads are reviewed
in chronological order.  The criteria for selecting these primary
reports were as follows:

     1.   The study must attempt to measure dust emissions from
          unpaved roads.

     2.   The report must provide raw data (i.e. exposure of each
          filter for the profiling method or upwind and downwind
          concentrations for the modeling method).

     3.   Emissions from the test road must be uncontrolled (i.e.
          no dust suppressing treatments should be effecting
          emissions).

     4.   The test road must be authentic or very nearly so (i.e.
          the road must reflect typical unpaved road surface
          conditions).

     The review of each primary report consists of a summary
description of the methodology, test site(s), measured
parameters, equipment configuration, sampling extent, quality
assurance, findings, and publication outlet.  Reports that are
relevant to the development of emission factors but which fail
one or more of the above criteria are considered of secondary
importance and are briefly summarized in chronological order in
the section following the primary study reviews.
                               2-3

-------
Studies of Primary Importance

Study 1— Jutze and Axetell, Investigation of Fugitive Dust
          Volume I - Sources, Emissions, and Control. EPA-450/3-
          74-036-a.  1974.

     Methodology--A beta gauge was used to determine particulate
concentrations at several heights and several distances from the
road.  These data points were substituted into the equation for a
continuously emitting infinite line source (Turner, 1970),
yielding an estimate for source strength (g/m/sec), which was
converted to an emission factor (Ib/veh-mile) by dividing by the
rate at which vehicles pass the sampling point (veh/minute).  It
was assumed that a vehicle passage rate of 5 per minute was
sufficient to qualify as a continuously emitting line source.
Several sets of measurements were made with traffic traveling at
different speeds for each test.  Traffic flow for this study was
regulated by the investigators.

     Regression analysis was used to estimate the relationship
between the measured emission factor and vehicle speed.  Because
the hi-vol is considered the standard equipment for measuring
total suspended particulates, and it includes in its "catch" a
much larger range of particle sizes than does the beta gauge, the
average ratio of concentrations measured with the hi-vol sampler
versus those obtained with the beta gauge was used as a
multiplier for the estimated relationship.

     Simple wind erosion of particulates from unpaved roads was
then estimated using a model developed by Woodruff and Siddoway
(1965).  Emissions were estimated in tons/mile/year and divided
by annual traffic volume (ADT * 365) to derive a factor in
Ibs/veh-mile.  This wind erosion factor was then added to the
factor representing vehicle travel on unpaved roads to obtain a
total emission factor for unpaved roads.

     Test Sites—Two pre-existing, unpaved roads in Sante Fe, New
Mexico were the sites for measuring plume concentrations using
the beta gauge.  Other characteristics of the roads, such as
percent silt or presence of gravel, were not documented.

     Parameters and Equipment—Plume concentrations were
estimated using both the beta gauge and the standard hi-vol
sampler.  Suspended particles were categorized as greater than or
less than 3.3 (J.m in diameter using a hi-vol fitted with an
Andersen impactor.  The investigators did not specify whether
this cut point was in aerodynamic or Stokes diameter.  Vehicle
speed was varied between sampling runs.  Automatic counters
tracked traffic volume.  Wind speed and direction were monitored
with continuous sensors.
                               2-4

-------
     Equipment Configuration--Beta gauge measurements were taken
at 50, 75, 125, 200, and 300 feet downwind from the unpaved road.
At each distance from the road, measures were taken at heights of
3, 6, and 10 feet.  Hi-voi samplers, filtering air at 6 feet
above ground, were also located at 75 and 200 feet from the road.
There was apparently only one beta gauge instrument used in the
study; therefore, measurements at the various distances from the
road were not taken simultaneously.

     Sampling Runs—Measurements were taken in six separate
intervals over a two-day period.  During each period, beta gauge
measurements were taken at the three heights at several of the
downwind stations  (75 ft, 125 ft, etc.), and hi-vol samples were
collected at the 75 and 200 foot stations.  The number of vehicle
passes varied between measurement periods.  The beta gauge
required between one and eight minutes to collect a sample and
measure its concentration.  Hi-vol samplers generally ran about
one hour for each test.

     Quality Assurance—Procedures for operating the hi-vol
sampling equipment and collecting samples were fully documented.
Careful attention was given to sample handling and
transportation.  However, detailed documentation of how the beta
gauge was used and how the hi-vol filter samples were weighed was
not provided.  Consequently, this study is generally not
reproducible.

     Finally, although the investigators claimed to use the usual
model for continuously emitting line sources, as found in the
Workbook of Atmospheric Dispersion Estimates  (Turner, 1970),
their equation differs significantly from the model presented in
that document.  The equation as it appears in the workbook is
                                                 1
         X (x,y,0;H)  -  	±-g-	  exp
                          s±n
The  formula applied by the investigators was
         X (x,y,0}H)   -  - 2-2 - 8±n $ exp
                           V/27I O  U
The  represents the angle between the mean wind direction and
the road.

     Furthermore, contrary to Turner's  (1970) recommendation, the
investigators collected much of their data during periods in
which <}> was less than 45°.

                               2-5

-------
     Findings—The regression analysis mentioned above produced
the following best fit line:

                        E = (0. 16) (1 .068)x

where

     E    =    dust emissions (Ib/veh-mile)
     x    =    vehicle speed  (mph)

Evaluating this formula for x - 30 produces an emission  factor of
1.15 Ib/veh-mile.

     The investigators found that plume concentration as measured
by the standard hi-vol method averaged 1.68 times higher than the
concentration estimated by the beta gauge, with a correlation of
.87.  Thus the coefficient in the above equation was multiplied
by 1.68, and the  equation was revised to  be:

                        E = (0.27)(1.068)x

Evaluating this formula  for x = 30 results in E =  1.94 Ib/veh-
mile .

     The estimated  emission factor for wind erosion on unpaved
roads was  1.54 Ib/veh-mile.   This was derived using a published
wind erosion model; no  field  data on concentrations downwind from
an unpaved road without  traffic were collected.  The sum of  these
two factors  is 3.7  Ib/veh-mile.

     Publication--This  study  was conducted and documented under
contract for the  Environmental Protection Agency, Office of  Air
and Waste Management, Office  of Air Quality Planning and
Standards, Research Triangle  Park, North  Carolina.  It was
published  in  1974 as  Publication No. EPA-450/3-74-036-a.

Study  2--  Cowherd et  al.   Development of  Emission Factors for
           Fugitive  Dust  Sources. EPA-450/3-74-037.  1974.

     Methodology--Under  a contract with the EPA the Midwest
Research Institute  (MRI)  developed and applied its exposure
profiling  methodology for measuring emissions for  certain
 fugitive dust  sources.   The basic procedure used  to measure
emissions  from vehicle  travel on unpaved  roads is  as  follows:

      1 .    Set  up  a  vertical array of  isokinetic filter  samplers
           downwind  from the unpaved road. Position a  standard
           hi-vol  sampler upwind from  the  road to  measure the
           background  dust level.

      2.    Collect filter samples of the dust plumes created  when
           vehicles  travel past  the array. Operate the  samplers

                                2-6

-------
          on the profile tower and the upwind,  hi-vol samplers
          simultaneously -

     3.   Determine the mass of dust on the profiler filters and
          the upwind filter.

     4.   Subtract the background mass from the dust mass on each
          of the profiler filters.

     5.   Calculate the exposure (mass/area)  of each sampling
          intake by dividing the dust mass on each filter
          (adjusted for background contribution)  by the intake
          area.

     6.   Calculate the integral of exposure with respect to
          height.  This is the plume mass per unit length of road
          during the test.

     7.   Divide by the number of vehicle passes to calculate the
          plume mass per vehicle mile.

     Test Sites—Two gravel roads and two dirt roads in Kansas
were used as sampling sites.  These public roads were considered
to be representative of the unpaved roads in the Dust Bowl area
of the Great Plains.

     Parameters and Eguipment--Table 2.1 shows the parameters
that were measured and the equipment used to measure them.  All
of these parameters were not measured in every sampling run.

   TABLE 2.1.  PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
Parameter
Wind speed
Wind direction
Cloud cover
Temperature
Relative humidity
Road surface type
Road surface texture
Road surface moisture content
Embankments
Traffic type
Traffic count
Plume dust exposure
Equipment
Unspecified
Unspecified
Direct observation
Sling psychrometer
Sling psychrometer
Direct observation
Sieves , scales
Oven, scales
Direct observation
Direct observation
Direct observation
Isokinetic exposure profiler
                               2-7
(continued)

-------
                     TABLE 2.1.  (continued)
Parameter
Plume particle size distribution
Plume concentration
Background concentration
Duration of sampling
Plume traveling time
Dust deposition
Ecfuipment
Cascade impactor in standard hi-vol
housing (Sierra impactor, Aerotec
cyclone, or Anderson impactor,
depending on the sampling run)
Standard High-volume sampler
Standard high-volume sampler
Timer
Timer
Dustfall buckets
     Equipment Confiquration--The exposure profiling tower was
positioned at least 20 feet downwind from the road.  Four
isokinetic exposure profiling heads were attached to the tower at
heights of 3, 5.5, 8, and 10 feet.

     Dustfall buckets were positioned in a line perpendicular to
the road at a height of one foot.  Their distance from the
downwind edge of the road varied between sampling runs and ranged
from 9 to  100 feet.  No more than three buckets were used in any
one sampling run.

     The upwind, background concentration was measured using a
standard high-volume sampler, positioned 3 feet above the ground
at an unspecified distance from the downwind edge of the road.
For some of the sampling runs, another standard high-volume
sampler was positioned 6 feet above the ground beside the
exposure profiler so that the large-particle trapping efficiency
of the standard high-volume sampler could be checked.

     When  the particle size classifier was used, it was located
beside the exposure profiler at a height of 6 feet.

     Meteorological conditions were deemed to be essentially
constant during each sampling run.  Wind direction and speed were
measured by unspecified instrumentation at a height of 12 feet.
Pasquill's Stability Classification was used to characterize the
prevailing meteorological conditions during each sampling run.

      Sampling Runs—A total of six one-hour sampling runs
employing  the exposure profiler were conducted. Three runs were
performed  on each of the two road types.  The number of vehicle
passes per sampling run ranged from 55 to 273.  The total number
of  vehicle passes was  1,018.
                                2-8

-------
     Quality Assurance—In the use of standard high volume
filtration, the investigators followed the procedures specified
by EPA in "Reference Method for the Determination of Suspended
Particulates in the Atmosphere (High Volume Method)," Federal
Register, 36, 28 Appendix B, 22388-22390, 25 November 1971.  For
the measurement of dust deposition, the investigators followed
the procedures set forth in "Standard Method for Collection and
Analysis of Dustfall," ASTM Method D 1739-62.

     Samples of the dust plume were collected only when the wind
speed was less than 20 mph, the maximum speed under which samples
could be collected isokinetically.  As noted above, wind
direction and speed were observed to be constant during each run.
Likewise, the intake velocity and the directional orientation of
the samplers in the exposure profiler were constant during each
sampling run.  Filter sampling under isokinetic conditions
assures a more accurate measure of dust exposure (the mass of
dust passing through a plane) than simple open-filter sampling.

     Filters were conditioned in a controlled temperature and
humidity environment before and after collection of dust samples.
Filter samples were transported to the laboratory in individual
folders.  The interior surfaces of the sampler heads were rinsed,
and the water was captured and later evaporated to determine the
mass of dust on the interior surfaces.

     The specific methodology followed in this study was
documented in sufficient detail to permit reproduction of the
study.  Indeed, the exposure profiling technique has been
utilized and adapted by several other organizations and
investigators.  However, documentation of quality assurance
practices such as collocation of samplers, processing of blank
profiler filters, and audits of profiler filter weights was not
provided.

     The investigators acknowledged two potential sources of
small particle bias in their measurement of particle size
distribution:  1) particles bouncing down through the cascade
impactor to smaller particle stages; 2) non-isokinetic sampling
which collects larger particles with lower efficiency than
smaller particles.

     Findings--The emissions data collected in this study are
presented in Table 2.2.

     The investigators developed the following equation for
estimating emissions of particles smaller than 100 |j.m in Stokes
diameter from vehicle travel on unpaved roads, based on the data
collected in this study:

                        e = 0.81 s  (S/30)


                               2-9

-------
where
     e
     s
     S
Emission factor  (pounds per vehicle-mile)
Silt content of the road surface material
Average vehicle speed  (miles per hour)
(percent!
This equation applies only to days with rainfall  less  than 0.01
inches.  Using this equation, the predictions  for the  six test
conducted were within 10% of the actual Ib/veh-mile.

     Publication--This study was conducted  and documented under  a
contract with the EPA Office of Air and Waste  Management, Office of
Air Quality Planning and Standards, Research Triangle  Park, North
Carolina.  It was published in  1974 as Publication No. EPA-450/3-
74-037.

           TABLE 2.2.  EXPOSURE DATA AND EMISSION FACTORS
Run





2
3
8
10
13
Height
(ft)
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3

Unit
Exposure
(mg/in.2/
vehicle)
0.082
0.289
0.591
0.619
0.162
0.357
0.552
0.843
0.172
0.459
0.753
1 .158
0.238
0.418
0.737
1 .56
0.150
0.242
0.281
0.511
0.866
1 .65
2.73
4.15
Emission Factors (Ib/vehicle-mile)
Total
Particulates
(Integrated
Unit
Exposure)
10.0
10.3
13.9
16.3


. u



DD . y

da > 30 \i.m
4.0 (40%)
3.5 (34%)
6.0 (43%)
8.2 (50%)
2.1 (35%)
24.0 (43%)
2 < da < 30 \JL-m
. 3.3 (33%)
3.7 (*36%)
4.3 (31%)
4.4 (27%)
2.1 (35%)
18.8 (34%)
da < 2 jim
2.7
(27%)
3.1
(30%)
3.6
(26%)
3.7
(23%)
1 .8
(30%)
13.1
(23%)
   a particle Stokes' diameter
                                 2-10

-------
Study 3-- Lemon et al.  Derivation of Suspended Particulate
          Emission Factor for Motor Vehicle Use of Unoaved Roads.
          1975.

     Methodology—Dust emissions from vehicle travel on unpaved
roads were estimated using dispersion modeling.  Standard hi-vol
samplers were used to measure upwind and downwind concentrations.
The usual equation for a continuously emitting line source
(Turner, 1970) was used to derive emission rates from the
calculated concentrations and other relevant field data.

     Test Site—An unpaved road in the Tucson metropolitan area
was the site for this field study.  The road was oriented east-
west and was considered to be representative of the unpaved roads
in the area in terms of emissions.

     Parameters and Equipment—Two unmodified hi-vol samplers
were used:  one for measuring background concentration and one
for measuring plume concentration.  A hand held wind instrument
was used to measure wind speed and direction.  Plume height at
the downwind sampler was estimated to be 3.05 meters (10 feet).
Traffic speed was also estimated at 30 mph.  The number and type
of vehicles passing the sampler were also recorded.

     Equipment Configuration--The upwind hi-vol was located 200
feet south of the road centerline.  The downwind hi-vol was 30.5
feet north of the centerline.  The wind instrument was held by
the investigators; wind speed and direction were continuously
monitored and recorded.

     Sampling Runs—Four test runs, each approximately three
hours long were conducted on separate days over a one week
period.  The number of. vehicle passes per run ranged from 33 to
84, and the total number of passes was 260.

     Quality Assurance—Quality assurance was lacking in this
study.  The specific procedures for sample handling and analysis
were generally not documented.  The investigators did not
correctly apply the  dispersion model for continuously emitting
infinite line sources  (Turner, 1970):  they repeated the mistake
made by Jutze and Axetell in 1974  (see review of that study).
Furthermore, the angle between the wind direction and the road
was less than 45° for a significant portion of the sampling time
for three of the four tests.  This is contrary to Turner's
recommendation.

     Findings—The investigators calculated emission factors of
8.14, 5.37, 6.81, and 3.25 Ib/veh-mile.

     Publication—This study was a follow-up of the field
investigation and analysis performed by PEDCo Environmental
Specialists, Inc. in 1974 (see Jutze and Axetell, 1974).  It was

                               2-11

-------
conducted by researchers at the Pima County Air Quality Control
District, based on their belief that PEDCo's findings were not
accurate.  No formal publication of this study has been found.

Study 4-- Struss and Mikucki,  Fugitive Dust Emissions from
          Construction Haul Roads.  1977.

     Methodology--Three different experiments were conducted.
For the first phase a fully enclosed 25 foot diameter test track
was used to evaluate the effect of soil type, soil moisture
potential, vehicle speed, and vehicle weight on emissions.  A
cross-arm spanning the track held one tire at each end and was
supported by a vertical center pole, which could be rotated at
various speeds.  The track was surfaced with gravel overlain with
6 inches of Gooselake clay.  This arrangement provided a physical
model for vehicle travel on unpaved roads.  Soil psychrometers
were placed 1/2 inch below the surface to measure soil water
potential.  Particulate concentration was measured using hi-vol
samplers, which collected a fraction of the air being vented from
the enclosure.  The emission rate was calculated as the
concentration times the flow rate divided by the tire speed.

     For the second phase, three test roads, each 300 feet long
and  10 feet wide, were constructed by plowing and tilling strips
of land near Champaign, Illinois.  No indication was given that
the roads were compacted prior to testing.  Soil psychrometers
were placed beneath the surface at several places on each road.
These roads all crossed at the middle and were arranged to allow
for testing with nearly perpendicular winds under various mean
wind directions.  Once the appropriate road was selected, a
single truck was driven back and forth on the track, while hi-
vols were operated downwind.  Vehicle speed and weight were
varied between tests.  The emission rate was calculated using the
Gaussian model for a continuously emitting, infinite line source.
The plume centerline height and the hi-vol sampling height were
judged to be equal.  Therefore, a simplified version of this
model could be used:

                          X	—-
                             v/27t uo Zsin4>


where

     q    =    source  strength
     u    =    wind speed
      az   =    standard deviation of the plume's vertical
               concentration distribution
      <|>    =    angle between wind direction and road

      For the  third phase,  soil from the outdoor test roads was
 applied  to  the enclosed  test track and tested in the same manner

                               2-12

-------
as described for the first phase.  The soil type on the outdoor
test track was Drummer silty clay loam.

     Data from the third and first phases were compared to
determine the effect of the change in soil type.  Data from the
third and second phases were compared to evaluate the model, i.e.
the test track, relative to the process being modeled, the
outdoor emission generation from outdoor haul roads.   The
authors felt that the outdoor test roads realistically
represented construction haul roads.  Emission rates were all
expressed in terms of mass per tire-mile, for ease of comparison.

     Test Sites—The indoor test track was at the University of
Illinois at Champaign. The outdoor test roads were constructed in
undeveloped land near Champaign.

     Parameters and Equipment—In both cases high volume samplers
were used to measure dust concentration.  The soils' water
potential was measure using soil psychrometers placed below the
surface.  Soil water potential was said to indicate how tightly
water was held in the soil.  The psychrometers were connected to
an instrument which automatically displayed and recorded the
data.  Other soil parameters for which data was collected
included soil plastic limit, liquid limit, silt content.  The
means for measuring these parameters were not documented.

     One parameter was measured only for the indoor test track.
The fraction of dust in the evacuation duct that was sampled by
the hi-vols was determined by using a probe apparatus to sample
the dust concentration at several points in the duct's cross
section.  The ratio of the probe concentration to the
concentration measured by the hi-vol was found to be  .9-

     In the outdoor testing operation, wind speed and direction,
atmospheric stability, and vehicle speed were also measured.  A
recording windvane anemometer provided data on wind speed and
direction, whereas atmospheric stability was estimated using the
method suggested by Pasquill  (1968), which involves noting the
frequency and magnitude of wind direction changes.

     Equipment Confiquration--For the indoor test track, hi-vol
sampling heads were fixed in a box connected by a duct to the
primary evacuation duct.  The hi-vol heads were connected by more
flexible ducts to their respective vacuum motors, which drew a
portion of the air from the evacuation duct into the box and
through the filters.  The soil psychrometers were placed
approximately one-half inch beneath the surface in the tire path.

     The equipment configuration and the methodology  for the
third phase was identical to that of the first.  The  only
difference was the soil type.
                               2-13

-------
     For the outdoor tests, three hi-vols were placed downwind
from the road most suitable for testing 'based on wind conditions.
Two were 50 feet downwind and the third was 100 feet.  Three
psychrometers were embedded in each of the roads in the tire
paths.

     Sampling Runs—Eighteen sampling runs were conducted in  the
first and third phases, randomly varying the tire speed and
weight between tests.  At high tire speeds the runs lasted
roughly three hours, compared to about eight hours for slow
speeds.  Hi-vol filters were changed and psychrometer readings
were taken every .5 to 2 hours.  Thus, every run yielded multiple
concentration measurements.

     Nine one-hour runs were conducted at the outdoor test site.
Again, vehicle speed and weight were varied between tests, but  it
was not practical in this instance to randomly vary the vehicle
weight.  Difficulties in weighting the tire arm prevented
randomization of weights.

     At the time this report was completed, only three sampling
runs in the third phase were completed.

     Quality Assurance—Quality assurance was not an important
issue for the authors of this report.  No documentation was
provided for calibration of hi-vol samplers, auditing of sample
weights, processing of blank filters, or other standard quality
assurance procedures.  However, documentation of raw data and
methodological procedures was thorough.

     Findings—The raw data generated by these three experiments
are presented in order in Tables 2.3-2.5 below.  Linear
regression analysis of the data from the first phase yielded  the
following relationship, with an R2 of 0.64:

           ln(D) = -5.37 + 0.12(T) + 0.21(S) + 0.90(W)

where

     D    =    emission rate, grains/veh-mile
     T    =    soil water potential, atmospheres of tension
      S    =    vehicle speed, mph
     W    =    vehicle weight, Ib/tire

Each  of the independent variables were significant at the  95
level.

     The following equation was produced using linear regression
analysis of the outdoor test road data:

            ln(D)  = 5.28  + 0.01(T)  + 0.06(S)  +  0.50(W)


                               2-14

-------
The R2 value for this model  was .66.

     Data from the third phase was pooled with that from the
first, and a regression analysis was again conducted, this time
with the addition of another variable, soil plastic limit.
Plastic limit, according to Struss and Mikucki, is "the percent
moisture, by weight, contained in a soil when it passes from a
plastic to a brittle state while drying."  The resulting equation
is shown below.  The R2 value  of this  equation was not
documented.

       InD = -13.05  + 0.12(T)  + 0.21(S)  + 0.90(W)  + 0.48(P)

In this equation P is the plastic limit, and all the other
variables are defined as before.

     The only comparison between the second and third phase
results was an acknowledgement that emissions from the outdoor
test site were consistently higher than  they were from the indoor
test track.  In fact, they were roughly  an order of magnitude
higher.  The investigators surmised that this was due primarily
to differences in the aerodynamics created by the moving vehicle
and the rotating arm.  Another factor which might lead to lower
concentration measurements for the test  track is the deposition
of dust particles on the interior surfaces of the ventilation
system before reaching the hi-vols.

     Finally, data from all three phases were pooled and
subjected to linear regression analysis.  The resulting equation,
for which no R2 value was given,  is shown below.

       InD = -5.28 + 0.01(T) + 0.06(S) + 0.50(W)  + 0.48(P)

     Publication—These experiments were performed and documented
by the U.S. Army Construction Engineering Research Laboratory
(CERL).  It is Special Report N-17, completed in February 1977.
                               2-15

-------
TABLE 2.3.  EMISSION FACTORS FOR FIRST  PHASE
Run
1



2


3
4
5
6
Duration
(hours)
2
2
1 .9
.9
1
.9
.5
.5
2.2
1 .9
1
.75
1
1 .1
.9
.8
1 .05
.5
.55
.45
.5
.45
.55
.25
1
1
.95
.8
1
.4
Filter
Weight
Increase
(grains)
0.5386
1 .8102
5.3333
4.0849
17.941
46.555
43.347
60.739
5.2314
25.704
32.244
28.322
0.9707
3.7700
11 .035
15.605
24.268
0.6713
3.6774
6.4382
14.043
18.847
26.157
19.065
1 .2639
1 .8796
4.6157
7.4907
14.954
9.8163
Emission Rate
(grains/tire-
mile)
1 .42
4.85
15.1
23.8
32.7
86.5
146
208
8.60
49.6
121
139
3.44
12.2
44.1
70.6
87.2
2.35
11 .8
25.4
50.5
76.3
87.8
137
4.51
6.75
17.2
33.9
54.5
88.2
Run
7
8
9

10
Duration
(hours)
1
1
1 .05
.95
1
.4
.95
.85
.9
.9
1 .05
.9
1
1
.9
1 .05
.5
.35
.5
.45
.5
.5
.9
1
.9
.95
.95
1
Filter
Weight
Increase
(grains)
0.6157
n.d.
1 .6759
4.4182
8.6234
5.1 543
0.2901
0.1096
0.1003
0.1111
0.2407
0.3395
0.3117
0.8287
0.9861
1 .9089
1 .0602
0.5093
5.1512
13.785
29.586
35.418
0.5447
0.2207
0.3827
0.7701
3 -.7099
3.3966
Emission Rate
(grains/tire-
mile)
2.23
n.d.
5.70
16.3
30.4
43.7
1 .57
0.66
0.58
0.64
1 .20
1 .97
1 .63
4.32
5.75
9.61
11 .0
2.51
17.8
53.6
106
131
3.14
1.14
2.20
4.23
20.5
17.5

                     2-16
(continued)

-------
TABLE 2.3.   (continued)
Run
11
12
13
14
Duration
(hours)
.5
.5
.5
.5
.45
.55
.45
1 .05
.95
.9
.95
.9
.5
.5
1 .05
1
1
.95
1
.5
1
1
.95
.9
.5
.5
Filter
Weight
Increase
(grains)
1 .0725
n.d.
3.3765
8.2253
13.475
20.000
22.741
0.875
2.6620
7.0169
14.400
22.125
13.048
19.952
0.5463
0.4367
1 .0833
2.4846
4.3410
3.0247
0.8318
2.6836
7.8827
16.268
13.275
15.241
Emission Rate
(grains/tire-
mile)
3.75
n.d.
11 .9
29.2
52.4
64.9
88.4
29.2
9.80
27.8
53.0
58.5
87.3
136
2.59
2.18
5.44
13.4
22.2
30.5
2.78
9.15
28.3
60.3
91 .7
107

Run
15
16

17
18
19
Duration
(hours)
1 .05
.95
1
.45
1
1
.95
.95
1
.95'
1
1
1
1 .1
.75
1 .05
.95
1
.95
.45
.5
.45
.5
.5
.55
1
.9
1 .1
Filter
Weight
Increase
(grains)
1 .9568
8.4783
16.378
11 .497
0.3688
0.1991
0.2562
0.5432
1 .3272
2.2700
1 .9630
0.4028
0.2269
0.2145
0.2160
0.5000
0.5880
1 .0772
1 .5555
0.8750
1 .4846
3.8287
7.2392
9.3194
14.992
0.8410
3.0802
11 .495
Emission Rate
(grains/tire-
mile)
6.16
29.9
56.2
85.5
1 .83
0.99
1 .34
2.83
6.66
12.0
9.67
2.03
1.15
0.99
1 .45
2.42
3.17
5.54
8.27
3.21
4.93
14.3
24.5
31 .6
46.5
1 .41
5.72
17.7
          2-17

-------
FABLE 2.4. TEST PARAMETERS AND EMISSION FACTORS FOR SECOND PHASE
Run
1
2
3
4
5
6
7
8
9
10
1 1
Duration
(hours)
1 .5
1
1
1
1
1
1
1
1
.9
1
Concentration
at 50'
(grain/ft3)
.000262
.000571
.00236
.00151
.00552
.00142
.000509
.00105
.000802
.00148
.00106
Concentration
at 100'
(grain/ft3)
.000154
.000342
.00128
.000910
.00287
.000941
.000278
.000602
.000633
.000833
.000880
0Z at
50'
(ft)
* .6
-: .6
3.9
6.2
6.2
4.6
2.6
3.9
2.3
2.3
2.9
az at
100'
(ft)
8.5
8.5
6.9
13.1
13.1
8.5
4.9
6.9
4.3
4.3
5.6
Wind
speed
(ft/sec)
10.5
9.5
11 .0
7.5
6.5
11 .0
13.0
13.0
19.0
14.0
17.0
Emission
Factor3
(grains/tire
-mile)
599
1060
2600
2470
5830
2530
723
1720
1850
1380
2270
a)  Whether this rate  is based on the measured concentration at 50 feet, 100
feet,  or  the average  was not noted.

           TABLE 2.5.   EMISSION  FACTORS FOR  THIRD PHASE
Run
1
2
3
Duration
(hours)
1
1
1
.95
.95
.95
.5
.5
.5
.45
.45
.5
.3
.5
.5
.5
.25
. Filter Weight
Increase (grains)
0.9537
1 .7731
2.8904
3.9815
5.1265
6.4382
2.4737
3.3580
7.7762
6.8750
10.356
12.185
8.4922
3.5123
9.1882
12.944
14.150
Emission Factor
(grains /tire-mile)
4.70
8.84
14.7
21 .3
27.4
34.7
16.8
22.9
53.0
52.8
80.0
85.3
97.8
11 .6
30.9
44.1
95.3
                                   2-1

-------
Study 5— Axetell.  Survey of Fugitive Dust from Coal Mines.
          EPA-908/1-78-003.  1978.

     Methodology--Upwind-downwind dispersion modeling was used to
measure emissions from haul roads at surface coal mines.
Concentrations and other field parameters were used in the model
shown below:
                         x-
                                 2q
                                    a zu
where
X

q

ff*


u
               plume centerline concentration at a distance x
               downwind from the source, g/m3
               source strength, g/sec-m
               angle between wind direction and line source
               the standard deviation of the plume's vertical
               concentration distribution at a downwind distance
               x, m
               mean wind speed, m/sec
Following the example set by Turner (1970, problem 23), the
investigators used a simplified version of the model for a
continuously emitting, infinite line source.  The effective
height of emission was taken to be zero; consequently the
expression
                           exp
                                   H
always evaluated to one and could therefore be omitted from the
equation.  The emission factor was calculated by solving for
source strength and dividing by the rate at which vehicles pass.

     An effort was made to measure particle fallout rates.
Concentrations at a series of downwind distances from the source
were measured, and corresponding emission rates at the source
were calculated using the above model, which assumes there is no
particle fallout.  Decreases in the "apparent emission rate" with
increasing distance from the source would serve as a measure of
fallout.

     Test Sites—Unpaved, uncontrolled haul roads at two western
surface coal mines were tested.
                               2-19

-------
     Parameters and Equipment—The parameters measured in this
study are listed in Table 2.6,  along with the tools used to
collect the data.

   TABLE 2.6.  PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
Parameters
Upwind concentration of TSP
Downwind concentration of TSP
Wind direction
Wind speed
Other atmospheric stability parameters
Particle size distribution
Equipment
Standard hi-vol
Standard hi-vol
Recording wind
Recording wind
anemometer
instrument
instrument, hand-held
Unspecified
Millipore filters on nuclepore filter
holder and pump, microscope
     Equipment Confiquration--For each sampling run at the first
mine, downwind hi-vols were set up at three downwind distances,
10, 20, and 30 meters;  for two of the runs an additional sampler
was operated 40 meters downwind.  These all collected air 1.2
meters above the ground.   A second hi-vol was placed at each of
the first three downwind distances 2.4 meters above the ground.
This configuration enabled the measurement of changes in
concentration with downwind distance and height.  Typically, a
pair of hi-vols were placed together at a site upwind from the
entire mine.  This was preferred over placing samplers
immediately upwind from the road because anticipated brief wind
direction reversals would be less likely to effect upwind
concentration measurements.  Other hi-vols were placed at 10, 20,
and 30 meters  (or a similar series of distances) downwind from a
particular source activity.  For about half of the tests,
downwind samplers were set up at both 1.2 and 2.4 meters above
the ground at these downwind locations to provide information on
the vertical dispersion of the plume.

     Sampling Runs—Four sampling runs were conducted on
uncontrolled haul roads at one mine; three runs were conducted at
the other.  The number of vehicle passes per run ranged from 32
to 46.  During each sampling run, which lasted between 45 and 60
minutes, four or six downwind concentration measurements were
taken, as described above.

     Quality Assurance—The guidelines in the Quality Assurance
Handbook for Air Pollution Measurement Systems  (EPA, 1976) were
followed in preparing filters, collecting and analyzing samples,
and auditing the data.  Hi-vol samplers were calibrated before
field work began at each mine.  One of every 25 filters was
treated as a blank.
                               2-20

-------
     Findings—The measured concentrations and other field
sampling data are shown in Table 2.7.  As was the case with the
other source activities in the mines, the apparent emission rate
from haul roads did not generally decrease with downwind
distance.

     For those sampling runs which included concentration
measurements at two consecutive downwind distances at the same
height, the modeled apparent emission rates decreased with
increasing distance in only about 35% of the sampling runs.  For
those tests in which concentrations were measured at downwind
distances differing by 10 meters, the concentration increased an
average of 19% between the two sampling points.

     When measured at two heights, 1.4 meters and 2.4 meters, at
the same downwind distance, the concentrations at the 'lower
height averaged 14% higher than those at the samplers 2.4 meters
above the ground.

     Publication—This study was conducted and documented under a
contract for the Environmental Protection Agency, Region VIII,
Office of Energy Activities, Denver, Colorado.  It was published
in February of 1978 as Publication No. EPA-908/1-78-003.

 TABLE 2.7.   CONCENTRATION MEASUREMENTS  AND ASSOCIATED FIELD  DATA
Mine/
Sample
B/1
B/2
B/3
Wind
Speed
(m/sec)
3.7
3.7
4.7
Stability
Class
C
C
C
Background
Concentration
(Hg/m3)
152
152
152
Net Plume
Concentration
(ng/m3)
1920
1829
1670 '
1798
1767
2034
1504
1773
1584
1156
3022
2321
1720
1617
Downwind
Distance
(m)
10
20
30
40
10
20
30
10
20
30
10
20
30
40
Height
(m)
1 .2
1 .2
1 .2
1 .2
1 .2
1 .2
1 .2
2.4
2.4
2.4
1 .2
1 .2
1 .2
1 .2
Apparent
Emission
rate
(Ib/veh-
mile)
11.5
12.7
14.0
18.0
10.6
14.2
12.6
8.5
9.9
8.7
21 .0
18.8
16.7
18.8
                               2-21
(continued)

-------
                       TABLE  2.7   (continued)
Mine/
Sample
B/4
E/1
E/2
E/3
Wind
Speed
(m/sec)
4.7
3.7
3.7
3.1
Stability
Class
C
B
B
B
Background
Concentration
(|ig/m3)
152
105
105
145
Net Plume
Concentration
(ng/m3)
3034
2516
1689
2336
1681
1318
693
794
748
427
575
571
500
877
619
509
727
788
1051
548
71 1
510
Downwind
Distance
(m)
10
20
30
10
20
30
8
17
26
8
17
26
8
17
26
34
8
17
26
8
17
26
Height
(m)
1 .2
1 .2
1 .2
2.4
2.4
2.4
1 .2
1 .2
1 .2
2.4
2.4
2.4
1 .2
1 .2
1 .2
1 .2
1 .2
1 .2
1 .2
2.4
2.4
2.4
Apparent
Emission
rate
(Ib/veh-
mile)
21 . 1
20.4
16.5
13.0
12.2
11 .6
2.3
3.7
4.4
1 .6
2.7
3.4
1 .7
4.1
3.7
3.7
1 .9
2.8
4.7
1 .6
2.5
2.3
Study 6— Bonn et al.   Fugitive Emissions from Integrated Iron and
          Steel Plants.   EPA-600/2-78-050.   1978.

     Methodology--Exposure profiling,  a technique  originally developed
by the Midwest Research Institute (MRI)  in 1974, was the basic method
used to measure emissions of particulates from vehicle travel on
unpaved roads within integrated iron and steel plants.

     Test Sites—Two plants provided sites for this investigation.
One plant in the dry western U.S. was used for measuring emissions
from vehicle travel on a road with a fine slag cover.  The other plant
was in the Great Lakes steel-producing area of the U.S.  It was used
for testing both paved roads and roads with a "hard-base" dirt cover.

     Parameters and Equipment--Table 2.8 provides  a list of parameters
documented in the study and the corresponding equipment used to take
the measurements.
                                 2-22

-------
     Equipment Configuration—In each sampling run the exposure
profiler was set up 5 meters from the edge of the unpaved road.  For
sites with light-duty traffic, sampling intakes were positioned at 1,
2, 3, and 4 meters above the ground.  For heavy-duty traffic, they
were set at 1.5, 3, 4.5, and 6 meters above the ground.  A standard
hi-vol sampler and a hi-vol cascade impactor with cyclone preseparator
were placed beside the profiler at approximately 2 meters above the
ground.  The wind instrument was located 3 meters downwind from the
edge of the road.  Dustfall buckets were placed at 1  and 3 meters from
the road's edge.  Another standard hi-vol sampler was placed five
meters upwind from the road.

     Sampling Runs—A total of nine sampling runs, were conducted on
unpaved roads, three on a road with fine slag cover and six on roads
with a hard-based dirt cover.  Each run lasted from 12 to 55 minutes.
The investigators recorded 319 vehicle passes while sampling plumes on
unpaved roads.

      TABLE 2.8.   PARAMETERS MEASURED AND  CORRESPONDING EQUIPMENT
Parameter
Wind speed
Wind direction
Cloud cover
Temperature
Relative humidity
Road surface condition
Dust loading
Dust texture
Traffic mix
Traffic count
Plume exposure
Plume particle size distribution
Downwind concentration
Background concentration
Duration of sampling
Deposition
Equipment
Anemometer
Anemometer
Direct observation
Sling psychrometer
Sling psychrometer
Direct observation
Dry vacuuming, scales
Sieves, standard shaker, scales
Direct observation
Automatic counters
Isokinetic hi-vol samplers on
profiler tower
Hi-vol cascade impactor with cyclone
preseparator
Hi-vol sampler
Hi-vol sampler
Timer
Dustfall buckets
     Quality Assurance—The extent to which methodological procedures
were documented in this report is sufficient to permit their
reproduction by other investigators.  The manner in which exposed
                                 2-23

-------
filters were collected and transported to the laboratory was described
in some detail.  However, documentation of specific quality assurance
procedures, such as auditing filter weights, processing of blank
filters, or calibration of sampling equipment, was generally not
provided.

     Findings—Exposure data and calculated emission factors for
suspended particulates  (less than 30 [o.m diameter) are presented in
Table 2.9.

     The predictive equation which was published in this study and
which is based on the cumulative database gathered by Midwest Research
Institute is shown below:

                 e = 5.9 (s/12)  (S/30)  (W/3)0-8 (d/365)
where

     e    =    suspended particulate emissions, Ib/veh-mile
     s    =    road surface silt content, %
     S    =    average vehicle speed, mph
     W    =    average vehicle weight, tons
     d    =    dry days per year

     Publication—This study was conducted and documented under a
contract with  the Environmental Protection Agency; Industrial Research
Laboratory; Office of Energy, Minerals, and Industry; Research
Triangle Park, North Carolina.  It was published as Publication No.
EPA-600/2-78-050 in March 1978.
                                  2-24

-------
TABLE 2.9. EXPOSURE DATA AND EMISSION FACTORS
Run
A- 7
A- 14
A-15
E-1
E-2
E-3
E-4
Sample
Height
(m)
1
2
3
4
1 .5
3
4.5
6
1 .5
3
4.5
6
1 .5
3
4.5
6
1 .5
3
4.5
6
1 .5
3
4.5
6
1
2
3
4
Filter
Exposure
(mg/cm2)
5.34
2.9
1 .54
0.28
17.9
6.33
5.11
1 .39
12.5
6.78
5.91
2.97
4.53
3.67
2.33
1 .24
4.43
3.16
2.92
1 .79
5.76
3.07
1 .70
0.95
4.24
2.94
1 .80
0.86
Integrated
Filter Exposure
(Ib/mile)
5.6
16
16
18
19
16
7.7
Emission Factor"
(Ib/veh-mile)
db < 30
4.9
27
29
17
16
19
13
                     2-25
(continued)

-------
                     TABLE 2.9.   (continued)
Run




E-5




E-6


Sample
Height
(m)
1
2

3
4
1
2

3
4
Filter
Exposure
(mg/cm2)
5.70
3.42

1 .82
0.68
8.15
2.25

2.47
0.76
Integrated
Filter Exposure
(Ib/mile)


1 1




14.2


Emission Factor"
(Ib/veh-mile)
db < 30


1 1




1 9


     a)  isokinetic
     b)  particle Stokes diameter
Study 7— Cowherd et al.  Iron and Steel Plant  Open  Source
          Fugitive Emission Evaluation.  EPA-600/2-79-103.
1979
     Methodology—  Exposure profiling was the primary technique
used to estimate emissions of fugitive dust  from unpaved roads.
For a full description of this methodology,  see the  summary of
Cowherd et al.,  1974.

     Test Sites—Tests were conducted at two sites in
unidentified iron and steel plants, one with a crushed slag road
and the other with a mixed dirt-slag road.

Parameters and Equipment—

     Table 2.10 lists the parameters measured and the  equipment
used to measure them.
                               2-26

-------
   Table 2.10.   Parameters Measured  and  Corresponding  Equipment
Parameters
wind speed
Wind direction
Cloud cover
Temperature
Relative humidity
Road surface condition
Road dust loading
Road dust % silt
Traffic mix
Traffic count
Exposure
Particle size distribution
Downwind concentration
Upwind concentration
Duration of sampling
Equipment
Recording anemometers
Unspecified
Direct observation
Sling psychrometer
Sling psychrometer
Direct observation
Vacuum/broom
200-mesh screen, shaker
Direct observation
Automatic counter
Isokinetic hi-vol filtration
(profiler)
Hi-vol cascade impactor &
cyclone preseparator
(directional)
Hi-vol sampler
Hi-vol sampler
Timer
     Equipment Configuration—The exposure profiler was
positioned 5 meters from the downwind edge of the road.  Four
sampler intakes were positioned at heights of 1.5, 3, 4.5,  and 6
meters.  A cascade impactor and a standard hi-vol sampler were
placed beside the profiler two meters above the ground.  For
three of the tests, a second downwind standard hi-vol sampler was
stationed about 17 meters from the edge of the road 2 meters
above the ground.  Wind measurements were taken between 7 and 14
meters upwind from the upwind edge of the road at two unspecified
heights.  A standard hi-vol sampler was also placed beside the
wind station with its intake at a height of 2 meters.

     Sampling Runs—Nine sampling runs were made on uncontrolled,
unpaved roads.  Three were on a mixed dirt-crushed slag surface,
and 6 were on a crushed slag surface.  The number of vehicle
passes per sampling run ranged from 40 to 74.  The total number
of vehicle passes was 533.

     Quality Assurance—An effort was made to apply or adapt the
American Society of Testing and Materials (ASTM) Standards in the
collection and analysis of road surface samples needed to
quantify the silt content of the surface material.  Except for
                               2-27

-------
the citation of this standard procedure, the authors documented
no normal quality assurance procedures.  However, documentation
of the basic methodological procedures in field operations,
sample handling, and data analysis was generally thorough.

     Dust samples were transported to the laboratory in
individual envelopes.  Filter samples were conditioned at
constant temperature and humidity for 24 hours before weighing.
This same procedure was followed in weighing the filters prior to
use.

     Careful attention was given to sampling under isokinetic
conditions.  The intake velocity of each sampler was set to match
the wind velocity prior to commencement of sampling.  Wind speed
was measured continuously during sample collection.  Isokinetic
correction factors were used to adjust exposures measured under
non-isokinetic conditions.

     An effort was also made to reduce small particle bias in
characterizing the particle size distribution of the dust plume.
One of the high volume cascade impactors was fitted with a
cyclone preseparator to reduce small particle bias caused by
large particles bouncing through the impactor stages to the back-
up filter.  The impactor/preseparator unit was calibrated to
determine the 50% cutoff diameters of the cyclone inlet and the
stages of the preseparator.  The investigators found that the
preseparator does eliminate much, but not all, of the small
particle bias.

     The procedure for accounting for background levels of dust
was not adequately documented.  Therefore, the emission factors
presented in the study could not be reproduced from the data
provided.

     Findings—This study evaluated emissions from three source
categories in iron and steel plants: vehicle travel on unpaved
roads, vehicle travel on paved roads, and storage pile stacking.
Data collected on unpaved road emissions, shown in Table 2.11,
was added to MRI's database of field collected emissions data!
and, using this larger database, the predictive emission factor
equation for vehicle travel on unpaved roads was revised.  A new
variable, the average number of wheels per vehicle traveling the
road segment, was added to the equation.  The addition of this
predictive variable is indicative of the fact that the emissions
database collected by MRI was expanded under this study to
include a wider range of test conditions.  The revised equation
is presented below.


         e =  5.9  (s/12)  (S/48)  (W/2.7)0-7  (w/4)°-5 (d/365)
                               2-28

-------
where
     s
     S
     w
     w
     d
Emission factor for particles < 30 |j.m in diameter,
pounds per vehicle-mile
Silt content, %
Average vehicle speed, miles per hour
Average weight per vehicle, tons
Average number of wheels per vehicle
Dry days per year
     Publication—This study was conducted and documented under a
contract for EPA, Industrial Environmental Research Laboratory,
Office of Energy, Minerals, and Industry, Research Triangle Park,
North Carolina.  It was published in 1979 as Publication No. EPA-
600/2-79-103.
         TABLE 2.11.  EXPOSURE DATA AND EMISSION FACTORS
Run
F91

F-) ~)

F*) o
^ J
G-27

GO fl
zo

Sampling
Height
(m)
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
Filter
Exposure
(mg/cm2)
0.79
0.39
0.17
0.18
0.60
0.47
0.24
0.17
1 .01
0.72
0.45
0.34
4.69
5.12
NA
1 .07
2.26
1 .60
NA
0.77
Integrated
Filter Exposure
(Ib/mile)
123
98.3
168
901
391
Emission Factor3
(Ib/veh-mile)
db < 30 ^m
3.0
1 .7
2.3
12.0
7.2
db < 5 |im
1 .0
0.53
0.75
3.1
2.4
                               2-29
                                        (continued)

-------
                    TABLE 2.11.   (continued)
Run

G-) Q
t.?
G-30
GO -1
J 1
Go •}

Sampling
Height
(m)
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
Filter
Exposure
(rag/cm2)
3.37
1 .83
NA
0.05
1 .98
1 .64
0.91
0.42
1 .71
0.96
1 .01
0.26
5.06
3.68
NA
1 .39
Integrated
Filter Exposure
(Ib/mile)
475
317
244
845
Emission Factor*
(Ib/veh-mile)
db < 30 (im
5.6
8.7
5.1
16.0
db < 5 (im
1 .2
1 .5
1 .1
3.4
  a)  isokinetically corrected
  b)  particle Stokes' diameter

Study 8-- Cuscino.  Taconite Mining Fugitive Emissions Study.
          1979.

     Methodology—Exposure profiling was used to measure
particulate emissions from haul trucks traveling on unpaved  roads
at a taconite mine.  As in the original exposure profiling study
(Cowherd et al. , 1974), point values of exposure were calculated
as the dust mass per unit area of the sampler intake.

     Test Sites—Tests were conducted on two different road
surface types  (sand/gravel and crushed rock) in the Erie Mining
Company's operation in the Mesabi Iron range in Minnesota.

     Parameters and Equipment—Several different pieces of
equipment were employed in measuring plume dust.  Exposure was
measured using a profiler with four sampling heads.  Air was
drawn in isokinetically and routed up through the horizontal
filter.  The plume's particle size distribution was measured
using a high volume cascade impactor with a cyclone preseparator
(to remove very large particles which tend to bounce through the
impaction substrates and cause fine particle measurement bias).
                               2-30

-------
Standard hi-vol samplers measured total suspended particulate
concentration at several downwind distances.

     The number and type of vehicles passing during each sampling
run were also documented.  Data on many test site conditions were
also collected:  wind direction and speed, temperature, cloud
cover, humidity, recent rainfall history, road surface material
density, and silt content.  The tools or methods of measuring
these parameters was not documented.

     Equipment Configuration—The high volume cascade impactor,
the profiler, and a standard high volume sampler were all set up
five meters from the downwind edge of the road.  The profiler had
sampling heads at four heights: 1.5, 3, 4.5, and 6 meters.
Anemometers were attached to the tower at heights of 1.5 and 4
meters.  Two other standard hi-vol samplers were placed 20 and 50
meters from the downwind edge of the road.  Wind speed and
direction were also measured at two weather stations, one five
meters upwind and the other 50 meters downwind.  The measuring
height of these weather stations was not documented.

     Sampling Runs—Eight sampling runs, ranging in duration from
29 to 68 minutes, were conducted in which emissions from the road
surface were not controlled.  Sampling runs generally consisted
of 15 haul trucks passing the sampling site where the fugitive
emissions were measured.  For two of the sampling runs, traffic
consisted partially of pick-up trucks.  Emissions from a total of
131 vehicle passes were measured.

     Quality Assurance—In collecting and analyzing road surface
samples for silt and moisture content, relevant standards set by
the American Society of Testing and Materials were followed with
some adaptation due to feasibility constraints.  The methods were
set forth in detail in an appendix.

     To the extent practical, exposure samples were collected
isokinetically.  Otherwise, correction factors were used to
adjust the sample to isokinetic conditions.

     The following relevant quality assurance procedures were not
documented for this study:  calibration of air samplers,
processing of blank filters, auditing of filter weights, and
collocation of samplers.

     It should also be noted that the haul road surfaced with
sand/gravel was not in use prior to emissions testing.  The
investigators judged that the process of fine dust  formation and
subsequent emission into the atmosphere did not reach  an
equilibrium state until after the first two sampling runs  (30
vehicle passes).  Furthermore, three other tests were  conducted
on the day after a heavy rain.


                               2-31

-------
     Findings—The measured exposures and corresponding emission
factors are shown in Table 2.12.  A regression analysis was
conducted on the data from this study and previous Midwest
Research Institute studies of unpaved road emissions (Cowherd e_t
al. ,  1974; Bohn, et al.,  1978; and Cowherd et al.,  1979).  Runs
1-1,  1-2, 1-6, 1-7, and 1-8, were excluded as they were believed
to be not representative of dry, uncontrolled, unpaved roads  (see
discussion in Quality Assurance).  The following equation was
developed using this analysis:

 TABLE 2.12.  MEASURED EXPOSURES AND CALCULATED EMISSION FACTORS
Run


I-1b

I-2b
1-3
1-4
1-5
I-6C
Sample
Height
(m)
1 .5
3.0
1 .5
6.0
1 .5
3.0
1 .5
6.0
1 .0
3.0
1 .5
6.0
1 .5
3.0
1 .5
6.0
1 .5
3.0
1 .5
6.0
1 .5
3.0
1 .5
6.0
Filter
Exposure
(mg/cm2)
0.76
0.51
0.45
0.32
0.65
0.55
0.54
0.46
1 .61
1.18
0.91
0.54
2.81
2.35
1 .60
0.76
3.74
2.72
2.55
1 .64
0.19
0.18
0.26
0.22
Integrated Filter
Exposure (Ib/mile)
138
153
285
501
742.5

69

Emission Factor"
(Ib/veh-mile)
9.2
10.2
19.0
33.4
49.5
2.3
                               2-32
(continued)

-------
                     TABLE 2.12.   (continued)
Run
1-7°
I-8C
Sample
Height
(m)
1 .5
3.0
1 .5
6.0
1 .5
3.0
1 .5
6.0
Filter
Exposure
(mg/cm2)
3.56
2.45
1 .76
0.98
1 .70
1.19
0.68
0.31
Integrated Filter
Exposure (Ib/mile)
582
260.7
Emission Factor3
(Ib/veh-mile)
38.8
23.7
     a) for particles with aerodynamic diameter < 30 jim, before isokinetic
     correction
     b) conducted on previously inactive road (see discussion in Quality
     Assurance)
     c) conducted on day after heavy rains
where
     EF

     S
     S
     W
     w
                     EF  =  0.00380 s S W0-7
Emission factor  for  particles with Stoke's
diameter < 30 |-im,  Ib/veh-mile
Silt content, %
Average vehicle  speed,  mph
Average vehicle  weight,  tons
Average number of  wheels per vehicle
     Publication—This study was conducted for  the  Minnesota
Pollution Control Agency,  Division of Air Quality,  Roseville,
Minnesota.   It  was  apparently never published.

Study 9— Reider.   Size Specific Particulate  Emission Factors for
          Uncontrolled Industrial and Rural Roads.   January 1983.

     Methodology--Exposure profiling was used to  measure
emissions from  unpaved roads in both rural and  industrial
settings.  An explanation of the specific procedures for
calculating  emission factors from the field data  was not provided
in the report.

     Test Sites—In addition to public rural  roads, tests were
conducted on unpaved roads in facilities of three different
industries:  stone  crushing, sand and gravel  processing, and
                                2-33

-------
copper smelting.  All of the industrial roads were two-lane,
whereas some of the public, rural roads were one-lane and others
were two-lane.  For the industrial roads, test sites were
selected according to three general criteria.  Each test site
needed to be: 1) suitable for the specific requirements of the
exposure profiling methodology, 2) representative of most
facilities in the industry, and 3) accessible via cooperation of
the facility personnel.

     Parameters and Equipment--The parameters measured in this
study and the equipment used to take these measurements are shown
in Table 2.13.

     The procedures for calculating point values of exposure from
these data were not included in this report.  They are presumably
similar to those discussed in Cowherd and Englehart, 1984, which
is reviewed in the paved road section.

     Equipment Configuration—Downwind air sampling equipment was
set up five meters from the edge of the road.  The exposure
profiler had sampling heads at heights of 1, 2, 3, 4, and 5
meters.  The wind speed is measured continuously at two sampling
heights and a logarithmic distribution of the vertical wind speed
profile is assumed in setting the sampling rates for the
remaining sampler heads.  A standard hi-vol sampler, a hi-vol
fitted with a 15 p.m size-selective inlet  (SSI), and two cascade
impactors with cyclone preseparators were also set up on the
downwind side.  The cascade impactors were positioned in a
vertical array at heights of 1 and 3 meters.  The hi-vols sampled
air at a height of 2 meters.

     Upwind sampling equipment was also generally set up five
meters from the road.  For all unpaved road tests, one of each of
the following samplers was set up with intakes two meters above
the ground:  standard hi-vol, hi-vol with SSI, and cascade
impactor with cyclone preseparator.
                               2-34

-------
   TABLE  2.13.   PARAMETERS  MEASURED  AND  CORRESPONDING EQUIPMENT
Parameter
Wind speed
Wind direction
Atmospheric pressure
Temperature
Relative humidity
Road surface condition
Road surface particulate loading
Road surface silt content
Traffic Mix
Traffic Count
Vehicle weight
Vehicle speed
Plume total particulate concentration
Plume inhalable particulate
concentration
Size distribution of inhalable
particulate
Equipment
Warm wire anemometer
Wind vane
Barometer
Sling psychrometer
Sling psychrometer
Direct observation
Dust pan, broom, scales
Sieves, scales
Direct observation
Direct observation
Interview plant operators
Interview drivers
Exposure profiler
Hi-vol with size-selective inlet
Cascade impactor with cyclone
preseparator (impactor cutpoints of
10.2, 4.2, 2.1, 1.4, and 0.73 urn)
     Sampling Runs—A total of 21 tests, ranging in duration
between 12 and 203 minutes, were conducted on unpaved roads for
this study.  Sampling run time was sufficient to produce a filter
weight gain of at least 5 mg on the top sampling head of the
profiler.  Tests were distributed among the various industries as
shown in Table 2.14.

     TABLE 2.14.  EXTENT  OF SAMPLING  FOR VARIOUS  INDUSTRIES
Industry
Stone crushing
Sand and gravel processing
Copper smelting
Public roads
crushed limestone
dirt
gravel
Number of
Tests
5
3
3
6
4
2
Total Vehicle
passes
225
80
151
591
244
68
Traffic Type
Medium Duty
Heavy Duty
Light Duty
Light Duty
Light Duty
Light Duty
                               2-35

-------
     Quality Assurance—Except for the absence of an explanation
of the manner in which values of concentration of IP and PM10 at
each profile height were calculated, quality assurance for this
study was good.  Calibration of the profiler, hi-vols, and
impactors was performed prior to testing at each site.  Sampling
filters and impactor substrates were equilibrated for 24 hours
prior to weighing.  Tare weights were given a 100% audit, and 10%
of loaded weights were audited.  Criteria for reweighing of the
entire batch were provided.  The orientation of the profile
sampling heads were adjusted if the 15-minute average wind
direction changed by more than 30 degrees.  The sampling rate of
the profile samplers was adjusted if the  15-minute average wind
speed changed by more than 20%.  Ten percent of all calculations
were also audited.

     Findings—The concentration measurements taken during the
field tests are presented in Table 2.15.  The investigators noted
that the data from the profiler is net of background dust levels.
The emission factors which were calculated from these data are
shown in Table 2.16.  Again, the specific procedures for
calculating the emission factors are not described in this
report.

     Publication—This report has been completed only as a draft
final report.  It has not been published.  However, it was cited
in the AP-42 section on unpaved roads.  The study was conducted
under EPA Contract No. 68-02-3158.
                               2-36

-------
TABLE 2.15.   MEASURED DUST  CONCENTRATIONS  FOR TOTAL PARTICULATES
                                AND PM,n

Run
U-1
U-2
U-3
U-4
U-5
U-6
AA-1
AA-2
AA-3
AA-4
AA-5
AB-1
AB-2
AB-3
AB-4
AC-1
AC-2
AC- 3
AE-1
AE-2
AF-1
AF-2
AF-3
Duration
(minutes)
211
79
119
196
240
147
65
58
96
77
73
173
266
266
162
143
143
50
295
295
153
193
227

	 — 	 Conce
ntration (|ig/m3
Total Particulate
1 m
56890
32040
54730
31340
17740
5680
15540
20220
3695
14290
12280
45410
121500
54580
15620
11130 •
6500
7802
4288
a
2487
1338
1290
2 m
27600
16080
34380
11380
8895
2876
10290
10320
1693
9809
8463
15710
17800
16090
6253
6534
4912
3828
2491
2193
a
a
a
3 m
10230
6650
18370
5503
4324
918
8163
4437
1081
11510
7239
6766
5350
9700
a
4348
3234
3525
1189
793
1026
826
1080
4 m
7720
1660
14940
4440
1566
182
6964
2003
556
7759
5600
3895
1167
2943
890
2422
2276
1668
b
141
1138
506
863
5 m
2180
810
7975
1400
428
65
5307
305
400
5654
4070
1918
162
1354
175
1773
1431
785
355
c
c
318
675
\
) ~ 	
PM,0
1 m
6793
4640
1165
452
811
466
2313
212
346
4078
2753
2308
827
2125
2308
995
1156
1542
571
502
130
317
602
3 m
2903
1803
766
424
365
38
1347
293
146
2244
1675
817
100
357
625
581
317
546
138
299
146
131
346
  a)  equipment malfunction or  failure
  b)  torn filter
  c)  net concentration resulted in negative
value
                                  2-37

-------
TABLE 2.16.  EMISSION FACTORS FOR TOTAL PARTICIPATES AND PM10
Run
U-1
U-2
U-3
U-4
U-5
U-6
AA-1
AA-2
AA-3
AA-4
AA-5
AB-1
AB-2
AB-3
AB-4
AC-1
AC-2
AC- 3
AE-1
AE-2
AF-1
AF-2
AF-3
Total Particulate
Emission Factor
44.8
17.9
20.6
27.0
22.0
14.1
9.36
15.3
4.83
35.2
30.3
112.6
42.1
32.5
11.1
9.36
7.62
10.0
5.43
7.96
15.3
9.80
8.28
PM,0 Emission
Factor
9.13
3.09
1 .75
1 .87
1 .97
1 .77
2.15
0.943
0.903
4.52
5.83
12.1
0.951
1 .99
1 .86
1 .63
1 .46
1 .91
0.713
0.957
2.60
2.34
3.26
                              2-38

-------
Study 10 —
Cuscino et al.  Iron and Steel Plant Open Source
Fugitive Emission Control Evaluation.  EPA-600/2-
83-110.  1983.
     Methodology—Exposure profiling was used to measure
emissions of particulate dust from vehicle travel on unpaved
roads in iron and steel plants.  The investigators made no
changes in the basic methodology developed and applied in 1974
(see summary of Cowherd et al., 1974).

     Test Sites—Tests of emissions from vehicular travel on
unpaved, uncontrolled roads  (i.e. not sprayed with a dust
suppressant) were conducted at two different sites at Armco
Steel, Incorporated's iron and steel works plant in Middletown,
Ohio.  Both sites were gravel roads.  One was used primarily by
heavy-duty vehicles, and the other was predominantly light-duty
traffic.

     Parameters and Equipment—Listed in Table 2.17 are the
parameters measured and the equipment used to take the
measurements in this study.

   TABLE 2.17.   PARAMETERS  MEASURED AND  CORRESPONDING  EQUIPMENT
Parameter
Wind speed
Wind direction
Temperature
Road surface silt content
Road surface moisture content
Traffic count
Plume exposure
Plume particle size distribution
Plume TSP concentration
Plume IP concentration
TSP background concentration
Duration of sampling
Meteorology
Recording anemometer
Recording anemometer
Unspecified
sieves, mechanical sieving
device, scales
Oven, scales
Direct observation
Isokinetic hi-vol samplers
(profiler)
Hi-vol cascade impactor with cyclone
preseparator
Standard Hi-vol sampler
Hi-vol sampler with size selective inlet
Standard Hi-vol sampler
Timer
                               2-39

-------
     Equipment  Configuration—

     Light-Duty Unpaved  Road—The  exposure  profiler was located 5
     meters  from the  downwind edge of  the light-duty unpaved
     road.   Sampling  intakes were  at heights of  1,  2,  3,  and 4
     meters.   Several other  instruments were positioned beside
     the profiler:  a standard  high-volume  sampler  at 2 meters
     above the  ground, a high-volume sampler fitted with a size-
     selective  inlet  (designed  to  collect only particles less
     than 15 jim in aerodynamic  diameter)  at 2 meters height, and
     two high-volume  samplers fitted with a cascade impactor and
     cyclone preseparator at 1  and 3 meters above the ground.

     Three instruments were  located 10 meters upwind from the
     upwind  side of the  road:  a high-volume sampler with a 15 (am
     size-selective inlet 3  meters high,  and two standard high-
     volume  samplers, one 3  meters high and the  other 1 meter
     high.

     Heavy-Duty Unpaved  Road—The  equipment for  the heavy-duty
     road site was configured the  same as for the light-duty road
     with two exceptions:  the  profiler had a fifth sampler
     intake  at 5 meters  height,  and there was no high volume
     sampler equipped with a size-selective inlet on the downwind
     side of the road.

     Sampling Runs—Seven sampling runs were conducted on
uncontrolled, unpaved roads  in  this study.   Four were on light-
duty roads,  and three were on heavy-duty roads.   Sampling run
time ranged  from 13 to 45 minutes, and the  number of vehicle
passes per run ranged from 10 to 101.   The  total number of passes
was 276.

     Quality Assurance--This study incorporated  a rigorous
quality control program.  Procedures followed in collecting and
analyzing samples were documented  in considerable detail.
Quality control measures were set  forth for the  sampling media,
sampling flow rates,  and sampling  equipment  (proper performance).
Criteria for interrupting sample collection were also documented.
Quality assurance practices included processing  of blank samples,
calibration of equipment, and auditing of sampling and analysis
procedures.

     The investigators note that their procedures met or
surpassed the requirements set  forth in the Quality Assurance
Handbook for Air Pollution Measurement Systems,  Volume II -
Ambient Air  Specific  Methods (U.S. EPA, 1977) and Ambient
Monitoring Guidelines for Prevention of Significant Deterioration
(U.S. EPA,  1978) .

     Careful attention was given to sampling under isokinetic
conditions.   Wind speed was monitored before and during sample

                               2-40

-------
collection.  Fifteen minute averages of the wind speed at two
monitoring heights and an assumed logarithmic vertical wind speed
profile were used to set the intake velocity for each sampler on
the profiler.

Findings—Because the purpose of this study was to evaluate
various methods of controlling open source fugitive emissions
from iron and steel plants, some experimental data was needed on
particulate emissions from vehicle travel on uncontrolled,
unpaved roads.  Table 2.18 presents the primary emissions data
collected for uncontrolled, unpaved roads in this study.

     The investigators did not estimate a new predictive equation
on the basis of their empirical findings.  However, they did
apply an equation estimated previously by MRI:

           e  =  5.9  (s/12)  (S/30)  (W/3)°-7  (w/4)°-5 (d/365)

where

     e    =    Mass of particulates < 30 |o.m in diameter Ib/veh-
               mile
     s    =    Silt content of road surface material %
     S    =    Average vehicle speed mph
     W    =    Average vehicle weight tons
     w    =    Average number of vehicle wheels
     d    =    Number of dry days per year

     The investigators interpolated from the three size
categories shown in Table 2.18 (assuming a log-normal
distribution of particle sizes) to estimate an emission factor
for particles less than 30 urn in diameter for each of the
sampling runs.  The ratio of predicted to actual (interpolated)
emission factors ranged from .34 to 1.21 for the 7 sampling runs
on uncontrolled, unpaved roads.

     Publication—This study was .conducted and documented for
EPA,  Office of Research and Development, Industrial Environmental
Research Laboratory, Research Triangle Park, North Carolina.  It
was published in 1983 as Publication No. EPA-600/2-83-110.
                               2-41

-------
     TABLE  2.18.   EXPOSURE  DATA AND  EMISSION FACTORS
Run
F-28
F-29


FT 1

F-68
F-69
F-70
Traffic
rpynp

T -i rrVt f
Duty


T -i rrVrh
Duty
Light
Duty
Light
Duty
Heavy
Duty
Heavy
Duty
Heavy
Duty
Sample
Height
fm)
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Net TPa
Exposure
( mn /cm^ )
3.52
3.58
1 .66
0.77
5.20
4.74
3.56
2.67
4.20
3.77
2.76 ,
1 .29
3.01
3.13
1 .81
0.92
12.0
15.3
14.6
12.7
9.6
10.7
10.5
10.8
6.82
4.44
8.60
7.52
6.00
5.76
3.63
Emission Factors11 (Ib/veh-mile)
TPa
10.7
14.2
9.98
12.4
129
133
133
dc < 1 5 urn
1 .05
4.25
2.99
3.90
33'. 5
25.9
32.8
dc < 2.5 nm
0.245
1 .27
0.898
1 .02
7.74
8.84
8.52
a)  total particulate,  i.e.  including mass in settling chamber
b)  isokinetically corrected
c)  particle aerodynamic diameter
                               2-42

-------
 Study  11—      Axetell  and  Cowherd.   Improved  Emission Factors
                for  Fugitive Dust  from Western  Surface  Coal  Mining
                Sources.  EPA-600/7-84-048.   1984.

     Methodology—This  study was  conducted  to  develop  emission
 factors  for  the various  activities at surface  coal mining
 operations in the Western U.S.  The review  will be limited  to
 summarizing  and discussing  research on uncontrolled emissions
 from vehicle travel on  unpaved roads.   Emission factors were
 developed for .four  particle size  categories:   total particulates,
 particles with aerodynamic  diameters  less than 2.5 urn,  those
 having diameters less than  15 urn, and total suspended
 particulates.

     Two different  testing  set-ups were used in measuring
 emissions from this source.  In the first,  exposure profiling
 alone  was employed.  For the second,  both the  profiling method
 and  the  upwind-downwind  method were used to determine  the
 comparability of the findings from these two disparate
 techniques.   The'standard general equation  for continuously
 emitting, infinite  line  sources  (Turner, 1970) was used to
 estimate source strength from measured concentrations.

     Three different tools  for measuring the particle  size
 distribution were also employed,  evaluated, and compared in this
 second testing configuration:  cascade impaction, dichotomous
 sampling, and microscopic examination of exposed millipore
 filters.  Distribution data from  the  cascade impactors was
 corrected to adjust  for  the particle  bounce problem (see reviews
 of Cowherd et  al..  1974  and Cowherd et al., 1979).  The effect of
 greasing the impaction substrates on  particle  bounce was also
 tested by operating impactors with grease for  some tests and
 without  grease  for  others.  The investigators  assumed  that the
 true particle  size  distribution was lognormal  and adjusted their
 experimental data to fit this form.   The data  generated by the
 dichotomous  samplers was corrected based on findings by Wedding
 (1980) which indicate that  the collection efficiency of the
 sampler  depends  on wind  speed.  The particle size measurements
 taken  using  microscopy were in terms  of physical diameters.  The
 investigators assumed an average particle density of 2.5 g/cm3
 and followed a  procedure for converting these  into aerodynamic
 diameters (U. S. Environmental Protection Agency, 1978).

     Test Sites—Tests were conducted at three unspecified mines,
 one in each  of  the following Western  coal fields:  Powder River
Basin,  Fort  Union,  and San  Juan River.  These  three fields were
targeted for the study because they produce high volumes of
strip-mined  coal and because they are  diverse  in character.  The
investigator's  intent with  this selection was  to maximize the
representativeness of the findings while satisfying budget and
time constraints.
                               2-43

-------
     Parameters and Equipment—Table 2.19 lists the parameters
measured and the corresponding equipment employed during the
field investigation of haul road emissions.

     TABLE  2.19.   PARAMETERS MEASURED AND EQUIPMENT  EMPLOYED
Parameters
Surface silt content
Vehicle speed
Vehicle weight
Total surface loading
Surface moisture content
Number of wheels
Solar intensity
Atmospheric pressure
Cloud cover
Temperature
Humidity
Exposure (downwind)
Wind speed (upwind)
Wind speed (downwind)
Concentration of particles < 1 5 |im diameter
(upwind and downwind)
Concentration of particles < 2.5 [im diameter
(upwind and downwind)
Concentration of total suspended particulates
(upwind and downwind)
Particle size distribution (downwind)
Dust deposition
Equipment
Oven, sieves, scales
Radar gun or timer
Truck scale
Broom, scales
Oven, scales
Direct observation
Pyranograph
Barometer
Direct observation
Thermometer
Sling psychrometer
Exposure profiler
Continuous wind monitor
Warm wire anemometers
Dichotomous sampler
Dichotomous sampler
Standard hi-vol
1 ) Hi-vol with cascade
impactor with cyclone
preseparator
2) Dichotomous samplers
3) Optical microscope and
millipore filters
Dustfall buckets
     Wind  speed  was monitored at two heights, and a  logarithmic
vertical distribution of the wind speed was assumed.  The
investigators  noted that although dust exposure could be measured
directly only  with the profiler, it could also be determined
using  dichotomous samplers by multiplying concentration, wind
speed,  and sampling time.
                               2-44

-------
Equipment Configuration—

Exposure Profiling Only—For all tests except those in which
dispersion modeling and profiling techniques were used
simultaneously, the following equipment was placed between
five and ten meters downwind from the road:  exposure
profiler with sampling heads at 1.5, 3.0, 4.5, and 6 meters
above the ground; a standard hi-vol sampler and a hi-vol
fitted with a cascade impactor and cyclone preseparator,
both having inlets 2.5 meters high; two dichotomous samplers
with intakes 1.5 and 4.5 meters above the ground; two
dustfall buckets at a height of 0.75 meters; and two warm
wire anemometers at heights of 1.5 and 4.5 meters.  The
height of these instruments was reduced for cases in which
the dust plume height was relatively low  (e.g. for light-
and medium-duty trucks).  Pairs of dustfall buckets were
also collocated 0.75 meters above the ground at 20 and 50
meters downwind, permitting measurement of deposition rates.
The following equipment was set up five meters upwind from
the road:  one dichotomous sampler 2.5 meters above the
ground; one standard hi-vol sampler, also at a height of 2.5
meters; two dustfall buckets 0.75 meters above the ground;
and one continuous wind monitor 4 meters high.

Exposure Profiling and Upwind-Downwind—The equipment
configuration for those tests in which both exposure
profiling and upwind-downwind modeling were used was very
complex.  Downwind air sampling was conducted primarily at
three downwind distances: 5, 20, and 50 meters.  Profiling
towers were set up at each of these stations so plume mass
depletion could be measured in addition to simple exposure.
The closest tower consisted of four sampling heads at
heights ranging from 1.5 to 6 meters.  The towers at 20 and
30 meters downwind both had five sampling heads, with the
highest heads at 9 and 12 meters, respectively.  This was
necessary due to the increased dispersion of the plume over
longer distances.  A vertical array of dichotomous samplers
was also set up five meters downwind.  The sampling heights
matched those of the nearby exposure profiler.  Two single
dichotomous samplers were also set up on either side of the
profiling towers at each of the three downwind monitoring
distances.  They had intakes 2.5 meters above the ground.

Two standard high volume samplers  (all sampling at a height
of 2.5 meters) were set up at each of the downwind
distances: 5, 20, 50, and 100 meters.  A third standard hi-
vol was used five meters downwind.  A total of three hi-vol
cascade impactors were used, two were placed five meters
from the road with intakes 1.5 and 4.5 meters above the
ground, and one 20 meters from the road with its intake 2.5

                          2-45

-------
     meters high.  Dustfall buckets were placed in pairs at each
     of the three downwind distances such that their sampling
     height was 0.75 meters.

     Sampling Runs—Of a total of 29 sampling runs in which
uncontrolled road emissions were measured, 19 had traffic
dominated by haul trucks, and 10 were dominated by light- and
medium-duty trucks.  The number of vehicle passes per run was
generally between 20 and 150.  Exposure profiling was used for
all sampling runs; for five of the haul truck runs, the upwind-
downwind method was also used.  Profile and hi-vol samplers were
generally operated between 45 minutes and 1-5 hours for each run.

     Quality Assurance—This study included a thorough quality
assurance program, which was subject to evaluation by a technical
review group (including the two EPA project officers,
representatives of the Bureau of Land Management, the Bureau of
Mines, and the mining industry).  Profilers, hi-vols, impactors,
and dichotomous samplers were calibrated on a regular basis.
Sampling media were conditioned under constant temperature and
humidity prior to sampling.  Seven percent of tare and final
filter weights were audited.  For every ten regularly processed
filters and substrates, at least one was processed as a blank.

     Regarding sampling isokineticity for the profiler, sampling
intakes were reoriented if the 15 minute average wind direction
changed by more than 30°, and the sampling rate was corrected
when the 15 minute average wind speed changed by more than 20%.

     The investigators recorded the total number of vehicle
passes as well as the number of "bad" passes, in which the wind
direction reversed and upwind filter weights were affected by
road emissions.  Data from one run had to be discarded because
the number of bad passes far outweighed the number of good
passes.  Several other sampling runs had a significant percentage
of bad passes.  Most runs had no bad passes.  For runs in which
bad passes occurred, the upwind dust concentration was estimated
by the average of the concentrations of the previous and
following sampling runs.  Bad passes were not counted when
calculating the emission factor (i.e. when dividing the
integrated exposure by the number of vehicle passes).

     Despite the attention given to normal quality assurance
practices in field data collection, the overall level of quality
assurance for this study is compromised by the paucity of
published raw field data.  For instance, measured exposures at
the various profiling heights were not reported.  Because of this
omission, the computations made by the investigators cannot be
repeated and verified.

     Findings—The emission factors measured for the uncontrolled
sampling runs are presented in Table 2.20.  Except where

                              2-46

-------
otherwise noted, data were collected using exposure profiling.
The total particulate emission factor is based on the catch of
each sampling head in the exposure profiler.  The factor for
particles smaller than 30 (o.m is based on an extrapolation of the
fractions smaller than 15 |xm and 2.5 ^im, which were measured
using the dichotomous sampler.

     The particle size distribution data collected using the
dichotomous sampler was found to be the most consistent of the
three methods.  The investigators cited three other reasons for
using the dichotomous sampler data in calculating emission
factors for the different particle size fractions:  convenience
of equipment, endorsement of dichotomous samplers by the EPA, and
the greater availability of dichotomous sampler data from this
particular equipment configuration.

     Regression analysis was used to determine if the method of
emissions measurement (profiling versus upwind-downwind modeling)
is a significant predictor of the emission factor.   Profiling
was found to produce statistically higher factors for both total
suspended particulates and inhalable particulates (< 15 p.m
aerodynamic diameter).  The average differences were 24% and 52%,
respectively.

     Multiple regression analysis was used to develop predictive
equations relating emission factors to various parameters.  The
estimated relationships for the different traffic types are shown
in Table 2.21.

     Publication--This study was conducted and documented under
contract with the EPA Office of Air Quality Planning And
Standards, Research Triangle Park, North Carolina, and the EPA
Industrial Environmental Research Laboratory, Cincinnati, Ohio.
It was published as Publication No. EPA-600/7-84-048 in March
1984.
                               2-47

-------
TABLE 2.20. MEASURED EMISSION FAnTTRS

Run
J-9
J-9b
J-10
J-10b
J-11
J-12
J-12b
J-20
J-20b
J-21
J-21b
K-1
K-6
K-7
J-13
J-18
J-19
K-2
K-3
K-4
K-5
P-11
P-12
P-13
	 Emission Factors (l>-,/v<=>h mi 1=0 	
Total
51 .4

54.1

67.2
16.5

36.6

76.4

23.2
8.0
4.6
7.0
9.5
7.1
5.0
3.1
3.0
2.7
12.8
12.8
9.7
da < 30 urn
15.2
14.1
33.0
12.0
30.2
12.9
3.6
12.3
6.4
14.2
15.0
8.2
2.2
3.9
5.5
8.2
6.7
0.64
0.76
0.60
0.93
8.5
9.0
7.8
da < 15 urn
7.4

17.7

15.4
7.9

5.4

6.0

3.3
1 .1
2.5
4.5
6.6
5.2-
0.33
0.39
0.34
0.52
4.5
5.1
4.1
da < 2.5 |am
0.41

0.54

0.69
0.26

0.14

0.21

0.05
0.07
0.07
0.50
1 .5
0.22
0.03
0.03
0.04
0.05
0.10
0.13
0.15
 a)  aerodynamic particle  diameter
 b)  upwind-downwind test  for comparison


TABLE 2.21.   EQUATIONS FOR PREDICTING  MEDIAN EMISSION  FACTORS'
Traffic Type
Light- and
Medium-Duty
Vehicles
Haul Trucks
TSP (Ib/veh-mile)
5.79/M4'0
0.0067 w3'4!,0-2
IP (Ib/veh-mile)
3.72/M'"3
0.0051 w3'5
 a) variable definitions:
    M - moisture content, %;  w -  number of wheels;  L - silt loading,  g/m2
                                  2-48

-------
Study 12—     Pyle and McCain.  Critical Review of Open Source
               Particulate Emission Measurements.  Part II -
               Field Comparison.  EPA-600/2-86-072.  1986.

     Methodology—This document describes and evaluates a side-
by-side field comparison of five organizations measuring
emissions from vehicle travel on an unpaved road.  The
organizations were Energy & Environmental Management, Inc. (EEM),
Midwest Research Institute (MRI),  PEI Associates, Inc.  (PEI), TRC
Environmental Consultants, Inc. (TRC),  and United States Steel
Corporation  (USS).  All five organizations used the same basic
methodology:  exposure profiling.   However, each organization
chose its own sampling equipment,  equipment configuration, and
analytical procedures.  Consequently, each organization's
implementation of profiling differed in some ways from the
others.  Therefore, this study provided the opportunity for
evaluation of the ease of implementation of the various
procedures,  their accuracy, and the comparability of the results.

     Each testing position was equipped with a standard high-
volume sampler set up three meters from the road's edge.  These
samplers were intended to serve as baseline monitors of
conditions at each site and,  therefore, did not rotate through
the testing  positions with the teams and the rest of the
equipment.

     Exposure and particle size distribution were measured using
the specific equipment and procedures explained below for each
group.  Some of the procedures used were common to all groups.
Every group  recorded measurements for wind speed; temperature;
silt content  (the percent of road surface material less
than 75 pun in diameter); and the number, type, and speed of the
passing vehicles.  For all teams,  sampling runs were considered
valid only when the ratio of mean sample inlet velocity to mean
wind speed fell between 0.8 and 1.3  (inclusive).
                           /
     Test Site—The test site was a paved slag haul road covered,
for the purpose of this study, with 5 to 10 centimeters of
aggregate.  This aggregate was composed of clay, iron ore, and
boiler ash, yielding an average silt content of 10%.  The site
was in a U.S. Steel facility in Gary, Indiana, about 50 meters
south of Lake Michigan.  Five test positions were established on
each side of the road.   Traffic on the road was a mixture of
service vehicles and dump trucks.

     Individual Group Implementation and Findings—

     Energy & Environmental Management—EEM used a vertical array
     of 5 directional high-volume sampling heads on the downwind
     side of the road to measure exposure.  Samplers were
     positioned at 2, 4,  6, 8, and 10 meters height.  Background
     exposure was measured using a tower with three directional

                               2-49

-------
hi-vols at 2, 4, and 6 meters height.  Filters for both
profilers were vertically oriented.

Wind speed was measured at one meter above the ground.  EEM
assumed a uniform vertical distribution of wind speed; the
intake velocity of each sampler on the profilers was set to
match the wind speed measured prior to commencement of
sampling.  The sampling velocity was not adjusted during
individual runs.  Sampler intakes were kept within 20
degrees of pointing directly into the wind.

Particle size distribution in the dust plume was estimated
using computer controlled scanning electron microscopy
 (CCSEM).  Particulates from selected portions of profiler
filters are  removed and recollected on SEM sample stubs for
analysis using a scanning electron microscope and an
automated image analysis system.  This system estimates a
physical diameter - weight distribution curve for the
sample.  Portions of the specific procedures followed in
using CCSEM  were considered proprietary by the investigators
and, therefore, were unavailable for review.

In addition  to those parameters collected by all parties, as
described above, EEM also measured the percent of road
surface material less than 45 [im in diameter and the percent
between 45 and 75 um in diameter.  Data was also collected
on wind direction, humidity, and soil moisture.  The
equipment and procedures for measuring these parameters were
not documented in the report.

EEM's exposure data and calculated emission factors are
presented in Table 2.22.

Midwest Research Institute--MRI used a vertical array of 5
directional  high-volume sampling heads positioned at 1.5, 3,
4.5, 6, and  7.5 meters above the ground to measure exposure
downwind of  the test road.  Background exposure was measured
using an array of 2 directional high-volume samplers at
heights of 3 and 6 meters.  Filters on both profilers were
oriented horizontally above the intakes.

Wind speed was measured at heights of  1.5 and 4.5 meters.  A
logarithmic  distribution of vertical wind speed was assumed
in setting each of the sampler intake velocities equal to
the wind speed.  Intake velocities were adjusted every 10
minutes if measured wind speed changed substantially.  The
 sampler intakes were kept within 20 degrees of the wind
direction.

For measuring  the size distribution of the particles in  the
dust plume,  MRI used a high-volume sampler equipped with a
cascade impactor and a cyclone.  Two of these samplers were

                          2-50

-------
located beside the downwind profiling tower at 1.5 and 4.5
meters above the ground.

MRI's exposure data and calculated emission factors are
presented in Table 2.23.

PEI Associates, Inc.—PEI used an array of 4 directional
high-volume samplers to measure both particle size
distribution and total exposure downwind from the road.
Samplers were placed at heights of 1, 2.5, 5, and 9 meters.
Filters were vertically oriented.  For tests 1-7 and 11, the
samplers on the profiler tower were fitted with a series of
stacked filters designed to provide cut diameters of 30 and
2.5 |J.m.  This equipment configuration was used to estimate
particle size distribution and emission factors for
particles less than 30 urn and for particles less than
2.5 jim.  For tests 8-10 the stacked filters were removed,
allowing measurement of total exposure and an emission
factor for total particulates.  An identical profiler was
used to measured upwind, background exposure and particle
size distribution.

Wind speed was measured at a height of 3 meters.  The
vertical change in wind speed was estimated by a power
function, and the intake velocity of each profile sampler
was set accordingly.  Sampler inlets faced within 15 degrees
of the wind direction.

In addition to the standard parameters, PEI also measured
and recorded humidity and soil moisture.

PEI's exposure data and calculated emission factors are
presented in Table 2.24.

TRC Environmental Consultants, Inc.—TRC used a vertical
array of 5 directional high-volume samplers to measure
exposure downwind from the road.  Sampling intakes were set
at heights of 1, 3, 5, 7, and 9 meters.  Exposure upwind
from the road was measured using a single directional high-
volume sampler at 2 meters above the ground.  Filters for
the upwind and downwind samplers were oriented horizontally
below the intakes.

Each sampler head was fitted with a thermal anemometer,
providing continuous monitoring and automatic adjustment of
intake velocity to match the wind velocity.  Thus, there was
no need to make assumptions regarding changes in wind speed
with height.  Sampler intakes faced within 30 degrees of the
wind direction.
                          2-51

-------
     Like  Energy &  Environmental Management,  Inc.,  TRC used
     computer  controlled scanning electron microscopy to measure
     the particle size distribution of the filter  samples.

     TRC's exposure data and emission factors are  presented in
     Table 2.25.

     United States  Steel Corporation—USS used a vertical array
     of  4  directional high-volume samplers to measure exposure
     downwind  from the road.   The samplers were positioned at 2,
     4.5,  6.5,  and 9 meters height.    Upwind exposure was
     measured  using a single,  directional high-volume sampler 2
     meters above the ground.   Filters were  oriented horizontally
     below the intakes.

     Wind  speed was monitored at each sampling head using a
     thermal anemometer, allowing continuous, automatic
     adjustment of the intake velocity to match the wind
     velocity.  Therefore,  assumptions regarding changes in wind
     speed with height were not necessary-  Sampling intakes
     faced no  more than 30  degrees from the  wind direction.  USS
     documented wind direction throughout the test runs.

     Computer  controlled scanning electron microscopy was used to
     measure the particle size distribution  of the dust plume.

     USS's exposure data and calculated emission factors are
     presented in Table 2.26.

     Sampling  Runs—Eleven  sampling runs were conducted in which
all five organizations simultaneously collected samples.  The
number of  vehicle passes per sampling run averaged about 45.  If
one vehicle pass is defined as the passage of a vehicle in front-
of a single testing position,  then a total of 2,321 vehicle
passes were logged during this study.  Sampling duration was not
recorded.

     Quality Assurance—Personnel from the Southern Research
Institute   (SoRI) supervised the field experiment and reviewed the
analytical procedures which were documented  in reports by each
organization.
     The organizations rotated through the test positions between
sampling runs  to reduce the possibility of bias arising from
variations in  the physical  characteristics of the road.

     SoRI  used Lundgren cascade impactors as a basis for
comparing  the  particle classifying devices used by the five
organizations.  For several tests the Lundgren impactors were
collocated with these other samplers  (i.e. MRI's cyclone and
impactor).
                               2-52

-------
     Quality assurance practices implemented by each team were
generally thorough and well documented.  All groups calibrated
their profiler heads before the field study began using various
methods.  TRC used a wind tunnel.  Some form of audit of the filter
weights was performed by each group.  The notable example in this
instance was MRI, who audited 10% of the tared filters and 100 % of
the exposed filters.  MRI and EEM also processed blank samples.

     Findings--Findings of the individual organizations, in terms
of emission factors for total particulates, are shown in Tables
2.22-2.26 at the end of the summary of this study- The SoRI
investigators, having supervised the field study and reviewed the
procedures and findings of the five organizations, came to the
following conclusions concerning implementation of exposure
profiling:

     •    Dust exposure is often significant as high as 9 meters
          above the ground; therefore, in future studies sampler
          intakes should be positioned at this height or higher.

     •    Dust exposure peaks at 1.5 to 2 meters above the ground;
          a sampler intake should be positioned at this level.

     •    Sampling isokineticity can be optimized using velocity
          sensors at each intake and a "servo system," enabling
          continuous adjustment of the intake velocity to match
          wind speed.

     •    The profiling systems used by the -five organizations
          produced comparable results for total emission factors.
     •    The different techniques for measuring particle size
          distribution in the dust plume yield dissimilar results
          in terms of distribution curves and size-specific
          emission factors.

     •    The cyclone/impactor system, which MRI utilized, is
          recommended over CCSEM and the stacked filter method for
          measuring particle size distribution in the dust plume.
          (SoRI recommended several minor changes in the way MRI
          implemented the system.)

     •    Differences in the methods of graphical integration of
          exposure versus height are unimportant relative to other
          variations in implementation of the methodology.

     Publication—This comparative review was prepared for the
Technical Support Office of the Environmental Protection Agency,
Air and Energy Engineering Research Laboratory, Research Triangle
Park,  North Carolina.  It was published in 1986 as Publication No.
EPA-600/2-86-072.  It is Part II of a two-part study comparing the
commonly used methods of measuring emissions and estimating
emission factors for vehicle travel on unpaved roads.

                                2-53

-------
   Data Tables--
TABLE 2.22  EEM'S EXPOSURE DATA AND CALCULATED EMISSION  FACTORS
EEM
Run
1
2
3
4
5
6
Sampling
Height (m)
2
4
6
8
9
2
4
6
8
9
2
4
6
8
9
2
4
6
8
9
2
4
6
8
9
2
4
6
8
9
Exposure
mg/cm2
7.61
5.15
1 .98
0.16
0.35
3.37
2.89
1 .10
0.46
0.05
8.48
4.42
2.35
0.93
0.21
7.75
6.57
2.77
1 .58
0.58
3.35
2.20
1 .44
0.71
0.43
3.72
3.15
2.15
1 .06
0.45
Total Particulate
Emission Factor
(kg/VKT)
10.24
6.52
1 1 .30
11 .20
5.00
5.49
                              2-54
(continued)

-------
TABLE 2.22  (continued)
EEM
Run
7
8
9
10
11
Sampling
Height (m)
2
4
6
8
9
2
4
6
8
9
2
4
6
8
9
2
4
6
8
9
2
4
6
8
9
Exposure
mg/cm2
7.76
5.71
2.78
1 .46
0.88
8.39
5.89
2.54
0.47
0.03
5.63
3.05
1 .35
0.25
0.14
7.88
6.39
2.95
0.67
0.08
5.86
3.69
2.06
0.64
0.08
Total Particulate
Emission Factor
(kg/VKT)
1 1 .58
7.46
5.71
9.06
8.16
          2-55

-------
TABLE 2.23.  MRI'S EXPOSURE DATA AND CALCULATED EMISSION FACTORS
MRI
Run
1




2
3
4
5
6
Sampling
Height
(m)
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
Exposure
mg/cm2
8.12
6.19
4.25
1 .45
0.30
3.22
2.26
1 .14
0.39
0.06
9.89
9.88
5.23
(3.1)
1 .01
13.50
11 .50
6.52
3.57
1 .50
0.60
2.15
1 .21
0.78
1 .14
5.08
6.49
6.00
4.04
3.46
Total Particulate
Emission Factor
(kg/VKT)
9.33
4.26
12.20
15.90
2.54
11 .50
                               2-56
(continued)

-------
TABLE 2.23  (continued)
MRI
Run
7
8
9
10
11
Sampling
Height (m)
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
1 .5
3
4.5
6
7.5
Exposure
mg/cm2
2.35
5.77
3.56
2.91
2.33
8.89
7.73
4.44
1 .73
0.55
6.99
5.88
3.58
2.50
0.40
6.99
5.88
3:78
2.50
0.40
5.32
6.52
5.19
2.40
0.74
Total Particulate
Emission Factor
(kg/VKT)
7.16
6.26
6.06
6.88
8.68
          2-57

-------
TABLE 2.24.   PEI'S  EXPOSURE DATA  AND CALCULATED  EMISSION FACTORS
PEIm
Run
8
9
10
Sampling
Height (m)
1
2.5
5
9
1
2.5
5
9
1
2.5
5
9
Exposure
mg/cm2
10.40
3.36
4.46
2.41
5.23
3.61
2.78
0.81
5.23
5.19
2.53
1 .01
Total Particulate
Emission Factor
(kg/VKT)
7.81
4.16
5.56
            a) PEI measured total particulate emissions only for
               these three tests.
                                  2-58

-------
TABLE 2.25.  TRC'S EXPOSURE DATA AND CALCULATED EMISSION FACTORS
TRC
Run
1
2
3
4
5
6
Sampling
Height (m)
1
3
5
7
9
1
3 '
5
7
9
1
3
5
7
9
1
3
5
7
9
1
3
5
7
9
1
3
5
7
9
Exposure
mg/cm2
5.21
5.10
3.31
0.64
-0.10
2.99
2.06
0.95
0.33
-0.05
6.02
4.53
2.53
0.43
-0.06
8.18
4.50
3.93
1 .28
0.21
4.49
3.92
2.56
1 .02
0.34
4.22
3.11
2.19
0.57
-0.32
Total Particulate
Emission Factor
(kg/VKT)
7.11
4.29
6.67
8.80
5.90
4.40
                              2-59
(continued)

-------
TABLE 2.25  (continued)
TRC
Run
7
8
9
10
1 1
Sampling
Height
(m)
1
3
5
7
9
1
3
5
7
9
1
3
5
7
9
1
3
5
7
9
1
3
5
7
9
Exposure
ing /cm2
4.18
3.29
2.20
1 .40
0.62
5.50
4.40
2.74
0.98
0.15
3.57
3.42
2.49
0.79
0.10
5.45
4.16
2.82
1 .06
0.42
4.78
3.72
1 .95
0.83
0.11
Total Particulate
Emission Factor
(kg/VKT)
5.85
4.54
4.06
5.65
5.74
         2-60

-------
TABLE 2.26.  USS'S EXPOSURE DATA AND CALCULATED EMISSION FACTORS
OSS
Run
1
2
3
4
5
6
7
Sampling
Height (m)
2
4.5
6.5
9
2
4.5
6.5
9
2
4.5
6.5
9
2
4.5
6.5
9
2
4.5
6.5
9
2
4.5
6.5
9
2
4.5
6.5
9
Exposure
ing /cm2
6.35
4.95
1.11
0.29
3.15
1 .70
0.43
0.11
6.88
5.01
1 .24
0.41
11 .74
7.93
2.45
0.59
3.51
2.08
1 .09
0.42
5.67
4.32
2.04
1 .10
7.49
5.03
2.97
1 .89
Total Particulate
Emission Factor
(kg/VKT)
8.56
5.29
9.20
15.92
4.84
7.79
12.28
                              2-61
(continued)

-------
TABLE 2.26.   (continued)
uss
Run
8
9
10
11
Sampling
Height
(m)
2
4.5
6.5
9
2
4.5
6.5
9
2
4.5
6.5
9
2
4.5
6.5
9
Exposure
mg/cm2
6.99
5.21
2.03
0.74
3.53
2.74
1 .07
0.22
6.69
5.34
2.65
0.72
7.14
4.35
1 .24
0.39
Total Particulate
Emission
Factor (kg/VKT)
6.59
4.19
8.22
9.61
          2-62

-------
Studies of Secondary Importance

Study 13—     Roberts.  The Measurement, Cost and Control of Air
               Pollution from Unpaved Roads and Parking Lots in
               Seattle's Duwamish Valley. 1973.

     The dust plume created by vehicle travel on a gravel road was
sampled using a towed rack to which a cascade impactor was attached.
The rack was designed as a vertical grid oriented perpendicular to the
car's path; the impactor was rotated among the various positions
between tests so that, after a series of tests, the average
concentration of dust in the plume could be determined.  In order to
derive an emission factor in Ib/veh-mile, the average plume
concentration was multiplied by the volume of air into which it was
emitted.  This volume was estimated in the following manner.  First,
the area of the plume behind the car was estimated by towing a
grid/rack of open impaction plates and examining the dust pattern on
the plates.  Second, this area — 70 square feet — was multiplied by
5,280 feet (1 mile) to obtain the air volume  (36,960 ft3)  into  which
the dust was emitted after 1 mile of travel.

     The impactor samples were also analyzed to determine the particle
size distribution of the plume.  This particle size breakdown was then
used to estimate emission factors for various particle size categories
based on the measured total particulate emission factors.

     Sample collection, handling, and analysis procedures were not
documented in detail.  For example,  specific procedures for handling
the impactor plates were not documented.  No mention was made of
problems with particle bounce in the impactor; this problem has been
documented by at least two other researchers  (Cowherd et al.,  1974 and
McCaldin,  1977).

     The distance between the vehicle and the towed rack was not
documented.  The investigator noted that it was difficult to certify
that sampling was conducted under isokinetic conditions, due to
turbulence in the wake behind the vehicle and changes in wind
direction and speed.

     Seventeen sampling runs were conducted at 20 mph on a gravel road
in the Duwamish Valley in Seattle.  Roberts noted that the test road
had been sprayed with oil two years before, and that emissions from
the road were still effectively suppressed at the time of the study.
Consequently,  the findings of this study are secondary in their
importance to the development of emission factor equations.

     For these 17 runs, the average concentration of total
particulates  in the air sampled by the cascade impactor was .133
grains/ft3.  Following  the procedure  explained above,  the emission
factor was calculated at 7 Ib/veh-mile.
                                 2-63

-------
     Limited testing was also conducted with vehicle  speeds  of 10 and
30 mph.  The cascade impactor collected dust samples  at  only two grid
positions on the sampling rack  (as compared to 6 positions 'at 20 mph).
The following proportional relationship was assumed in estimating an
emission factor for the 10 mph  tests:
where
     X
     Y
                                   = X/Y
average concentration at grid sampling point D at  20
mph
average concentration at grid sampling point D at  10
mph
total average  (of all grid sampling points) at 20  mph
total average  (of all grid sampling points) at 10  mph
The same procedure was used to estimate an emission  factor  for  30  mph.

     The calculated factors for the various  speeds are  shown  below in
Table 2.27.  Roberts also presented emission factors for  particles
smaller than 10 pirn and for particles  smaller than 2  |j.m  in diameter,
though he did not explain how the  fraction of the total particulate
sample consisting of particles smaller than  10 [im was determined;  i.e.
he did not indicate whether the cascade impactor provided a cut point
at that size.
           TABLE  2.27.  EMISSION FACTORS FOR VARIOUS  SPEEDS

Speed
(mph)
10
20
30
20
Emission Factor (Ib/veh-mile)
Total Participates
3.5
7.0
22.2
7.3
PM10
.58
1 .9
9.0
2.0
da < 2 |im
.10
.24
.77
nd
                a) particle diameter
     This  study was  conducted and documented as  partial fulfillment of
 requirements  for  a master's  degree in engineering at  the University of
 Washington.   Portions  of  this work and some follow-up studies were
 published  in  the  Journal  of  the  Air Pollution Control Association in
 1975.
                                  2-64

-------
Study 14—     Handy et al.  "Unpaved Roads as Sources for Fugitive
               Dust."  Transportation Research News. 1975.

     Deposition of dust from ten unpaved roads was estimated using
dustfall buckets mounted one meter above the ground in lines
perpendicular to the road.  The buckets were half filled with water to
prevent resuspension of dust deposited in the bucket.   Buckets were
left in the field for 3 to 4 weeks before being sealed and brought to
the laboratory for analysis.  Contaminants such as insects and chaff
were removed by hand before samples were dried and weighed.  Particle
size analysis was performed using sieves.  Average deposition was
plotted against distance from the road.  Total dust deposited was
calculated by integrating the area under this curve.

     The dustfall buckets provided deposition data in terms of
kg/hectare-month.  The total dust mass deposited per road kilometer
was calculated by integrating this with respect to distance from the
road.  Extrapolations were made from this field data to estimate
kg/km/vpd/year, based on the estimated number of vehicles per day
(vpd) traveling the road.

     Deposited road dust was found to fall into two categories:
roadside and distributed.  The distinction is based on the change in
the differential mass deposited as distance from the road is
increased.  The change occurs at about 10 meters from the centerline,
roughly the same as the right-of-way for most secondary roads.

     Total deposited dust ranged from 70 kg/km/month for a chemically
treated surface to 56,640 kg/km/month for one untreated road.  These
numbers translated into 3.3 kg/km/vpd/year for the best road and 2185
kg/km/vpd/year for the worst.

     The authors implied an assumption that dust deposition as
measured with the dustfall buckets minus background deposition equals
dust emitted from the road.  Dust emission rates were not directly
measured in this study.  Traffic volume was estimated from a secondary
data source, rather than measured precisely.
                                 2-65

-------
Study 15—     Dyck and Stukel, "Fugitive Dust Emissions from Trucks
               on Unpaved Roads." Environmental Science & Technology.
               1976.

     The upwind-downwind method was used to estimate emission factors
for a truck traveling on unpaved roads in a construction access area.
One upwind hi-vol sampler, 50 feet from the road, and four downwind
hi-vol samplers, ranging from 50 to 250 feet from the road, were used
to measure the road's contribution to dust concentration.  This data
was applied to a modification of the model for a continuously
emitting, infinite line sources found in the Workbook of Atmospheric
Dispersion Estimates (Turner, 1970) to yield estimated emission
factors.  The truck's weight and speed were varied by the
investigators, as was the road surface type.  Testing was conducted
over several days.  On each day, the choice for a test road was
dictated by the prevailing wind direction.  Each test lasted about one
hour.

     Measured emission factors for various combinations of vehicle
weight, percent silt on the road surface, and road type (silty-sand
and clay) were tabulated.  The investigators used multiple linear
regression to develop an equation for estimating the emission factor
as a function of the experimental variables:

                E  =  5.286  -  3.599(R)  +  0.00271 (V) (W) (S)

subject to
                             10 < V < 25

where
     E    =    emission factor, Ib/veh-mile
     R    =    road type, silty-sand or clay
     V    =    velocity, mph
     W    =    weight, thousands of pounds

     The investigators did not document the specific analytic
procedures followed in calculating the percent silt in the road
surface material, conditioning and weighing filter samples, or
measuring the moisture content of the road surface.  Data on measured
concentrations were not published.

     The emission factor equation published in this article was an
incorrect adaptation of the model cited in Turner, 1970.  The model
for a continuously emitting, infinite line source as it is presented
by Turner (1970) and by the present study is


                                              f  if  H V
              C(x,y,0,H) --  - exp
                             sin
a
                                                     z
                                 2-66

-------
The investigators incorrectly solved this equation for source
strength,  q,  as follows:
                                            1 I  H
    C sin<{> v/27i o _J7
q	1	z— exp   2
The 'exp' term should be in the denominator in this equation.  Because
field data were not presented, it was not possible to determine
whether this was a typographical error or a misapplication of the
model.

     Environmental Science & Technology (Volume 10,  Number 10,  October
1976, pp. 1046-48) is apparently the only outlet through which this
study was published.

Study 16—     Cuscino et al.  Fugitive Dust from Vehicles Travelling
               on Unpaved Roads. 1977.

     The investigator proposed a refinement of the dispersion model
for a continuously emitting infinite line source (Turner, 1970) to
allow for dust deposition between the source and the location at which
concentration is measured.  A parameter representing the fraction of
particles which touch the ground that are not deposited was added to
the equation.  This parameter is believed to be an unknown function of
downwind distance, atmospheric conditions, and particle fall velocity.

     Field experiments were conducted to determine the accuracy of the
new model.  Dust concentrations at several heights and distances
downwind from an unpaved road on which vehicles traveled were measured
using open-faced filters.  Dust emissions were not measured.  The
source strength was taken from Roberts' (1973) estimates of emission
factors for unpaved road.  Measured concentrations were compared
graphically with predicted concentrations.

     Predicted concentrations were within a factor of 2 of the
experimentally determined concentrations for most of the measurements,
and within a factor of 3 for the remainder.  No comparison was made
between the accuracy of this proposed model and the accuracy of the
standard model in the Workbook of Atmospheric Dispersion Estimates
(1970) .

     This work was published in 1977 by the USDA Forest  Service as
General Technical Report NE-25, ODC907.3: 187: 111: 273: 425.1.
                                 2-67

-------
Study 17—     McCaldin. Fugitive Dust Study for Pima Country Air
               Quality Control District, Tucson, Arizona.   1977. Data
               on upwind and downwind dust concentrations and on mean
               wind speed and direction were collected and  applied to
               a dispersion model to estimate an emission factor for
               vehicle travel on unpaved roads.

     Five pre-existing dirt roads in Pima County, Arizona were the
test sites for this study.

     Upwind and downwind concentrations were measured using standard
hi-vol samplers and "standard gravimetric methods."  Sierra Impactors
were used to measure particle size distribution.  Vehicle speed was
controlled by the investigator for most of the passes.  The speed of
the remainder were estimated visually.  The number of passes was
likewise controlled mostly by the investigator.  Composite  samples of
road surface material were analyzed for percent silt using  a 200 mesh
screen.  The Belfort Model 443 anemometer, held at 2 meters above the
ground, was used to measure wind speed and direction.

     Normally, two standard hi-vol samplers were located 50 feet from
the center of the road, one on the upwind side and one on the downwind
side.  For some early tests two side-by-side hi-vols were placed at 50
feet upwind to determine the random variation in upwind concentration,
and hi-vols were run at heights of 1 ,  2, 3, and 3.6 meters  50 feet on
the downwind side to ascertain the plume's vertical concentration
distribution.  The investigator did not specify whether these downwind
hi-vols were operated simultaneously in a vertical array.   The
sampling duration was between ten minutes and five hours for each run.
For some of the final tests, Sierra Impactors were set up at the
upwind station and at 50, 150, and 250 feet from the center of the
road on the downwind side.  They were operated for about four hours
each test.  The heights of the samplers were not specified.

     Data were collected on 46 sampling runs.  Raw data for unpaved
road emissions were not provided with this report.

     Detailed documentation of sample collection and handling
procedures was not provided.  Data reduction could not be verified, as
raw data was not included in the report.  However, a sample
calculation of an emission factor from concentration and wind data was
provided.

     Collocated hi-vol samples were collected, as noted above, but a
comparison of the measured concentrations was not provided.  Other
quality assurance measures such as processing of blank samples and
calibration of sampling equipment were not documented.

     Emission rates were found to vary directly with percent silt and
exponentially with vehicle speed.  The average emission rate for
                                 2-68

-------
vehicles traveling 30 mph on a road with a 10% silt surface was about
3 Ib/veh-mile.  An equation relating the emission rate to percent silt
and vehicle speed was estimated from the collected data as follows:

                          E =  (s) 0.035  (S)2

where

     E    =    TSP emission factor,  Ib/veh-mile
     s    =    mass percent silt
     S    =    traffic speed,  mph

     Attempts to estimate particle size distribution failed due to
particles bouncing down through the stages and dust settling on other
surfaces within the hi-vol unit.

     This study was conducted and documented under contract with the
Pima County Air Quality Control District (AQCD).   EPA Financial Grant
Number A0090055-77-2, awarded to the Pima County AQCD, provided
funding for the study.

Study 18—     Pinnick et al.   "Dust Generated by Vehicular Traffic on
               Unpaved Roadways:  Sizes and Infrared Extinction
               Characteristics."  Aerosol Science and Technology.
               1985.

     This study is of secondary importance in the development of
emission factors for vehicle travel on unpaved roads for three
reasons: 1) no attempt was made to measure the total mass of dust
emitted from the road or to measure emissions in mass/veh-mile
2) unusual vehicles (i.e. Army tanks) were used in the testing and 3)
the test roads were constructed for this study and apparently did not
reflect typical unpaved road conditions.

     Two roads were constructed for this study of emissions from
vehicle travel on unpaved roads.  Army tanks, armored personnel
carriers, and five-ton trucks served as the test vehicles.  A battery
of optical particle counters as well as a hi-vol sampler were used to
estimate dust concentration in the plume of the passing vehicles.

     A series of light-scattering aerosol counters and an optical
array probe measured the mass and number of particles of several size
categories, covering the range from sub-micron to 150 jam in diameter.
These instruments measure particle size and number concentration.
Considerable attention was given to calibrating these instruments.

     The aerosol counters were mounted in a cage which was held up by
a crane, enabling easy movement of the instruments from one side of
the road to the other in the event of changes  in wind direction.  A
series of hi-vol samplers were also stationed beside the road to
measure aerosol "mass dosage."


                                 2-69

-------
     Particle size distributions in the plume were found to be
consistently bi-modal, with mass mean radii at 4 and 45 (am for  roads
made of sandy soil, and 10 fim and 35 (J.m for roads made of silty soil.
For each test two separate concentrations were reported, one for  each
mode.  Total measured concentrations ranged from .282 grams/m3 for a
single 5-ton truck, to 2.084 grams/m3  for  an armored personnel
carrier.
                                  2-70

-------
VEHICLE TRAVEL ON PAVED ROADS

Introduction

     The earliest field-measured emission factor for vehicle travel on
paved roads was published in a master's thesis in 1973.  Emissions
from both paved and unpaved roads were measured in that study-  The
methodology, which consisted of towing a dust sampling rack behind a
car to measure the plume concentration, was not utilized in any other
field research.

     From its beginning in this initial study, the historical
development of field research on paved road emissions has closely
paralleled that of unpaved roads.  This, of course, is due to the
similarity of the sources.  Vehicle travel on paved roads has not
received as much research attention as unpaved roads.

     Most of the work falls into one of two methodological categories:
upwind-downwind dispersion modeling or exposure profiling.  For this
particular source, the latter method has been utilized more than the
former.  The improvements in the exposure profiling methodology
discussed in the section on unpaved roads are .also true regarding its
application to paved road emissions.

     The reviews of field research reports are organized like those in
the unpaved road section.  Those reports which are of primary
importance to the development of paved road emission factors are
reviewed in chronological order.  The review of each primary report
consists of a summary description of the methodology, test site(s),
measured parameters, equipment configuration, sampling extent, quality
assurance, findings, and publication source.

     Reports of secondary importance are briefly reviewed following
the primary report reviews.  Only two of the reviewed reports were
considered secondary in importance:  Roberts' 1973 master's thesis,
and McCaldin's 1977 field study for the Pima County Air Quality
Control District.  Neither one of these documents provided raw field
data or a detailed description of the field sampling or analysis
procedures.  However, because both of these studies involved the
measurement of emissions from vehicle travel on paved roads, they were
judged to be relevant to a review of the development of a paved road
emission factor.
                                 2-71

-------
Studies of Primary Importance

Study 1 — Sehmel, G.A.  "Particle Resuspension from an Asphalt  Road
          Caused by Car and Truck Traffic."  Atmospheric Environment.
          Vol. 7.  1973.

     Methodology — This study employed a methodology similar  to  that
used in the original Midwest Research Institute  (MRI) profiling study
(Cowherd et al . ,  1974).  A known mass of zinc sulfide  (ZnS)  particles
was applied to one lane of a paved road.  Vehicles were driven  either
through or beside the ZnS at speeds ranging from 5 to 50 mph.
Vertical filter arrays were set up downwind from the road  to measure
particle exposure at various heights and distances from the  road.
Point values for exposure were calculated as the mass of ZnS per unit
filter area.  Deposition was also measured.  The fraction  of ZnS
emitted  (or "resuspended") from the road was measured as the
integrated exposure at a particular downwind distance plus the
integrated deposition between the road and that distance.  The
following assumed mass balance relationship was the basis  for this
calculation:
   where
                         ZnSlost = ZnSexp + ZnSdep
     ZnSiost    =    mass of ZnS lost from road surface
     ZnSexp     =    mass of ZnS passing through a vertical plane
                    parallel to the road at a given downwind distance
     ZnSdep     =    mass of ZnS deposited between the road and thesame
                    downwind distance

The emitted mass was expressed as a fraction of the ZnS particle mass
on the road prior to the vehicle pass.  This fraction was called the
resuspension rate.

     Test  Site — An asphalt road was used as the test site.  The road
did not appear dusty, but it was not cleaned prior to this experiment.
The ZnS dust was screened such that particles applied to the road
would be generally smaller than 25 |im.  Cylindrical particle
generators were used to apply about 0.5 gram of material per square
foot of road.

Parameters and Equipment — The two principal parameters measured were
exposure and deposition.  The exposure samplers used were described
not explicitly but by reference to previous published studies by the
same author.  He did note that the flow rate was controlled by a one
cfm critical orifice, and that the exposed membrane filter had a
diameter of 1.6 inches.  Special real time samplers were also used to
sample ZnS at 2 feet above the ground.  These samplers enabled the
investigator to know when sufficient mass was collected on the
filters.
                                 2-72

-------
     Deposition was measured using membrane filters mounted on filter
holders; these were recessed into the ground so that the filter was
approximately level with the surface.

     Three-cup anemometers were used to measure wind speed at two
heights; a vector vane indicated the wind direction.  Friction
velocity was calculated using the wind speed profile, as measured with
the anemometers.

     Equipment Configuration—Exposure samplers were attached to
towers at heights of one, three, six, and eight feet.  The towers were
set up in a square matrix on the downwind side of the road.  Three
towers were placed 10 feet apart at each of 3 perpendicular distances
downwind:  10, 20, and 30 feet.  For a portion of the study, similar
towers, with added filters at 11 feet above the ground, were also
placed at the edge of the road.

     Deposition samplers were placed at the base of every exposure
sampler tower and at downwind distances of 3.5, 60, and 100 feet.
Three samplers were placed at each of these distances, each sampler
being 10 feet apart.

     The anemometers were attached to a tower at heights of one and
seven feet, and the vector vane was attached at three feet above the
ground.  The tower was positioned approximately three feet away from
the road on the downwind side.

     Sampling Runs--Twenty-one sampling runs were documented in this
report.  Vehicle type (car versus 3/4 ton truck), vehicle pass type
(driving through the ZnS versus driving in the lane beside the ZnS)
and speed were held constant during individual runs and varied between
runs.  The number of vehicle passes per sampling run was not
documented.

     Quality Assurance—Quality assurance for this study was poor.  As
the author noted, wind erosion losses from the road surface between
tests were not measured; therefore, the actual mass of ZnS on the road
prior to each test was not known with certainty-  Consequently,
calculated resuspension rates  (mass suspended/source mass prior to
test) could be lower than the true rate.

     The description of equipment employed and procedures followed was
very general.  The interval at which vehicles passed the test area was
not indicated.  The manner in which filters were handled, transported,
conditioned, and weighed was not discussed.

     Exposure samples were not collected isokinetically, nor were any
corrections made to the data to adjust them to isokinetic conditions.

     The investigator's data reductions could not be checked because
raw data, including measured exposures or deposited masses, were not
published.

                                 2-73

-------
     Findings—As noted above, emissions were expressed as a fraction
of the total source mass prior to each run that was "resuspended" per
vehicle pass;  it was referred to as the "resuspension rate".  This
fraction was found to increase with vehicle speed.  More particularly,
it increased with the square of vehicle speed for those tests in which
the car drove through the ZnS.  On the first day of testing  (on which
the ZnS was originally applied),  vehicle speeds through the ZnS ranged
from 5 to 50 mph, and the corresponding resuspension rate ranged from
0.000019 to 0.0109-  A similar relationship was apparent for those
runs in which vehicles passed by the ZnS, except the resuspension rate
was about an order of magnitude smaller.

     Truck passes through the ZnS at five mph produced a higher
resuspension rate than cars; at 50 mph, cars produced a higher
resuspension rate than trucks.  Sehmel suspected that this reversal
might be due to inaccurate sampling.

     The resuspension rate for the ZnS particles was found to decline
over time.  The investigator noted that this could be due to losses of
the source ZnS due to wind erosion  (as discussed under Quality
Assurance).

     Dust deposition between the road and any of the downwind sampling
distances was maximized in tests with cars traveling at 15 mph.

     Publication--This study was documented in an article in the
journal Atmospheric Environment,  Volume 7, pages 291-309.
                                 2-74

-------
Study 2— Cowherd et al.  Quantification of Dust Entrainment from
          Paved Roadways.  EPA-450/3-77-027.  1977.

     Methodology—Exposure profiling was used to measure emissions
from paved roads.  Standard high volume (hi-vol) samplers were also
used to measure the decrease in suspended dust concentration with
'increasing distance downwind from the road.  For one of the three
sites at which emissions were measured, pulverized soil and gravel
were applied to the road in measured amounts.

     Test Sites— Three test sites were selected in the Kansas City
Area.  Two of the sites were four-lane roads in areas known to have
problems with particulate levels.  Both of these sites were bordered
by unpaved parking lots.  One was surfaced with asphalt and the other
was concrete.  The third site, at which dust was artificially applied,
was also a four-lane road.  It was in an undeveloped portion of an
industrial park, and was closed to public traffic during emissions
testing.  All three roads had curbs.

     Parameters and Equipment—Table 2.28 lists the parameters
measured and corresponding equipment employed in this study.

     Equipment Configuration—Certain aspects of the equipment
configuration varied between the test sites.  The basic configuration,
common to all three sites, was as follows.  The exposure profiler
consisted of four isokinetic sampling heads positioned on the tower at
heights of one, two, three, and four meters.  For all three sites the
profiler was set back three to five meters from the edge of the road.
Dustfall buckets were attached to the profile tower at heights of one
and four meters.  A hi-vol cascade impactor with its intake at 2
meters above the ground was placed beside the profiler.  Except where
otherwise noted, all standard hi-vol samplers had intake heights of
two meters.

     The placement of the standard hi-vol samplers, the wind
instrument,  and dustfall buckets varied from site to site and, in some
instances, from one test to the next.  These variations are explained
below.
                                 2-75

-------
     TABLE 2.28.  MEASURED PARAMETERS AND  CORRESPONDING EQUIPMENT
Parameter
wind speed
Wind direction
Cloud cover
Temperature
Relative Humidity
Pavement Type
Road surface condition
Road surface dust loading*
Road surface dust texture3
Traffic mix
Traffic count
Plume exposure
Plume particle size distribution
Downwind concentration
Background concentration
Sampling duration
Dust deposition
Equipment
Unspecified
Unspecified
Direct observation
Sling psychrometer
Sling psychrometer
Direct observation
Direct observation
Vacuum, scales
Sieves , scales
Direct observation
Automatic counters, direct
observation
Isokinetic hi-vol samplers (on
profile tower)
Hi-vol cascade impactor, cutpoints at
0.4, 0.8, 1.8, and 3 \im
Standard hi-vol sampler
Standard hi-vol sampler
Timer
Dustfall buckets
 a) Measured only for site at which dust was artificially applied

     At the first site  (37th Street), when  the  wind  was  from the
south, background dust concentration was measured  with a standard hi-
vol sampler 19.9 meters from the upwind curb, with its intake at a
height of 2 meters.  Downwind concentration was sampled  from beside
the profiler at the same height.  Wind speed  and direction were
measured on the downwind side 4.6 meters from the  curb at a height of
about 2.5 meters.

     When the wind was from the north during  sampling at the first
site, upwind dust concentration was measured  at five meters from the
curb.  Downwind concentration was measured  with the  standard hi-vol
sampler beside the profiler.  Wind speed and  direction were measured
at 0.5 to 1 meter downwind from the road and  4  meters above the
ground.

     At the test site in the industrial park  (Stillwell  site),  the
configuration did not depend on the wind direction.   The upwind
standard hi-vol sampler was positioned four meters from  the curb.
                                 2-76

-------
Wind speed and direction were also measured here, at a height of 4
meters.  During each test at this site, a standard hi-vol sampler and
a dustfall bucket were placed 20 meters .downwind.  For some tests
concentration and deposition were also measured at 10 or 50 meters
downwind.  Some tests at this site included dustfall measurements at
one meter from the downwind curb.

     At the third site (Fairfax Trafficway)  dustfall measurements were
not taken, except for the two buckets attached to the profile tower.
Wind speed and direction were monitored at 7.3 meters downwind, 4
meters above the ground.   Background concentration was sampled five
meters upwind from the road.  Downwind concentration was sampled at
two positions:  beside the profile tower at the usual height and 50
meters downwind at approximately 7 meters above the ground.

     Sampling Runs—Emission factors were collected for 13 sampling
runs:  3 at 37th Street,  8 at the industrial park, and 2 at Fairfax
Trafficway-  The industrial park was the site at which the dust on the
road was artificially controlled.  Pulverized topsoil was applied
before the first four runs at this site, and gravel was used for the
last four.  At the 37th Street site the number of vehicle passes per
4-j-hour run ranged from 1,880, to 2,440.  At the industrial park each
run lasted between 30 and 90 minutes and included' from 100 to 600
passes.  At Fairfax Trafficway one run had 3,791 passes, and the other
had 4,146 passes.  Both of these were four-hour sampling runs.  A
total of 16,331 vehicle passes were logged during this study.

     Quality Assurance—The guidelines set forth in the "Reference
Method for the Determination of Suspended Particulates in the
Atmosphere (High Volume Method)"  (1971) were followed in measuring
dust concentrations with the standard hi-vol sampler.  The "Standard
Method for Collection and Analysis of Dustfall"  (American Society of
Testing and Materials) was followed in collecting dust deposition
data.  The methods of handling, transporting, and processing exposed
filters were thoroughly documented; sample integrity was assured to
the extent practical.

     In the exposure profiling operation quality assurance measures
such as processing of blank filters, collocation of samplers,
calibration of exposure profiling equipment, or auditing of tared or
exposed filter weights were not documented.

     The methods of measuring wind speed and direction and for setting
the sampling rate and direction were not documented.  The
investigators noted that for the test site on 37th Street, the
isokinetic flow ratio (sampling intake velocity divided by wind
velocity) was much greater than one.  Tabular data indicate ratios  of
5, 6, and 7.5 for the three tests at this site.  Sampling under these
conditions was necessary due to light wind conditions and low dust
concentrations.  Isokinetic correction factors were used to adjust  the
measured exposures and concentrations to isokinetic conditions.


                                 2-77

-------
     Findings—Data on  exposure  of each filter on the profiler  were
not published.  Concentrations at the different profiler heights  were
published only in graphic  form.   Emission factors calculated  from
field measurements of exposure are shown below in Table 2.29.   Runs 3,
5, and 6 were performed on 37th  Street; runs 7 through 14 were  from
the industrial park, at which the surface loading was artificially
controlled; and runs  15 and 16 were from Fairfax Trafficway.

               TABLE 2.29.   CALCULATED EMISSION FACTORS

Run
3
5
6
7
8
9
10
1 1
12
13
14
15
16
Emission Factors"
(Ib/veh-mile)
Totalb
0.015
0.020
0.012
34.7
26.7
12.2
6.9
10.0
6.8
5.3
1 .1
0.019
0.010
dc < 30 urn
0.013
0.019
0.012
19.4
9.6
3.7
2.1
4.8
3.7
2.2
0.46
0.017
0.0092
dc < 5 \J.m
0.007
0.013
0.008
6.2
3.2
1 .1
0.62
1 .6
0.95
0.74
0.14
0.008
0.0042
                 a) corrected to isokinetic conditions
                 b) based on total "catch" by each sampling
                 head on the
                   profiler
                 c) particle Stokes' diameter

     Data  from both  dustfall buckets and standard hi-vol samplers
positioned at  various  distances downwind from the road indicate that
the rate of dust  deposition  decreases very rapidly with increasing
distance from  the road.

     The investigators developed the following linear relationship
between emission  factors  and silt loadings less than 20 g/m2 (280
kg/km, upper limit of  typical loading range) based on a scatter plot
of the data from  the eight sampling runs in which surface  loading was
controlled artificially:

                                e =  KLs
                                  2-78

-------
  where

     e    =    Emission factor, kg/veh-km
     K    =    Proportionality constant, vehicle"1
     L    =    Surface loading excluding curb area, kg/km
     s    =    Silt content, fraction

The value for K was estimated from this graph at 0.00098.  This number
was then used to estimate silt loadings at the other two test sites
based on the measured emission factors.  The estimates were compared
with measurements made in a previous study on the same streets:
Sartor and Boyd (1972) measured total loading as part of a water
pollution study.  To make the comparison, Cowherd et al. assumed that
10% of the total loading was silt.  The resulting numbers were within
about 37% of the values estimated by Cowherd et al.

     Publication — This study was conducted and documented under
contract for the Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, North
Carolina.  It was published as Publication No. EPA-450/3-77-020 in
July 1977.

Study 3 — Axetell and Zell.  Control of Reentrained Dust from Paved
          Streets.  EPA-907/9-77-007 .  1977.

     The primary purpose of this report was to document methods of
controlling dust emissions from paved roads.  Several field studies
were conducted in support of this objective, three of which pertain to
uncontrolled emissions from paved roads.  Only one actually estimates
emission rates from paved roads.  It is reviewed below, whereas the
other two are considered of secondary importance and are therefore
discussed briefly at the end of the paved road section.

     Methodology — Normal upwind-downwind dispersion modeling was used
to estimate emission rates from traffic on paved streets.  The model
used is shown below.
                            x-
                               sinv/27f ozu

where

     X    =    plume centerline concentration at ground level at
               distance x downwind from the source, g/m3
     q    =    line source strength, g/sec-m
     4>    =    angle between wind direction and the road
     0Z    =    the standard deviation of the vertical distribution of
               the plume concentration at a distance x downwind from
               the source, m
     u    =    mean wind speed, m/sec
                                 2-79

-------
It is a simplified version of the standard model for a  continuously
emitting,  infinite line source  (Turner, 1970).  As in the  example  set
by Turner (p. 53),  the term
                             exp


was omitted because the effective height of emission, H, was  taken  to
be zero; thus, the 'exp' term always evaluated to be one.  Also,  the
researchers estimated an initial oz (i.e. az  when x = 0)  according to
the recommendation by Turner  (1970).  This initial  az was estimated at
1.5 meters.

     Test Sites—Four separate test sites in unspecified locations
were used.  Each site was located within a different land use area:
undeveloped, park, residential/commercial, and extensive
commercial/campus.

     Parameters and Equipment—Upwind and downwind  concentration  of
total  suspended particulates was measured using  standard high-volume
samplers.  Wind speed and direction were also recorded every  two
minutes during sampling; the  type of equipment was  not specified.  An
automatic counter was used to measure traffic volume for each test.
The average traffic speed was determined by driving a vehicle with  the
flow of the traffic and noting the speedometer reading.

     Equipment Confiquration--Three standard high volume samplers were
placed downwind 10, 20, and 30 meters from the street.  One background
hi-vol was placed 10 meters from the upwind edge of the road.

     Sampling Runs—A total of 60 runs were conducted at the  four
different sites.  Of these, 35 were performed under stable wind
conditions.  Each run lasted  one to two  hours.   The traffic volume  per
run was not documented.

     Quality Assurance—The documentation for this  field study was
particularly  scant.  No indication was given that  any published
quality assurance guidelines  (e.g. Quality Assurance Handbook for Air
Pollution Measurement Systems.  Volume II - Ambient Air  Specific
Methods  [U.S. EPA, 1977], or  Ambient Monitoring  Guidelines for
Prevention of Significant Deterioration  [U.S. EPA,  1978]) were
followed.  Documentation was  not provided for such quality assurance
.measures as auditing of filter weights or data reduction calculations,
calibration of equipment, processing of  blank samples, or collocation
of samplers.

     Furthermore, data  for the other dispersion  modeling parameters
such as wind  speed and  atmospheric  stability were  not  included in the
report.
                                  2-80

-------
     Samples collected during periods of widely variable wind
direction were not considered valid;  the results of these, however,
were listed in a separate table.

     Findings—The average TSP emission factor for the 35 test runs
was 3.7 g/veh-mile.  The standard deviation was 3.3 g/veh-mile.  A
comparison of emission factors for streets in areas of differing land
uses did not indicate a strong relationship between land use and
emission factors.   The data collected during periods of consistent
wind direction are presented in Table 2.30.

     TABLE 2.30.  PARTICULATE CONCENTRATIONS AND EMISSION FACTORS



Run
No.
3
4
7
10
12
13
16
18
20
22
23
25
26
27
28
29
30
31
34
37
38
39
40
41
42
43
45
47
49
50
51
52
54
56
57
avg.
Net downwind concentration
Hg/m3
Apparent Emission Factor
g/veh-mi
--Distance from Downwind Edge of Street--
10m

76.9
46.7
85.3
90.5
34.1
40.2
35.9
16.8
9.6
62.9
42.9
48.7
72.3
73.2
132.1
36.5
80.3
62.1
30.1
46.1
85.5
44.8
166.1
54.7
30.7
73.3
61 .2
64.6
71 .0
76.4
90.4
90.3
148.6
99.1
98.5
67.9
20m

40.5
40.6
59.3
54.4
7.8
5.7
20.9
5.5
1 .6
37.8
37.0
48.4
40.7
59.9
119.7
23.7
71 .3
53.4
2.7
27.6
62.5
28.9
132.4
20.9
15.9
44.0
39.8
29.5
29.6
58.0
57.7
58.4
95.7
51 .5
81 .0
44.7
30m

33.3
59.7
35.0
39.6
2.1
0
20.2
2.9
16.6
32.0
41 .2
15.3
24.5
36.8
61 .2
22.4
68.5
41 .6
0
10.9
26.9
6.9
74.4
38.8
9.6
30.2
43.4
9.5
22.5
33.4
13.3
36.8
75.4
38.1
36.5
30.3
10m

2.98
2.83
1 .65
7.99
2.13
3.38
1 .99
1 .32
0.42
1 .62
6.66
1 .10
5.36
2.87
16.67
5.86
6.44
1 .46
1 .54
5.15
4.31
3.15
3.38
7.84
2.38
5.68
5.01
1 .84
6.88
1 .23
4.45
1 .89
8.12
7.69
4.34
4.21
20m

2.10
3.27
1 .54
6.41
0.64
0.64
1 .54
0.58
0.09
1 .30
7.66
1 .45
4.03
3.13
20.16
5.06
7.62
1 .68
0.19
4.12
4.18
2.72
3.95
4.00
1 .64
4.57
4.33
1 .12
3.84
1 .25
3.79
1 .63
6.98
5.33
4.76
3.63
30m

2.16
5.98
1 .13
5.83
0.21
-
1 .87
0.38
1 .20
1 .37
10.66
0.57
3.04
2.42
12.88
5.99
9.15
1 .63
-
2.04
2.25
0.81
2.52
9.28
1 .24
3.90
5.92
0.45
3.63
0.90
1 .09
1 .29
6.87
4.93
2.68
3.26
                                 2-81

-------
     Publication--Preparation of this document and the supporting
field work described above were sponsored by the Environmental
Protection Agency, Region VII - Air Support Branch, Kansas City,
Missouri.  It was published as Publication No. EPA-907/9-77-007 in
July of 1977.

Study 4— Bohn et al.  Fugitive Emissions from Integrated Iron and
          Steel Plants.  EPA-600/2-78-050.  1978.

     Methodology—Exposure profiling, a technique originally developed
by the Midwest Research Institute  (MRI) in 1974, was the basic method
used to measure emissions of particulates from various specific
activities within integrated iron and steel plants.

     Test Sites—Paved road emissions were measured at an unidentified
iron and steel plant in the Great Lakes steel-producing area of the
U.S.

     Parameters and Equipment—Table 2.31 provides a list of
parameters documented in the study and the corresponding equipment
used to  take the measurements.

     TABLE 2.31.  PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
Parameters
Wind speed
Wind direction
Cloud cover
Temperature
Relative humidity
Road surface condition
Dust loading
Dust texture
Traffic mix
Traffic count
Plume exposure
Plume particle size
distribution
Downwind concentration
Background concentration
Duration of sampling
Deposition
Equipment
Unspecified
Unspecified
Direct observation
Sling psychrometer
Sling psychrometer
Direct observation
Dry vacuuming, scales
Sieves, standard shaker, scales
Direct observation
Automatic counters
Isokinetic hi-vol samplers on
profiler tower
Hi-vol cascade impactor with cyclone
preseparator
Hi-vol sampler
Hi-vol sampler
Timer
Dustfall buckets
                                  2-82

-------
     Equipment Configuration—In each sampling run the exposure
profiler was set up 5 meters from the edge of the road.  For sites
with light-duty traffic, sampling intakes were positioned at 1, 2, 3,
and 4 meters above the ground.  For heavy-duty traffic, they were set
at 1.5, 3, 4.5, and 6 meters above the ground.  A standard hi-vol
sampler and a hi-vol cascade impactor with cyclone preseparator were
placed beside the profiler at approximately two meters above the
ground.  The wind instrument was located three meters downwind from
the edge of the road at a height of about 5 meters.  Dustfall buckets
were placed one and three meters from the road's edge.  Another
standard hi-vol sampler was placed five meters upwind from the road.

     Sampling Runs—Three, one-hour sampling runs were conducted on
paved roads.  However, emissions data were provided for only two of
these tests.  Winds were apparently too light to collect valid data
for the third test.  For the two valid tests, the number of vehicle
passes was 127 and 104; most of these were light-duty vehicles.

     Quality Assurance—The issue of quality assurance was not
directly addressed in this report.  Documentation of specific quality
assurance procedures, such as auditing filter weights, processing of
blank filters, or calibration of sampling equipment was not provided.


     The extent to which methodological procedures were documented in
this report is sufficient to permit their reproduction by other
investigators.  The manner in which exposed filters were collected and
transported to the laboratory was described in some detail.

     Findings—Exposure data and calculated emission factors for
suspended particulates  (less than 30 (am diameter) are presented in
Table 2.32.  Based on the data collected during this and a previous
study of emissions from paved roads (Cowherd et al. , 1977), a
predictive equation for paved road emissions was developed:

                   e = 0.45  (s/10)  (L/5000)  (W/3)°-8
   where

     e    =    suspended particulate emissions, Ib/veh-mile
     s    =    road surface silt content, %
     L    =    dust loading on traveled part of road, Ib/mile
     W    =    average vehicle weight, tons

The specific procedures followed in developing this equation were not
specified.  The coefficient was based on test results from traffic on
paved arterial highways with assumed values for percent silt and
surface dust loading.  The first two correction parameters,  (i.e. s/10
and L/5000) were derived from data collected in Cowherd et al.. 1977.
The (W/3)°-8 was added by analogy to the unpaved road emission factor.
No precision factor value was provided for this formula.
                                 2-83

-------
           TABLE 2.32.  EXPOSURE DATA AND  EMISSION FACTORS
Run
E-7
E-8
Sample
Height
(m)
1
2
3
4
1
2
3
4
Filter
Exposure
(mg/cm2)
0.22
0.15
0.24
0.20
0.30
0.28
0.41
0.37
Integrated
Filter
Exposure
(Ib/mile)
0.42
1 .1
Emission
Factor*
(Ib/veh-mile)
db < 30 jam
0.8
1 .1
            a) isokinetic
            b) particle Stokes' diameter
     Publication—This study was conducted and documented under a
contract with the Environmental Protection Agency; Industrial Research
Laboratory; Office of Energy, Minerals, and Industry; Research
Triangle Park, North Carolina.  It was published as Publication No.
EPA-600/2-78-050 in March 1978.

Study 5-- Cowherd, et al.   Iron and Steel Plant Open Source Fugitive
          Emission Evaluation.  EPA-600/2-79-103.  1979.

     Methodology--Exposure profiling was used to measure emissions
from paved road traffic.

     Test Sites—Two paved roads at unidentified iron and steel plants
were the sites for these measurements.

     Parameters and Eguipment—Table 2.33 lists the parameters
measured and the equipment used to measure them.

     Eguipment Configuration—The sampling configuration was different
for each of the two sites.  They are explained separately below.

     First Site—For one of the sites, the exposure profiler was set
     up four meters from the downwind edge of the road  with sampler
     heads at 1, 2, 3, and 4 meters above the ground.   A cascade
     impactor was positioned beside the profiler.  Three downwind  hi-
     vol samplers were set up, one each at 5, 20, and 50 meters from
     the road.  Two wind stations were set up, one on each side of the
     road, to monitor wind conditions at a height of four feet.  The
     upwind station was about  15 meters from the road,  and the
     downwind station was placed beside the farthest downwind hi-vol,
                                 2-84

-------
     about  20  meters  from the  road.   Upwind dust  concentration was
     measured  by a standard  hi-vol placed  beside  the  upwind  wind
     station.

     Second Site—For the other  site  a  taller  profile tower  was set  up
     four meters from the road with sampler heads at  heights of 1.5,
     3,  4.5, and 6 meters.   A  cascade impactor was placed beside the
     profiler.   Hi-vols  were placed 3 and  20 meters from the road on
     the downwind side.   Upwind  concentration  was measured at 10 to  15
     meters from the  road with a hi-vol sampler.   Two wind stations
     were set  up on the  upwind side of  the road,  one  at about 5 meters
     and the other at about  12 meters from the road.   They also
     monitored wind conditions at four  meters  above the ground.
     TABLE 2.33.   PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
Parameters
Wind speed
Wind direction
Cloud cover
Temperature
Relative humidity
Road surface condition
Road dust loading
Road dust % silt
Number of traffic lanes
Traffic mix
Traffic count
Exposure
Particle size distribution
Downwind concentration
Upwind concentration
Duration of sampling
Equipment
Recording anemometers
Unspecified
Direct observation
Sling psychrometer
Sling psychrometer
Direct observation
Vacuum/broom
200-mesh screen, shaker
Direct observation
Direct observation
Automatic counter
Isokinetic hi-vol exposure profiler
Hi-vol cascade impactor & cyclone
preseparator (directional)
Hi-vol sampler
Hi-vol sampler
Timer
     All  hi-vol  samplers and cascade impactors had sampling heights of
two meters.

     Sampling  Runs—Six sampling runs were conducted,  three at each of
the two sites.   The number of vehicle passes per run ranged from 47 to
123.   A total  of 481  passes, were logged.
                                 2-85

-------
     Quality Assurance—An effort was made to apply or adapt  the
American Society of Testing and Materials  (ASTM) Standards  in the
collection and analysis of road surface samples needed to quantify the
silt content of the surface material.  Except for the citation of this
standard procedure, the authors documented no normal quality  assurance
procedures.  However, documentation of the specific methodological
procedures in field operations, sample handling, and data analysis was
generally thorough.

     Dust samples were transported to the laboratory in separate
envelopes.  Filter samples were conditioned at constant temperature
and humidity for 24 hours before weighing.  This same procedure was
followed in weighing the filters prior to use.

     Careful attention was given to sampling under isokinetic
conditions.  The intake velocity of each sampler was set to match the
wind velocity prior to commencement of sampling.  Wind speed  was
measured continuously during sample collection.  Isokinetic correction
factors were used to adjust exposures measured under non-isokinetic
conditions.

     An effort was also made to reduce small particle bias in
characterizing the particle size distribution of the dust plume.  One
of the high volume cascade impactors was fitted with a cyclone
preseparator to reduce small particle bias caused by large particles
bouncing through the impactor stages to the back-up filter.   The
impactor/preseparator unit was calibrated to determine the  50% cutoff
diameters of the cyclone inlet and the stages of the preseparator.
The investigators found that the preseparator does eliminate  much, but
not all, of the small particle bias.

     Findings--The emission factors measured in this study are
presented in Table 2.34.

     A predictive equation was developed from the findings of  this
study and previous research by the Midwest Research Institute  (Cowherd
et al., 1977; Bohn et al., 1978).

               e = 0.090 I (4/n)  (s/10)  (L/1000)  (W/3)°-7
   where

     e    =    Mass of suspended particulates, Ib/veh-mile
     I    =    industrial road factor
     n    =    number of traffic lanes
     s    =    road surface silt content, %
     L    =    road surface dust loading, Ib/mile
     W    =    average vehicle weight

   The precision factor for this  equation when predicting the emission
   factors measured during these  three studies was 3.31.   The
                                   2-86

-------
precision factor  (f)  is defined such that the 68% confidence
interval for each  predicted emission factor  (P) is bounded by the
values P/f and Pf.

     Publication—This study was conducted and documented under a
contract for EPA,  Industrial Environmental Research Laboratory,
Office of Energy,  Minerals, and Industry, Research Triangle Park,
North Carolina.  It  was published in 1979 as Publication  No.  EPA-
600/2-79-103.

          TABLE  2.34.  EXPOSURE DATA AND EMISSION FACTORS
Run
F-13
F-14
F-15
F-16
F-17
F-18
Sample
Height
(ml
1
2
3
4
1
2
3
4
1
2
3
4
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
1 .5
3.0
4.5
6.0
Filter
Exposure*
(ma/cma)
0.24
0.17
0.16
0.16
0.18
0.12
0.10
0.08
0.12
0.04
0.02
0.07
1 .57
1 .09
0.66
0.33
1.29
0.98
0.60
0.27
0.30
0.29
0.23
0.20
Integrated
Filter Exposure
(Ib/raile)
37.2
21 .8
14.7
244
209
67.0
Emission Factors' (Ib/veh-raile)
dc < 30 (im
0.58
0.20
0.16
2.5
1 .7
0.48
d° < 5 im
0.16
0.11
0.66
0.78
0.24
0.17
      a)  net of background dust
      b)  isokinetically corrected
      c)  particle Stokes' diameter
                                2-87

-------
Study 6— Reider.   Size Specific Particulate Emission Factors for
          Uncontrolled Industrial and Rural Roads.  January 1983.

     Methodology--Exposure profiling was used to measure
emissions from paved roads in industrial settings.  An
explanation of the specific procedures for calculating emission
factors from the field data was not provided in the report.

     Test Sites--Tests were conducted on paved roads in
facilities of four different industries:  asphalt batching,
concrete batching, sand and gravel processing, and copper
smelting.  The test sites for the asphalt batching and the sand
and gravel processing facilities were one-lane roads, whereas
two-lane roads were tested for the concrete batching and the
copper smelting facilities.  Test sites were selected according
to three general criteria; each test site was found to be: 1)
suitable for the specific requirements of the exposure profiling
methodology, 2) representative of most facilities in the
industry, and 3) accessible via cooperation of the facility
personnel.

     Parameters and Equipment--The parameters measured in this
study and the equipment used to take these measurements are shown
in Table 2.35.
                              2-88

-------
   TABLE  2.35.   PARAMETERS MEASURED AND  CORRESPONDING  EQUIPMENT
Parameter
Wind speed
Wind direction
Atmospheric pressure
Temperature
Relative humidity
Road surface condition
Road surface particulate loading
Road surface silt content
Traffic Mix
Traffic Count
Vehicle weight
Vehicle speed
Plume total particulate concentration
Plume inhalable particulate concentration
Size distribution of inhalable particulate
Equipment
Warm wire anemometer
Wind vane
Barometer
Sling psychrometer
Sling psychrometer
Direct observation
Dust pan, broom, scales
Sieves, scales
Direct observation
Direct observation
Interview plant operators
Interview drivers
Exposure profiler
Hi-vol with size-selective inlet
Cascade impactor with cyclone
preseparator (impactor cutpoints
of 10.2, 4.2, 2.1, 1.4, and 0.73
jim)
     The procedures for calculating point values of exposure from
these data were not included in this report.  They are presumably
similar to those discussed in Cowherd and Englehart, 1984, which
is also reviewed in this section.

     Equipment Configuration—Downwind air sampling equipment was
set up five meters from the edge of the road.  The exposure
profiler had sampling heads at heights of 1 , 2,  3, 4,  and 5
meters.  The wind speed is measured continuously at two sampling
heights and a logarithmic distribution of the vertical wind speed
profile is assumed in setting the sampling rates for the
remaining sampler heads.  A standard hi-vol sampler, a hi-vol
fitted with a size-selective inlet  (SSI), and two cascade
impactors with cyclone preseparators were also set up on the
downwind side.  The cascade impactors were positioned in a
vertical array at heights of 1 and 3 meters.  The hi-vols sampled
air at a height of 2 meters.

     Upwind sampling equipment was also generally set up five
meters from the road.  For most of the tests, one of each of the
following samplers was set up with intakes two meters above the
                               2-89

-------
ground:  standard hi-vol, hi-vol with SSI, and cascade impactor
with cyclone preseparator.   Background dust levels at the
concrete and asphalt operations were judged to be too low to
collect sufficient mass on the different stages of the cascade
impactor, so the upwind impactor was not used in tests at these
facilities.

     Sampling Runs—A total of 13 tests, ranging in duration
between 13 and 344 minutes, were conducted on paved roads for
this study.   Sampling run time was sufficient to produce a filter
weight gain of at least 5 mg on the top sampling head of the
profiler.  These tests were distributed among the various
industries as shown in Table 2.36.

      TABLE  2.36.   EXTENT OF  SAMPLING  FOR  VARIOUS INDUSTRIES
Industry
Asphalt batching
Concrete batching
Copper smelting
Sand and gravel
processing
Number of
Tests
4
3
3
3
Total Vehicle
passes
373
372
123
47
Traffic Type
Medium Duty
Medium Duty
Medium Duty
Heavy Duty
     Quality Assurance--The investigator noted that the sampling
 and analysis procedures met or surpassed the guidelines set forth
 in the Quality Assurance Handbook for Air Pollution Measurement
 Systems, Vol. II - Ambient Air Specific Methods  (U.S. EPA, 1977)
 and the Ambient Monitoring Guidelines for Prevention of
 Significant Deterioration  (U.S. EPA, 1978).  Except for the lack
 of a description of some data reduction procedures, quality
 assurance for this study was good.  The profiler sampling heads,
 hi-vols, and impactors were calibrated prior to testing at each
 site.  Sampling filters and impactor substrates were equilibrated
 for 24 hours prior to weighing.  Tare weights were given a 100%
 audit, and 10% of loaded filter weights were audited.  Criteria
 for reweighing of the entire batch were provided.  The
 orientation of the profile sampling heads were adjusted if the
 15-minute average wind direction changed by more than 30 degrees.
 The sampling rate of the profile samplers was adjusted if the  15-
 minute average wind speed changed by more than 20%.  Ten percent
 of all calculations were also audited.

     Findings--The concentration measurements taken during the
 field tests are presented in Table 2.37.  The investigator noted
 that the data from the profiler is net of background dust levels.
                               2-90

-------
The emission factors which  were  calculated from these data were
shown in Table 2.38.  Again,  the specific procedures for
calculating the emission  factors are not described in this
report.

     Publication—This  report has been completed only as a draft
final report.  It has not been published.  However, it was cited
in the AP-42 section on paved industrial roads.  The study was
conducted under EPA Contract  No. 68-02-3158.
 TABLE 2.37.   MEASURED DUST CONCENTRATIONS  FOR TOTAL PARTICULATES
                              AND  PM10

Run
Y-1
Y-2
Y-3
Y-4
Z-1
Z-2
Z-3
AC-4
AC- 5
AC-6
AD-1
AD- 2
AD-3
Duration
(minutes)
367
443
200
192
348
313
313
76
58
74
103
71
41


entration (\ig/r
Total Particulate (Exposure Profiler)
1 m
432
411
1698
7992
1352
3214
4214
7226
3261
6746
1273
832
1065
2 m
118
223
791
2753
806
1775
2409
5893
1864
5314
1036
1173
788
3 m
99
112
562
1638
454
1364
1750
4812
1644
4144
974
600
504
4 m
a
77
355
1001
366
919
1256
2755
1007
a
720
347
b
5 m
37
b
156
490
215
641
71 1
1863
857
1524
588
177
b
n3\
" I
PM10 (Cascade Impactor)
1 m
18
58
82
292
425
520
810
2399
1567
1367
251
165
80
3 m
16
51
52
191
252
285
404
599
591
367
110
82
57
 a) torn filter
 b) net concentration resulted in negative value
                               2-91

-------
  TABLE 2.38.  EMISSION FACTORS  FOR TOTAL PARTICULATES AND PM10
Run
Y-1
Y-2
Y-3
Y-4
Z-1
Z-2
Z-3
AC-4
AC- 5
AC- 6
AD-1
AD- 2
AD-3
Total Particulate
Emission Factor
(Ib/veh-mile)
1 .43
1 .48
0.75
3.65
2.25
7.23
17.5
15.74
10.8
7.07
19.3
6.64
4.35
PM,0 Emission
Factor
(Ib/veh-mile)
0.257
0.401
0.0801
0.441
0.699
1 .63
4.01
3.86
3.13
1 .35
3.27
0.753
0.513
Study 7— Cuscino et al.  Iron and Steel Plant Open Source
          Fugitive Emission Control Evaluation.  EPA-600/2-83-
          110.  1983.

     Methodology—Exposure profiling was used to measure
emissions from vehicle travel on paved roads in iron and  steel
plants.  The method of calculating exposure at the various
profile sampler heights differs significantly from that used  in
previous Midwest Research Institute (MRI) profiling studies.
Prior to this study, exposure was calculated as the dust  mass on
the filter divided by the intake area.  For this study the
equation below was used to compute exposures for particle size
categories of interest:
where
     E
     C
     U
     t
                          E  =  1CT7 C U t
exposure (mg/cm2)  for particle size of interest
concentration (jj.g/m3)  of same particle size
mean wind speed (m/s)
sampling duration (s)
No explanation was given for this change in methodology.
                               2-92

-------
     Test Site—Tests of emissions from vehicular travel on
paved, uncontrolled roads (i.e. not sprayed with a dust
suppressant) were conducted at six different sites at Armco
Steel, Incorporated's iron and steel works plant in Middletown,
Ohio.

     Parameters and Equipment—Listed in Table 2.39 are the
parameters measured and the equipment used to take the
measurements in this study.

     Equipment Configuration—For each test all downwind
monitoring equipment were placed five meters from the edge of the
road, and all upwind samplers were placed ten meters from the
road.  Four or five isokinetic sampling heads, depending on the
expected plume height, were attached to the profiling tower at
one meter intervals.  Two hi-vol impactors with cyclone
preseparators were set up beside the profile tower one meter and
three meters above the ground.  A standard hi-vol sampler was
also placed beside the profile tower with its intake two meters
above the ground.  For some of the tests, a hi-vol with a size-
selective inlet  (SSI) was also stationed on the downwind side.

   TABLE  2.39.   PARAMETERS MEASURED AND  CORRESPONDING  EQUIPMENT
Parameter
Wind speed
Wind direction
Temperature
Road width
Number of lanes in Road
Road surface silt loading
Road surface silt content
Road surface moisture content
Traffic count
Plume exposure
Plume particle size distribution
Plume TSP concentration
Plume IP concentration
TSP background concentration
Duration of sampling
Equipment
Recording anemometer
Recording anemometer
Unspecified
Unspecified
Direct observation
Broom , vacuum
Sieves, mechanical sieving
device, scales
Oven, scales
Direct observation
Isokinetic hi-vol samplers
(profiler)
Hi-vol cascade impactor with cyclone
preseparator
(cutpoints not provided)
Standard Hi-vol sampler
Hi-vol sampler with size selective inlet
Standard Hi-vol sampler
Timer
                               2-93

-------
     A standard hi-vol sampler was placed with its intake two
meters above the ground on the upwind side for each test.  Hi-
vols with SSIs were also used on the upwind side; for some of the
tests two were set up with intakes at heights of one and three
meters, and for others, a single hi-vol/SSI was set up with its
intake at two meters above the ground.

     Sampling Runs—Eleven sampling runs were conducted on
uncontrolled, paved roads in this study.  Most samples were
collected over one or two hours, but one run lasted over four
hours.  The number of vehicle passes per run ranged from 79 to
301.  The total number of passes was 1,279.

     Quality Assurance—This study incorporated a rigorous
quality control program.  Procedures followed in collecting and
analyzing samples were documented in considerable detail.
Quality control measures were set forth for the sampling media,
sampling flow rates, and equipment maintenance.  Criteria for
interrupting sample collection were also documented.  Quality
assurance practices included processing of blank samples,
calibration of equipment, and auditing of sampling and analysis
procedures.

     The investigators note that their procedures met or
surpassed the requirements set forth in Quality Assurance
Handbook for Air Pollution Measurement Systems, Volume II -
Ambient Air Specific Methods  (U.S. EPA, 1977) and Ambient
Monitoring Guidelines for Prevention of Significant Deterioration
 (U.S. EPA, 1978).

     Careful attention was given to sampling under isokinetic
conditions.  Wind speed was monitored before and during sample
collection.  Fifteen minute averages of the wind speed at two
monitoring heights and an assumed logarithmic vertical wind speed
profile were used to set the intake velocity for each sampler on
the profiler.

     Findings--Because the purpose of this study was to evaluate
various methods of controlling open source fugitive emissions
from iron and steel plants, some experimental data was needed on
particulate emissions from vehicle travel on uncontrolled, paved
roads.  Table 2.40 presents the primary emissions data collected
for this source.

     The investigators did not estimate, in this report, a new
predictive equation on the basis of their empirical findings.
However, they did test predictions of the equation which was
developed from field data collected by MRI prior to this study
 (Cowherd et al., 1977; Bohn et al., 1978); and Cowherd et al. .
1979).  The equation is presented in the review of Cowherd
                               2-94

-------
et al. .  1979.  The precision factor of this equation when
predicting emissions for sampling runs conducted in the present
study and the study by Cowherd et al. (1979)  was 2.14.

     Publication—This study was conducted and documented for
EPA, Office of Research and Development, Industrial Environmental
Research Laboratory, Research Triangle Park,  North Carolina.  It
was published in 1983 as Publication No. EPA-600/2-83-110.

         TABLE 2.40.  EXPOSURE DATA AND EMISSION FACTORS

Run
F-27
F-32
F-34
F-35
F-45
F-57
F-58

Sample
Height (m)
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Net TP3
.Exposure (mg/cm2)
1.14
0.94
0.66
0.00
0.683
0.523
0.385
0.346
1 .24
0.82
0.66
0.42
3.18
2.02
1.12
0.00
3.44
2.50
2.01
1 .41
1 .45
1 .18
1 .39
1 .09
0.605
0.439
2.00
0.569
0.805
0.431
0.300
Emission Factors'3 (Ib/veh-mile)
Total*
Particulates
0.848
0.292
1 .73
2.18
2.75
2.86
2.90
dc < 1 5 urn
0.357
0.144
0.536
0.849
0.608
0.554
1 .08
dc < 2.5 ^m
0.106
0.0503
0.147
0.207
0.173
0.148
0.197
                               2-95
(continued)

-------
                            TABLE  2.40    (continued)

Run
B-59
B-60
B-61
B-62
Sample
Height (m)
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Net TPa
Exposure (mg/cm2)
1 .93
0.597
0.887
0.433
0.379
1 .34
1 .51
0.803
0.603
0.430
2.95
2.60
1 .97
1 .66
0.987
2.66
2.58
2.07
1 .29
0.00
Emission Factors" (Ib/veh-mile)
Total3
Particulates
2.95
3.72
4.65
3.50
dc < 15 p.m
0.993
1.18
1 .35
0.929
dc < 2.5 ^im
0.334
0.432
0.327
0.245
a)  total particulate,  i.e. including mass  in settling chamber
b)  isokinetically corrected
c)  particle aerodynamic diameter
                                          2-96

-------
Study 8— Cowherd and Englehart.  Paved Road Particulate
          Emissions:  Source Category Report.  EPA-600/7-84-077.
          1984.

     Methodology—Exposure profiling was used to measure
particulate emissions from paved roads.

     Test Sites—Tests were conducted at three sites in the
Kansas City area during the winter of 1980 and at five sites in
the St. Louis / Granite City, Illinois area the following spring.
Streets in these areas were selected on the basis of the
following criteria:

     •    traffic volume and road surface particulate mass must
          be sufficient to generate adequate filter loading
          within a four-hour test run

          space must be available for upwind and downwind
          sampling equipment and for staff

          expected wind direction must be within 45° of
          perpendicular to the road

     •    wind fetch upwind from the road should be large

     Each test site was on one of the following four road types:
commercial/industrial, commercial/residential, expressway, and
rural town.

     Parameters and Equipment—The parameters for which direct
field measurements were available (i.e. no extrapolation or
interpolation of data needed) and the equipment used to collect
the data are listed in Table 2.41.

     The concentration of particles with aerodynamic diameters
less than 2.5 (am and of those smaller than 10 [im was estimated in
the following manner:

     1 .   The measured concentrations of each of the particle
          size categories listed in Table 2.41 were converted to
          percentages of the total suspended particulate
          concentration.

     2.   These data points were graphed with particles sizes in
          logarithmic scale on the Y-axis and the corresponding
          percent of the total particulate mass smaller than the
          given size in probability scale on the X-axis.
                               2-97

-------
  TABLE  2.41.   PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
Parameter
Wind speed
Wind direction
Cloud cover
Temperature
Relative humidity
Pavement type
Road surface condition
Surface particulate loading
Surface particulate texture
Vehicle mix
Vehicle count
Plume total particulate concentration
Plume total suspended particulate concentration
Plume concentration of particles < 1 5 [im in
aerodynamic diameter
Plume concentration of particles in the
categories :
< 7.2 (j.m aerodynamic diameter
< 3.0 ^m "
< 1 . 5 urn "
< 0.95 urn "
< 0.49 urn "
Equipment
Warm wire anemometer
Wind vane
Direct observation
Sling psychrometer
Sling psychrometer
Direct observation
Direct observation
Dry vacuum, scales
Sieves, scales
Direct observation
pneumatic tube axle counters
Profiler
Standard high volume sampler
High volume with size-
selective inlet
Cascade impactor with
greased substrates attached
to the above hi-vol sampler
     3.
Graphic interpolation is used to derive percentages  of
the total particulate mass smaller than 2.5  ^im  and 10
Urn.
     The equation below was used to compute point  values of
exposure at the sampling heads for any of  the  particle size
categories:
where
     E
     C

     U
     t
                          E = 1CT7 C U t
     exposure  for particle size of interest,  mg/cm2
     concentration ((j,g/m3)  (measured or interpolated)
     of same particle size, |o,g/m3
     mean wind speed, m/s
     sampling duration, s
                               2-98

-------
     Equipment Configuration—The sampling equipment
configuration varied in some details over the course of the
study.  The basic arrangement is described below, and the
components which did change are described in the following
paragraphs.

     The downwind configuration was the same throughout the
study.  The exposure profiler, which was placed about 2.5 meters
from the downwind edge of the road, consisted of four isokinetic
sampling heads fixed at heights of one, two, three, and four
meters.  A standard hi-vol sampler and two hi-vols with attached
size-selective inlets  (SSI) and cascade impactors were placed
beside the profile tower at respective heights of two, one, and
three meters.  Also in every test, an upwind standard hi-vol
sampler was positioned about four meters from the road with its
intake at the usual height of two meters.

     For the tests in the Kansas City area, two hi-vols with SSIs
were set up at heights of two and four meters.  For some of the
tests in the St. Louis/Granite City area, a single upwind hi-vol
fitted with attached SSI and cascade impactor was set up 2 meters
above the ground.  For the remainder of the tests, two upwind hi-
vols with SSIs  (but no cascade impactors) were placed at one and
three meters above the ground.  All upwind air samplers were
about four meters from the road.

     Sampling Runs—Nine sampling runs were conducted at the
three sites in the Kansas City area.  Seven of these runs passed
the investigators' quality control criteria (see discussion under
Quality Assurance).  Of the ten tests conducted in the St.
Louis/Granite City area, only three passed the quality control
criteria.  Runs which failed these criteria were not included in
the multiple regression analysis.  Most samples were collected
over a period of two to four hours, but one rural road test
lasted almost six hours.  The number of vehicle passes per
sampling run ranged from about 1,900 to about 15,000.

     Quality Assurance—Quality assurance was thorough and well
documented.  Calibration schedules and acceptable variations were
presented for air samplers and laboratory balance.  Exposed
filters were conditioned for 24 hours prior to weighing.  Some of
each type of filter or substrate were processed as blanks to
determine necessary corrections for the effects of filter
handling.  Filter weights were audited regularly.  To assure
isokinetic sampling, the orientation of the intake direction was
adjusted when the 15-minute average wind direction changed by
more than 30°, and the sampling velocity was adjusted any time
the wind speed  (15 minute average) changed by more than 20%.
These quality assurance procedures met or surpassed applicable
requirements found in EPA's Quality Assurance Handbook for Air
Pollution Measurement Systems  Volume II - Ambient Air Specific
Methods (U.S. EPA, 1977), and Ambient Monitoring  Guidelines  for

                               2-99

-------
Prevention of Significant Deterioration  (U.S. EPA,  1978).   In
addition,  the investigators specified the following conditions
under which sampling results were excluded from regression
analysis:

          Mean angle between wind direction and profiler
          orientation > 20°

          Mean angle between profiler orientation and
          road orientation > 45°

          Wind speed < 4 mph

     •    Background concentration relative to downwind
          measurements deemed "acceptable"

     •    Results not based on average of data from other
          runs

     Findings--Except for a sample calculation in an appendix, data
were not published on concentrations or exposures at individual
profile sampler heads.  The calculated emission factors for those
sampling runs which were included in a multiple regression analysis
are listed in Table 2.42.

              TABLE 2.42.  CALCULATED EMISSION FACTORS

Site
M-1
M-2
M-3
M-9
M-6
M-7
M-1 5
M-11
M-1 2
M-8
Road Type
Commercial/
Industrial
Commercial/
Industrial
Commercial/
Industrial
Commercial/
Industrial
Commercial/
Residential
Commercial/
Residential
Commercial/
Residential
Expressway
Expressway
Rural Town
Emission Factors (Ib/veh-mile * 10*) by
Particle Size Category
da < 1 5 ^m
125.0
35.7
84.8
99.3
32.9
117.0
35.8
7.8
2. 1
311 .0
da < 1 0 (im
110.0
34.0
78.1
71 .2
30.4
92.8
32.3
7.0
1 .9
247.0
da < 2.5 urn
63.2
30.4
52.0
40.5
20.9
36.8
22.0
3.4
1 .4
50.4
        a) aerodynamic diameter
                               2-100

-------
     As can be seen from the Table 2.42, the  expressway had the
lowest emissions, and the rural town road had the  highest
emissions.

     The fraction of the total suspended particulate mass  in
particles smaller than 15 \im was found  to be  greater on the
upwind side of the road than the downwind side.  The same  was
true for PM10 and fine particulates.

     Stepwise linear regression was used to build  predictive
models for emissions of total suspended particulates  (TSP),
inhalable particulates  (IP), PM10/  and fine particulates (FP).
The candidate predictor variables were  total  loading  (g/m2) , silt
loading (g/m2) ,  average vehicle speed (kph),  and average vehicle
weight (Mg).  The resulting equations are shown below:
     'TSP

     -IP
      -PM10

      'FP
where
     sL
5.87 (sL/0.5)0-9
2.54 (sL/0.5)0-8
2.28 (sL/0.5)0-8
1.02 (sL/0.5)0-6
Emission factor for particle size category i,
g/veh-km
Silt loading, g/m2
Speed was not selected as a predictor variable because of  its
high correlation with silt loading.  The precision  factors for
these equations when predicting emissions  in  the  ten  sampling
runs which passed the QA screening are, respectively, 2.4,  2.0,
2.2, and 2.2.

     Publication—This study was published as Publication  No.
EPA-600/7-84-077 in July of 1984.  It was  conducted and
documented under contract for the Environmental Protection
Agency, Industrial Environmental Research  Laboratory, Research
Triangle Park, North Carolina.
                               2-101

-------
Studies of Secondary Importance

Study 9— Roberts.  The Measurement, Cost and Control  of  Air
          Pollution from Unpaved Roads and Parking Lots in
          Seattle7s Duwamish Valley. 1973.

     The dust plume created by a vehicle traveling on  a paved
road was sampled using a towed rack to which a cascade impactor
was attached.  The rack was designed as a vertical grid oriented
perpendicular to the car's path; the impactor was rotated among
the various positions between tests so that, after a series of
tests, the average concentration of dust in the plume  could be
determined.  In order to derive an emission factor in  Ib/veh-
mile, the average plume concentration was multiplied by the
volume of air into which it was emitted.  This volume  was
estimated in the following manner.  First, the area of the plume
behind the car was estimated by towing a grid/rack of  open
impaction plates and'examining the dust pattern on the plates.
Second, this area, 70 square feet, was multiplied by 5,280 feet
 (1 mile) to obtain the air volume  (36,960 ft3)  into which the
dust was emitted after 1 mile of travel.

     The impactor samples were also analyzed to determine the
particle size distribution of the plume.  This particle size
breakdown was then used to estimate emission factors for  various
particle size categories.

     Sample collection, handling, and analysis procedures were
not documented in detail.  For example, specific procedures for
handling the impactor plates were not documented.  No  mention was
made of problems with particle bounce in the impactor; this
problem has been documented by at least two other researchers
 (Cowherd et al.,  1974 and McCaldin, 1977).

     The distance between the vehicle and the towed rack  was not
documented.  The investigator noted that it was difficult to
certify that sampling was conducted under isokinetic conditions,
due to turbulence in the wake behind the vehicle and changes in
wind direction and speed.

     An unspecified number of sampling runs were conducted on two
paved roads in the Duwamish Valley in Seattle.  One of the roads
had curbs and was regularly swept, and the other had no curbs and
was visibly dusty.  Following the procedure explained  above,
total emissions were measured at an average of 0.14 Ib/veh-mile
for the clean road and 0.83 Ib/veh-mile for the dusty  road.  For
particles smaller than 10 (am, emissions were measured  at  0*0055
Ib/veh-mile and 0.17 Ib/veh-mile for the clean and dusty  road,
respectively.  A factor for particles with diameter less  than'two
^m was also measured for the dusty paved road: 0.022 Ib/veh-mile
The investigator did not explain how the fraction of the  total
particulate sample consisting of particles smaller than 10 ^m was

                              2-102

-------
determined; i.e. he did not indicate whether the cascade impactor
provided a cut point at that size.

     This study was conducted and documented as partial
fulfillment of requirements for a master's degree in engineering
at the University of Washington.  Portions of this work and some
follow-up research were published in the Journal of the Air
Pollution Control Association in 1975.

Study 10—     Axetell and Zell.  Control of Reentrained Dust
               From Paved Streets.  EPA-907/9-77-007.  1977.

     Of the several field studies which were documented in this
publication, two were considered secondary in importance to the
development of reliable emission factors for vehicle travel on
uncontrolled paved roads.  They were carried out in support of
the primary purpose of this report, which was to document methods
of controlling dust emissions from paved roads.  One involved
measurement of TSP concentrations near the mud carry-out area
associated with a building construction site, and in the other
zinc sulfide dust was distributed in the same mud carry-out area
to trace the distance mud is tracked away from a construction
site.  These two studies are described briefly below.

     In the first study, four high volume samplers were set up
adjacent to the mud carry-out area associated with a single
access point for a construction site in Kansas City.  Separate
mean concentrations were calculated and presented for each of
four different road cleaning programs.  One of these was
essentially no control at all.  Mean concentrations for the four
cleaning programs were also compared with the mean TSP
concentrations of 15 regional monitoring sites.  This regional
average was used as a measured of background TSP concentration.
For the no-control tests, the concentrations were roughly 20 to
40 ng/m3 higher than the regional background level of 84.1
Corresponding emission rates were not estimated from these
concentration measurements.

     The second study was conducted at this same site.  Zinc
sulfide (ZnS)  mixed with sand was sprinkled on the mud track-out
area.  Samples of street dust were collected at several distances
from the site entrance one day and again eight days after the ZnS
was applied.  On the first day after the application, increased
ZnS levels were detected along 1500 feet of the access road.  On
the eighth day higher than normal levels of ZnS were detected up
to 2000 feet from the entrance.

     No indication was given that any normal quality assurance
procedures were followed for either of these studies.  The
authors acknowledged that concentration measurements could have
                              2-103

-------
been effected by activities within the construction site itself.
The duration of each hi-vol sample was not documented.  Raw data
were not published.

     Preparation of this document, and the supporting field work
described above, were sponsored by the Environmental Protection
Agency, Region VII - Air Support Branch, Kansas City, Missouri.
It was published as Publication No. EPA-907/9-77-007 in July of
1977.

Study 11--     McCaldin.  Fugitive Dust Study for Pima County Air
               Quality Control District, Tucson, Arizona.  1977.

     The investigator used upwind-downwind dispersion modeling to
measure emissions from vehicular traffic on paved roads.  Two
types of paved roads were tested:  one with dirt shoulders and no
curbs, and one with curbs.  Upwind and downwind dust
concentrations were measured using one standard high volume
sampler on each side of the road, 50 feet from the centerline.

     The road without curbs had four lanes and about 45 feet of
shoulder on either side.  Traffic on the paved lanes varied from
500 to 700 vehicles per hour, and vehicles using the shoulders
passed at a rate of 6 to 12 per hour.  Thus,  traffic on the
unpaved portion ranged approximately from 0.8% to 2.3% of the
total traffic volume on the roadway.  Six tests, each lasting
between three and four hours, were conducted at this site over a
period of 21 days.  Calculated emission factors, as well as
measured upwind and downwind concentrations,  were documented in
this report.  Emission factors for the uncurbed road were
calculated at .003, .006, .007, .022, .026 and .068 Ib/veh-mile.

     The three curbed road test sites had traffic volumes of
about 650, 130, and 220 vehicles per hour.  Each test road had
identifiable sources of road dust nearby.  A total of seven
tests, each lasting from three to five hours,  were conducted on
these three roads.  One emission factor was reported for each
curbed road test site:  0.004, 0.02, and 0.05 Ib/veh-mile.
Upwind and downwind concentrations at this site were not
published.

     No quality assurance procedures were documented for the
paved roads field research.  Details regarding the manner in
which filters were tared, transported, and equilibrated (if at
all), were not given.   The measured angle between the wind and
the road,  a key parameter for the model, was not published.

     This study was conducted and documented under contract with
the Pima County Air Quality Control District (AQCD).   EPA
Financial Grant Number A0090055-77-2, awarded to the Pima County
AQCD, provided funding for the study.


                              2-104

-------
MINING

Introduction

     Mining and quarrying activities are significant sources of
fugitive dust in some areas.  Four studies have been conducted to
measure emission rates in this source category, two at western
surface coal mines and two at quarries.  In the first mining
study  (Axetell, 1978), upwind-downwind dispersion modeling was
used to estimate emission rates for several activities at five
strip mines.  In the other mining study  (Axetell and Cowherd,
1984), the different mining activities were tested using the
upwind-downwind method, the quasi-stack method, or the exposure
profiling method, depending on the character of the source.  In
both of the quarry studies, emissions were estimated using the
upwind-downwind method.  These four studies are reviewed in
chronological order.

     Two additional field studies involved measurement of dust
concentration and other parameters in western strip mines.
Because they did not measure emission rates, they are considered
secondary in importance.  In one of these  (Cook et al., 1980),
the relationship between dust concentration and other variables,
including activity type and intensity near the sampler, was
examined statistically-  In the other  (Marple et al., 1980), the
performance of a mobile air sampling and analysis vehicle was
documented, and the findings from the collected data were
discussed.  These two studies are summarized following the review
of the primary studies.
                               2-105

-------
Studies of Primary Importance
Study 1— Axetell.   Survey of Fugitive Dust from Coal Mines.
          EPA-908/1-78-003.  1978.

     Methodology—The investigators used upwind-downwind
dispersion modeling to measure emission factors for several
mining operations:   topsoil removal, drilling, blasting,
dragline, shoveling/truck loading, and fly-ash dumping.  The
following area source model, from Turner, 1970, was used:
               X -
           exp
                                  y
exp
where
     X

     Q
     u
     Y

     H
plume centerline concentration at a distance x
downwind from the source,  g/m3
source strength, g/sec
the standard deviation of the horizontal
distribution of the plume concentration at a
distance x downwind from the source, m
the standard deviation of the vertical
distribution of the plume concentration at a
distance x downwind from the source, m
mean wind speed, m/sec
horizontal distance from the sampler to the plume
centerline, m
average vertical distance from plume centerline to
samplers, m
     An effort was made to measure particle fallout rates.
Concentrations at a series of downwind distances from the source
were measured, and corresponding emission rates at the source
were calculated using the above model, which assumes there is no
particle fallout.  Decreases in the "apparent emission rate" with
increasing distance from the source would serve as a measure of
fallout.

     Test Sites—Operations at five western coal mines were
tested.  Except for one lignite mine, all of the operations
extracted sub-bituminous coal.

     Parameters and Equipment—Most of the parameters measured in
this study are listed in Table 2.43, along with the tools used to
collect the data.  In addition, estimates were made for the
initial dimensions and dispersion of the dust plume, the
receptor's distance from the source, the receptor's vertical and
horizontal distance from the plume centerline, and the length of
time the receptor is in the plume.  These estimates were all made
visually.
                              2-106

-------
  TABLE  2.43.   PARAMETERS  MEASURED  AND CORRESPONDING EQUIPMENT
Parameters
Upwind concentration of TSP
Downwind concentration of TSP
Wind direction
Wind speed
Other atmospheric stability parameters
Particle size distribution
Equipment
Standard hi-vol
Standard hi-vol
Recording wind instrument
Recording wind instrument , hand-
held wind speed anemometer
Unspecified
Millipore filters on nuclepore
filter holder and pump, microscope
     Equipment Configuration—The equipment configuration varied
between sources and mines.  It was not completely described for
each test run.  Typically, a pair of hi-vols were placed together
at a location upwind from the entire mining operation.  This was
preferred over placing samplers immediately upwind from the
activity because anticipated brief wind direction reversals would
be less likely to effect upwind concentration measurements.
Other hi-vols were placed at 10, 20, and 30 meters (or a similar
series of distances) downwind from a particular source activity.
For about half of the tests, downwind samplers were set up at
both 1.2 and 2.4 meters above the ground at these downwind
locations to provide information on the vertical dispersion of
the plume.

     Sampling Runs—The number of sampling runs conducted on each
activity is indicated in Table 2.44.  Each sampling run consisted
of several concentration measurements, which ranged in duration
primarily between 30 and 90 minutes.

  TABLE  2.44.  EXTENT  OF  SAMPLING AT  VARIOUS  MINING  ACTIVITIES
Activity
Topsoil removal
Drilling
Blasting
Dragline
Number of
Sampling Runs
10
5
13
30
Activity
Shovel /Truck
loading
Front -end loader
Train loading
Total
Number of
Sampling Runs
26
1
9
89
                              2-107

-------
     Quality Assurance—The guidelines in the Quality Assurance
Handbook for Air Pollution Measurement Systems (U.S. EPA, 1976)
were followed in preparing filters, collecting and analyzing
samples, and auditing the data.  Hi-vol samplers were calibrated
before field work began at each mine.  One of every 25 filters
was treated as a blank.

     The investigators listed several problems experienced in the
field sampling program which indicate a decrease in the
reliability of the field data.  Foremost was the fact that much
of the data collected in the field was subject to the ability of
the field staff to make visual estimates.  Plume dimensions and
distance from the plume centerline are two examples of this.  In
addition, one avoidable problem greatly decreased the integrity
of the data:  the two separate field crews did not follow the
same procedures in collecting samples in several critical
respects.  There was no way to fully correct the data for these
differences.

     Findings—The measured concentrations and other field
sampling data are shown in Table 2.45.  Emission factors for
these sampling runs were presented separately in the report.
Average emission factors by operation and mine are shown in Table
2.46.

     The empirical data collected in this study did not support
the supposition that the apparent emission rate decreases with
distance due to particle fallout.  For those sampling runs which
included concentration measurements at two consecutive downwind
distances at the same height, the modeled apparent emission rates
decreased with increasing distance in only about 35% of the
sampling runs.  For those tests in which concentrations were
measured at downwind distances differing by 10 meters, the
concentration increased an average of 19% between the two
sampling points.

     When measured at two heights, 1-4 meters and 2.4 meters, at
the same downwind distance, the concentrations at the lower
height averaged 14% higher than those at the samplers 2.4 meters
above the ground.

     Publication—This study was conducted and documented under a
contract for the Environmental Protection Agency, Region VIII,
Office of Energy Activities, Denver, Colorado.  It was published
in February of 1978 as Publication No. EPA-908/1-78-003.
                              2-108

-------
TABLE 2.45.  DISPERSION MODELLING DATA
Source
Drag-
line












Mine/
Sample
A/1
A/ 2
A/ 3
A/4
A/ 5
A/ 6
A/7
A/8
B/1
B/2
B/3
B/4
B/5
Wind
Speed
(m/sec)
.4
.4
.4
.4
1.8
1.8
.4
.4
3.6
3.6
5.8
5.8
5.4
Stability
Class
B
B
B
B
B
B
B
B
C
C
D
D
D
Back-
ground
Concentr .
((ig\m5)
88
88
88
88
64
64
64
64
131
131
131
131
131
Net
Plume
Concentr.
(ng/mj)
1476
825
1376
1324
1247
1020
1658
1145
1234
500
303
322
408
337
416
466
337
305
954
828
1155
795
351
990
569
944
414
272
258
174
290
334
38
52
83
47
0
51
20154
18010
15633
21087
3935
20542
24846
19034
17520
23216
27025
26753
7429
22000
25867
21860
26017
7476
7594
18664
17497
Downwind
Distance
(m)
30
40
50
30
40
50
30
40
50
50
60
70
50
60
70
50
60
70
40
55
70
40
55
40
55
70
75
90
105
75
90
105
75
90
105
75
90
105
50
65
80
50
65
80
50
65
80
70
80
90
70
80
90
70
80
90
70
85
100
Vertical
Distance
from Plume
Center line
(m)
5.0
5.0
5.0
6.3
6.3
6.3
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
6.3
6.3
6.3
-3.0
-3.0
-3.0
-1.8
-1 .8
-3.0
-3.0
-3.0
11.0
11 .0
11 .0
12.2
12.2
12.2
11.0
11.0
11.0
12.2
12.2
12.2
-1.3
-1 .3
-1 .3
-.1
-.1
-.1
-1 .3
-1.3
-1.3
-1.3
-1.3
-1 .3
-.1
-.1
-.1
-1.3
-1 .3
-1 .3
-1.3
-1.3
-1 .3
Horizontal
Distance
from Plume
Centerline
(m)
5.5
8.9
12.3
4.5
7.9
11 .3
3.5
6.9
10.3
11 .3
16.7
23.6
9.3
14.7
21 .6
10.3
15.7
22.6
3.0
3.0
3.0
2.0
2.0
1.0
1.0
1.0
46.0
55.0
64.0
45.0
54.0
63.0
46.0
55.0
64.0
45.0
54.0
63.0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
                 2-109
(continued)

-------
TABLE 2.45.   (continued)



Source
Drag-
line



























































Mine/
Sample


B/6





B/7




B/8



B/9




B/10





C/1




C/2



C/3




C/4




C/5




C/6



D/1




Wind
Speed
(m/sec)


5.4





3.1




3.1



3.6




3.6





3.6




3.6



4.0




4.0




5.4




5.4



6.3




Stability
Class


D





C




C



C




C





B




B



B




B




C




C



D



Back-
ground
Concentr.
(ngXm1)


131





125




125



125




125





89




89



89




89




89




89



94



Net
Plume
Concentr.
((ig/mj)
7722
12809
18765
7197
12686
14131
5184
3304
5222
6564
4899
3681
5848
4574
4449
4769
12853
3499
3509
3619
4393
3266
3396
3090
3982
3180
96
0
210
125
133
89
37
78
127

208
262
287
72
12
170
113
100
161
128
139
236
297
323
144
217
241
287
394
289
1475
1585
1577
1631


Downwind
Distance
(m)
70
85
100
70
85
100
70
80
90
70
80
90
70
80
90
100
70
80
90
100
70
80
90
70
80
90
70
70
79
79
87
87
70
79
87
96
70
79
87
70
79
87
70
79
87
96
70
79
87
70
79
87
70
79
87
96
75
84
92
100
Vertical
Distance
from Plume
Centerline
(ra)
-1 .3
-1 .3
-1 .3
-.1
-.1
-.1
-3.8
-3.8
-3.8
-2.6
-2.6
-2.6
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-2.6
-2.6
-2.6
-6.5
-5.0
-6.5
-5.0
-6.5
-5.0
-6.5
-6.5
-6.5
void
-6.5
-6.5
-6.5
-5.0
-5.0
-5.0
-6.5
-6.5
-6.5
-6.5
-6.5
-6.5
-6.5
-5.0
-5.0
-5.0
-6.5
-6.5
-6.5
-6.5
-3.8
-3.8
-3.8
-3.8
Horizontal
Distance
from Plume
Centerline
(ra)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
o
0
          2-110
(continued)

-------
TABLE 2.45.   (continued)
Source
Drag-
line




Shovel/
Truck
Loading







Mine/
Sample
D/2
D/3
D/4
D/5
D/6
A/1
A/2
A/ 3
A/4
A/5
A/6
B/l
B/2
Wind
Speed
(in/sec)
7.2
7.2
7.2
6.3
5.8
.5
.5
.4
.4
1.3
1 .3
.6
.6
Stability
Class
D
D
D
D
D
B
B
B
B
A
A
B
B
Back-
ground
Concentr.
(ng\mj)
94
94
101
101
101
88
88
88
88
87
87
153
153
Net
Plume
Concentr.
(ug/m1)
898
1498
1095
1389
763
1258
1047
1062
608
610
808
911
892
914
1128
1821
925
1078
1082
838
672
1475
668
756
660
1273
583
786
643
618
980
705
528
3104
1217
790
1786
1965
1477
3528
1916
1227
4135
2449
1699
1399
1940
1474
1562
1268
5276
3706
2766
2149
4924
3910
2662
4763
3712
2576
Downwind
Distance
(m)
75
75
84
84
92
92
75
84
92
100
75
75
84
84
92
92
75
84
92
100
75
84
92
100
30
45
60
30
45
60
30
45
60
15
30
45
15
30
45
15
30
45
16
26
36
46
16
26
36
46
35
45
55
65
35
45
55
35
45
55
Vertical
Distance
from Plume
Centerline
(m)
-3.8
-2.6
-3.8
-2.6
-3.8
-2.6
-3.8
-3.8
-3.8
-3.8
-3.8
-2.6
-3.8
-2.6
-3.8
-2.6
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-3.8
-2.3
-2.3
-2.3
-1 .1
-1.1
-1 .1
-2.3
-2.3
-2.3
-2.3
-2.3
-2.3
-2.3
-2.3
-2.3
-1 .1
-1.1
-1 .1
-2.3
-2.3
-2.3
-2.3
-2.3
-2.3
-2.3
-2.3
-2.8
-2.8
-2.8
-2.8
-2.8
-2.8
-2.8
-1.6
-1 .6
-1 .6
Horizontal
Distance
from Plume
Centerline
(m)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-3.0
-3.0
-3.0
-2.0
-2.0
-2.0
-1 .0
-1 .0
-1 .0
-1 .0
-1 .0
-1 .0
-3.0
-3.0
-3.0
-2.0
-2.0
-2.0
0
0
0
0
0
0
0
0
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
          2-111
(continued)

-------
TABLE 2.45.   (continued)


Source
Shovel/
Truck
Loading


































Shovel/
Truck
Loading -
Coal


















Mine/
Sample

B/3



B/4




B/5



B/6



C/1




C/2




C/3




C/4




E/l




E/2




E/3




E/4




Wind
Speed
(m/sec)

.4



.4




np '



np



3.6




3.6




3.6




3.6


2.5






2.5




2.3




2.3




Stability
Class

B



B




np



np



C




C




C




C




B




B




B




B



Back-
ground
Concentr.
(fig\m>)

153



153




np



np



59




59




59




59




64




64




64




64



Net
Plume
Concentr.
((ig/m])
1449
2607
2185
3337
2194
1806
3907
2223
2096
1859
1569
2351
1743
1470
554
1466
1739
1410
2450
1516
341
320
293
296
287
183
184
244
269
290
226
130
192
134
238
164
141
120
44
154
1191
1031
1059
655
954
311
652
1163
911
537
855
822
806
817
859
762
880
558
1061
735


Downwind
Distance
(m)
30
40
50
30
40
50
30
40
50
60
40
50
60
40
50
60
40
50
60
70
50
50
59
59
67
67
50
59
67
76
50
50
59
59
67
67
50
59
67
76
20
20
28
28
37
37
20
28
37
46
20
20
28
28
37
37
20
28
37
46
Vertical
Distance
from Plume
Center line
(m)
-2.8
-2.8
-2.8
-1 .6
-1 .6
-1 .6
-2.8
-2.8
-2.8
-2.8
1 .7
1 .7
1 .7
2.9
2.9
2.9
1 .7
1 .7
1 .7
1 .7
-3.8
-2.6
-3.8
-2.6
-3.8
-2.6
-3.8
-3.8
-3.8
-3.8
-3.8
-2.6
-3.8
-2.6
-3.8
-2.6
-3.8
-3.8
-3.8
-3.8
6.3
5.1
6.3
5.1
6.3
5.1
6.3
6.3
6.3
6.3
6.3
5.1
6.3
5.1
6.3
5.1
6.3
6.3
6.3
6.3
Horizontal
Distance
from Plume
Centerline
(m)
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5

b
b
b
b
b
b
b
b
b
1 .0
0
1 .0
0
1 .0
0
1 .0
1 .0
1 .0
1.0
1 .0
0
1 .0
0
1 .0
0
1 .0
1 .0
1 .0
1 .0
1 .0
0
1 .0
0
1 .0
0
1 .0
1 .0
1 .0
1 .0
1 .0
0
1 .0
0
1 .0
0
1 .0
1 .0
1 .0
1 .0
          2-1 12
(continued)

-------
TABLE 2.45.  (continued)



Source
Shovel/
Truck
Loading -
Over-
burden
































Over-
burden
Blast


















Mine/
Sample



E/1






E/2





E/3





E/4





E/5





E/6


A/1


C/1




C/2





E/1




Wind
Speed
(m/sec)



3.6






3.6





3.1





3.1





2.7





2.7


2.4


3.6




3.6





3.7





Stability
Class



B






B





B





B





B





-B


B


B




B





C



Back-
ground
Concentr .
(figW)



112






112





112





112





112





112


88


61




61





77



Net
Plume
Concentr .
(ug/m1)
2569
2510
1232
2426
1361
1825
679
317
2255
2633
2676
2668
646
4572
2298
3406
2263
2841
2802
262
205
4117
2369
4427
3231
308
2739
2106
1548
1839
1172
1558
267
240
2531
2053
2730
2178
403
5340
3222
2002
9085
8799
5782
5751
9930
7810
6297
4503
5924
5531
1094
1502
1359
740
394


Downwind
Distance
(m)
20
20
25
25
30
30
142"
142"
20
25
30
35
142"
20
20
25
25
30
30
142"
142"
20
25
30
35
142b
22
22
28
28
33
33
158b
158"
22
28
33
39
158"
100
110
120
30
39
47
56
30
30
39
39
47
47
74
81
89
96
197"
Vertical
Distance
from Plume
Centerline
(m)
-1 .3
-.1
-1 .3
-.1
-1.3
-.1
-1 .3
-1.3
-1.3
-1 .3
-1 .3
-1 .3
-1 .3
-1 .3
-. 1
-1 .3
-.1
-1 .3
-.1
-1 .3
-1 .3
-1 .3
-1.3
-1.3
-1.3
-1 .3
-1.3
-.1
-1.3
-.1
-1.3
-.1
-1 .3
-1 .3
-1 .3
-1 .3
-1 .3
-1 .3
-1 .3
-47.6
-47.6
-47.6
-3.8
-3.8
-3.8
-3.8
-3.8
-2.6
-3.8
-2.6
-3.8
-2.6
-16.3
-16.3
-16.3
-16.3
-16.3
Horizontal
Distance
from Plume
Centerline
(m)
2.7
3.7
4.1
5.1
5.6
6.6
16.0
41 .0
4.7
6.1
7.6
9.0
66.0
1.2
2.2
2.2
3.2
3.4
4.4
5.0
30.0
3.2
4.2
5.4
6.4
55.0
6.3
7.3
8.6
9.6
11 .0
12.0
41 .0
66.0
8.3
10.6
13.0
15.3
91 .0
0
0
0
0
0
0
0
0
0
0
0
0
0
29.0
31 .0
34.0
36.0
78.0
          2-113
(continued)

-------
TABLE 2.45.  (continued)




Source
Over-
burden
Blast





Coal
Blast



































Drilling


















Mine/
Sample



E/2






B/1




B/2



C/1




C/2



D/1



D/2




E/1




E/2




A/1





C/1




C/2




Wind
Speed
(ra/sec)



3.7






3.0




3.0



5.4




5.4



4.0



4.0




2.6




2.6




.9





3.6




3.6





Stability
Class



C






B




B



C




C



B



B




B




B




B





C




C



Back-
ground
Concentr.
((ig\mj)



77






153




153



89




89



115



115




128




128




88





89




89



Net
Plume
Concentr .
^g/m5)
1327
1064
1097
1265
1323
616
479
445
.
66611
76174
5448
50274
67913
59125
67093
74570
3079
1137
2721
3307
2669
2189
2967
3156
2381
2254
1186
668
1149
733
1340
1004
810
628
469
2289
2194
1456
1851
1939
1587
2627
3347
2203
2485
146
292
247
29
39
-
461
244
403
222
274
129
269
171
175
214


Downwind
Distance
(m)
74
81
89
74
81
89
197"
197"
11
22
33
11
22
33
11
22
33
111
111
121
121
130
130
111
121
130
139
100
100
109
109
117
100
109
117
126
200
208
217
200
208
217
100
108
117
126
16
26
36
16
26
36
6
6
15
15
24
24
6
15
24
34
Vertical
Distance
from Plume
Center line
(m)
-16.3
-16.3
-16.3
-15.1
-15.1
-15.1
-16.3
-16.3
0
0
0
0
0
0
0
0
0
-3.8
-2.6
-3.8
-2.6
-3.8
-2.6
-3.8
-3.8
-3.8
-3.8
-13.8
-12.6
-13.8
-12.6
-13.8
-13.8
-13.8
-13.8
-13.8
-3.8
-2.6
-3.8
-2.6
-3.8
-2.6
-13.8
-13.8
-13.8
-13.8
.2
.2
.2
1 .4
1 .4
1 .4
1.0
2.2
1 .0
2.2
1.0
2.2
1.0
1 .0
1.0
1 .0
Horizontal
Distance
from Plume
Centerline
(m)
24.0
26.0
29.0
19.0
21 .0
24.0
64.0
53.0
0
0
0
0
0
0
0
0
0
-.5
.5
-.5
.5
-.5
.5
1.5
1 .5
1 .5
1 .5
0
0
0
0
0
0
0
0
0
43.0
44.0
46.0
42.0
43.0
45.0
21 .0
23.0
25.0
27.0
c
c
c
c
c
c
.3
1 .3
4.2
5.2
8.2
9.2
2.3
6.2
10.2
14.2
          2-114
(continued)

-------
TABLE 2.45.  (continued)




Source
Drilling









Train
Loading










































Topaoil
Renoval
(scraper)






Mine/
Sample

E/1




E/2





C/1




C/2




C/3




C/4



E/1



E/2




E/3





E/4




E/5




D/1





Wind
Speed
(m/sec)

4.1




4.1





4.9




4.9




4.5




4.5



11. d.



11. d.




ii. d.





n.d.




ii. d.




5.8






Stability
Class

C




C





B




B




B




B



n.d.



n.d.




11. d.





n.d.




ii. d.




C




Back-
ground
Concentr .
(ng\m3)

676




676





89




89




89




89



28



28




28





43




43




158




Net
Plume
Concentr.
(ng/m1)
'2723
1862
1127
717
2116
1049
255
1307
132
0
457
167
50
48
221
11
77
172
103
25
404
186
158
119
328
220
115
132
76
84
3136
1474
943
1919
1416
2050
383
1041
2047
2022
1340
1155
1041
1223
281
372
127
105
123
117
329
254
263
150
1704
2310
1583
466
2035
1390


Downwind
Distance
(m)
5
14
22
30
5
14
22
5
14
22
12
12
21
21
29
29
12
21
29
38
12
12
21
21
29
29
12
21
29
38
10
20
30
40
10
20
30
40
10
10
20
20
30
30
10
10
20
20
30
30
10
20
30
40
30
34
38
30
34
38
Vertical
Distance
from Plume
Centerline
(m)
.7
.7
.7
.7
.7
.7
.7
1 .9
1 .9
1 .9
-3.4
-2.2
-3.4
-2.2
-3.4
-2.2
-3.4
-3.4
-3.4
-3.4
-3.4
-2.2
-3.4
-2.2
-3.4
-2.2
-3.4
-3.4
-3.4
-3.4
-1 .3
-1 .3
-1 .3
-1 .3
-1 .3
-1 .3
-1.3
-1 .3
-1.3
-.1
-1.3
-.1
-1 .3
-.1
-1.3
-.1
-1.3
-.1
-1.3
-.1
-1 .3
-1 .3
-1 .3
-1 .3
-.3
-.3
-.3
.9
.9
.9
Horizontal
Distance
from Plume
Centerline
(m)
1.0
1 .0
1.0
1 .0
1 .0
1 .0
1 .0
0
0
0
1 .0
2.0
1 .0
2.0
1 .0
2.0
3.0
3.0
3.0
3.0
1.0
2.0
1 .0
2.0
1.0
2.0
3.0
3.0
3.0
3.0
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
np
np
np
np
np
np
          2-1 15
(continued)

-------
                          TABLE 2.45.    (continued)




Source
Topsoil
Renoval
(scraper)















Front-end
Loader





Mine/
Sample

D/2




D/3




D/4



D/5



D/1




Wind
Speed
(ra/sec)

6.2




7.2




7.2



7.6



2.7





Stability
Class

c




c




c



c



B



Back-
ground
Concentr.
(jig\m3)

158




158




158



158



122



Net
Plume
Concentr.
^g/m>)
6914
3055
7075
8363
12149
11800
16507
5944
7385
7672
5415
7556
6178
8107
4797
5925
5658
4597
1812
2149
2539
1972


Downwind
Distance
(m)
30
34
38
42
30
34
38
30
34
38
30
34
38
42
30
37
41
45
80
88
97
106
Vertical
Distance
from Plume
Center line
(m)
-.3
_ 3
-!3
-.3
-.3
-.3
-.3
.9
.9
.9
.3
-.3
-.3
-.3
-.3
-.3
-.3
-.3
-1 .3
-1 .3
-1 .3
-1 .3
Horizontal
Distance
from Plume
Centerline
(m)
np
np
np
np
np
np
np
np
np
np
np
np
np
np
np
np
np
np
0
0
0
0
" np - not published
b Sampled by mining company
e Samplers not in plume
" n.d. - not determined
                                       2-1 16

-------
    TABLE 2.46.  AVERAGE EMISSION RATES BY OPERATION AND MINE3
Operation
Dragline
Shovel/Truck
loading
coal
overburden
Blasting
coal
overburden
Truck dump
bottom dump
end dump
overburden
Drilling
coal
overburden
Fly-ash dump
Train loading
Topsoil removal
scraping
dumping
Front-end loader
Units
lb/ydj
Ib/ton
Ib/blast
Ib/ton
Ib/hole
Ib/hr
Ib/ton
Ib/yd1
Ib/ton
Mine
A
N.W.
Colorado
.0056
.014
1690
.014

3.9



B
S.W.
Wyoming
.053
.007

.020





C
S.E.
Montana
.0030
.002
25.1
14.2
.005
1 .5

.0002


D
Central
N. Dakota
.021

78.1
.027



.35
.03
.12
E.
N.E.
Wyoming

.0035
.037
72.4
85.3
.007
.002"
.22




  a) The authors advise that these factors should be used only in
    factors.
        conjunction with theoretical fallout
Study 2— Chalekode et al.   Emissions  from the Crushed Granite
          Industry:  State of the Art.   EPA-600/2-78-021.   1978.

     Methodology—The investigators used upwind-downwind
dispersion modeling to estimate emission factors for several
quarrying and  rock processing operations.   Only two of these
operations, drilling and blasting, are pertinent to the  "mining
and quarrying" category.  Turner's  (1970)  equation for a ground
level source with no plume rise was used to model the source
strength from  drilling:
For blasting,  the following model was  used to estimate the  total
mass emitted:
                      DT-
exp
                               2-1 17

-------
where

     DT   =    total dose, g-sec/m3
     QT   =    total release, grams and all other variables  are
               defined as normal

     The source strength for each sampling run was calculated  as
the average of the emission rates calculated from concentration
measurements at several downwind stations.  This was then divided
y the production rate to give an emission factor in mass emitted
per unit mass of product.

     Test Sites—Two granite quarrying and processing facilities
served as test sites for this study.  The operations were said to
be representative of the granite industry.

     Parameters and Equipment—The report only partially
documents the use of field equipment, particularly instruments
which measure dust concentration.  The summary below is based on
limited textual explanation as well as tabular data on sampling
durations.

     A GCA portable respirable dust monitor, which uses
"electronic measurement of the beta absorption of the collected
sample" of air, was used to measure dust concentrations downwind
from the drilling operations.  The monitor, which normally
captures particles smaller than 50 |u.m in diameter, was equipped
with a cyclone separator, providing a cut point of 10 |o.m.  Wind
speed was recorded automatically every 15 seconds by an
anemometer attached to the respirable dust monitor.  Mean wind
speed for each sample was computed by averaging the 15 second
readings.

     High volume samplers were used to measure concentrations
upwind and downwind from quarry blasts and apparently some
drilling operations.  Wind speed, wind direction and temperature
were measured from a meteorological station.  An average wind
speed was computed at the end of each 15 minutes.  These values
were then averaged to yield an average wind speed for the
sampling run.  Stability class was determined on the basis of
cloud cover, wind speed, and time of day.  Sampler positions
relative to the source were estimated by pacing.

     Equipment Confiquration--Aqain, documentation of equipment
deployment was inadequate.  The authors did not describe how the
portable dust monitor was positioned in the field.  For blasting,
five hi-vols were used, one upwind and four downwind.  Three
downwind samplers were configured in an arc of roughly equal
distance from the source, and the fourth was closer to the source
and near the plume centerline (i.e. y = 0).  The distances of  all
of these samplers from the source were not published.
                              2-1 1

-------
     Sampling Runs—Only one concentration measurement from
downwind of the blasting operation was reported.  It was a 45
minute sample.  The remaining seven published concentrations were
from four-minute samples taken downwind of drilling activities.
Water was applied to the drill face in six of these.  The number
of blasts, the number of holes drilled, and the depth of the
holes drilled during sampling were not documented.

     Quality Assurance—The issue of quality assurance was not
addressed in this report.  No documentation was provided for
calibration of sampling equipment, auditing of filter weights, or
use of reference methods.

     The explanation of field sampling procedures and conditions
was unclear.  Although the authors noted that the hi-vols were
used for all concentration measurements, use of the respirable
dust monitor was also mentioned.  It was capable of collecting
samples over a short time period.  Thus, for this review it was
assumed that the four minute samples were taken using the
respirable dust monitor and that the 45 minute blasting sample
was collected with the hi-vol.  It should be noted that although
five hi-vols were employed, only one concentration reading for
the blasting was published.  The hi-vol sampler from which it was
derived was not reported.

     The method used to measure emissions from blasting may have
been inappropriate for the source.  It is likely that the dust
plume from a single blast would remain in the vicinity for only a
fraction of the 45 minutes during which the air sample was
collected, especially given the measured wind speed of 7 mph.  As
was noted earlier, the number of blasts set off during the
sampling time was not reported.  However, it seems unlikely that
multiple blasts could be set off at the same distance upwind from
the sampler during a 45 minute period, particularly given the
authors' indication that a single blast typically provided enough
raw material for the processing facility for several days.
Consequently,  the available evidence indicates that a significant
portion of the air sampled by the downwind hi-vol was not from
the blast plume.  If this was the case, then the measured
downwind concentration would not have been representative of the
average concentration in the plume, resulting in an inaccurate
estimate of the emission factor.

     The investigators did not use an appropriate model to
estimate emissions from drilling.  Their reported modeling
parameters indicate that downwind concentration measurements were
not made on the plume centerline.  In the modeling framework
presented by Turner (1970), such a source should be modeled using
                              2-1 19

-------
an equation which includes the following correction factor to
adjust for concentration measurements taken away from the plume
centerline:
                           exp
where

     y    =    distance between plume centerline and receptor
     0y   =    standard deviation of the plume's crosswind
               concentration distribution

     Findings—The published dispersion modeling data and the
resulting source strengths are shown in Table 2.47.  Note that
the source strength, even for blasting, is given in mass per unit
time.  In the text of the report, the authors noted that the
tabulated emission from blasting was actually mass per blast.
This is yet another discrepancy in the report.

     The investigators learned from plant personnel that one
blast typically provides material for 26,250 tons of product.
The resulting emission factor, after converting to metric tons,
was .0796 kg of "total particulate"  (< 100 ^im in diameter) per
metric ton of product.  Based on the findings of Blackwood et
al., 1978, which indicate that the average ratio between
respirable particulates and total particulates in  (traprock)
quarrying emissions is 0.169, the emission factor for respirable
particulates (< 10 jam) was calculated as .0135 kg/metric ton.

     The investigators averaged the four wet drilling, total
particulate emission rates to derive a single source strength:
0.015 grams/sec.  In converting this to'an emission factor, it
was assumed that most plants use wet drilling and that the
drilling time averages 176 hours per blast.  The resulting factor
for total particulates was 3.99 x 10~A kg/metric ton.  The
average of the two respirable emission rates 0.0015 grams/sec, or
10% of the average total particulate source strength.  Therefore,
the emission factor for respirable particulates was taken to be
10% of the factor for total particulates:  3.99 x  10"5 kg/metric
ton.
                              2-120

-------
                TABLE  2.47.  DISPERSION MODELING  DATA
Activity
Blasting
Drilling,
Dry
Drilling,
Wet
Drilling,
Wet
Drilling,
Wet
Drilling,
Wet
Drilling,
Wet
Drilling,
Wet
Wind
Speed
(mph)
7
2
2
2
2
2
2
2 i
Monitoring Position (feet)
X
2300
78
90
90
90
90
90
90
Y
0
20
22
22
0
22
0
0
Z
230
0
0
0
0
0
0
0
Concentration3
(ng/m3)
763.4 total
1 540 respire
70 total
130 total
560 total
130 total
1 20 respire
1 30 respire
Source
Strength
(grams/sec)
1908000
.3562
.01159
.02152
.006728
.02152
.001442
.001562
Stability
Class
D
C
D
D
D
D
D
D
a) some measurements were for total particulates and others were for respirable
particulates (i.e. smaller than 10 urn in diameter)

     Publication—This study was conducted and documented under
contract for  the EPA,  Industrial Environmental Research Laboratory,
Cincinnati, Ohio.   It was published as Publication No. EPA-600/2-78-
021 in February of 1978.
                                 2-121

-------
Study 3— Blackwood and Chalekode.  Source Assessment:  Crushed
          Stone.  EPA-600/2-78-004L.  1978.

     Methodology—This study was very similar to the source
assessment of crushed stone production (Chalekode et al. , 1978).
In fact, much of the documentation was identical to that of
Chalekode et al.,  1978.  It should be noted that this report was
published about two months after Chalekode et al.,  and that
Chalekode and Blackwood were co-authors of both reports.

     As in the earlier study, the investigators used upwind-
downwind dispersion modeling to estimate emission factors for
rock excavation operations.  Blasting and quarrying (i.e.
gathering and loading blasted material into haul trucks) were
monitored, and Turner's (1970)  equation for a ground level point
source with no plume rise was used to model the source strength:


                      v- ——— exp
     Test Sites—Field monitoring was conducted at two
unidentified stone quarry and processing operations.  Both
produced crushed traprock.  They were considered representative
of the crushed stone industry, particularly because 68% of
crushed stone produced (in 1978) was traprock.


     Parameters and Equipment--High volume samplers were used to
test emissions from blasting and quarrying.  They collected
particles smaller than 100 |^m in diameter.  Wind speed, wind
direction, and temperature were measured from a meteorological
station.  An average wind speed was computed at the end of each
15 minutes.  These 15-minute averages were then averaged to give
a mean wind speed for the sampling run.  Stability class was
determined and reevaluated every two or three hours on the basis
of cloud cover, wind speed, and time of day.

     Equipment Configuration—Five hi-vols were distributed in
the same manner described for Chalekode et al., 1978.  The
distances of all of these samplers from the source were not
published.  In fact, the reported field data only provide
concentrations at one, or at most two, monitoring stations, which
were identified by their coordinates relative to the source.
Neither the sampling height of the monitor nor the location of
the wind or temperature measurements were documented.

     Sampling Runs—The extent of sampling conducted for this
study is unclear.  Two concentration measurements were presented
for blasting and two for quarrying.  The two blasting samples
were collected over 16 and 55 minute periods.  Both quarrying

                              2-122

-------
samples were collected over 45 minute periods.  Concentration
measurements were not reported for all of the downwind monitoring
sites.  The authors did not indicate which,  if any, of the
tabulated concentration measurements were made simultaneously -

     Quality Assurance--The documented quality assurance
practices for this study are limited.  The hi-vols were
calibrated prior to sampling,  and filters were inspected and
desiccated prior to weighing.

     A lack of clear documentation of the methodology and
findings brings the quality of the published data into question.
As was noted under "Sampling Runs," the tabulated findings do not
coincide with the described methodology.  The applied dispersion
model assumes that monitors are at ground level.  However, one of
the samples was given a z-coordinate (height) of 30 feet.

     Findings--The dispersion modeling parameters for the
blasting and quarrying operations are presented below in Table
2.48.

           TABLE 2.48.  DISPERSION MODELING PARAMETERS
Activity
Blasting
Blasting
Quarrying
Quarrying
Wind Speed
(mph)
8
8
17
17
Monitoring Position (feet)
X
204
204
615
791
y
0
0
0
0
z
0
0
30
0
Concentration
(fig/m3)
393
678
169
135
Source
Strength
179.8
grams
1066
grams
1 .602
grams /sec
1 .595
grams /sec
Stability
Class
C
C
C
C
     The emission factors for blasting were reported at 8.8
mg/metric ton for "respirable particulates" (< 10 ^.m in diameter)
and 52.2 mg/metric ton for "total particulates"  (< 100 |xm) . For
quarrying,  the respective factors are 1,050 mg/metric ton and
10,500 mg/metric ton.  The method of deriving these factors was
not described.

     Publication—This report is Publication No. EPA-600/2-78-
004L,  prepared for EPA's Industrial Environmental Research
Laboratory in Cincinnati, Ohio.  It was published in May of 1978.
                              2-123

-------
Study 4— Axetell and Cowherd.   Improved Emission Factors £°r
          Fugitive Dust From Western Surface Coal Mining Sources.
          EPA-600/7-84-048.   1984.

     Methodologies—Seven different mining operations were tested
to determine emission factors for fugitive dust.  The
sampling/measurement method depended upon the character of the
source.  Upwind-downwind dispersion modeling was used for dozer,
scraper, dragline, and coal loading (with shovels or front-end
loaders) operations.  The investigators used the same dispersion
models as were described in the review of Axetell, 1978.  The
scraper was tested as a captive line source.  It was driven back
and forth along a test strip without scraping or dumping, because
emissions from these two actions were judged to be insignificant
when compared to emissions from the movement of the vehicle
across the ground.  The dozer was treated as a line source when
it could be found operating in that mode; when it could not it
was tested as a captive area source.  The dragline and coal
loading operations were modeled as area sources.

     Normal exposure profiling was used for grader and scraper
operations.  These sources were considered line sources, and, as
such, a horizontal array of plume samplers was not needed.
Sample masses were converted to concentration by dividing by the
flow rate and the sampling duration.  Exposure at each intake is
then calculated as the product of concentration, sampling rate,
and duration, divided by the area of the sampler intake.

     The scraper operation was tested using both profiling and
modeling to determine the comparability of the results of these
two methods.  The results of this comparability study, which also
included testing of unpaved road emissions using the two methods
and an evaluation of several particle sizing devices, are
described in the review of this document in the unpaved roads
section.

     Blasting of overburden was tested using a variation of
exposure profiling which merits elaboration here.  An array of
five samplers was suspended from a tethered balloon to enable
sampling over the vertical extent of the plume.  Each sampler was
fitted with a wind vane which kept the intake pointing into the
wind.  Flexible tubing ran from each sampler to a pump on the
ground.  The samplers operated isokinetically at a wind speed of
5 mph.  Sampler pairs consisting of a standard hi-vol and a
dichotomous sampler were distributed in an arc at the same
distance from the blast area as the balloon.  They provided data
on concentration across the horizontal extent of the plume.
Plume boundaries were determined from these measured
concentrations and photographs.  The mass of dust passing the
downwind sampling location was calculated by integrating measured
concentrations with respect to horizontal position and height and
multiplying by the wind speed and sampling time.

                              2-124

-------
     Emissions from the drilling operation were measured using
the quasi-stack method.  A temporary enclosure was placed over
the drill and hole,  and dust emissions were vented through a
single outlet.  The outlet area was divided evenly into four
equal parts.  The wind velocity was measured and air samples were
collected at the center of each of these areas simultaneously-
To approximate isokinetic conditions, the intake velocity of the
samplers was adjusted every two to three minutes.  The mass
emitted was calculated using the following equation:
                            E-
where

     E    =    emitted mass, g
     Xi    =    concentration measured at sampler i, g/m3
     VA    =    total volume of air filtered through sampler i, m3

This mass was then converted to an emission factor by dividing by
the number of holes drilled.

     Test Sites--Tests were conducted at three unspecified mines,
one in each of the following Western coal fields:   Powder River
Basin, Fort Union, and San Juan River.  These three fields were
targeted for the study because they produce high volumes of
strip-mined coal and because they are diverse in character.  The
investigator's intent with this selection was to maximize the
representativeness of the findings while satisfying budget and
time constraints.

     Parameters and Equipment--The plume sampling parameters
measured for each operation are discussed in the following
section.  Those source characterization and meteorological
parameters which were monitored for each operation are listed in
Table 2.49.
     TABLE 2.49
SOURCE CHARACTERIZATION AND METEOROLOGICAL
         PARAMETERS
Operation
Drilling
Blasting
Parameter
Silt content
Moisture content
Depth of hole
Number of holes
Size of blast area
Moisture content
Equipment /Source
Oven, sieves, scales
Oven, scales
Given by drill operator
Direct observation
Direct observation
Given by mining company
                              2-125
                                     (continued)

-------
                     TABLE 2.49-   (continued)
Operation
                        Parameter
                        Equipment/Source^
Coal loading
Silt  content
Moisture content
Bucket capacity
Equipment operation
Oven, sieves scales
Oven, scales
From equipment
specifications
Field notes on
variations
Dozer
Silt content
Moisture content
Speed
Blade size
Oven, sieves,  scales
Oven, scales
Time/distance
From equipment
specifications	
Dragline
Silt content
Moisture content
Bucket capacity
Drop distance
Oven, sieves,  scales
Oven, scales
From equipment
specifications
Direct observation
Scraper & Grader
Surface silt content
Vehicle speed
Vehicle weight
Total surface loading
Surface moisture content
Number of wheels
Oven, sieves,  scales
Radar gun
Truck scale
Broom, scales
Oven, scales
Direct observation
All sources
Wind speed
Wind direction
Temperature
Solar intensity
Humidity
Atmospheric pressure
Cloud cover
Anemometer
Anemometer
Thermometer
Pyranograph
Sling psychrometer
Barometer
Direct observation
     Equipment  Configuration—

     Quasi-Stack Method—The equipment configuration for the
     drilling operation  was relatively simple.  A wooden
     enclosure  was constructed with 4X6  foot openings at  two
     ends.  This enclosure was set up on the downwind side  of the
     drill base.  Four profile samplers were fixed  horizontally
     across the downwind outlet  of the enclosure.   No other air
     sampling equipment  was used for this  operation.  Deposition
     was not monitored.

     Exposure Profiling  Using Tethered Balloon—A vertical  array
     of samplers was suspended from a balloon about 100 meters
     downwind from the edge of the blast zone.  Five sampler
     heals were positioned at heights of 2.5, 7.6,  15.2, 22.9,
     and 30.5 meters.  Five hi-vol/dichot  sampler pairs were each
                                 2-126

-------
20 to 30 meters apart.  The dichot samplers provided data  on
the particle  size distribution of the plume.  Dust
deposition was not measured for this operation.

Standard Exposure Profiling—For those tests which employed
exposure profiling as the only emissions measurement
technique, the following equipment was placed between five
and ten meters downwind from the source:  exposure profiler
with sampling heads at 1.5, 3.0, 4.5, and 6 meters above the
ground; a standard hi-vol sampler and a hi-vol fitted with a
cascade impactor and cyclone preseparator, both having
inlets 2.5 meters high; two dichot samplers with intakes 1.5
and 4.5 meters above the ground; two dustfall buckets at a
height of 0.75 meters; and two warm wire anemometers at
heights of 1.5 and 4.5 meters.  Pairs of dustfall buckets
were also collocated 0.75 meters above the ground at 20 and
50 meters downwind, permitting measurement of dust
deposition rates.

The following equipment was set up five meters upwind from
the road:  one dichot sampler 2.5 meters above the ground;
one standard hi-vol sampler, also at a height of 2.5 meters;
two dustfall buckets 0.75 meters above the ground; and one
continuous wind monitor 4 meters high.


Upwind-Downwind Method—The configuration of equipment for
this method depended on whether the mining operation was
treated as a line source or a area source.  For area sources
an array of about 14 air samplers, 9 hi-vol and 5 dichot,
was set up on the downwind side.  Each sampler was mounted
on a tripod stand such that it sampled at a height of 2.5
meters.  Two hi-vol/dichot sampler pairs were placed about
30 meters downwind.  Three of these pairs were placed in an
arc about 60 meters downwind.  Three hi-vols were
distributed at a downwind distance of about 100 meters.
When the layout of the field sites permitted, one or two hi-
vols were placed about 200 meters from the source.  Upwind
air samples were collected with a hi-vol and a dichot
sampler placed between 30 and 100 meters from the source.
They also sampled at a height of 2.5 meters.

For line sources, two hi-vol/dichot sampler pairs were
placed about 20 meters apart at each of the following
downwind distances:  5, 20, and 50 meters.  Two hi-vols were
placed 100 meters downwind.  These all sampled air 2.5
meters above the ground.

In addition to the dichot samplers, millipore filters were
used to provide a measure of particle size distribution.
The height and downwind distance at which these filters were
exposed were not documented.

                         2-127

-------
     Simultaneous Exposure Profiling and Upwind-Downwind—The
     equipment configuration for those tests in which both
     exposure profiling and upwind-downwind modeling were used
     was very complex.   Downwind air sampling was conducted
     primarily at three downwind distances: 5,  20,  and 50 meters.
     Profiling towers were set up at each of these stations so
     plume mass depletion could be measured in addition to simple
     exposure.  The closest tower consisted of four sampling
     heads at heights ranging from 1.5 to 6 meters.  The towers
     at 20 and 30 meters downwind both had five sampling heads,
     with the highest heads at 9 and 12 meters, respectively.
     This was necessary due to the increased dispersion of the
     plume over longer distances from the source.  A vertical
     array of dichot samplers was also set up five meters
     downwind.  The sampling heights matched those of the nearby
     exposure profiler.  Two single dichot samplers were also set
     up on either side of the profiling towers at each of the
     three downwind monitoring distances.  They had intakes 2.5
     meters above the ground.

     Two standard high volume samplers (all sampling at a height
     of 2.5 meters) were set up at each of the downwind
     distances: 5, 20,  50, and 100 meters.  A third standard hi-
     vol was used five meters downwind.  A total of three hi-vol
     cascade impactors were used, two 5 meters from the road with
     intakes 1.5 and 4.5 meters above the ground, and one 20
     meters from the road with its intake 2.5 meters high.
     Dustfall buckets were placed in pairs at each of the three
     downwind distances such that their sampling height was 0.75
     meters.

     Sampling Runs—The number of sampling runs and the sampling
method(s) for each operation are listed in Table 2.50.

          TABLE 2.50.  SAMPLING RUNS FOR EACH OPERATION
Source
Drilling (overburden)
Blasting
Coal loading
Dozers
Dragline
Scraper
Grader
Total
Method(s)
Quasi-stack
Exposure profiling
with tethered balloon
Upwind-downwind
Upwind-downwind
Upwind-downwind
Exposure profiling
and Upwind-downwind
Exposure profiling

Sampling
Runs
30
18
25
27
19
5
7
131
Approximate Duration
per Run (min)
15-90
3-30
15-90
15-145
30-75
15-110
30-120

                              2-128

-------
     Quality Assurance—This study included a thorough quality
assurance program, which was subject to evaluation by a technical
review group (including the two EPA project officers, and
representatives of the Bureau of Land Management, the Bureau of
Mines, and the mining industry).  Profilers, hi-vols, impactors,
and dichot samplers were calibrated on a regular basis.  Sampling
media were conditioned at constant temperature and humidity prior
to weighing.  Seven percent of tare and final filter weights were
audited.  For every ten regularly processed filters and
substrates, at least one was processed as a blank.

     Regarding sampling isokineticity of the profiler, sampling
intakes were reoriented if the 15 minute average wind direction
changed by more than 30°, and the sampling rate was corrected
when the 15 minute average wind speed changed by more than 20%.

     For tests of scraper and grader emissions, the investigators
recorded the total number of passes as well as the number of
"bad" passes, in which the wind direction reversed and upwind
filter weights were affected by road emissions.  For one of the
scraper tests,  only about half of the passes were judged to be
good.  There were no bad passes for the grader tests.  For runs
in which bad passes occurred, the upwind dust concentration was
estimated by the average of the concentrations of the previous
and following sampling runs.  Bad passes were not counted when
calculating the emission factor (i.e. when dividing the
integrated exposure by the number of vehicle passes).

     Despite the attention given to normal quality assurance
practices -in field data collection, the overall level of quality
assurance for this study is compromised by the paucity of
published raw field data.  For instance, measured exposures at
the various profiling heights were not reported.  Because of this
omission, the calculations of base emission factors made by the
investigators cannot be repeated and verified.

     Findings--The emission factors measured for each sampling
run are shown in the Tables 2.51-2.55.  Multiple linear
regression analysis yielded the predictive equations for TSP and
IP emissions shown in Table 2.56.  With the exception of four
predictor variables, all are significant at the 0.05 level.  For
the drilling operation, the only variable which proved useful  in
predicting TSP was % silt.  However, contrary to expectations,
the relationship between % silt and TSP was inverse; therefore,
it was eliminated from the equation, leaving only the geometric
mean as the emission factor.

     Publication—This study was conducted and documented under
contract with the EPA Office of Air Quality Planning And
Standards, Research Triangle Park, North Carolina,  and the EPA


                              2-129

-------
Industrial Environmental Research Laboratory,  Cincinnati, Ohio,
It was  published as  Publication  No.  EPA-600/7-84-048  in March
1984.
 TABLE 2.51.   EMISSION FACTORS DERIVED USING EXPOSURE PROFILING:
                          SCRAPER AND GRADER

Source
Scraper

Grader
Run
J-1
J-2
J-3
J-4
J-5
K-15
K-16
K-17
K-18
K-22
K-23
L-5
L-6
P-15
P-18
K-19
K-20
K-21
K-24
K-25
P-16
P-17
Emission Factor (Ib/veh-mile)
Total3
41 .4
66.5
125
27.5
96.7
126
206
232
179
58.4
118
360C
184
383
18.8e
31 .3
29.0
22.5
13.1
19.5
53.2
73.9
db < 30 Jim
8.6
9.4
50.2
3.9
17.7
16.2
29.2
74.3
43.0
10.3
24.5
355C
163
d
4.0e
4.0
4.3
1 .8
3.2
7.3
34.0
8.6
db < 15 nm
4.2
4.0
26.1
1 .7
10.0
7.2
15.6
35.6
19.3
4.8
11.1
217=
94
d
1 .4e
2.3
1 .7
0.89
1 .9
4.1
15.4
2.9
db < 2.5 urn
0.27
0.19
1 .5
0.09
1 .4
0.39
1 .8
1 .6
0.81
0.29
0.54
0.72C
1 .0
d
0.02e
0.33
0.46
0.08
0.29
0.38
0.09
0.04
         a) total particulates  (i.e.  all sizes)
         b) aerodynamic diameter
         c) profiler samplers malfunctioned
         d) only one dichot sampler and only four good
         e) only two profilers  operational
passes
                                 2-130

-------
TABLE  2.52.  EMISSION FACTORS  DERIVED FROM QUASI-STACK TESTING:
                              DRILLING

Mine /Run
1/1
1/2
1/3
1/4
1/5
1/6
1/7
1/8
1/9
1/10
1/11
ivr/1
1W/2
1W/3
1W/4
Emission Factor
(Ib/hole)
Filter-
1 .18
0.20
0.24
0.04
0.17
0.11
0.33
1 .56
1 .98
2.43
0.95
0.76
3.38
2.57
1 .95
Totalb
6.75
0.75
0.81
0.28
0.47
1 .92
7.61
24.31
50.31
41 .01
12.69
5.80
43.46
144.3
23.52

Mine /Run
1W/5
1W/6
1W/7
1W/8
1W/9
1W/10
1W/11
1W/12
3/1
3/2
3/3
3/4
3/5
3/6
3/7
Emission Factor
(Ib/hole)
Filter"
2.54
2.91
3.35
3.05
2.23
0.53
0.06
0.45
3.06
7.29
4.65
6.48
4.04
1 .79
5.84
Total"
111 .72
44.34
68.50
40.71
34.86
2.09
1 .04
3.89
21 .07
35.23
12.72
22.18
15.92
9.96
26.47
          a)  calculated using only the mass collected on the filter
          b)  calculated using mass on the filter and in the
          settling chamber
          c)  winter sampling at mine 1
  TABLE 2.53.
EMISSION FACTOR DERIVED USING  EXPOSURE PROFILING
  WITH TETHERED BALLOON:   BLASTING
Mine/Material
1 /Coal
1 /Coal
1 /Coal
1 /Overburden
1 /Overburden
2 /Coal
2 /Coal
2 /Coal
Run
1
2
3
1
2
1
2
3
Emission factor (Ib/blast)
da < 30
Urn
32.5
2.7
51 .7
40.4
79.4
8.8
1 .1
10.7
da < 15
(j.m
44. 9b
1 .56
17.3
32.9
48.9
1 .55
0.62
3.57
da < 2.5 urn
3.62
0.32
1 .23
0.79
0.09
0.10
0.06
0.80
                                2-131
                                           (continued)

-------
                   TABLE 2.53.    (continued)
Mine/Material
2/Coal
2/Coal
2/Coal
3/Coal
3 /Coal
3/Coal
3 /Coal
3/Coal
3 /Overburden
3 /Overburden
Run
4
5
6
2
3
4
5
6
1
2
Emission factor (Ib/blast)
da < 30
(im
1 .6
40.3
11 .8
401
514
148
113
206
35.2
270
d1 < 15
^m
0.45
15.30
1 .99
123.4
142.8
87.9
35.3
71 .3
16.9
93.9
da < 2.5 nm
0.10
1 .27
0.01
10.4
12.3
13.0
2.1
19.8
3.5
16.2
       a) aerodynamic diameter
       b) this value represents mass of particles with
       aerodynamic diameter < 20.5  |j.m.
TABLE  2.54.  EMISSION FACTORS DERIVED FROM  UPWIND DOWNWIND
                   MODELING:   COAL LOADING

Mine /Run
1/1
1/2
2/1
2/2
2/3
2/4
2/5
2/6
2/7
2/8
3/1
3/2
3/3
Emission Factor (Ib/ton)
TSP
0.0069
0.0100
0.044
0.068
0.0147
0.0134
0.0099
0.0228
0.0206
0.0065
0.120
0.082
0.051
IP
.002
.003
.005
.022
.003
.005
.004
.017
.008
.004
.044
.008
.016
FP
.0001
.0002
.0002
.0008
.0001
.0018
.0007
.0029
.0008
.0002
.0038
.0005
.0022

Mine /Run
3/4
3/5
3/6
3/7
3/8
3/9
3/10
3/11
3/12
3/13
3/14
3/15
Emission Factor (Ib/ton)
TSP
0.0105
0.0087
0.0140
0.035
0.062
0.058
0.193
0.095
0.042
1 .09
0.358
0.188
IP
.002
.001
.006
.008
.012
.014
.038
.020
.011
.378
.121
nd
FP
.0002
.Q001
.0001
.0012
.0012
.0005
.0033
.0005
.0021
.0054
.0035
nd

                              2-132

-------
TABLE 2.55.  EMISSION FACTOR DERIVED FROM UPWIND-DOWNWIND
          METHOD:  DOZER, DRAGLINE, AND SCRAPER

Mine /Source
1 /Dozer
(Over-
burden)
2/Dozer
(Over-
burden)
3/Dozer
(Over-
burden)
1 /Dozer
(Coal)
2/Dozer
(Coal)
3 /Dozer
(Coal)
Run
1
2
3
4
1
2
3
4
5
6
7
1
2
3
4
1
2
3
4
1
2
3
1
2
3
4
Emission Factor
TSP
16.2
12.6
2.6
3.0
0.9
1 .8
2.6
1 .3
9.2
1 .0
1 .0
5.4
5.2
18.0
20.7
16.1
40.1
19.0
21 .3
9.1
6.2
3.0
289
222
439
323
IP
3.18
2.18
2.85
c
2.12
5.88
1 .00
0.48
1.14
0.68
1 .22
.98
.781
4.57
32.6
4.49
39.9
4.73
13.0
2.26
2.26
0.92
177
178
236
176
FP
.436
.322
1 .01
c
.583
.091
.790
.065
.680
.421
.536
.356
.089
.925
1 .73
.243
.730
1 .00
2.68
.252
.199
.138
3.50
2.25
4.49
3.28
Units
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr
Ib/hr

Mine /Source
3/Dozer
(Coal)
1 /Dragline

2/Dragline
3 /Dragline


1 /Scraper

Run
5
1
2
3
4
5
6
1
2
3
4
5
1
2
3
4
5
6
7
8
J1
J2
J3
J4
J5

Emission Factor
TSP
224
.024
.029
.004
.048
.070
.400
.042
.026
.003
.016
.068
.184
.133
.192
.099
.060
.068
.104
.105
10.6
18.6
35.6
5.7
20.0

IP
82.2
.006
.012
.002
.006
.0165
.061
.003
.007
.001
.015
.035
.018
.016
.058
.043
.038
.028
.024
.017






FP
3.50
.0009
.0002
.0001
.0001
.0009
.0087
.0002
.0008
.0003
.0010
.0110
.0017
.0011
.006
.005
.0001
.0017
.0023
.0004






Units
Ib/hr
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
lb/yd3
Ib/veh-
mile
Ib/veh-
mile
Ib/veh-
mile
Ib/veh-
mile
Ib/veh-
mile

                          2-133

-------
  TABLE 2.56.   PREDICTIVE EQUATIONS DEVELOPED THROUGH REGRESSION
                      ANALYSIS OF FIELD DATA3
Operation
Drilling
Blasting
Coal loading
Dozer, coal
Dozer , overburden
Dragline
Scraper
Grader
TSP
1 .3
(961 A°-8)/(Du8 M1-9)
1 .16/M1'2
78.4 s'-2/MU3
5.7 s'-VM'-3
0,0021 d'-VM0'3
(2.7 * 10-5) S '-3
w2-"
0.040 S2-5
. IP
NA
(2550 A°-6)/(DK5
M2-3)
0.119/M0'9
18.6 s'-VM1-4
1 .0 s'-5/M'-4
0.0021 d°-7/M°-3
(6.2 * 10'6) s1'4 W2-5
0.051 S2-0
Units
Ib/hole
Ib/blast
Ib/ton
Ib/hour
Ib/hour
lb/yd3
Ib/veh-mile
Ib/veh-mile
  a) A - area blasted, ft2;  D - depth of holes, ft;  M - moisture content, %;
  s - silt content, %;  d = drop distance,  ft;  W - vehicle weight, tons;  S
  - vehicle speed, mph


Studies of Secondary  Importance

Study 5— Cook  et al.  Fugitive  Dust  from Western Surface Coal
          Mines.  EPA-600/7-80-158.   1980.

     The methodology  for this  study was relatively straight-
forward.  Data  were collected  on a  series of environmental
parameters at several strip mines in  the Western U.S.:  total
suspended particulate  (TSP) concentration,  duration of mining
operation (for  12 different operations),  wind speed, and
precipitation.  An analysis of variance (ANOVA)  and regression
analysis were conducted  to determine  which variables are
statistically related to TSP.  Emission factors for the various
activities were not calculated.

     Measurements were collected at four mines  in three separate
visits per mine.  These  visits took place during late spring,
fall, and winter.  Two of the  mines had rolling terrain with
predominantly grassland  vegetation, and two had semi-rugged
terrain with sagebrush vegetation.

     The field  sampling  program  was performed by a subcontractor.
Four high volume samplers fitted with Anderson heads were used to
                               2-134

-------
measure TSP-  The investigators noted that the hi-vols  (General
Metal Works GMWL 2000) collected particles smaller than 100 [am in
diameter.  The mass on each filter was combined to give the TSP
mass.  The samplers were set up between 250 and 500 meters
downwind of each operation.  Sampling periods corresponded
roughly with the eight-hour work shifts at the mines.  It should
be noted that the wind direction sometimes changed, causing
nominal downwind samplers to be outside of the plume for a
period.

     Detailed procedures for filter and dust sample handling were
well-documented.  Plastic gloves were always used when handling
filters to prevent changes in weight due to contamination.
Filters were individually packed in plastic bags before and after
sample collection.

     The relationship between TSP and the predictor variables was
estimated by the equation below:

                  TSP = 30.3 Q/-13 Q2°-10 Q8°-10 S°-40

where

     TSP  =    average concentration of particles with diameter <
               100 |xm, ng/m3
     QT   =    dragline operating minutes per shift
     Q2   =    coal truck hauling trips per shift
     Q8   =    number of vehicle passes on nearby public road per
               shift
     S    =    wind speed, mph

     The investigators noted that of the four predictor
variables, wind speed had the strongest influence on TSP, as
indicated by the fact that doubling S would increase TSP by 32%,
whereas doubling each of the other variables individually would
increase TSP by no more than 10%.

     This study was published in August of 1980 as Publication
No. EPA-600/7-80-158.  It was conducted for the Environmental
Protection Agency, Industrial Environmental Research Laboratory
in Cincinnati, Ohio.

Study 6— Marple et al.  Fugitive Dust Study of an Open Pit Coal
          Mine with the University of Minnesota Mobile
          Laboratory.  1980.

     The emphasis of this study was on determining what
instruments could feasibly be used in monitoring fugitive dust
from a mobile laboratory.  The concentration and particle size
distribution of dust from the mine were the main parameters of
interest.  The report consists primarily of 1) documentation  of
the contents and design capabilities of the University  of

                              2-135

-------
Minnesota Mobile Laboratory (UMML),  2) a description of  the  field
sampling activities, 3) data analysis and findings, and  4)
conclusions and recommendations.

     Field experimentation was conducted at several sites  in a
strip mine in the western United States.  The mining activities
occurring during the field tests were similar to those described in
Axetell et al.,  1984, which is also summarized in this section.

     Particle size distribution was measured in parts with several
different instruments.  An electrical aerosol analyzer  (EAA)
counted particles in the size range 0.01 ^m to 1 |j.m.  A  modified
optical particle counter (OPC) measured the concentration  of
particles in the 0.5 to 5.6 (am range, and a second modified  OPC
measured particles from 5.6 urn to 15 [am in diameter.  In addition,
an Aitken nuclei counter (General Electric GE-1) and a rate  meter
attached to the large particle OPC provided real time data on the
concentration of particles in the respective ranges 0.01 ^.m  to 1.0
(j.m and larger than 5.6 |j.m.  The measurement height for all of these
instruments was about 3 meters.  Open faced Millipore filters were
exposed and analyzed microscopically as a check on the ability of
the above equipment to detect very large particles.  A special
cascade impactor was used in the field tests, though it  was  not
part of the UMML.  The impaction plates rotated during sample
collection so that dust was distributed evenly around the  impaction
substrates.

     In addition to dust particle size and concentration,  data were
collected for meteorological parameters, gas concentrations,  and
aerosol chemistry from instruments on the UMML.

     Conclusions regarding plume dust which were drawn by  the
investigators as a result of this study are listed below:

     •    The instruments in the UMML could be used to study and
          measure dust plumes in any location the UMML could
          access.

          For plumes from passing vehicles, the concentration
          increased as vehicle size increased, as vehicle  speed
          increased, and as the distance from the road decreased.

          Dust emissions were measured in terms of mass  per  unit
          length of road per unit of plume height.  Plume  height
          was never measured or estimated, so emission factors
          could not be calculated.

     This study was conducted for the U.S. Department of Interior,
Bureau of Mines, under contract J0295071.  The researchers worked'
under the auspices of the University of Minnesota, Particle
Engineering Laboratory, Minneapolis, Minnesota.  The report  date is
August, 1980.  It was apparently never published.

                                2-136

-------
STORAGE PILES

Introduction

     The transfer of aggregate material into and out of storage
piles and the maintenance of those storage piles generates
fugitive dust.  Emissions have been measured or modeled in five
field studies, four of which were performed by the Midwest
Research Institute.  The first two, which were documented in the
original exposure profiling report (Cowherd et al.,  1974),
examined emissions from a sand and gravel storage area and from
one particular operation, load out of aggregate into dump trucks.
The third study (Bohn et al.,  1978) measured emissions from
several types of aggregate transfer operations at two iron and
steel facilities.   The study by Blackwood and Chalekode (1978)
involved upwind-downwind dispersion modeling of emissions from
the dumping of traprock gravel in a quarry storage area.  In the
fourth study  (Cowherd et al.,  1979) emissions were measured from
the loading of iron pellets and coal into storage piles using
long conveyor stackers.  Each of these studies is reviewed here
in chronological order.

     A sixth study (Vekris, 1971) included measurement of
emissions from a heavy duty vehicle traveling across a large coal
storage pile at a power plant.  Although the fugitive dust source
was not the transfer of material into or out of the storage area,
but rather the action of the vehicle against the pile surface,  it
is reviewed briefly at the end of this section because it
involves a storage pile and it fits into no other defined
category-

     A study by Axetell  (1978) included measurement of dust
downwind from storage piles.  However, because the emission rate
from this source was expressed as a function of wind speed, and
because the intensity of activity in the storage pile was not
recorded, it is reviewed in the section on wind erosion.

Studies of Primary Importance

Study 1— Cowherd et al., Development of Emission Factors for
          Fugitive Dust Sources.  EPA-450/3-74-037.  1974.

     This report documented two distinct field studies  on dust
emissions from storage piles.   These two studies are reviewed
separately below.   Conclusions from the findings of both  studies
are then summarized.

     Total Emissions from Aggregate Storage Operations—

     Methodology—The mass of dust emitted from a storage
     operation was measured using a rather simple methodology.
     The site's contribution to the downwind ambient dust

                              2-137

-------
concentration was estimated using upwind and downwind hi vol
samplers.  The mass of dust emitted per unit time was
calculated as the average net concentration times the
atmospheric ventilation rate (which is in this case the
volumetric rate at which air passes through the cross-
sectional area defined by the width and height of the
storage area).

Sampling runs were conducted during 12- and 24-hour periods.
Four parameters were evaluated as possible factors
influencing the dust emission rate:  rainfall, wind speed,
aggregate size, and intensity of activity.  For the rainfall
factor, sampling results were divided into two groups:  dry
period results and wet period results.  Runs were deemed wet
if any precipitation occurred during sampling or if more
than a trace occurred on the day before.  The average net
concentration and emission rate for each group was compared
to determine if emissions were higher during dry periods
than wet periods.

To check for effects of wind speed and aggregate size, these
variables were each plotted against downwind concentration.
For aggregate size, this was accomplished by matching
storage piles of various sizes of aggregate with particular
samplers downwind from the piles.

Accurate data were not available on the level of activity in
the storage area during each sampling run.  Thus, it was not
possible to estimate a relationship between the intensity of
activity and concentration.  However, after some
manipulation of the data, it was possible to compare
concentrations corresponding to periods when the storage
area was active with those for non-working periods.   The
12-hour  sampling runs were conducted exclusively during non-
working  hours, and most 24-hour samples were conducted
during periods including both working and non-working hours.
 (The rest covered exclusively non-working periods.)  A
concentration for working hours was calculated using the
following relationship, in which A is the 24-hour average
concentration for a day including 8 to 12 hours of activity,
B is the average concentration during hours of activity in
the storage area, and C is the average concentration for
inactive hours:

                     A =  (B + C) /2

Test Sites—Tests were conducted at a sand and gravel quarry
and processing center near Cincinnati, Ohio.  The
investigators considered it representative of operations at
many medium and large aggregate sites.  Although the gravel
pit was  adjacent to the storage area, the samplers used in
this study were judged to be sufficiently isolated from any

                         2-138

-------
dust generated in the pit itself by the difference in
elevation between the pit and the storage area.  Fifteen
storage piles, ranging in height from 5 to 30 feet, were
maintained in this area.  The investigators computed an
average pile height, weighted on the basis of the pile
surface area, of 23 feet.  Each pile was for a different
size aggregate.  The turnover rate for these piles was said
to be high.  No processing of aggregate was conducted in
this area.

Parameters and Equipment—Upwind and downwind concentration
of total suspended particulates was measured using standard
hi-vol samplers.  These were automatically activated by wind
sensors when the wind direction was within 90° of South.
They were also equipped with timers to record the duration
of the sample.  A high volume cascade impactor was used to
measure the particle size distribution of the dust downwind
from the storage area.  Meteorological data, including cloud
cover, temperature, and precipitation, were acquired from a
nearby Federal Aviation Administration Weather Station.  The
size of the aggregate in each pile was noted so the effect
of aggregate size on downwind concentration could be
analyzed.  Data were also collected on the height and
configuration of each pile.  The equipment operator's
records documented the tonnage of material excavated, sized,
and loaded onto trucks for transport.  This provided only a
very rough indication of the level of activity at the site
during a given day.

Equipment Configuration—The five downwind samplers were
scattered on the downwind side of the storage area.  Three
of them were set up among the storage piles, and two were
immediately downwind of the entire storage area.  The intake
height of these samplers ranged from 3 to 20 feet.  The
height and upwind distance of the background sampler was not
documented, nor was the position of the cascade impactor.
Wind speed and direction were measured continuously on a
pole about 25 feet above the ground.

Sampling Runs—Eleven 24-hour and seven 12-hour sampling
runs were conducted.  Four of the 24-hour runs were on
weekends when there was no activity at the storage area.
The remainder ran from noon one work day until noon the
next.  All of the 12-hour samples ran from 6:00 p.m. until
6:00 a.m. the next day.

Quality Assurance—None of the normal quality assurance
procedures, such as processing of blank samples, calibration
of equipment, or auditing of measurements or calculations,
were documented for this field study.
                         2-139

-------
Presentation of raw field data was complete.   Measured
upwind concentrations and net downwind concentration for
each site and sampling period were documented.  Evidence
that the net downwind concentration measured the full
contribution of the storage area alone was presented in
three separate comparisons:  the upwind concentration was in
line with typical regional ambient levels, the upwind
concentration was less than the downwind concentrations in
almost every instance, and the average upwind concentrations
during working periods were close to those of non-working
periods.

Findings—The concentrations measured for each sampling run
are shown in Table 2.57.  The run time for each sampler is
also shown for every test.

The average concentration for working days and non-working
days was 182.7 ug/m3 and 47.4  (j.g/m3, respectively.  Using
the methodology described above, corresponding emission
rates were calculated at 103 and 26.8 kg/day.  The
concentration calculated for working hours was 318 [j.g/m3,
which converts to an emission rate of 7.5 kg/hr.

Neither the calculations for converting these numbers into
emission factors nor the activity rate (e.g. tons stored per
day) were presented in the report.  The factor for an active
storage area with eight to twelve hours of activity per 24
hours was given as 0.42 Ib/ton placed in storage (13.2
Ib/storage acre/day).  For periods of inactivity, the
emission factor was calculated as 3.5 Ib/storage acre/day
(0.11 Ib/ton placed in storage).  For a normal mix of five
workdays per week, the emission factor was calculated as
0.33 Ib/ton placed in storage.

Rainfall, as recorded in the manner described above, was
found to reduce emissions by roughly 50%.  Neither the size
of the aggregate in the storage piles nor the wind speed was
found to have a significant influence on TSP emissions.

The particle size distribution of dust downwind from the
storage area was not discussed, although the cascade
impactor was listed as an instrument used in the field
study-
                         2-140

-------
   TABLE 2.57.  CONCENTRATION MEASUREMENTS AND SAMPLE RUN TIMES

Date
6/9
6/11
6/12
6/13
6/14
6/16
6/18
6/19
6/20
6/21
6/23
6/25
6/26
6/30
7/2
7/3
7/5
7/6

Test
Period
(hr)
24
12
24
12
24
24
12
24
12
24
24
12
24
24
12
12
24
24

Upwind
Concent .
(Hg/m3) /
Sample
Time
(min)
94/1130
95/484
60/1009
65/276
139/695
75/1126
71/532
49/1149
61/410
7/1205
67/1087
86/586
58/1181
61/1233
64/611
50/1139
95/770
124/1093
Net Downwind Concentration (jag/m3)
(minutes)
Hi-vol 2
8/1140
107/403
85/1074
215/70
575/424
3/1192
21/340
93/1127
Voidc
152/1032
8/1011
55/578
121/1440
16/1119
20/613
28/1058
231/508
362/734
Hi-vol 3
23/1082
152/415
113/1103
125/73
134/347
Ob/1128
16/381
57/1160
Voidc
140/1301
6/1440
19/721
134/1365
31/1066
17/620
24/1031
138/420
170/842
Hi-vol 4
49/1074
184/413
252/1039
15/80
239/360
OV1082
37/406
105/1134
48/201
249/940
Voidc
89/301
50/1290
Ob/1190
11/596
28/869
146/1311
332/751
Hi-vol 5
13/1165
172/423
208/1073
OV280
175/661
7/1 168
42/619
74/1440
2/719
154/1423
9/1352
33/719
202/240
31/982
71/378
22/1249
150/1280
183/1432
/ Sampling
Hi-vol 6
4/1064
76/355
147/1090
125/62
259/285
26/378
Voidc
170/940
Voidc
108/1009
27/1024
210/510
Voidc
42/1032
Voidc
19/1054
40/375
241/706
Time
Average"
19
138
161
96
276
7
29
100
25
161
12
81
127
24
30
24
141
258
a) Average downwind sampling time was not calculated
b) A net concentration of zero was assumed when the upwind concentration
   slightly higher than the downwind
c) No explanation was given for the voided samples
was
     Emissions  from Aggregate Load Out—

     Methodology—Exposure profiling was used to measure  dust
     generated  by a high loader dropping aggregate into a dump
     truck.  A  vertical, two-dimensional matrix of isokinetic
     high volume samplers was positioned downwind from the truck
     being loaded.  The dust  emitted when  the aggregate was
                                2-141

-------
dumped passed through the sampling array.  Total exposure
was calculated by integrating the mass collected per unit
area with respect to height and horizontal position.

Test Site—An asphalt producing facility in Kansas City
provided the site for this study.  The operation used four
different types of aggregate in producing asphalt.  Testing
was conducted on a weekend when the plant was not in
operation.

Parameters and Equipment--The profiler consisted of six hi-
vol sampling heads with vertically oriented filters.  The
samplers were pre-set to operate isokinetically with wind
speeds of 10 mph.   A high volume cascade impactor was used
to collect data on plume particle size distribution.  It
provided cut points at diameters of approximately 0.7, 1.25,
2.1, and 4.4 [im.  Upwind and downwind TSP concentrations
were measured with standard hi-vol samplers.

Meteorological parameters for which data were recorded
included wind speed and direction, cloud cover, temperature,
relative humidity, and atmospheric stability.  Several
source characterization parameters were recorded.  The size
distribution of aggregate in the pile was determined by
scooping 12 samples from the pile and dry sieving them.
Moisture content was determined by measuring the sample
weight loss after oven drying.  The age and configuration of
the storage pile, and the load capacity of the high loader,
about 15 tons, were also recorded.

Equipment Configuration—The distance between the dump truck
and the profile grid was not documented.  From a figure
depicting the sampling configuration, it appears to be
roughly five meters.  The profiler sampling heads were
arranged as follows:  one each at the top and bottom of a
vertical support pole, and four others spaced evenly on a
horizontal support pole, which bisected the vertical one.
This structure was fixed on the top of a van.  The top
sampler was about 12.5 feet above the ground over the center
of the truck.  The two samplers on both sides of the
horizontal support were 1.5 and 4.5 feet from the vertical
support.  They were about 9.5 feet above the ground.  The
bottom sampler was approximately 6.5 feet high.  Thus, the
profiler spanned about nine feet horizontally and six feet
vertically, with its center roughly 9.5 feet above the
ground.

The cascade impactor was also on top of the truck, about 8
feet above the ground and 1 foot from the vertical support
pole of the profiler.  The positions of the two standard hi-
                         2-142

-------
vol samplers were not documented in this report.  Wind speed
and direction were continuously monitored at a height of 12
feet.

Sampling Runs—Only two sampling runs were conducted in this
field experiment.  One run consisted of 86 dumps, and the
other 80.  In both cases about 150 tons of aggregate were
loaded.

Quality Assurance—In the use of standard high volume
filtration, the investigators followed the procedures
specified by EPA in "Reference Method for the Determination
of Suspended Particulates in the Atmosphere (High Volume
Method)" (1971).  For the measurement of dust deposition,
the investigators followed the procedures set forth in
"Standard Method for Collection and Analysis of Dustfall,"
ASTM Method D 1739-62.

The wind speeds for the two tests were 12.6 and 14.0 mph.
Thus the isokinetic ratios were, respectively, .79 and .71.
No corrections were made to the calculated emission factors
to account for the sub-isokinetic sampling.

Filte'rs were conditioned in a controlled temperature and
humidity environment prior to weighing both before and after
collection of dust samples.  Filter samples were transported
to the laboratory in individual folders.  The interior
surfaces of the sampler heads were rinsed, and the water was
captured and later evaporatea to determine the mass of dust
on the interior surfaces.

The investigators acknowledged two potential sources of
small particle bias in their measurement of particle size
distribution:  1) particles bouncing down through the
cascade impactor to smaller particle stages; 2) non-
isokinetic sampling which collects larger particles with
lower efficiency than smaller particles.

It should be noted that, although this work was described as
determining an emission factor for the aggregate load out
process, emissions were measured only from the dropping of
aggregate from the high loader into the dump truck.
Emissions generated when the loader scooped the aggregate
from the pile or when it moved between the pile and the
truck were not measured or discussed.

Findings—The exposure measurements and calculated emission
factors for the two sampling runs are presented in Table
2.58.  The investigators concluded that an average TSP
emission factor for this operation was about 0.05 Ib/ton
based on the following:
                         2-143

-------
      1.    The TSP emission factor for these two tests were 0.053
           and 0.063 Ib/ton.

      2.    Emissions from these particular tests were thought to
           be near the maximum for this type of operation because
           the tested aggregate had been crushed less than one
           week earlier and had remained dry,  the wind velocity
           was high, the aggregate sizes tested were small, and
           the amount of fines in the piles was substantial.

  TABLE 2.58.  MEASURED EXPOSURES AND  CALCULATED EMISSION FACTORS

Run
15
16

Sample
Time
(min)
61 .2
59.1
Sampler Position
Height
Above
Grade
(feet)
6.5
9.5
9.5
9.5
9.5
12.5
6.5
9.5
9.5
9.5
9.5
12.5
Distance
from
Center
of Truck
(feet)
0
4.5
1 .5
1 .5
4.5
0
0
4.5
1 .5
1 .5
4.5
0

Exposure
(mg/in2/ton)
2.59
2.75
2.92
3.74
2.60
1 .64
1 .07
1 .48
2.66
3.32
2.28
3.68
Emission Factors (Ib/ton)
Total
0.11
0.1 1
d" > 30 (im
0.057
0.047
2 < da < 30 ^m
0.018
0.021
d* < 2
|im
0.035
0.042
a) particle Stokes diameter
      Conclusions From The Two Studies—The investigators assumed
 that the total emissions from an aggregate storage pile area
 equaled the sum of the emissions from the following four sources:

      1.    Loading of aggregate into storage piles
      2.    Equipment travel in the storage area
      3.    Wind erosion
      4.    Load out of aggregate for shipment

      The first study measured the total emissions from the
 storage area (0.33 Ib/ton) and emissions produced by wind erosion
 (0-11  Ib/ton); the second study measured emissions from aggregate
                               2-144

-------
load out (0.05 Ib/ton).   Emissions from source #1 were apparently
assumed to be similar to those from source #4:  the factor for
loading aggregate into storage piles was taken to be 0.04 Ib/ton.
Thus, estimates were available for the total operation and every
component except equipment travel.  It was calculated by
subtraction as 0.13 Ib/ton.

     The precipitation-evaporation index was found to be the most
useful parameter in characterizing regional variability in total
emissions from aggregate storage operations.  The corrected
emission factor for each ton of aggregate placed in storage was
presented as

                          e-    °'33
                               (P£-/100)2

where

     e    =    TSP emission factor, Ib/ton placed in storage
     PE   =    precipitation index

The method of estimating this relationship between the emission
factor and the PE index was not discussed.

     Publication—These two field studies were conducted and
documented under a contract with the EPA Office of Air and Waste
Management, Office of Air Quality Planning and Standards,
Research Triangle Park, North Carolina.  The report was published
in 1974 as Publication No. EPA-450/3-74-037.

Study 2— Bonn et al.  Fugitive Emissions From Integrated Iron
          and Steel Plants.  EPA-600/2-78-050.  1978.

     Methodology—An exposure profiler similar to the one used by
Cowherd et al. (1974) was used to measure the mass flux of dust
downwind from two operations related directly to aggregate
storage piles.  A third source, the transfer of aggregate between
perpendicular conveyor belts, was also tested.  Although this
source is not directly associated with storage piles, the action
is very similar to that seen in storage pile operations.

     Test Sites—Sampling was conducted at two integrated iron
and steel plants.  One was in the dry western U.S.  It included
tests of emissions from the load-out of slag from a storage pile
into a dump truck, and from the addition of pelletized or lump
iron ore to existing storage piles by a mobile stacking conveyor
belt.  At the second facility, in the eastern steel producing
region of the country, tests were conducted on emissions from a
transfer point between two perpendicular conveyor belts carrying
sinter.
                              2-145

-------
     Parameters and Equipment—The exposure profiling grid
employed two different types of air filtering devices.  Most
prevalent was the isokinetic high volume sampler developed by
Midwest Research Institute.  This sampler differed slightly from
the kind employed in 1974  (Cowherd et al.) in that the filter was
positioned horizontally and air was drawn in through a settling
chamber and up through the filter.  The settling chamber captured
particles larger than about 50 |u.m in diameter.  Smaller, lower
capacity ("auxiliary") samplers were used at the two ends of the
horizontal rod.  These also had intakes facing into the wind to
allow isokinetic sampling, but they did not have settling
chambers.

     Standard hi-vol samplers were used to determine upwind and
downwind concentrations of TSP.  A high volume cascade impactor
with a cyclone preseparator provided data on the plume's particle
size distribution.  It sampled isokinetically at a wind speed of
10 mph.  The cut points for this' impactor were not specified.

     The transferred aggregate was characterized by recording the
material type, its moisture content, its texture (including
percent silt), and the throughput rate.  Moisture content was
determined by weight loss after oven drying, and the texture was
measured by dry sieving.

     Meteorological parameters for which data was collected
included wind speed and direction, cloud cover, temperature, and
relative humidity.  Anemometers monitored wind speed and
direction prior to commencement of testing and continuously
recorded wind conditions during the runs; thus, the intake
velocity of the samplers could be set to match the wind speed
and, after testing was completed, exposures could be corrected
for deviations from isokinetic sampling.

     Equipment Configuration—For the storage pile load-out
operation, the profiling grid was placed two meters downwind from
the dump truck.  The grid consisted of a vertical array of two
isokinetic hi-vol samplers and a horizontal array of two
isokinetic hi-vols  (closest to the vertical support) and two
auxiliary isokinetic samplers  (at each end of the horizontal
support).  The vertical boundaries of the grid were typically 2.5
meters high on the bottom and 6.25 meters high on the' top; the
width of the grid was normally 4.2 meters (i.e. 2.1 meters on
each side of the vertical support).  The horizontal array of
samplers had intakes 4.5 meters above the ground.

     For the conveyor belt stacking operation, the profiling grid
was about five meters from the center of the pile on the downwind
side.  The vertical array consisted of four isokinetic hi-vol
samplers fixed at heights of 1, 2, 3, and 4 meters.  The
horizontal support had only two auxiliary isokinetic samplers
with lateral displacements from the center support of 1.4 meters.

                              2-146

-------
     Emissions from the conveyor transfer station were measured
using a smaller scale sampling grid placed an unspecified
distance downwind from the transfer point.  This grid had three
profiling samplers 1.6 meters above the ground and one each at
heights of 1.1 and 2.2 meters.  The two outside samplers on the
horizontal support were about one meter from the center support.
Only auxiliary type samplers were used on this profiler.

     For all three of these operations, the hi-vol cascade
impactor and a standard hi-vol were attached to the horizontal
support.  The upwind hi-vol sampled at a height of 2 meters; its
upwind distance was not noted.  Anemometers measured wind speed
at two heights on the profile grid.  Any assumptions regarding
the vertical distribution of wind speed were not made explicit.

     Sampling Runs—Three sampling runs were conducted for the
conveyor transfer station, and six each were conducted for the
aggregate load out and conveyor stacking operations.  Each 30- to
40-minute test of the load out operation was conducted while
about 150 tons of material were loaded.  During each 15-minute
test of the conveyor transfer process, 52 tons of sinter were
transferred between the two belts.  The mass stacked by the
conveyor stacker per sampling run ranged from 216 to 500 tons.
The run time for testing of this operation was between 13 and 30
minutes.

     Quality Assurance—The issue of quality assurance was not
explicitly addressed in this report.  Standard quality assurance
procedures such as calibration of air samplers, processing of
blank filters, and auditing of sample weights were not
documented.

     Field data presentation was complete.  The procedures for
determining the dust mass on the filters were described in
detail.  Air filters and impaction substrates were stored and
transported in separate envelopes and were allowed to equilibrate
in a constant temperature and humidity environment before
weighing.

     Some of the computational procedures employed were not
described in detail and, therefore, could not be repeated.  The
authors did not explain how they integrated the point values  of
exposure when there were two different types of exposure samplers
measuring two different particle size ranges.  Recall that  the
isokinetic hi-vol samplers have a settling chamber, whereas the
auxiliary samplers do not.  Thus, the filters in the hi-vol heads
collect dust particles smaller than about 50 urn, while  the
filters in the auxiliary samplers collect particles with no
maximum particle diameter.  No indication was given that these
                              2-147

-------
disparate sample types were treated any differently in computing
what the investigators refer to as the "integrated filter
exposure."

     Findings--The measured point values of exposure and the
calculated emission factors for material load-out, conveyor
stacking, and conveyor transferring are shown respectively in
Tables 2.59, 2.60, and 2.61.  Those sampling heads for which no
filter exposure is given are the smaller capacity auxiliary
samplers.  Again,  the method used to integrate exposure with
respect to horizontal position is unclear,  due to the lack of
"filter exposures" for each sampling head.   Hence, it is unclear
how the values for "integrated filter exposure" were calculated.

     Two predictive equations were developed from this data and
the findings of a field study of emissions from a sand and gravel
storage area (Cowherd et al.,  1974).  The model developed for the
conveyor stacker is


                        EF - 0.0018 \ 5 '' 5
where

     EF   =    emission factor for suspended particulates,  Ib/ton
               transferred
     s    =    aggregate silt content,  %
     U    =    wind speed, mph
     M    =    aggregate moisture content, %

The model for aggregate load-out from a high loader to a dump
truck is basically the same, except for the addition of a new
correction factor:  Y is the effective capacity of the loader in
cubic yards.  The equation is
                        EF - 0.0018
                                    -V-
                                    55
The method used to develop these equations was not described.

     Publication—This study was conducted and documented under a
contract with the Environmental Protection Agency; Industrial
Research Laboratory; Office of Energy, Minerals, and Industry;
Research Triangle Park, North Carolina.  It was published as
Publication No. EPA-600/2-78-050 in March 1978.

                              2-148

-------
TABLE 2.59.
MEASURED EXPOSURES AND CALCULATED
        FOR MATERIAL LOAD-OUT
                                               EMISSION FACTORS
Run
A1
A2
A3
A4
A5
Duration
(min)
30
40
30
30
40


Height
(meters)
3
4.5
4.5
4.5
4.5
6
2.5
4.37
4.37
4.37
4.37
6.25
2.5
4.37
4.37
4.37
4.37
6.25
2.5
4.37
4.37
4.37
4.37
6.25
2.5
4.37
4.37
4.37
4.37
6.25
Distance
from
Centerline
(meters)
0
2.1 R
0.7 R
0.7 L
2.1 L
0
0
2.4 R
0.7 R
0.7 L
2.4 L
0
0
2.4 R
0.7 R
0.7 L
2.4 L
0
0
2.4 R
0.7 R
0.7 L
2.4 L
0
0
2.4 R
0.7 R
0.7 L
2.4 L
0
Total
Exposure
(mg/cm2)
274
41 .2
99.1
182
76.0
74.1
88.8
16.4
77.8
80.9
12.5
34.0
454
51 .6
169
285
104.7
134
63.4
23.9
31 .4
35.9
24.2
10.8
20.5
9.1
13.0
12.0
7.3
n.d.
Filter
Exposure
(mg/cm2)
51 .0

22.7
40.4

23.8
14.09

14.7
25.5

12.3
52.2

29.5
47.6

27.2
8.0

4.4
3.1

3.1
3.7

1 .9
2.9

n.d.
Integrate
d Filter
Exposure*
(Ib/ton)
0.15
0.062
0.16
0.032
0.013
Emission
Factor"
(Ib/ton)
0.056
0.028
0.059
0.030
0.011
                             2-149
                                         (continued)

-------
                      TABLE 2.59.  (continued)
Run
A6
Duration
(min)
40
Height
(meters)
2.5
4.37
4.37
4.37
4.37
6.25
Distance
from
Centerline
(meters)
0
2.4 R
0.7 R
0.7 L
2.4 L
0
Total
Exposure
(mg/cm2)
61 .2
14.9
21 .7
41 .0
32.7
5.9
Filter
Exposure
(mg/cm2)
9.0

5.5
11 .0

3.0
integrate
d Filter
Exposure1
(Ib/ton)
0.017
Emission
Factor"
(Ib/ton)
0.011
a)  corrected to isokinetic sampling conditions
b)  for particles with Stokes' diameters < 30 [i
 TABLE 2.60.
MEASURED EXPOSURES  AND CALCULATED EMISSION
         FOR CONVEYOR STACKING
FACTORS
Run
A8
A9
Duration
(min)
30
15
Sampling
Height
(meters)
1
2
2
2
3
4
1
2
3
3
3
4
Distance
from
Centerline
(meters)
0
1 .4 R
0
1 .4 L
0
0
0
0
1 .4 L
0
1 .4 R
0
Total
Exposure
(mg/cm2)
113
18.1
21 .7
12.6
11
3
51
48
45.0
62
46.8
26
Filter
Exposure
(mg/cm2)
25.5

5.8

2.4
0.8
19.7
14.6

16.7

6.2
Integrated
Filter
Exposure'
(Ib/ton)
0.0041
0.024
Emission
Factor"
(Ib/ton)
0.0040
n.d.
                                2-150
                                            (continued)

-------
                       TABLE 2.60.    (continued)
Run
A10
A11
A12
A13
Duration
(min)
13
22
25
28
Sampling
Height
(meters)
1
2
3
3
3
4
1
2
2
2
3
4
1
2
2
2
3
4
1
2
3
3
3
4
Distance
from
Centerline
(meters)
0
0
1 .4 R
0
1 .4 L
0
0
1 .4 L
0
1 .4 R
0
0
0
1 .4 R
0
1 .4 L
0
0
0
0
1 .4 L
0
1 .4 R
0
Total
Exposure
(mg/cm2)
70
61
31 .0
58
30.3
8
38.5
15.1
14.7
9.9
11 .5
4.0
10.5
8.0
5.5
1 .7
3.72
1 .78
1 .39
1 .65
2.09
2.05
3.62
1 .59
Filter
Exposure
(mg/cm2)
20.6
12.6

15.7

8.5
5.4

2.1

1 .3
0.8
0.9

0.6

0.4
0.4
0.3
0.5

0.5

0.3
Integrated
Filter
Exposure*
(Ib/ton)
0.038
0.0038
0.00058
0.00031
Emission
Factor"
(Ib/ton)
0.010
0.00099
0.00066
; 0.00046
a)  corrected to isokinetic  sampling conditions
b)  for particles with Stokes' diameters < 30  urn
                                  2-151

-------
 TABLE 2.61.  MEASURED EXPOSURES AND CALCULATED  EMISSION FACTORS
                       FOR CONVEYOR TRANSFER
Run
E10
E11
E12
Duration
(min)
15
15
15
Sample
Height
(meters)
2.2
1 .6
1 .6
1 .6
1 .1
2.2
1 .6
1 .6
1 .6
1 .1
2.2
1 .6
1 .6
1 .6
1 .1
Distance
from
Centerline
(meters)
0
1
0
1
0
0
1
0
1
0
0
1
0
1
0
Total
Exposure
(mg/cm2)
16.8
17.2
39.5
51 .0
32.2
45.6
26.8
31 .2
57.1
30.4
16.1
31 .2
20.3
14.6
18.6
Integrated
Filter
Exposure8
(Ib/ton)
0.043
0.084
0.038
Emission
Factor"
(Ib/ton)
0.036
0.064
0.037
  a) corrected to isokinetic sampling conditions
  b) for particles with Stokes' diameters < 30 (i
Study 3— Blackwood and Chalekode.  Source Assessment:   Crushed
          Stone.   EPA-600/2-78-004L.  1978.

     Methodology--The investigators used upwind-downwind
dispersion modeling to estimate emission factors  for unloading of
processed rock  from trucks into storage piles.  Turner's (1970)
equation for a  ground level source with no plume  rise was used to
model the source strength:
                       x-
exp
                                         y
                               2-152

-------
The source strength for the activity was calculated as the
average of those emission rates calculated at several downwind
stations.

     Test Sites—Field monitoring was conducted at two
unidentified stone quarry and processing operations.  Both
produced crushed traprock.  They were considered representative
of the crushed stone industry, particularly because 68% of
crushed stone produced (in 1978) was traprock.

     Parameters and Equipment—A GCA portable respirable dust
monitor was used to measure concentrations downwind from the
loadout site.  When used with a cyclone separator, it measured
concentration of "respirable" (i.e. smaller than 10 ^.m)
particles.  The equipment used to monitor upwind concentration
was not noted.  Wind speed as measured by an anemometer was
recorded automatically every 15 seconds.  The average wind speed
during each sampling run was calculated as the average of these
15-second averages.  The downwind distance from the source was
estimated by pacing between the source and the monitor.
Stability class was determined and reevaluated every two or three
hours on the basis of cloud cover, wind speed, and time of day.

     Equipment Configuration—Little documentation is provided
regarding the manner in which the respirable dust monitor was
deployed.  When used for testing various rock processing
operations, it was used to collect several concentration
measurements in a traverse downwind from the source.  Although
this was probably not possible for a quick batch operation like
truck unloading, no explanation was given for the positioning of
the sampler downwind.from this activity.  However, tabulated
field data indicate that the sampler's downwind distance ranged
from 40 to 210 feet.  The sampling height of the monitor was not
published.

     Sampling Runs--A total of 13 concentration measurements for
truck unloading operations were reported in this document.  The
number of dumps was not recorded.  Sampling duration was four
minutes for each run,  except one eight minute test.

     Quality Assurance—No documentation was provided for
calibration of the dust monitor or the cyclone separator that was
used with it.  Field procedures were not thoroughly explained.
As was noted earlier,  the method of measuring the upwind
concentration was not discussed.

     The method of measuring emissions is clearly inappropriate
for the source, because unloading of a truck only requires a
fraction of the time required to collect the air sample.
Consequently, concentration measurements are biased low.  This
                              2-153

-------
issue was not addressed by the authors, and it was in effect
hidden by the lack of documentation of activity level during each
sample collection.

     Findings—The dispersion modeling data published in this
report is shown in Table 2.62.  Note that source strength is
sometimes given as a dose and sometimes as an emission rate.  No
justification was given for this.   The respirable particulate
emission factors calculated for plants B and A, respectively,
were 0.0746 and 0.033 g/metric ton transferred.  The mean
emission factor of 0.0538 g/metric ton was given a 95% confidence
interval of ± 0.264 g/metric ton.  The total particulate emission
factors for plants B and A were 0.209 and 0.0442 g/metric ton,
respectively-  The mean total particulate factor, 0.127 g/metric
ton, was given a 95% confidence interval of ± 1.05 g/metric ton.
The method of deriving the emission factors for each plant was
not described.

     Publication—This report is Publication No. EPA-600/2-78-
004L, prepared for EPA's Industrial Environmental Research
Laboratory in Cincinnati, Ohio.  It was published in May of 1978.

    TABLE 2.62.  DISPERSION MODELING DATA FOR UNLOADING TRUCK
Plant
B
B
B
B
B
B
B
B
B
A
A
A
A
Wind
Speed
(mph)
5.3
5.3
5.3
5.3
5.3
5.3
5.3
5.3
5.3
8
8
8
8
Monitoring Position (feet)
X
123
70
90
100
90
70
90
100
90
210
70
70
210
Y
0
0
0
0
0
0
0
0
0
0
0
0
0
z
0
0
0
0
0
0
0
0
0
0
0
0
0
Concentration
(Hg/m3)
140
10
25
28
20
22
672
27
72
10
80
204
4
Source
Strength
8.0 g
.2043 g
.8081 g
2.194 g
.6465 g
.4495 g
21 .72 g
2.116 g
2.327 g
.005021
g/sec
.00551 1
g/sec
.01405
g/sec
.002008
g/sec
Stability
Class
B
B
B
B
B
B
B
B
B
C
C
C
C
                               2-154

-------
Study 4— Cowherd et al._  Iron and Steel Plant Open Source
          Fugitive Emission Evaluation.  EPA-600/2-79-103.  1979.

     Methodology—Dust emissions from a mobile conveyor stacker
were measured using the exposure profiling methodology.  Because
the source was treated as a moving point rather than a fixed
source, there was no need to measure exposure at several points
in the horizontal plane.  Total exposure was calculated as the
mass emitted per stacker-mile.  This was then converted to an
emission factor by multiplying by the stacker's velocity  (i.e
m/sec) and the inverse of the stacking rate  (hr/ton).

     Test Site—The stacker was used to form elongated storage
piles of iron pellets and coal.  The length of the conveyor
stacker boom was about ten meters for the coal stacking operation
and 67 meters for the iron pellet stacking operation.  The drop
distance for the iron pellets ranged from 9 to 12 meters; for the
coal it was 5 meters.  The facility at which these tests were
conducted was not identified.

     Parameters and Equipment—Upwind and downwind concentrations
of total suspended particulates were measured using standard hi-
vol samplers.  A high volume cascade impactor was used to measure
the particle size distribution of the dust downwind from the
storage area.  A profiling tower consisting of four isokinetic
hi-vol samplers was used to measure point values of exposure at a
series of heights above the ground.

     Wind speed and direction were continuously monitored with
recording anemometers.  In setting the sampling velocity to match
the local wind speed, the vertical distribution of the wind speed
was assumed to be logarithmic.  Other meteorological parameters
for which data was collected included cloud cover, temperature,
precipitation, and relative humidity.

     In order to characterize the stored material, data on
several relevant parameters was collected.  The material type
(coal versus iron pellets) and the throughput rate were noted.
Aggregate samples were taken and analyzed using the usual methods
to determine the moisture and silt content.

     Equipment Configuration—For each sampling run the profiler
was placed between 4.5 and 11.5 meters downwind from the pile
being formed.  The sampling heads had intakes 1.5, 3,  4.5, and 6
meters above the ground.  A high volume cascade impactor and a
standard hi-vol were set up beside the profiler; they  both
sampled at a height of 2 meters.  For some tests additional
downwind standard hi-vols were placed 20 and 50 meters from the
source.  The upwind hi-vol, which also sampled air two meters
                              2-155

-------
above the ground,  was placed 30 to 50 meters upwind or crosswind
from the source.  Only one measurement height was given for wind
speed, four meters.  For some tests this was done 10 meters
upwind from the source and in others 50 meters downwind.

     Sampling Runs—Emission factors were calculated for four
sampling runs, three of which were for iron pellet stacking, and
the fourth was for coal stacking.  For iron pellets, the number
of stacker passes in front of the profiler varied between 7 and
11.  Thirty passes were logged in the coal stacking test.

     Quality Assurance—An effort was made to apply or adapt the
American Society of Testing and Materials (ASTM)  Standards in the
collection and analysis of samples needed to quantify the silt
content of the surface material.  Except for the citation of this
standard procedure, the authors documented no normal quality
assurance procedures.  Documentation of the basic methodological
procedures involved in field operations, sample handling,  and
data analysis was generally thorough.

     Dust samples were transported to the laboratory in
individual envelopes.  Filter samples were conditioned at
constant temperature and humidity for 24 hours before weighing.
This same procedure was followed in weighing the filters prior to
use.

     The procedure for accounting for background levels of dust
was not adequately documented.   Therefore,  the emission factors
presented in the study could not be reproduced from the data
provided.

          Findings—Table 2.63  shows the measured exposures and
resulting emission factors for  each sampling run.  The duration
of the runs was not documented.
                              2-156

-------
TABLE 2.63  EXPOSURE MEASUREMENTS AND CALCULATED EMISSION FACTORS
Run
H-10
H-11
H-12
F-19
Sample
Height
(meters)
1 .5
3
4.5
6
1 .5
3
4.5
6
1 .5
3
4.5
6
1 .5
3
4.5
6
Total
Exposure
(mg/cm2)
12.1
5.88
3.18
4.13
0.92
0.74
0.50
0.10
3.45
1 .15
1.11
1 .82
0.82
0.34
0.35
0.27
Filter
Exposure
(mg/cm2)
2.43
1 .43
0.89
2.56
0.42
0.62
0.46
0.09
1 .38
0.35
0.80
1 .59
0.42
0.21
0.19
0.15
Integrated
Filter
Exposure
(Ib/mile)
319
90.8
202
65.6
TSP
Emission
Factor
(Ib/ton)
0.0023
0.0029
0.0023
0.00014
FP
Emission
Factor
(Ib/ton)
0.00012
0.00087
0.00069
0.000011
     A revised predictive equation for continuous stacking was
developed from this field study and previous research (Bohn et
al.,  1978):
where
     EF

     S
     M
     U
     H
                       EF - 0.0018
                  MM
                     (I)2
TSP emission factor, Ib/ton of aggregate
transferred
silt content, %
moisture content, %
mean wind speed, mph
drop height, meters
                              2-157

-------
The equation is the same as the one published in the earlier
report except for the addition of the correction factor for drop
height .

     Publication- -This study was conducted and documented under a
contract for EPA, Industrial Environmental Research Laboratory,
Office of Energy, Minerals, and Industry, Research Triangle Park,
North Carolina.  It was published in 1979 as Publication No. EPA-
600/2-79-103.

Studies of Secondary Importance

Study 5-- Vekris.  "Dispersion of Coal Particles from Storage
          Piles."  Hydro Research Quarterly.  Vol. 23.  No. 2.
          1971 .

     The emission rate from a heavy duty vehicle moving across a
coal storage pile was calculated using the equation

                                wuhs
                                 nV

where

     Q    =    mass emitted per unit time
     w    =    weight gain of high volume filter
     u    =    wind speed
     h    =    height of the dust plume
     s    =    vehicle speed
     n    =    number of vehicle passes
     V    =    volumetric flow rate of sampler
No
   explanation was given for how this equation was developed.

     A high volume sampler was positioned at the edge of the
pile, downwind from the vehicle, and three feet above grade.  A
total of six samples were collected.  The duration of these
samples was not published.  Calculated dust generation rates
ranged from 13.42 g/sec to 66.38 g/sec.

     The source studied in this work does not strictly fit the
definition of the storage pile source category.  The author noted
that the emissions are from the action of the vehicle against the
surface of the storage pile, rather than from the transfer of
material into or out of the pile.  Neither does it fit the
definition of the unpaved road source category.  However, it was
felt that the study is worthy of mention in this report.
                              2-158

-------
CONSTRUCTION ACTIVITIES

Introduction

     Field measurement of particulate emissions related to
construction activities has been performed at three sites.  Dust
emissions from two building construction sites and a road
construction site were modeled using the upwind-downwind
dispersion modeling method.  Two different reports (Jutze and
Axetell, 1974; Cowherd et al.,  1974) document the field research
and findings for the building construction sites.  Both reports
cover both sites, but, because they do not provide the same
documentation and data analysis, they are reviewed separately -
The report on road construction emissions (Kinsey and Englehart,
1983) included, in addition to upwind-downwind dispersion
modeling of emission rates, a regression analysis relating TSP
concentration to other field variables.  These three reports are
reviewed here in chronological order.

Studies of Primary Importance

Study 1-- Jutze and Axetell.  Investigation of Fugitive Dust
          Volume I: Sources, Emissions, and Control.   EPA-450/3-
          74-036-a.  1974.

     Methodology—Upwind-downwind dispersion modeling was used to
estimate emissions from building construction sites.   Individual
construction operations, such as grading or materials unloading,
were not tested separately.  The investigators used the Gaussian
dispersion model for concentrations along a plume centerline from
a ground level source with no effective plume rise:
                            x-


     Two adjustments were made to this equation.  First, because
the construction sites were considered area sources, the initial
standard deviation of the crosswind distribution of the plume
concentration, ayo, was assumed to be  equal  to  the length of  the
side of the construction area divided by 4.3.  Second, because
the sampling time was longer than three minutes, the above
formula for concentration was multiplied by 0.36.  Both of these
adjustments follow the recommendation of Turner  (1970).  The
resulting formula for source strength, Q (g/sec), is shown below:

                         Q - 2.787ioyozux
                              2-159

-------
where

     ay   =    standard deviation of the crosswind distribution
               of the plume's concentration at the downwind
               measurement site (a function of atmospheric
               stability, downwind distance, and the assumed
               initial standard deviation, 0yo) .
     oz   =    standard deviation of the vertical distribution of
               the plume's concentration at the downwind
               measurement site (a function of the atmospheric
               stability and downwind distance),  m
     u    =    mean wind speed, m/sec
     X    =    measured concentration of particulates at the
               downwind measurement site  (minus the background
               concentration),  g/m3

Calculated source strengths were converted into tons/year and
into tons/acre-month.

     Test Sites—Data were collected from two building
construction sites:  a 100 acre site in Las Vegas, and a 90 acre
site in Maricopa County  (Phoenix), Arizona.

     Parameters and Equipment—Standard high volume samplers were
used to measure concentration of TSP both upwind and downwind of
the construction sites.  Wind velocity and direction were
measured using continuous windvane/anemometer sensors.

     Equipment Confiquration--The specific positioning of the
samplers was not documented, except that the downwind measurement
distance was given in a table showing the dispersion model
parameters.  At the Las Vegas construction site,  concentration
measurements were made at downwind distances of 650 and 525
meters.  At the Maricopa County site, concentration measurements
were made at three downwind distances, 315, 758,  and  1575 meters.
The upwind distance of the background sampler(s)  was not
published.

     Sampling Runs—Four concentration measurements were taken
for the Las Vegas site, two at each of the downwind distances.
For the Maricopa County site,  12 concentration measurements were
taken, five at the closest sampling distance, 3 at the
intermediate distance, and 4 at the farthest distance.  Samples
were collected over 24-hour periods.

     Quality Assurance—The investigators provided little
assurance of the quality of the data collected for this study.
Although the procedures for operating the hi-vol samplers and for
handling and transporting the  filters were thoroughly documented
the procedures used to determine the filtered dust mass were not'
described.  Normal quality assurance procedures, such as
calibration of samplers, processing of blank filters, or auditing

                               2-160

-------
of weight measurements  or data reduction calculations, were not
documented for this  study-

     Findings—The field data collected in the  study of fugitive
dust emissions from  construction sites is shown in Table 2.64.
The average emission rate for the Las Vegas  site was 1162
tons/year, which  converts into about 1 ton/acre/month.  The
average for the Maricopa County site was 1970 tons/year; this is
equivalent to  1.8 tons/acre/month.
   TABLE 2.64.
DISPERSION MODELING  DATA AND CALCULATED  SOURCE
            STRENGTHS
Site
Las
Vegas
Las
Vegas
Las
Vegas
Las
Vegas
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Downwind
distance
(meters)
650
650
525
525
315
315
315
315
345
758
758
758
1575
1575
1575
1575
Stability
Class
C
B
C
C
B
B
B
B
B
B
B
B
B
B
B
B

-------
Standards, Research Triangle Park, North Carolina.  It was
published in 1974 as Publication No. EPA-450/3-74-036-a.

Study 2— Cowherd et al.  Development of Emission Factors for
          Fugitive Dust Sources.  EPA-450/3-74-037.  1974.

     Methodology—This report documents the field research and
findings of the same study described by Jutze and Axetell (EPA-
450/3-74-036a, 1974).  However, more concentration measurements
were available to Cowherd et al. for analysis and modeling than
were included in that report.  It should be noted that both Jutze
and Axetell were also co-authors of this document (EPA-450/3-74-
037) .  This report sheds additional light on the positions of the
high volume samplers relative to the construction sites and to
the wind direction.  It also includes more extensive data
analysis and interpretation than Jutze and Axetell, 1974.

     Prior to using the dispersion model, the investigators
prepared and analyzed pollution roses for each of the hi-vol
stations to determine if the receptors were being influenced by
other local dust sources.  Some receptors were excluded from
formal analysis on the basis of this cursory evaluation.  Source
strengths were modeled separately for several different wind
directions and compared subjectively before a final emission
factor was selected.

     The dispersion modeling approach employed in this report is
similar to that used by Jutze and Axetell.  It includes the same
assumptions regarding initial plume dispersion and concentrations
for long sampling periods.  The primary difference between the
method used by Jutze and Axetell and that of Cowherd et al.  is
that Cowherd et al. averaged net downwind concentrations for a
particular wind direction prior to calculating a corresponding
average source strength.

     Test Sites—Emissions were modeled for a 100 acre building
construction site in Las Vegas and for a 80 acre building
construction site in Maricopa County, Arizona.  Note that the
size of the Maricopa site is ten acres less than the size stated
in Jutze and Axetell's report.  Both of these sites were watered
on most air sampling days.

     Parameters and Equipment—Upwind and downwind concentrations
of total suspended particulates were measured with standard hi-
vol samplers.  Wind direction and velocity were monitored using a
continuous windvane/anemometer.  Atmospheric stability was not
recorded or discussed.  Construction activity levels were
subjectively rated as "no activity," "light to moderate
activity," or "heavy activity" for each sampling period.

     Equipment Configuration—For the Maricopa County site,  six
standard high volume samplers were distributed as follows.'  One

                              2-162

-------
sampler was placed adjacent to the construction site on the east
side (typically the downwind side).   This was about 315 meters
from the center of the construction site.  Three others were set
up in the northeast quadrant (relative to the construction site),
in anticipation of predominantly southwesterly winds.  One of
these samplers was later judged to provide an unrepresentative
sample, and measurements collected with it were excluded from the
analysis.  The downwind distances from the center of the
construction site to the remaining samplers were 731 meters and
1021 meters.

     Two samplers were set up on the anticipated upwind side of
the site, but one of these was also judged to provide a poor
measure of background dust levels and was consequently excluded
from the analysis.  The remaining upwind sampler was roughly 1520
meters from the center of the site.   The intake height of the hi-
vols was not noted.

     Five hi-vols were used for the Las Vegas site.  One was
adjacent to the southwest corner of the site.  The other four
were distributed to the north and east of the site at distances
ranging from about 8.1 km to 13 km from the center of the site.
They were intended to measure downwind concentration.  The
placement of the wind instruments was not documented for either
site.

     Sampling Runs—Each sample was collected over a 24-hour
period.  For the Maricopa County site, 24 sampling runs were
conducted over a 7-^ week period.  A total of 88 concentration
measurements made during these sampling runs were published,
including 24 background samples.  For the Las Vegas construction
site, 30 sampling runs were conducted, from which 125
concentration measurements were published.  These samples spanned
a nine week period.

     The term sampling run should be considered loosely for this
report, because concentrations measured on a given day were not
always used to generate a single corresponding source strength
from the model.  Rather, average net downwind concentrations were
calculated for various wind directions and converted into
corresponding emission factors using the dispersion model
described above.

     Quality Assurance—The investigators provided  little
assurance of the quality of the data collected for  this study.
Although, the procedures for operating the hi-vol samplers and
for handling and transporting the filters were thoroughly
documented in the report by Jutze and Axetell, the  methods used
to determine the filtered dust mass were not described.  Normal
quality assurance procedures, such as calibration of samplers,
processing of blank filters, or auditing of weight  measurements
or data reduction calculations were not documented  for  this

                              2-163

-------
study.  As was noted earlier,  not all of the data needed for the
dispersion equation were presented in this report.

     Several problems regarding the configuration of the samplers
diminish the quality of the data gathered in the project.  No
information was given on the intake height of the samplers.  They
could have been on roof-tops or on the ground.  It is apparent
that the hi-vols did not always collect samples from the plume
centerline; this is contrary to the assumptions of the dispersion
model applied.  The report also indicates that other significant
dust sources between the upwind and downwind samplers probably
affected the concentration measurements, particularly at the Las
Vegas site.

     Findings—During the field work at the Maricopa County site
the winds were generally from the south, west, or southwest.
Recall that the three downwind samplers were all northeast of the
site.  To determine which wind direction yielded the most
accurate measure of source strength, the investigators used the
data presented in Table 2.65.   They concluded that, because the
two estimates of source strength for southwesterly winds were
closest together, they must be closer to the true source
strength.  Thus, the emission factor for this site was taken to
be 1.4 tons/acre/month.
    TABLE 2.65,
SOURCE STRENGTHS CALCULATED FOR VARIOUS WIND
          DIRECTIONS

Wind
direction
Southwest
South
West
Modeled Source strength (tons/acre/month)
Based on only one downwind
sampler, adjacent to the
site on the east side
1 .37
1 .13
0.42
Based on average of
all three downwind
samplers
1 .41
1 .51
0.65
     For the Las Vegas site separate source strengths were
calculated for southwest winds and for north winds.  They were,
respectively, 0.6 and 0.96 tons/acre/month.  The investigators
judged that the receptors were affected least by other local dust
sources during north winds, and concluded that emissions from the
construction site were about 1.0 tons/acre/month.

     For neither site was a strong correlation detected between
source strength and intensity of construction activity.

     Publication—This study was conducted and documented under a
contract with the EPA Office of Air and Waste Management, Office
of Air Quality Planning and Standards, Research Triangle'Park,
                              2-164

-------
North Carolina.  It was published in 1974 as Publication  No.  EPA-
450/3-74-037.

Study 3— Kinsey and Englehart.  Study of Construction Related
          Dust Control.  1983.

     Methodology—Dust concentration measurements were taken
upwind and downwind from a road construction operation.   This
data was used in two different analytical frameworks.  For one,
several other variables for which field data was collected were
regressed on concentration to estimate the relationship between
these variables and concentration.  This procedure was performed
for measured concentrations of several particle size categories.
In the second framework, a generalized atmospheric dispersion
model was applied to the data to estimate TSP emission factors
for each set of upwind and downwind measurements.  The equation
used to calculate source strength is shown below.  It was derived
from the model for a continuously emitting, infinite line source
 (Turner, 1970).

                         0.5xsin (27i)°-5o.u
                            exp
                -1/2
where
     q
     X

     *
     CT,
     u
     z
line source strength, g/m-sec
plume centerline concentration for TSP measured 50
meters downwind from the source, g/m3
angle between the wind direction and the road
standard deviation of the plume's vertical
concentration distribution, a function of downwind
distance, stability, and an assumed initial plume
dispersion, azo, m
mean wind speed, m/sec
vertical distance of sampler inlet from the
ground,  m
The initial standard deviation of the plume's vertical
concentration distribution was assumed to be 4 meters.

     Test Site—Field data were collected at an active road
construction site in a rural area near Minneapolis/St. Paul,
Minnesota.  The road was oriented east to west; prevailing winds
were from the north.  The land on both sides of the road were  in
cover crops.  Separate tests were conducted for three different
construction operations:  1)topsoil removal, 2) cut and fill
operations,  and 3)  final grading and preparation of the road
base.
                              2-165

-------
     Parameters and Equipment—Two types of air filtering
equipment were used for this study:  a standard high volume
sampler and a hi-vol fitted with a size-selective inlet  (SSI) and
a cascade impactor.  The standard hi-vol measured TSP
concentration.  The SSI had a theoretical cut point of 15 (am, and
the cascade impactor measured particles in the size range between
0.5 M-m and 7 \m..  The PM-10 and IP concentrations were determined
by interpolation or extrapolation of data points derived from
these devices.

     Two "automated meteorological stations" collected data on
wind speed and direction (at two heights, 5 and 8 meters),
temperature, and precipitation.  Atmospheric stability was also
evaluated and recorded.  Records were kept on the number and type
of vehicles passing the monitoring stations.  The predominant
vehicle type for each operation is shown in Table 2.66.  In
addition, the quality of each pass was rated "good," "marginal,"
or "bad," depending on the wind/road angle and the presence or
absence of a visible dust plume impacting the downwind samplers.
Soil moisture and silt content were also measured in the normal
manner.
    TABLE 2.66.
PRIMARY EQUIPMENT FOR EACH OPERATION OF ROAD
     CONSTRUCTION PROJECT
Operation
Site Clearing & Topsoil Removal
Cut & Fill
Final Grading & Preparation of Road Base
Primary Vehicles
Bulldozer, Scraper Pan
Bulldozer, Scraper Pan
Vibratory Drum Roller
Road Grader
Dump Truck
     Equipment Confiquration--Four dust monitoring stations were
 set up,  two on the north side at perpendicular distances of 25
 and 50 meters, and two on the south side at the same distances.
 Each monitoring  station consisted of a standard hi-vol sampler
 and a hi-vol cascade impactor, as described above.  Intake
 heights  of these samplers were normally two meters above the
 ground;  however,  toward the end of the sampling program, the
 downwind samplers had to be elevated because the cover crop,
 corn, would have otherwise effected concentration measurements.
 Two directional  wind activators and event recorders were set up,
 one on each side of the road, about half way between the two
 monitoring stations, to turn off all the dust monitoring
 equipment when the wind direction shifted to within 67.5° of
 north or south.

     Sampling Runs—A total of 12 sampling runs were conducted
 during this study.  The number of vehicle passes per run ranged
                               2-166

-------
from 16 to 105.  The sample run time ranged from 14 to 116
minutes,  except for one run which lasted 742 minutes.

     Quality Assurance—The quality assurance for this study was
generally good.  One out of every ten filters was processed as a
blank.  All of the filter tare weights and 20% of final filter
weights were audited by a second analyst.   The hi-vol samplers
were calibrated before and after each sampling run.  Before
weighing, filters were equilibrated at a constant temperature and
humidity.  The accuracy of the balance was checked with Class S
weights before each weighing.

     Findings—The data for the dispersion modeling calculations
are shown in Table 2.67.  Raw upwind and downwind concentrations
were not published.  The computed emission factors for each
sampling run and the construction activity occurring are also
presented in this table.  Calculated emission factors are based
on the concentration measurement 50 meters downwind, because the
plume was believed to be better defined at greater distances.
Topsoil removal was found to produce more TSP emissions than the
other activities.

     Multiple regression analysis of the data yielded the
equations shown in Table 2.68.  The amount of variation in field-
measured concentrations that is explained by each formula is
indicated by the R2 value.

     Publication—This study was conducted for the Minnesota
Pollution Control Agency, Roseville, Minnesota.  The report was
prepared in April of 1983.  It has not been published.
                              2-167

-------
TART.E 7.67. nT.qPER.qTON MODELING PARAMETERS AND RESULTS
Run
AH-1
AH-2
AH-3


AH-4
AH-5
AH-6

AH-7
AH-9

AH- 10
AH-1 1


AH-1 2


AH- 14

Stability
Class
D
D
C


B
D
D

C
B

D
C


C


C

o2a
(m)
6.01
6.01
7.49


9.12
6.01
6.01

7.49
9.12

6.01
7.49


7.49


7.49

Mean
Wind
Speed
(m/sec)
4.4
5.1
4.1


3.1
3.8
8.0

4.9
2.8

6.7
5.5


5.8


3.4

Net
Downwind
Concent r .
(ng/m3)
13292
16996
595


7642
3281
292

124
676

977
604


2448


845

Passes
Per
Minute
1 .03
1 .57
0.47


1.12
1 .26
0.94

0.07
0.86

0.88
0.21


0.38


0.68

TSP
Emission
Factor
(Ib/veh-
mile)
75.5
73.4
8.41


41 .5
13.2
3.31

14.1
4.29

9.86
25.8


61 .0


6.88

Activity
Topsoil
removal
Topsoil
removal
Pan
scraper
traffic
(empty)
Cut/Fill
Cut/Fill
Earth
Hauling
Cut/Fill
Earth
Hauling
Cut/Fill
Aggregat
e
Hauling
Aggregat
e
Hauling
Aggregat
e
Hauling
a)  standard deviation  of  the plume's vertical
   the downwind measurement distance
concentration distribution at
                                    2-168

-------
     TABLE  2.68.   FINDINGS OF  REGRESSION ANALYSIS

Particle
Size
Category
(Hg/m3)
TSP
IP
PM,0
Formula3 for Predicted Concentration (|ig/m3)
25 meters downwind
575(s)0-87(Td)°-89(M)-°-S6
R2 = 0.81
142(s)°-8B(Td)°-93(M)-°-55
R2 - 0.78
112(s)°-87(Td)°-95(M)-°-52
R2 = 0.78
50 meters downwind
374(s)0.eo(Td)o.97(M)-o.47
R2 - 0.74
87(s)°-86(Td)1 (M)-°-<4
R2 - 0.76
60(s)o.88(Td)1.o4(M)-o.40
R2 = 0.75
a)  variable definitions:
   s  - surface silt content,  %
   Td - traffic density, vehicle passes per minute
   M  - surface moisture, %
                           2-169

-------
AGRICULTURAL ACTIVITIES

Introduction

     Agricultural operations have received limited attention
regarding the fugitive dust emissions they produce.  Emission
rates have been measured in two field studies, both conducted by
the Midwest Research Institute using the exposure profiling
methodology.  They are reviewed here in chronological order.

Studies of Primary Importance

Study 1 — Cowherd et al.  Development of Emission Factors for
          Fugitive Dust Sources.  EPA-450/3-74-037.  1974.

     Methodology—This field research was part of the original
study which used exposure profiling to measure fugitive dust
emissions.  The methodology is explained fully in the section on
unpaved roads under this same reference.  The only difference is
that the emission factor is expressed as mass per unit area (e.g.
Ib/acre) rather than mass per unit length (e.g. Ib/veh-mile).   To
derive these units, the mass generated per unit length of tilling
(which is measured the same as in the case.of unpaved roads) is
multiplied by the number of passes required to till an acre of
land.

     Test Sites—A total of four sites, two each in Morton County
and Wallace County, Kansas were tested.  The soils at the sites
in Morton County, referred to as sites A1 and A2,  were
Dalhart/Richfield fine sandy loam and Ulysses/Richfield silt
loam, respectively.  Site A1,  fallow prior to tilling, had a
slight vegetative cover.  Site A2 had no vegetation; it was also
fallow.  The soils at sites A3 and A4 in Wallace County were
respectively Ulysses/Colby silt loam and Keith/Colby silt loam.
Both of these were fallow with light vegetative cover.  The land
at all of these sites was level to gently sloping (up to 2%) .

     Parameters and Equipment—Table 2.69 lists the parameters
measured and the corresponding equipment used in studying
emissions from tilling.

     For each test the tilling equipment was either a one-way
disk plow or a sweep-type plow.  These were considered
representative of the equipment widely used in dry land farming
in the Great Plains.  The width of the equipment ranged from 12
to 30 feet.  Equipment speed ranged from four to seven mph.  Soil
samples were collected to depths between four and six inches.

     Emission factors were calculated for three particle size
categories:  larger than 30 \im, between 2 and 30 (am, and smaller
than 2 [am.  According to Muleski et al.  (1983), these categories
were based on Stokes' diameter.  Though not detailed explicitly

                              2-170

-------
the percentage of the emission  plume which fell into the  less-
than-2 p.m category was apparently determined by translating the
cascade impactor cut points  (see  footnote for Table 2.69)  into
equivalent Stokes' diameters  and  interpolating.

   TABLE  2.69.   PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
Parameters
Wind speed
Wind direction
Cloud cover
Temperature
Relative humidity
Soil texture
Soil moisture
Vegetative cover
Tillage equipment type
Tillage equipment width
Tillage equipment speed
Tillage equipment passes
Plume total dust exposure/concentration
Plume particle size distribution
Plume TSP concentration
Background concentration
Sampling duration
Dust deposition
Dust saltation
Equipment
Unspecified
Unspecified
Direct observation
Sling psychrometer
Sling psychrometer
Hydrometer
Oven, scales
Direct observation
Direct observation
Direct observation
Timer, reference points
Direct observation
Isokinetic exposure profiler
High volume cascade impactor
(Anderson impactor, Sierra Impactor3)
Standard high volume sampler
Standard high volume sampler
Timer
Dustfall buckets
Saltation catcher
 a) An Aerotec cyclone was used for one test.  Particle size
   data were presented only for the Anderson impactor, which
   had aerodynamic cut diameters at 0.17, 1.25, 2.1, and 4.5 ^m.

     Equipment Configuration—All downwind sampling equipment was
placed about 20 feet from  the downwind edge of the tilling path.
The exposure profiler was  set up with sampling heads  3,  5.5,  8,
and 10.5 feet above the  ground.   The particle size classifier
(Anderson impactor, Sierra impactor, or Aerotec cyclone,
depending on the sampling  run)  had its intake six feet  above the
ground.   Dustfall  buckets  were  used in only two of the  seven
sampling runs.  Saltation  catchers sampled at heights ranging
from 1 to 2.5 feet.  The downwind standard hi-vol sampler was
used in only one of the  seven tests; it sampled at a  height of
                               2-171

-------
six feet.  The position of the upwind standard hi-vol was not
documented.  Wind speed and direction were measured at a height
of 12 feet.

     Sampling Runs—A total of seven sampling runs were conducted
on tilling operations.  The number of machinery passes per run
ranged from 10 to 16.  Sampling runs lasted between 13 and 35
minutes.  Plumes from a total of 87 passes were sampled.

     Quality Assurance—In the use of standard high volume
filtration, the investigators followed the procedures specified
by EPA in "Reference Method for the Determination of Suspended
Particulates in the Atmosphere (High Volume Method)," (1971).
For the measurement of dust deposition,  the investigators
followed the procedures set forth in "Standard Method for
Collection and Analysis of Dustfall," ASTM Method D 1739-62.

     Samples of the dust plume were collected only when the wind
speed was less than 20 mph, the maximum speed under which samples
could be collected isokinetically.  As noted above, wind
direction and speed were observed to be constant during each run.
Likewise, the intake velocity and the directional orientation of
the samplers in the exposure profiler were constant during each
sampling run.

     Filters were conditioned in a controlled temperature and
humidity environment before and after collection of dust samples.
Filter samples were transported to the laboratory in individual
folders.  The interior surfaces of the sampler heads were rinsed,
and the water was captured and evaporated to determine the mass
of dust on the interior surfaces.

     Documentation of quality assurance practices such as
collocation of samplers, processing of blank profiler filters,
and audits of profiler filter weights was not provided.

     The investigators acknowledged two potential sources of
small particle bias in the measurement of particle size
distribution:  1) particles bouncing down through the cascade
impactor to smaller particle stages; 2)  non-isokinetic sampling
which collects larger particles with lower efficiency than
smaller particles.

     Findings—The exposures and emission factors measured in
this study are presented in Table 2.70.  A predictive equation
was developed based on the limited data collected in this study:

                        e -  1-4 s (5/5.5)
                               (PS/50)2
                              2-172

-------
 where
     e     =    emission factor, Ib/acre
     s     =    silt content  (i.e. percent between 2 and 50
                diameter)
     S     =    implement speed, mph
     PE    =    Thornthwaite's precipitation-evaporation index

The predictions of  this model were within 16%  of the measured
Ib/acre  for these  seven tests.

          TABLE 2.70.  EXPOSURE DATA  AND EMISSION FACTORS
                                                                      in


Pi in
5


6


7


9


11


12


14



Ht
Ift- }
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3
10.5
8
5.5
3
Unit
Exposure
(mg/in.2/
equivalent
-r,acc»\
0.804
1 .23
3.77
7.27
0.537
1 .92
3.60
10.8
0.256
1 .24
3.60
10.8
1 .30
1 .91
2.96
4.29
1 .76
2.35
4.35
6.53
2.31
3.34
5.35
9.06
2.53
3.74
5.59
8.38

Integrated
Exposure
fTh/tn-i 1 »bt

81 .4


75.4


86.6


50.5


92.4


124


114



Total


55.9


51 .9


59.6


41 .6


63.6


85.2


78.1

dc > 30 urn


5.6 (10%)


5.2 (10%)


6.0 (10%)


4.2 (10%)


15.9 (25%)


21.3 (25%)


19.5 (25%)

Factors (Ib/acre
2 < dc < 30 (im


28.2 (50%)


26.2 (50%)


30.0 (50%)


21.0 (50%)


27.7 (44%)


31.9 (37%)


36.3 (46%)



dc < 2 urn


22.1
(40%)


20.5
(40%)


23.6
(40%)


16.4
(40%)


20.0
(31%)


32.0
(38%)


22.3
(29%)

a)  In order to facilitate comparison with data  from unpaved roads, which generally
   have lanes about  12 feet wide, data describing exposures from  12-, 20-, and 30-
   foot wide tillers were normalized to represent the mass per square inch per 12-ft
   wide tilling pass.
b)  i.e. per mile of  12-foot wide tilling
c)  particle Stokes' diameter
                                   2-173

-------
     Publication—This study was conducted and documented under a
contract with the EPA Office of Air and Waste Management, Office
of Air Quality Planning and Standards, Research Triangle Park,
North Carolina.  It was published in 1974 as Publication No. EPA-
450/3-74-037.

Study 2:  Cuscino et al.   The Role of Agricultural Practices in
          Fugitive Dust Emissions.  1981.

     Methodology--The investigators used exposure profiling to
measure fugitive dust emissions from several agricultural
operations.  The eolation E = C U t, where E is the exposure
(mass/area),  C is the concentration (mass/volume),  U is the wind
speed (length/time), and t is the sampling time, was used to
calculate exposure at each sampling height.

     Test Sites—Sampling was conducted at three sites in
California.  Several soil preparation and maintenance operations
were tested at the Norman Clark Farm in Fresno County in the San
Joaquin Valley, at which hay and alfalfa were being grown, and at
the Rice Experimental Station in Butte County in the Sacramento
Valley.  Sugar beet harvesting operations were tested at the
Hamatani Farm in Sacramento County in the Sacramento Valley.

     Parameters and Equipment—Table 2.71 below lists the
parameters which were measured and the equipment employed in
making those measurements.  Exposures for various particle size
categories were calculated using the method described in Cowherd,
et al.  (1984); an explanation of this method is provided in the
review of that report in the section on paved roads.

     Equipment Configuration—The exposure profiler was kept five
meters from the downwind edge of the equipment path.  Sampling
heads on the profiler were fixed at heights of 1, 2, 3, and 4
meters.  The cassette-mounted filter was positioned five meters
downwind and two meters above the ground.  Wind direction and
speed were measured four meters above the ground at an
unspecified distance upwind.  A standard hi-vol sampler and a hi-
vol cascade impactor were also operated on the upwind side at an
unspecified distance.

     For the tests of soil preparation and maintenance
operations, downwind inhalable particulates  (IP, particles with
aerodynamic diameter less than 15 |^m)  concentration was measured
at distances of 5, 50, and 100 meters with the intakes 2 meters
above the ground.  For some tests this series was duplicated at
another height, 1 or 3 meters, to measure the change in IP
concentration with height.

     Sampling Runs—A total of 17 sampling runs were conducted to
measure emissions from agriculture operations.  Eleven of these
were soil preparation or maintenance activities.  Emissions were

                              2-174

-------
   TABLE  2.71.   PARAMETERS MEASURED AND CORRESPONDING EQUIPMENT
Parameters
Wind speed
Wind direction
Soil moisture
Soil erodibility
Soil silt content
Equipment type
Equipment width
Concentration of TSP upwind
Plume concentration of TSP
Plume concentration of particles with
aerodynamic diameter < 1 5 ^im
Plume concentration of particles in
the categories:
< 7.2 |im aerodynamic diameter
< 3 jim "
< 1 . 5 \Jirn "
< 0.95 urn "
< 0.49 urn "
Plume concentration of total
particulates
Decay in plume IP concentration with
downwind distance3
Largest particle size
Equipment
Recording anemometer
Unspecified
Oven , scales
Oven, sieves, scales
Oven, sieves, mechanical sieving
device
Direct observation
Unspecified
Standard hi-vol sampler
Standard hi-vol sampler
Hi-vol with size-selective inlet
Cascade impactor with greased
substrates attached to the above hi-
vol sampler
Exposure profiler
Hi-vol/SSI in series of increasing
downwind distance
Cassette-mounted 37 mm filter,
microscope
 a)  Measured only for soil preparation / maintenance operations

measured from the towed implement and the  towing tractor,  which
was either on tracks or on six wheels.  Of the five runs  at  the
Norman Clark Farm, three tested emissions  for land  planing (using
an implement to level the land), and two were a disc operation.
All of the runs at the Rice Experimental Station measured
emissions from the disc operation.

     Six sampling runs were conducted on a beet harvesting
operation at Hamatani Farm.  This harvest  consisted of  two
separate field operations:  leaf beating  (removal of tops of
plants)  and beet digging.  Both operations were tested  in three
sampling runs.  Both implements were on two wheels  and  where
towed behind a four-wheel tractor.  During the digging  operation
an 18-wheel truck followed beside the digger.   The duration of
each sampling run was not reported; however, it was long  enough
                              2-175

-------
to collect a measurable mass of dust and allow the averaging of
the emitted dust mass over several implement passes.

     Quality Assurance—Ample assurance was provided for the
quality of the data gathered in this study -

     The quality of each machine pass was judged according to the
angle between the wind direction and the machine's path.  The
vast majority of the runs were of good quality.

     The procedures for collecting and analyzing soil samples for
moisture content, erodibility, and silt content were presented in
detail; the laboratory scale's zero was checked prior to each
measurement.  All filter tare weights were audited.  The measured
weights of ten percent of the used filters were audited.
Equipment was calibrated before testing at each site and at two
week intervals.

     Exposure samples were considered isokinetic if the ratio of
sampling velocity over wind velocity was within the range of 0.8
to 1.2.  Most samples met this criteria.  Those samples which
were not collected isokinetically were corrected to isokinetic
conditions.

     Findings—Field emissions data and calculated emission
factors are shown in Table 2.72.  Data on wind speed and sampling
duration were not published.  Runs with an "N" in the run
identification column were conducted during soil preparation and
maintenance operations, whereas those with an "R" were conducted
on harvesting operations.

     In developing predictive equations for the two kinds of
operations tested, the investigators considered two variables:
soil silt content, and soil moisture content.  To develop an
equation for total particulate emissions from soil preparation
and maintenance operations, data from the tests at the Norman
Clark Farm and the Rice Experimental Station  (runs N-3 through N-
13) and from 7 tests conducted in Kansas in 1974  (Cowherd et al.,
1974) were used in a regression analysis.  For equations for
inhalable and fine particulate emissions, only data from those
tests conducted in this study could be used, because these
parameters were not measured in the 1974 study.

     Inclusion of moisture content did not improve the predictive
accuracy of the equation.  To explain this the investigators
theorized that moisture content effects emissions only when the
moisture content was near a critical level, above which emissions
would be greatly reduced, and that the farmers were all tilling
when soil moisture was below that level.  Table 2.73 shows the
relationships estimated for each of the particle size categories
as well as some summary statistics.  Due to the limited number of
sampling runs conducted on beet harvesting  (including leaf

                              2-176

-------
beating and beet digging),  only a very limited statistical
analysis was possible.  Both silt content and moisture content
were found useful in predicting emission factors.  However,
specific predictive equations were not published.

     Publication—This study was prepared for the California Air
Resources Board (GARB) in Sacramento, California.  It was
published in June of 1981 as (CARB's) Report No. ARB/r-81/138.

    TABLE  2.72.  EMISSION MEASUREMENTS AND  CALCULATED EMISSION
                             FACTORS
Run
N-3
N-4
N-5
N-6
N-7
N-8
Ht
(m)
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Concentr .
(Hg/m3)
8504
7195
4023
2564
8113
7960
4594
3388
5293
4544
3588
2373
5806
2310
806
184
10417
5130
1642
1359
3036
1462
581
293
Integrated Exposure
(mg/cm)
Total
1700
1900
1840
1440
1060
382
IPb
226
213
169
299
313
48.5
FPC
59
85
100
106
78.4
17.9
Emission Factor
(kg/km2)
Total
2320
2600
2520
2960
2720
522
Ipb
309
291
231
613
797
66.2
FPC
80.5
116
137
217
200
24.4
                              2-177
(continued)

-------
TABLE 2.72.   (continued)
Run
N-9



N-10
N-11
N-12
N-1 3
R-1
R-2
Ht
(m)
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Concentr .
(ng/m3)
2405
1060
363
127
3621
1597
579
216
5692
2124
598
187
6710
2051
552
213
5303
2207
687
193
2833
6064
1164
1482
3539
2054
1203
707
Integrated Exposure
(mg/cm)
Total
591
1120
1180
953
764
590
322
IPb
251
442
395
264
282
21 .1
149
FPC
94.0
214
182
125
137
11.1
66.3
Emission Factor
(kg/km2)
Total
865
1630
1610
1300
1120
3940
2150
jpt
379
646
540
361
412
141
997
FP°
138
314
249
170
200
74.3
444
          2-178
                                  (continued)

-------
                      TABLE 2.72.   (continued)
Run
R-3
R-4
R-5
R-6
Ht
(m)
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Concentr .
(Hg/m3)
6615
2703
1645
1149
3220
1147
966
782
2178
1796
1316
1292
3285
2955
2518
1647
Integrated Exposure
(mg/cm)
Total
512
816
366
454
Ipb
122
160
103
200
FPC
59.2
74.3
43.9
108
Emission Factor
(kg/km2)
Total
3430
3560
11 .40
3300
IPb
816
699
321
1460
FP°
396
325
137
786
      a) gross measurements, i.e.  including background
      b) inhalable particulates
      c) fine particulates
TABLE 2.73.
 PREDICTIVE EMISSION FACTOR EQUATIONS  FOR  SOIL
PREPARATION AND  MAINTENANCE  OPERATIONS
Emission
kg/km2
kg/km2
kg/km2
of
of
of
Factor Equation
TP3 -
IPb -
FPC -
538 0
135 0
53.8
fc silt)0-6
k silt)0-6
(% silt)0-6
R-squared One-sigma precision
factor
.88
*
not published
not published
1
2
2
.29
.17
.33
 a) Total particulates;  i.e., no limit on particle size
 b) Inhalable particulates; i.e., particles with aerodynamic diameter < 15
     (im
 c) Fine particulates;  i.e., particles with aerodynamic diameter < 2.5 urn
                                  2-179

-------
LANDFILLS

     Particulate emission rates from sanitary landfills have  not
been field measured.  An emission inventory was prepared by
Muleski and Hecht (1987) for two landfills near Chicago.  Field
data on soil silt and moisture content, vehicle weight, number of
wheels, travel distance, and speed, and traffic density were
applied to the AP-42 emission factor for unpaved roads to
estimate emissions from the haul road within the landfills.   The
AP-42 emission factor for a bulldozer working coal mine
overburden was used to estimate emissions from a dozer and a
compactor operating in the landfill.  The old AP-42 factor for
batch drop operations was used to estimate emissions from
materials handling at the landfill.  The three emission factor
equations are shown below:

                            Haul Roads

                         s
where
     e    =    PM-,0 emission factor, Ib/veh-mile
     s    =    silt content, %
     S    =    vehicle speed,  mph
     W    =    vehicle weight, tons
     w    =    number of wheels per vehicle
     p    =    number of day with more than 0.1 inches of
               precipitation

                  Dozer and Compactor Operation

                                   (s)1-2
                          e - 5.69
                                     1.3
where the variables are defined as above, except e is the TSP
emission factor in Ib/veh-mile.

                        Materials  Handling
                      e - 0.00065
                                   M2i r\°-33
                              2-180

-------
 where

     U    =    mean wind speed, mph
     H    =    drop height, ft
     Y    =    dumping device capacity, yd3

and the other variables are as defined above.

Total calculated PM10 emissions from the two landfills were  37
tons/year and 13 tons/year.
                              2-181

-------
CATTLE FEEDLOTS

     Very little field data is available on dust emissions from
cattle feedlots.  The source assessment conducted by Peters and
Blackwood (1977) found only two studies, both sponsored by the
California Cattle Feeders Association, in which particulate
concentrations at cattle feedlots . were measured  (Elam et al . ,
1971; Elam et al. ,  1972).  However, emission rates were not
measured or modeled in either of these two studies.  In fact,
data on other field parameters, such as upwind and downwind
sampling distance,  wind speed, atmospheric stability, or feedlot
size, were not collected.

     Peters and Blackwood took the 24-hour concentration data
reported by Elam et al.  (1972), corrected it to a  10 minute
sampling time using the method described by Turner  (1970), made
assumptions for each of the deficient parameters, and modeled an
emission rate for each measured concentration.  The downwind edge
of the feedlot was treated as an apparent line source; thus the
equation for a continuously emitting line source  (Turner, 1970)
was used:

                           2QL
                      X -- _    exp
where

     X    =    concentration at downwind distance x, g/m3
     QL   =    emission rate per length of a line source, g/s-m
     u    =    average wind speed, m/s
     H    =    effective height of emission, m
     crz   =    standard deviation of the plume's vertical
               concentration distribution, m

     The downwind distance was taken to be 50 meters.  The
national average wind speed of 4.47 m/s and the average  stability
class, C, were assumed.  The emission height was assumed to be
3.05 meters, and az was assumed to be 4 meters.   Table 2.74 shows
the field measured concentrations, the corrected concentrations,
and the modeled source strengths.  The average source strength
was 0.0361 g/s-m.

     Additional data were collected on the average  size  and
density of California cattle feedlots to permit conversion of the
average source strength to units of mass emitted per second for
the average California feedlot, which was found to  contain 8,000
head on 27.5 acres.  The resulting rate, 11.9 g/s,  was then
converted to 36.7 |j.g/s-m2.   As presented in AP-42,   11.9 g/s also
equates to 128 kg/day/ 1,000 head capacity.
                               2-182

-------
 TABLE 2.74.  CONCENTRATIONS AND CORRESPONDING SOURCE STRENGTHS
Feedlot
Number
19
25
4
8
9
12
13
15
17
20
2
3
5
6
7
10
14
22
23
24
1
11
16
26
18
Measured
Concentration
(Hg/m3 )
453.3
977.9
418.3
1034.8
108.6
348.4
534.5
959.8
716.0
53.7
1046.4
379.8
1184.7
660.3
860.3
1267.8
703.6
268.5
1161 .0
276.1
279.4
263.7
1129.8
216.2
660.6
Corrected
Concentration
(Hg/m3)
1056.2
2278.5
974.6
2411 .1
253.0
811 .8
1245.4
2236.3
1668.3
125.1
2438.1
884.9
2760.4
1538.5
2004.5
2954.0
639.4
625.6
2705.1
643.3
651.0
614.4
2632.4
503.7
1539.2
Source Strength
(g/s-m)
0.0256
0.0553
0.0237
0.0585
0.0061
0.0197
0.0302
0.0543
0.0405
0.0030
0.0592
0.0215
0.0670
0.0373
0.0487
0.0717
0.0398
0.0152
0.0657
0.0156
0.0158
0.0149
0.0639
0.0122
0.0373
     Documentation of the absence or presence of dust control
techniques at these feedlots was not provided.  The field testing
occurred exclusively during California's dry season.
                              2-183

-------
UNPAVED PARKING LOTS

     McCaldin (1977) attempted to measure emissions from vehicle
travel in an unpaved parking lot using the upwind-downwind
method.  Four 1.5 to 3 hour tests were conducted at an unpaved
parking lot in Tucson, Arizona.  Standard high volume samplers
were used to measure dust concentration.  There was no
significant difference in background and downwind concentrations
The number of "vehicle movements" in the lot during sampling
totaled 38.

     McCaldin estimated emissions from a hypothetical one-acre
parking lot using the emission factor equation he developed from
unpaved road field test results:

                        E = (s)  0.035 (S)2
where

     E    =    TSP emission factor,  Ib/veh-mile
     s    =    mass percent silt
     S    =    traffic speed,  mph

He assumed a vehicle speed of 1 5 mph and a- resulting emission
factor of 1 Ib/mile.  The assumed percent silt was not given.
The total travel distance for a vehicle entering and exiting the
parking lot was taken to be 300 feet.  The resulting emission
factor was 0.05 Ib/use of the lot.

     In a similar manner, the Midwest Research Institute (1988)
used the AP-42 emission factor equation for unpaved roads and
several large assumptions to calculate the mass emitted per use
of a parking lot.  However, the emission rate was expressed as a
function of the dimensions of the parking lot:

                      EiQ - 0.2 365'P (L + W)
                                JD D

where

     E10   =    PM10  emissions, g/use  of the lot
     p    =    number of days/year with rain
     L    =    dimension of parking lot perpendicular to aisles,
               m
     W    =    dimension of parking lot parallel to aisles, m

In developing the above equation from the unpaved road emission
factor, the following assumptions were made:

     1)   silt content is 12%
     2)   average number of wheels per vehicle is 4
     3)   average weight of vehicles is 3 tons
     4)   vehicles  travel at 10 mph in parking lots

                              2-184

-------
WIND EROSION

Introduction

     Emissions estimates for wind erosion have typically dealt
with either emissions from open areas (such as agricultural
fields) or from storage piles during periods when they are not
actively being utilized.

     Development of wind erosion emissions estimates in early
emission inventory efforts typically involved using an equation
developed from data collected in the late 1940s and early 1950s
that was not intended for use in evaluating suspended particulate
emissions estimates, but rather was developed for evaluating
strategies for minimizing the horizontal flux (translational
movement) of particulates from field to field.  This early work
was performed by Dr. W.S. Chepil.  The majority of this early
work reflected measurements made using wind tunnels.  The early
form of the wind erosion equation was a simple exponential
expressing the amount of soil loss in a wind tunnel as a function
of soil cloddiness, amount of surface residue and degree of
surface roughness.  This equation has been continuously modified
as new research data has become available, and is now a complex
equation indicating the relation between potential soil loss from
a field and a number of primary field and climatic variables.
The form of the wind erosion equation most commonly utilized in
developing TSP emission inventories was proposed in 1965 by
Woodruff and Siddoway.  As put forward in their 1965 publication,
the wind erosion equation can only be utilized to develop wind
erosion soil loss estimates on an annual basis,  and as previously
stated is not intended to produce estimates of the vertical flux
of particles, but rather total soil loss from an upwind field to
some downwind location.

     In utilizing the wind erosion equation for determining wind
erosion particulate emission estimates, an assumption regarding
the fraction of the horizontal material that is actually
suspended has been made.  Typically, this value (for TSP) has
been chosen to be 2.5%, however, the range has been indicated as
between 2-10%.

     There are few actual emission measurement studies from wind
erosion sources.  Typically, the measurements that have been made
can be grouped into one of three types.  First, measurements of
aerosol particle numbers were taken at two or more heights above
ground in order to determine the vertical flux of particles.
Second, measurements of wind erosion emissions from a variety of
surface types were made using wind tunnels  (either by bringing
into the lab samples of the material, or utilizing portable wind
tunnels in the field).  Finally, a few emissions measurements
were made in fields using upwind-downwind sampling  techniques to
try and determine the emissions from a particular field.

                              2-185

-------
     Details of those studies that did report emissions data of
one of the three types indicated above are reported below.  The
studies evaluated include wind erosion emissions measured for
open areas as well as wind erosion emissions measurements
determined for sources such as aggregate storage piles, mining
and quarrying and other fugitive sources considered in the
previous sections.

Studies of Primary Importance

Study 1— Gillette et al.  "Measurements of Aerosol Size
          Distributions and Vertical Fluxes of Aerosols on Land
          Subject to Wind Erosion." Journal of Applied
          Meteorology. Vol. 11. pp. 977-987. 1972.

     Methodology—Vertical fluxes of aerosol particulates were
determined based on considerations of both the aerosol and
momentum fluxes and by making some simplifications concerning the
coefficient of exchange for aerosols and the eddy viscosity.  In
practice, the vertical flux of particles was determined using the
following equation:

                       Fa=-pCU!2 (n2-n,) /u2-u,

where Fa is the vertical aerosol  flux,  p is the density of air,  C
is the drag coefficient, u, is the wind speed taken at height  z^,
u2 is the wind speed taken at  height z2, and n, and n2  are  the
number of particles measured at heights z,  and z2 respectively.
In practice, aerosols within the size range 0.3-6 microns were
measured, and the resulting particle fluxes as functions of size
were expressed in the conventional dN/d log r notation where N(r)
is the total number of particles having radii £ r.

     Test Site—All samples were taken in an eroding field at the
University of Nebraska Northwest Experimental Laboratory located
near Alliance, Nebraska.

     Parameters and Equipment—Aerosol samples were collected
using single-stage jet impactors.  The aerosol impaction surface
was a microscope cover slip coated with filtered silicone oil and
mounted on a standard microscope slide.  Size distributions of
the collected aerosol were determined from photomicrographs taken
using a Zeiss TGC-3 particle sizer and counter.  All samples were
collected isokinetically.

     In addition to the aerosol collection, wind speed was
determined using a three cup Belfort anemometer.  Soil erosion
parameters were also measured.  Soil particle creep was measured
using 15 err long cylinder buried flush with the soil surface.
Total soil flow  (soil creep and saltation) was measured using a
Bagnold catcher.  Soil condition parameters were also measured.
These parameters included water content, cloddiness, clod

                              2-186

-------
stability, crop residue and ridge roughness.  Water content was
measured by drying containers of soil at 120 C until three
successive weights were the same.  Soil cloddiness was determined
as the percentage of soil passing through a rotary sieve with
0.84 mm square openings.  Clod stability was also determined
using the rotary sieve.  Crop residue was measured using standard
USDA methods.  Ridge roughness was measured by comparison to
photographs.  The particle size distribution of the soil
particles were determined by wet sieving and scanning electron
microscopy.

     Equipment and Configuration—Aerosol samples collected using
the jet impactors were collected at two heights, 1.5 and 6 meters
above the ground.  The wind speed was measured at 1.5, 3, and 6
meters above the ground.

     Sampling Runs—A total of thirteen samples were obtained,
however, several of the experiments were repeated twice.  The
data for the duplicates was not presented.  Samples were obtained
on March 17, 18 and 31, April 1,4,15,28-30, May 13, and July 12,
1971.  Sampling times ranged from 30 minutes to 2 hours.

     Quality Assurance—As indicated above, duplicate aerosol
size distributions were obtained for use in determining the
aerosol size flux.  The results of these duplicate samples were
used as an indicator of the sensitivity of the technique.  The
error associated with the particle size counting was determined
using a standard Poisson distribution.  Variability resulting
from the laboratory procedure was determined by examining the
results for laboratory repeatability-  Two independent
determinations were made on three aerosol collections, and the
percentage standard deviations for the four particle size classes
used in the field data were calculated.  The results of these
standard deviation calculations were presented.

     Findings—Of the thirteen samples taken, only four indicated
positive vertical fluxes (i.e., emissions of particles for the
conditions examined in the study).  For those samples indicating
positive fluxes, the number fluxes were converted to mass fluxes
by assuming that the soil and aerosol bulk densities were the
same and that the particles were spherical.  The results
indicated that the vertical flux of aerosols £10 microns were
between 0.6 x 10"5 and  1 x 10"4  micrograms/cmVsec.

     Horizontal fluxes were also determined using the wind
erosion equation and the actual measured flux determined by
summing soil creep and saltation measurements.  These
measurements were in rough agreement.  One interesting point  in
these measurements is that it provides an indication of the
actual percentage of the total horizontal flux that is
represented by the vertical  (suspended) flux.  As noted above,
for most previous TSP wind erosion emission inventories, it was

                              2-187

-------
assumed that 2.5% of the total wind erosion flux was the
suspended component.  However, the ratio of the vertical fluxes
determined in this study to the horizontal fluxes measured yields
values between 0.0067 and 0.01%, values that are much lower than
the 2.5% that has been utilized in the past.

     Publication—This study was documented in the Journal of
Applied Meteorology. Volume 11, pages 977-987, 1972.

Study 2— Jutze and Axetell.  Investigation of Fugitive Dust
          Volume I - Sources,  Emissions, and Control.  EPA-450/3-
          74-036-a.  1974.

     Methodology--Upwind-downwind sampling using directional hi-
vol samplers was carried out to measure emissions from
agricultural sites at two locations.  A diffusion equation for
ground-level sources with no effective plume rise was used to
estimate the source strength from measured concentrations.
                            x-
     The resulting formula for source strength, Q (g/sec),  is
shown below:

                         Q -  2.78rcoyosrux
where
     u
     X
standard deviation of the crosswind distribution
of the plume's concentration at the downwind
measurement site (a function of atmospheric
stability, downwind distance, and the assumed
initial standard deviation, ayo) .
standard deviation of the vertical distribution of
the plume's concentration at the downwind
measurement site (a function of the atmospheric
stability and downwind distance),  m
mean wind speed, m/sec
measured concentration of particulates at the
downwind measurement site  (minus the background
concentration),  g/m3
     Test Site—Two sites located in agricultural areas were
utilized to try and determine the emissions resulting from wind
erosion of agricultural land.  One of the test sites was located
near the University of Arizona in Mesa, Arizona.  The second site
was located in the San Joaquin Valley of California, close to the
University of California Westside Agricultural Field Station.

     Parameters and Equipment—Equipment utilized in this study
included directional hi-vols, hi-vols equipped with Andersen

                              2-188

-------
cascade impactors, standard hi-vols and a meteorological tower.
However, even though all of these samplers were utilized at the
two test sites, only the information from the directional hi-vols
was utilized to develop the emissions estimates.  Wind velocity
and direction were measured using continuous windvane/anemometer
sensors.

     Equipment and Configuration—No detailed information was
given concerning the location of the sampling equipment.  Maps
showing the locations of samplers were provided, but didn't
differentiate between the different types of samplers.  The
downwind measurement distance used in the dispersion modeling was
given in a table in the report.  At the Mesa site this distance
was 30 meters and 315 meters for the two sampler locations for
which measurements were made.  Two sets of measurements were made
for each of the sampler locations at the Mesa site.  For the San
Joaguin site, the distances for the four measurements were 250,
150, 150, and 250 meters downwind.  A single sample was obtained
at each downwind distance at the San Joaquin site.

     Sampling Runs—Four concentration measurements were taken
for each of the two sites.  Samples were collected over either a
24-hour or 48- hour period.

     Quality Assurance—The investigators provided little
assurance of the quality of the data collected for this study.
Although, the procedures for operating the hi-vol samplers and
for handling and transporting the filters were thoroughly
documented, the procedures used to determine the filtered dust
mass were not described.  Normal quality assurance procedures,
such as calibration of samplers, processing of blank filters,  or
auditing of weight measurements or data reduction calculations,
were not documented for this study.

     Findings—The field data collected in this study of fugitive
dust emissions from agricultural wind erosion is shown in Table
2.75.  The emission rate calculated for the San Joaquin valley
agricultural area was 0.6 tons/acre/year.  For the Mesa
agricultural site, the emission rate was determined to be 2.1
tons/acre/year.  The investigators calculated emissions using
these emission factors and compared those emissions with those
calculated from the wind erosion equation.  The results showed
that emissions calculated using the wind erosion equation would
only be 22 and 21  percent of the San Joaquin and Mesa emissions
calculated using the emission rates determined from the field
studies.  The investigators attributed this overprediction using
the developed emission factors to the fact that the emission
factors do not account for the fact that some of the acreage used
to develop the total emissions estimates had continuous ground
cover,  which is accounted for in the wind erosion equation, but
not in the emission factors developed in this study.  In spite of
the fact that emission factors were developed using field

                              2-189

-------
information,  the actual emissions estimates produced for this
study were made using the wind erosion equation (with the
assumption that 2.5% of the total mass eroded was emitted
vertically).

     TABLE 2.75.  DISPERSION PARAMETERS - AGRICULTURAL SITES
Site
San
Joaguin
San
Joaguin
San
Joaquin
San
Joaguin
Mesa
Mesa
Mesa
Mesa
X
(m)
250
150
150
250
30
30
315
315
Stability
Class
B
B
B
C
C
C
B
B
°y
(m)
125
116
116
115
186
186
228
228
o*
(m)
87
75
75
67
108
108
172
172
Wind
Speed
(m/sec)
3.1
2.2
3.1
4.0
3.1
3.6
2.7
2.7
Concentr .
(mg/m3)
.039
.026
.029
.035
.019
.037
.023
.072
Source
Strength
(g/sec)
11 .8
5.0
7.9
9.5
10.3
23.3
21 .3
66.6
     Publication--This study was conducted and documented under
contract for the Environmental Protection Agency, Office of Air
and Waste Management, Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina.  It was
published in 1974 as Publication No. EPA-450/3-74-036-a.

Study 3— Cowherd et al.,  Development of Emission Factors for
          Fugitive Dust Sources.  EPA-450/3-74-037.  1974.

     This report documented two distinct field studies on dust
emissions from storage piles:  total emissions from aggregate
storage operations and emissions generated during aggregate load-
out operations.  However,  emissions measurements made during the
total emissions monitoring program could be further subdivided
into two components:  measurements made during active operations
within the storage area and measurements made during non-active
periods.  The investigators considered the measurements made
during the non-active periods to be indicative of wind erosion
emissions from storage piles.

     Methodology—The mass of dust emitted from a storage
operation was measured using a rather simple methodology.  The
site's contribution to the downwind ambient dust concentration
was estimated using upwind and downwind hi-vol samplers.  The
                              2-190

-------
mass of dust emitted per unit time was calculated as the average
net concentration times the atmospheric ventilation rate (which
is in this case the volumetric rate at which air passes through
the cross-sectional area defined by the width and height of the
storage area).

     Sampling runs were conducted during 12- and 24-hour periods.
Four parameters were evaluated as possible factors influencing
the total dust emission rate for aggregate storage operations:
rainfall, wind speed, aggregate size, and intensity of activity.
For the rainfall factor, sampling results were divided into two
groups:  dry period results and wet period results.  Runs were
deemed wet if any precipitation occurred during sampling or if
more than a trace occurred on the day before.  The average net
concentration and emission rate for each group was compared to
determine if emissions were higher during dry periods than wet
periods.

     To check for effects of wind speed and aggregate size, these
variables were each plotted against downwind concentration.  For
aggregate size, this was accomplished by matching storage piles
of various sizes of aggregate with particular samplers downwind
from the piles.

     Accurate data were not available on the level of activity in
the storage area during each sampling run.  Thus, it was not
possible to estimate a relationship between the intensity of
activity and concentration.  However, after some manipulation of
the data, it was possible to compare concentrations corresponding
to periods when the storage area was active with those for non-
working periods.   The 12-hour sampling runs were conducted
exclusively during non-working hours, and most 24-hour samples
were conducted during periods including both working and non-
working hours.   (The rest covered exclusively non-working
periods.)  A concentration for working hours was calculated using
the following relationship, in which A is the 24-hour average
concentration for a day including 8 to 12 hours of activity, B is
the average concentration during hours of activity in the storage
area, and C is the average concentration for inactive hours:

                          A =  (B + C)/2

     Test Site—Tests were conducted at a sand and gravel quarry
and processing center near Cincinnati, Ohio.  The investigators
considered it representative of operations at many medium and
large aggregate sites.  Although the gravel pit was adjacent to
the storage area,  the samplers used in this study were judged to
be sufficiently isolated from any dust generated in the pit
itself by the difference in elevation between the pit and the
storage area.   Fifteen storage piles, ranging in height from  5  to
30 feet, were maintained in this area.  The investigators
computed an average pile height, weighted on the basis of  the

                              2-191

-------
pile surface area, of 23 feet.  Each pile was for a different
size aggregate.  The turnover rate for these piles was said to be
high.  No processing of aggregate was conducted in this area.

     Parameters and Equipment—Upwind and downwind concentration
of total suspended particulates was measured using standard hi-
vol samplers.  These were automatically activated by wind sensors
when the wind direction was within 90° of south.   They were also
equipped with timers to record the duration of the sample.  A
high volume cascade impactor was used to measure the particle
size distribution of the dust downwind from the storage area.
Meteorological data, including cloud cover, temperature, and
precipitation, were acquired from a nearby Federal Aviation
Administration Weather Station.  The size of the aggregate in
each pile was noted so the effect of aggregate size on downwind
concentration could be analyzed.  Data were also collected on the
height and configuration of each pile.  The equipment operator's
records documented the tonnage of material excavated, sized, and
loaded onto trucks for transport.  This provided only a very
rough indication of the level of activity at the site during a
given day.

     Equipment Confiquration--The five downwind samplers were
scattered on the downwind side of the storage area.  Three of
them were set up among the storage piles, and two were
immediately downwind of the entire storage area.  The intake
height of these samplers ranged from 3 to 20 feet.  The height
and upwind distance of the background sampler was not documented,
nor was the position of the cascade impactor.  Wind speed and
direction were measured continuously on a pole about 25 feet
above the ground.

     Sampling Runs—Eleven 24-hour and seven 12-hour sampling
runs were conducted.  Four of the 24-hour runs were on weekends
when there was no activity at the storage area.  The remainder
ran from noon one work day until noon the next.  All of the 12-
hour samples ran from 6:00 p.m. until 6:00 a.m. the next day.
For the purposes of estimating the wind erosion component of the
storage pile emissions, the samples collected on the weekends and
from 6:00 p.m. until 6:00 a.m. were considered inactive and thus
representative of the wind erosion component.

     Quality Assurance—None of the normal quality assurance
procedures, such as processing of blank samples, calibration of
equipment, or auditing of measurements or calculations, were
documented for this field study.

     Presentation of raw field data was complete.   Measured
upwind concentrations and net downwind concentration for each
site and sampling period were documented.  Evidence that the net
downwind concentration measured the full contribution of the
storage area alone was presented in three separate comparisons:

                              2-192

-------
the upwind concentration was in line with typical regional
ambient levels,  the upwind concentration was greater than the
downwind concentrations in almost every instance, and the average
upwind concentrations during working periods were close to those
of non-working periods.

     Findings—The concentrations measured for each of the
sampling runs made during inactive periods are shown in Table
2.76.  The run time for each sampler is also shown for every
test.
     The average concentration on non-working days was 47.4
^ig/m3.   Using the  methodology described above,  the corresponding
emission rate for the .wind erosion component of storage piles was
calculated to be 26.8 kg/day.

     Neither the calculations for converting these numbers into
emission factors nor the activity rate  (e.g. tons stored per day)
were presented in the report.  For periods of inactivity, the
emission factor was calculated as 3.5 Ib/storage acre/day (0.11
Ib/ton placed in storage).

     Rainfall, as recorded in the manner described above, was
found to reduce total aggregate storage pile emissions by roughly
50%.  However, when the data for inactive periods only was
examined, the reduction due to precipitation was slightly less
than 25%.  Neither the size of the aggregate in the storage piles
nor the wind speed was not found to have a significant influence
on wind erosion emissions from storage piles.

     The particle size distribution of dust downwind from the
storage area was not discussed, although the cascade impactor was
listed as an instrument used in the field study.

     Wind erosion emissions from storage piles were found to
contribute approximately 25% of the total emissions from
aggregate storage piles based on the data collected as part of
this study.  The investigators also presented a procedure for
estimating windblown dust as an appendix to this report.  In that
appendix they detailed how to use the USDA wind erosion equation
and indicate that the fraction of total eroded material that is
suspended is 2.5%.  They also indicated that one of the drawbacks
of using the wind erosion equation is the assumption of a
constant percentage of total soil losses becoming suspended and
that this assumption was made without any substantiating data.

     Publication—This field study was  conducted and documented
under a contract with the EPA Office of Air and Waste Management,
Office of Air Quality Planning and Standards, Research Triangle
Park, North Carolina.  The report was published in  1974 as
Publication No.  EPA-450/3-74-037.
                              2-193

-------
  TABLE  2.76.   CONCENTRATION MEASUREMENTS  AND SAMPLE RUN TIMES

Date
6/9
6/11
6/13
6/16
6/18
6/20
6/23
6/25
6/30
7/2
7/3

Test
Period
(hr)
24
12
12
24
12
12
24
12
24
12
12

Upwind
Concent.
(ng/m3) /
Sample
Time
(min)
94/1 130
95/484
65/276
75/1 126
71/532
61/410
67/1087
86/586
61/1233
64/611
50/1139
Net Downwind Concentration (^g/m3) / Sampling
(minutes)
Hi-vol 2
8/1140
107/403
215/70
3/1192
21/340
Voidc
8/1011
55/578
16/1119
20/613
28/1058
Hi-vol
3
23/1082
152/415
125/73
OV1128
16/381
Voidc
6/1440
19/721
31/1066
17/620
24/1031
Hi-vol
4
49/1074
184/413
15/80
OV1082
37/406
48/201
Voidc
89/301
Ob/1 190
11/596
28/869
Hi-vol 5
13/1165
172/423
Ob/280
7/1168
42/619
2/719
9/1352
33/719
31/982
71/378
22/1249
Hi-vol
6
4/1064
76/355
125/62
26/378
Voidc
Voidc
27/1024
210/510
42/1032
Voidc
19/1054
Time
Avga
19
138
96
7
29
25
12
81
24
30
24
 a) Average downwind sampling time was not calculated
 b) A net concentration of zero was assumed when the upwind concentration was
    slightly higher than the downwind
 c) No explanation was given for the voided samples

Study 4— Axetell.  Survey of Fugitive Dust from Coal Mines.
          EPA-908/1-78-003.   1978.

     Methodology—The investigators used upwind-downwind
dispersion modeling to measure emission factors for several
mining operations:  topsoil removal,  drilling, blasting,
dragline, shoveling/truck loading, and fly-ash dumping.
Concentrations resulting  from emissions from exposed areas were
also measured.  However,  emission rates were not determined  for
exposed areas.  During most of the sampling periods for exposed
areas,  the particulate concentrations increased with distance
downwind from the source.   These  values would have produced
highly variable emission  rates at the various downwind distances,
using the area source dispersion  model employed in this study -

     Test Sites—Operations at five western coal mines were
tested.  Except for one lignite mine, all of the operations
extracted sub-bituminous  coal.

     Parameters and Equipment—Most of the parameters measured in
this study are listed in  Table 2.77 along with the tools  used to
                               2-194

-------
collect the data.  In addition,  estimates were made for the
initial dimensions and dispersion of the dust plume, the
receptor's distance from the source, the receptor's vertical and
horizontal distance from the plume centerline, and the length of
time the receptor is in the plume.  These estimates were all made
visually.

   TABLE  2.77.   PARAMETERS MEASURED  AND  CORRESPONDING EQUIPMENT
Parameters
Upwind concentration of TSP
Downwind concentration of TSP
Wind direction
Wind speed
Other atmospheric stability parameters
Particle size distribution
Equipment
Standard hi-vol
Standard hi-vol
Recording wind instrument
Recording wind instrument, hand-held
wind speed anemometer
Unspecified
Millipore filters on nuclepore
holder and pump, microscope
filter
     Equipment Configuration—The equipment configuration varied
between sources and mines.   It was not completely described for
each test run.  Typically,  a pair of hi-vols were placed together
at a location upwind from the exposed area.  This was preferred
over placing samplers immediately upwind from the activity
because anticipated brief wind direction reversals would be less
likely to effect upwind concentration measurements.  Other hi-
vols were placed at 10, 20, and 30 meters  (or a similar series of
distances)  downwind from the exposed area activity.  Downwind
samplers were set up at both 1.2 and 2.4 meters above the ground
at these downwind locations to provide information on the
vertical dispersion of the plume.

     Sampling Runs—The number of sampling runs and the measured
concentrations for the exposed areas measured are in Table 2.78.
Each sampling run consisted of several concentration
measurements.

     Quality Assurance—The guidelines in the Quality Assurance
Handbook for Air Pollution Measurement Systems  (U.S. EPA, 1976)
were followed in preparing filters, collecting and analyzing
samples, and auditing the data.  Hi-vol samplers were calibrated
before field work began at each mine.  One of every 25 filters
was treated as a blank.

     The investigators listed several problems experienced in the
field sampling program which indicate a decrease in the
reliability of the field data.  .Foremost was the fact that much
of the data collected in the field was subject to the ability of
                              2-195

-------
 TABLE 2.78.  MEASURED CONCENTRATIONS  DOWNWIND OF EXPOSED AREAS


Mine
A



B






C




D


E




Sampling
period
Dist
1
2
3
4
1
2
3
4
5
6
Dist
1
2
3
4
Dist
1
2
3
1
2
3
4
Downwind concentrations, |ag/m3
At 1 .2 m ht
10m 20m 30m 40m
103
377
423
490
1376
1480
3846
4842
1398
1602
253
603
322
240
1865
1944
4716
6199
2310
2203
171
313
347
338
1979
2139
9309
6433
2565
2446


349
297
2431


6407

2771
At 2.4 m ht
10m 20m 30m





689
815

398






742
1325

850






1111
1640

991

Samplers within exposed area
82
86
99
106
Om
213
457
598

119
1 44
65
77
11 1
168
1 17
10m
100
317
567
298
134
66
92
109
171
1 17
165
20m
172
474
270
143
153
104
94

81
134

30m
109

559
228
153

1 17
120


64
Om

429

91

98

32


42
10m

409

74

105

74


1 17
20m

1705

110

66


Background
concentration
64
64
88
88
84
84
84
84
84
84

32
32
32
32

87
105
105
a
a
a
a

Wind
Speed
(m/sec)
.8
1 .0
.4
.4
7.2
7.2
8.2
8.2
7.5
7.5

2.2
2.2
2.2
2.2

5.7
9.7
8.8
4.8
4.8
6.2
6.2
 a Wind reversal at start of sampling period; no upwind samplers.

the field staff to make visual estimates.  Plume dimensions  and
distance from the plume centerline are  two examples of this.   In
addition, one avoidable problem greatly decreased the integrity
of the data:  the two separate field crews did not follow  the
same procedures in collecting samples in several critical
respects.  There was no way to fully correct the data for  these
differences.

     Findings—As indicated above, the  data did not allow  for  the
calculation of emission rates due to wind erosion of exposed
areas.  In fact the data indicated that the concentrations
increased downwind of the exposed areas.

     The investigators initially assumed that the emission rates
would decrease with distance downwind due to particle deposition.
However, despite the fact that emission rates were not calculated
for exposed areas, the empirical data collected in this  study  did
not support the supposition that the apparent emission rate
decreases with distance due to particle fallout.  This is  readily
                              2-196

-------
apparent from examination of the concentration information which
shows an increase in concentration with distance downwind.

     The data also indicate that,  when measured at two heights
(1.4 meters and 2.4 meters) at the same downwind distance, the
concentrations at the lower height were generally 20-50% higher
than those measured at the sampler highest above ground.

     Publication—This study was conducted and documented under a
contract for the Environmental Protection Agency, Region VIII,
Office of Energy Activities, Denver,  Colorado.  It was published
in February of 1978 as Publication No. EPA-908/1-78-003.

Study 5— Cowherd et al.,  Iron and Steel Plant Open Source
          Fugitive Emission Evaluation.  EPA-600/2-79-103.  1979.

     Methodology—Wind erosion emissions were generated using a
portable wind tunnel outfitted with a sampling train that
consisted of a tapered probe, cyclone precollector,  parallel-slot
cascade impactor, back-up filter,  and high-volume sampler motor.
Interchangeable probe tips were utilized on the probe to
facilitate isokinetic sampling.

     Emission factors developed from the data collected as part
of this study were presented, but the methodology used to
calculate the emission factors was not documented.

     Test Sites—Preliminary tests were conducted on disturbed
prairie land and a coal storage pile to determine the threshold
friction velocities for wind erosion and to gather other data
required to design the sampling module.

     Emissions tests were made a total of 12 times.   Eight of
these tests were conducted on the upper flat surface of an
inactive coal storage pile.  These eight tests were subdivided
into five tests of undisturbed (crusted) areas and three tests of
disturbed areas.  In addition to these eight tests,  two tests
were performed on a flat undisturbed area adjacent to a dolomite
storage pile and two tests were conducted on disturbed prairie
soil.

     Parameters and Equipment—A portable wind tunnel with an
open-floored test section measuring 15 cm x 2.4 m was utilized to
generate wind erosion emissions on a coal storage pile, ground
adjacent to a dolomite storage area,  and disturbed prairie soil
within iron and steel plants.

     In addition to the wind tunnel,  particulate samples were
collected using an isokinetic probe attached to a cyclone
precollector mounted to a cascade impactor.
                              2-197

-------
     Soil parameters were also determined.  Samples were obtained
both prior to and subsequent to sampling using a small broom.
These samples were analyzed gravimetrically in the laboratory.
In addition,  the moisture content and the silt content were also
determined.  The silt content was determined using a conventional
shaker.

     Dust that collected in the probe tip and in the cyclone
precollector were also collected by rinsing the probe and
collector with distilled water and then drying until a dry
residue remained.

     Equipment Configuration—A sampling train that consisted of
a tapered probe, cyclone precollector, parallel-slot cascade
impactor, back-up filter, and high-volume sampler motor was
mounted so that the probe was oriented at right angles to the
flow in the tunnel and along the centerline.  Interchangeable
probe tips to facilitate isokinetic sampling were utilized on the
probe.  These tips enabled isokinetic sampling at average cross-
sectional velocities of 7,  12, 17, and 27 m/sec.

     Sampling Runs—One sampling run was conducted for each test
site detailed above, resulting in a total of 12 sampling runs.

     Quality Assurance—A minimal amount of information
concerning the handling of filters and dust samples collected in
the field was provided, however, detailed procedures concerning
data manipulation, emission factor calculations, treatment of
blanks,  calibration of samplers and other similar items was not
provided.

     Findings—Table 2.79 details the emissions data collectfed
as part of this study.  As can be seen from this table, emission
rates ranged from 0.023 to 6.5 g/sec/m2.   Undisturbed  surfaces
emitted much lower levels than did disturbed surfaces of the same
material, verifying the capability of surface crusts to inhibit
wind erosion.  In addition, the authors note that wind erosion
emissions tend to decay over the collection period.  No effort
was made to develop emission factors from this data.

     Publication—This study was conducted and documented under a
contract for the Environmental Protection Agency, Office of
Research and Development, Industrial Environmental Research
Laboratory, Research Triangle Park, North Carolina.  It was
published in May of 1979 as Publication No. EPA-600/2-79-103.
                              2-198

-------
          TABLE 2.79.  WIND EROSION SAMPLING PARAMETERS
Run
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
C-9
C-10
C-11
C-1 2
Surface Type
Undisturbed
coal
Undisturbed
coal
Undisturbed
coal
Disturbed coal
Disturbed coal
Disturbed coal
Disturbed coal
Disturbed coal
Undisturbed
dolomite
Undisturbed
dolomite
Disturbed
prairie soil
Disturbed
prairie soil
Cross-
Sectional
Average
Velocity
(m/sec)
15.6
25.0
25.0
8.49
16.1
16.1
19.2a
15.6a
10.3
14.8
9.83
10.7
Volume
Sampled
(m3)
5.66
2.83
5.23
5.66
1 .13
3.40
0.850
0.378
5.66
5.66
5.66
1 .08
Total
Mass
Collected
(g)
0.8680
2.7565
0.2176
0.2995
2.4418
0.4106
2.6867
3.0931
0.3773
4.2370
0.6034
8.1764
Suspended
Particulate
Emission
Factor
(g/sec/m2)
0.12
1 .2
0.053
0.023
1 .7
0.097
3.1
6.5
0.035
0.56
0.052
4.1
      Estimated value
Study 6— Cuscino et al.,  Iron and Steel Plant Open Source
          Fugitive Emission Control Evaluation.  EPA-600/2-83-
          110.   1983.

     Methodology—Wind erosion emissions were generated using a
portable wind tunnel outfitted with a sampling train that
consisted of a tapered probe, cyclone precollector,  parallel-slot
cascade impactor, back-up filter, and high-volume sampler motor.
Interchangeable probe tips were utilized on the probe to
facilitate isokinetic sampling.

     Particulate emission rates were determined using the
following equation:
                             E-
                              2-199

-------
where

     E  = particulate emission rate, g/m2-sec
     Cn = net particulate concentration, g/m3
     Qt = tunnel flow rate, m3/sec
     A  = exposed test area, m2 = 0.918

     No attempt was made to determine emission factors based on
the soil properties measured as part of this study and the
emission rates determined.

     Test Sites—Test sites were formed by plant personnel at two
facilities.  Test sites were prepared by either a front-end
loader or a bulldozer.  Tests were performed on coal storage
piles or on coal yard exposed areas.  Preliminary tests were
conducted to determine the threshold friction velocities for wind
erosion and to gather other data required to design the sampling
module.

     Emissions tests were made a total of 29 times.  Fourteen of
these tests were conducted on uncontrolled surfaces of coal
storage areas, twelve on controlled coal storage piles, two on
active exposed areas, and one test on an inactive exposed area.

     Parameters and Equipment—A portable wind tunnel with an
open-floored test section was utilized to generate wind erosion
emissions on a coal storage pile, or on exposed areas within iron
and steel plants.  Tunnel wind speed was measured using a pitot
tube at the downstream end of the working section and was related
to wind speed at the standard 10 meter height by means of a
logarithmic profile.

     In addition to the wind tunnel, particulate samples were
collected using an isokinetic probe attached to a cyclone
precollector mounted to a cascade impactor.

     Soil parameters were also determined.  Soil samples were
obtained both prior to and subsequent to air sampling using a
small broom.  These samples were analyzed gravimetrically in the
laboratory.  In addition, the moisture content and the silt
content were also determined.  The silt content was determined
using a conventional shaker.

     Dust that collected in the probe tip and in the cyclone
precollector were also collected by rinsing the probe and
collector with distilled water and then drying until a dry
residue remained.

     Equipment Configuration—A  sampling train that consisted of
a tapered probe, cyclone precollector, parallel-slot cascade
impactor, back-up filter, and high-volume sampler motor was
mounted so that the probe was oriented at right angles to the

                              2-200

-------
flow in the tunnel and along the centerline.  Interchangeable
probe tips were utilized on the probe.  These tips enabled
isokinetic sampling over the desired tunnel wind speed range.

     Sampling Runs—One sampling run was conducted for each test
site detailed above, resulting in a total of 29 sampling runs.

     Quality Assurance—The sampling and analysis procedures
utilized in this study were subject to certain quality control
guidelines.  Procedures followed in collecting and analyzing
samples were documented in considerable detail.  Quality control
measures were set forth for the sampling media, sampling flow
rates, and sampling equipment (proper performance).   Criteria for
interrupting sampling were also documented.  Quality assurance
practices included processing blank samples, calibration of
equipment, and auditing of sampling and analysis procedures.

     The investigators note that their procedures met or exceeded
the requirements set forth in Quality Assurance Handbook for Air
Pollution Measurement Systems, Volume II - Ambient Air Specific
Methods (U.S. EPA, 1077) and Ambient Monitoring Guidelines for
Prevention of Significant Deterioration (U.S. EPA,  1978).

     Filter substrates were coated with grease to. help eliminate
known particle bounce problems associated with cascade impactors.
However, correction routines were still necessary to try and
correct for particle bounce problems.

     Findings—Table 2.80 details the emissions data collected as
part of this study.  Only emissions data collected for
uncontrolled tests are presented in this table.  Emission rates
range from less than 0.1 mg/m2/sec  to 288  mg/m2/sec.  This entire
range was found for the three tests on exposed areas and
encompasses the results for coal storage piles.  The authors note
that wind erosion emissions tend to decay over the collection
period.  For those results where the emission rates were found to
decay by more than 20% during back-to-back tests, an additional
calculation was performed in order to determine the erosion
potential.  They proposed a technique for determining the erosion
potential.  The calculation technique assumes that there is an
exponential decay in the emission rate that is based upon the
material available for erosion.  The erosion potential was
calculated from the measured loss rates for two different erosion
tests of the same surface.  The equation for determining the
erosion potential is:
                           In	
                           In
                              (M0-L2)   tz
                              2-201

-------
where:

     M0 = Erosion potential  (quantity of erodible material     ^
          present on the surface before onset of erosion),  g/m
     L! = measured loss rate during time period 0 to  t,,  g/m  =

          Eifci                                               2
     L2 = measured loss rate during time period 0 to  t2,  g/m  - L,
          + E2(t2-t1)

     Although emission rates were determined in this  study,  no
emission factor for wind erosion was developed.

               TABLE  2.80.   WIND EROSION PARAMETERS
Run
F-46
F-47
F-48
F-49
F-50
F-51
F-52
F-53
F-54
F-55
H-20
H-21
H-22
H-23
H-24
H-25
H-26
Material
Exp. area
Exp. area
Exp. area
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Flow
Rate
(mVhr)
3,570
2,800
3,890
2,030
2,790
2,790
3,740
3,740
1 ,840
2,760
3,220
4,180
4,320
2,770
2,820
3,710
4,390
Volume
Sampled
(m3)
13.0
6.60
1 .68
12.6
5.15
20.6
10.4
41 .5
10.4
14.3
9.85
13.1
13.4
0.889
8.00
12.1
14.2
Total Mass
Collected
(mg)
4.72
309
413
5.06
22.6
38.2
82.9
73.7
5.30
13.2
232
459
105
9.43
43.0
135
1 ,770
Net Emission Rate
TP
(mg/m2/s)
0.0843
39.1
288
0.215
3.65
5.14
9.06
2.01
0.258
0.753
2.14
43.9
9.96
8.89
4.47
12.4
166
IP
(mg/m2/s)
-
5.19
17.0
0.0685
0.489
0.569
0.0543
0.0806
0.163
0.166
0.265
3.42
0.911
1 .51
0.113
0.285
2.64
FP
(mg/m2/s)
-
1 .75
5.31
0.0238
0.163
0.179
0.0181
0.0161
0.0635
0.0516
0.125
1 .01
0.303
0.308
0.0844
0.140
0.319
     Publication—This  study was  conducted and documented under a
 contract  for  the  Environmental  Protection Agency,  Office of
 Research  and  Development,  Industrial Environmental Research
                               2-202

-------
Laboratory, Research Triangle Park, North Carolina.  It was
published in October of 1983 as Publication No. EPA-600/2-79-103

Study 7— Axetell and Cowherd.  Improved Emission Factors for
          Fugitive Dust from Western Surface Coal Mining Sources.
          EPA-600/7-84-048.  1984.

     Methodology--Wind erosion emissions were generated using a
portable wind tunnel outfitted with a sampling train that
consisted of a tapered probe, cyclone precollector, parallel-slot
cascade impactor, back-up filter, and high-volume sampler motor.
Interchangeable probe tips to facilitate isokinetic sampling were
utilized on the probe.

     Particulate emission rates were determined using the same
methodology described above for the Cuscino et al., 1983.

     Test Sites—A total of 37 wind tunnel tests were conducted
as part of this study.  Tests were performed on either coal
storage piles or on exposed ground areas.  27 tests were
performed on coal storage piles and 10 tests were performed on
exposed ground areas at three different mines.  Emissions tests
on exposed surfaces were made on topsoil, subsoil,  overburden and
scoria.

     Parameters and Equipment—A portable wind tunnel with an
open-floored test section was utilized to generate wind erosion
emissions on coal storage piles, or on exposed areas within
western surface coal mines.  Tunnel wind speed was measured using
a pitot tube at the downstream end of the working section and was
related to wind speed at the standard 10 meter height by means of
a logarithmic profile.

     In addition to the wind tunnel, particulate samples were
collected using an isokinetic probe attached to a cyclone
precollector mounted to a cascade impactor.  Particle size
distributions were determined for particles less than or equal to
30 microns in diameter.

     Soil parameters were also determined.  Soil samples were
obtained both prior to and subsequent to air sampling using a
small broom.  These samples were analyzed gravimetrically in the
laboratory-  In addition,  the moisture content and the silt
content were also determined.  The silt content was determined
using a conventional shaker.

     Dust that collected in the probe tip and in the cyclone
precollector were also collected by rinsing the probe and
collector with distilled water and then drying until a dry
residue remained.
                              2-203

-------
     Equipment Configuration—A sampling train that consisted of
a tapered probe, cyclone precollector,  parallel-slot cascade
impactor, back-up filter, and high-volume sampler motor was
mounted so that the probe was oriented at right angles to the
flow in the tunnel and along the centerline.  Interchangeable
probe tips were utilized on the probe.   These tips enabled
isokinetic sampling over the desired tunnel wind speed range.

     Sampling Runs--One sampling run was conducted for each test
site detailed above, resulting in a total of 37 sampling runs.
However, because many of the samples exhibited decaying emission
rates, several of these runs represent a second run on the same
surface so that the erosion potential could be calculated.

     Quality Assurance—This study included a thorough quality
assurance program, which was subject to evaluation by a technical
review group  (including the two EPA project officers, and
representatives of the Bureau of .Land Management, the Bureau of
Mines, and the mining industry).  All samplers were calibrated on
a regular basis.  Sampling media were conditioned at constant
temperature and humidity prior to weighing.  Seven percent of
tare and final filter weights were audited.  For every ten
regularly processed filters and substrates, at least one was
processed as a blank.

     The investigators note that their procedures met or exceeded
the requirements set forth in Quality Assurance Handbook for Air
Pollution Measurement Systems, Volume II - Ambient Air Specific
Methods  (U.S. EPA, 1977).

     Filter substrates were coated with grease to help eliminate
known particle bounce problems associated with cascade impactors
only at the third mine sampled.  However, correction routines
were still necessary to try and correct for particle bounce
problems even for samples obtained using coated filters.  In
tests conducted for other fugitive sources at the mines,
dichotomous samplers were also utilized, and it was determined
from collocated samplers that the dichotomous samplers yielded
more reproducible results than cascade impactors for determining
the particle size distribution.  Unfortunately, for wind erosion,
the experimental setup did not include dichotomous samplers.

     Despite the attention given to normal quality assurance
practices in field data collection, the overall level of quality
assurance for this study is compromised by the paucity of
published raw field data.  For instance, measured exposures at
the various profiling heights were not reported.  Because of this
omission, the calculations of base emission factors made by the
investigators cannot be repeated and verified.

     Findings--Table 2.81 details the emission rates determined
for this study.  Unfortunately, the actual mass catch on the

                              2-204

-------
filters was  not presented  in this report  and thus is not  reported
in the table.   Emission rates were determined for both  suspended
particulates (< 30 |j.m) and inhalable particulates (< 15 ^im) .
Emission rates ranged from approximately  1  mg/m2/sec to 254

                   TABLE  2.81.  WIND  TESTING DATA


Surface
Type

Coal Storage
Piles

























Exposed
Ground Areas

















Run

J-24
J-25
J-26
J-27
K-39
K-40
K-41
K-42
K-43
K-45
K-46
P-20
P-21
P-22
P-23
P-24
P-25
P-26
P-27
P-28
P-29
P-30
P-31
P-32
P-33
P-34
P-35
J-29
J-
30"
K-35
K-36
K-37
K-49
K-50
P-36
P-37
P-38
P-
39b
P-40
P-
41b
Wind speed
at tunnel
centerline
(m/s)

14.3
14.2
11 .7
15.6
16.7
15.0
14.8
16.9
16.9
13.6
13.6
11 .6
13.1
13.1
14.2
14.8
16.0
16.2
16.0
15.8
17.3
16.9
11 .8
12.0
14.5
14.4
14.5
18.1
16.6
15.1
14.8
15.1
15.8
15.8
10.3
10.3
10.3
6.3
8.1
10.7



Emission Rate

Suspended
particulate
(mg/m2-sec)
3.40
5.20
254
74.8
170
111
4.54
93.1
4.36
59.8
7.41
12.7
9.66
1 .08
2.32
1 .76
3.92
9.48
38.6
5.78
16.1
1 .68
19.1
2.31
27-. 4
6.05
2.78
1 .60
-
36.8
1 .20
6.93
33.7
0.782
16.1
30.5
60.2
-
116
-



Inhalable
particulate
(mg/m2-sec)
2.26
3.44
157
47.2
119
72.2
2.96
62.6
2.79
43.6
5.48
8.11
4.14
0.597
1 .39
1 .07
2.31
5.33
20.2
3.43
11 .2
0.97
10.1
0.943
15.7
3.03
1 .85
1 .08
-
24.5
0.822
4.58
22.2
0.652
10.1
19..0
37.7
-
7.55
_



        a) No particle size data available
        b) Emissions consisted entirely of particles larger than 11.6
          (am aerodynamic  diameter.
                                2-205

-------
mg/m2/sec for  suspended particulates for coal piles.  The range
for inhalable particulates was from approximately 0.6 mg/m /sec
to 157 mg/m2/sec.   For exposed areas the ranges for suspended and
inhalable particulates were from 0.8-116 mg/m2/sec and 0.6-25
mg/m2/sec,  respectively.

     Although emission factors for other fugitive dust sources
were determined in this study using multiple linear regression
analysis, the authors decided not to perform this analysis on the
data for wind erosion.  This decision was based largely on the
fact that there were a number of samples that indicated emission
rates that decayed with time  (i.e. the  soils tested displayed
erosion potential).  As a consequence,  sequential tests on the
same surface were required which had the effect of reducing the
number of independent data points from  32 to 16.  The authors
felt that, because of the large number  of potentially significant
correction parameters in relation to the total number of
independent samples, regression analysis should not be conducted.
As a consequence, no attempt was made to determine emission
factors based on the soil properties measured as part of this
study and the emission rates determined.

     Publication—This study was conducted and documented under a
contract for the Environmental Protection Agency, Office of
Research and Development, Industrial Environmental Research
Laboratory, Cincinnati, Ohio.  It was published in March of 1984
as Publication No. EPA-600/7-84-048

Study 8— Connor et al., Wind Erosion Testing by Portable Wind
          Tunnel at an Iron and Steel Plant.  Paper 86-22.1, 79th
          Annual Meeting of the Air Pollution Control
          Association, Minneapolis, MM.  1986.

     Methodology--Wind erosion emissions were generated using a
portable wind tunnel outfitted with a sampling train that
consisted of a tapered probe, cyclone precollector, parallel-slot
cascade  impactor, back-up filter, and high-volume sampler motor.
Interchangeable probe tips to facilitate isokinetic sampling were
utilized on the probe.

     Particulate emission rates were determined using the same
equation utilized by Cuscino et al., 1983.

     No  attempt was made to determine emission factors based on
the soil properties measured as part of this study  and the
emission rates determined, although the relationships between the
erosion potential  and several of the soil properties measured
were examined.

     Test Sites—Tests were performed at the Dofasco iron and
steel plant in Hamilton, Ontario, Canada.  Tests were performed
on coal  storage piles or on exposed areas.

                              2-206

-------
     Emissions tests were made a total of 24 times.   Twelve of
these tests were conducted on uncontrolled surfaces  of coal
storage areas,  six on controlled coal storage piles,  and six on
exposed areas.

     Parameters and Equipment—A portable wind tunnel with an
open-floored test section was utilized to generate wind erosion
emissions on a coal storage pile,  or on exposed areas within the
iron and steel plant.  Tunnel wind speed was measured, but the
method used to measure it was not detailed.

     Soil parameters were also determined.  Samples  were analyzed
to determine the soil moisture, silt content and surface
compaction, but the details of the sampling and analysis
procedures were not published.

     Equipment Configuration—A sampling train that  consisted of
a tapered probe connected to a high-volume sampler was mounted so
that the probe was oriented at right angles to the flow in the
tunnel and along the centerline.  Interchangeable probe tips to
facilitate isokinetic sampling were utilized on the  probe.  These
tips enabled isokinetic sampling over the desired tunnel wind
speed range.
                                                 i
     Sampling Runs—It was not clear how many sampling runs were
conducted for each sample.  Since the erosion potential requires
a minimum of two samples back-to-back at the same site, it would
seem logical that a minimum of two sampling runs/sample were
conducted.

     Quality Assurance—With the exception of stating that all
filters were conditioned at constant temperature and humidity for
24 hours both prior to and following sampling, quality assurance
procedures were not detailed in this study.

     Findings—Only a limited amount of emission rate information
was presented in this paper.  The data presented was utilized to
compare the results of this study with that of earlier work
conducted at iron and steel facilities by Midwest Research
Institute.  The authors indicated that the data were largely not
comparable due to differences in the test conditions.  Only two
of the six measurements compared were similar.  The  other
measurements exhibited large differences.

     The authors did determine the erosion potential for each of
the test sites.  In addition, they examined the relationship to
the erosion potential of the silt content and the wind speed.
Graphs of these relationships were presented and indicated that
the erosion potential was roughly exponentially related to both
the silt content and the wind speed.
                              2-207

-------
     Publication—This study was conducted by personnel employed
by Dofasco, Inc. and was published in June of 1986 as Paper
number 86-22.1 in the Proceedings of the 79th Annual Meeting of
the Air Pollution Control Association.

Studies of Secondary Importance

     As indicated earlier in this section, the wind erosion
equation as described by Woodruff and Siddoway (1965) has
typically been the "emission factor" utilized for the development
of emission inventory estimates for wind erosion.  Despite the
number of studies examined above that made direct field
measurements, no emission factor has been developed as a direct
result of the those studies.  However, this has not prevented the
development of additional emission factors for calculating
emissions from wind erosion.

     Cowherd et al. (1977) propose a method for estimating
particulate emissions generated by wind erosion in an appendix to
their report.  The emission factor equation suggested is a
hybridization of the Woodruff and Siddoway wind erosion equation
coupled with some information obtained from work performed by
Dale Gillette.  The equation proposed is as follows:


                        £-0.0089   esr
                                 (PE/50)2
where
     E =  emissions of suspended dust, tons/acre/year
     e =  soil erodibility, tons/acre/year
     s =  silt content of surface soil, %
     f =  fraction of time wind exceeds the threshold value for
          wind erosion, assumed to be 12 mph
     r =  mitigative fractional reduction in wind erosion due to
          vegetative cover, derived from a graph
     PE=  Thornwaite's Precipitation-Evaporation index

As indicated above, this equation is based solely on analogy to
the wind erosion equation, and has no reliance upon any measured
data.

     Bohn et al. (1978) further perpetuate this equation and the
utilization of analogy in building emission factor equations in
their assessment of fugitive emissions from integrated iron and
steel plants.  In that report, they present two emission factor
Bquations for wind erosion.  One for storage pile wind erosion
emissions and one for wind erosion of exposed areas.  The
equation presented for wind erosion of exposed areas is of the
same basic form as that presented in Cowherd et al.. 1977 with

                              2-208

-------
the exception that the emission factor is now in Ibs/acre/year,
the coefficient is now 3400 instead of 0.0089,  the credibility
term is now e/50 rather than e, the silt term is now s/15 rather
than s, the fraction of time that the wind exceeds the threshold
is now f/25 rather than f,  and the vegetative cover term r is
eliminated altogether.  No explanation of the process used to
make these modifications was given.  Additionally, even though
the authors note that "it is known that above the wind speed
threshold of 12 mph for wind erosion, the erosion rate increases
with the cube of the wind speed,  the wind speed correction term
was simplified to reflect an average value of 15 mph for periods
of erosion."

     The equation given by Bohn et al. (1978) for wind erosion of
storage piles is as follows:
where

     EF=  suspended particulate emissions,  Ib/ton material stored
     s =  silt content of aggregate,  %
     D =  duration of storage,  days
     d =  dry days per year
     f =  percentage of time wind speed exceeds 12 mph

With the exception of the coefficient, none of the corrective
parameters are based on measured data.  The coefficient is based
on the measurements made during the Bohn et al. study-  However,
it has been adjusted to one half the measure value based on the
authors' estimate that the average wind speed through the
emission layer was only half the value measured above the top of
the piles.

     By 1984, the equation proposed above for storage pile wind
erosion emissions had been replaced by a new equation which
relates the wind erosion emissions to the erosion potential
corresponding to the fastest mile of wind for the period between
disturbances and the frequency of disturbances.  The equation is:


                           EF-fP(u\5)
                              2-209

-------
where

     f =       frequency of disturbance per month
     P(u+15)  =  erosion potential corresponding to the observed
               (or probable) fastest mile of wind, Ib/acre

The fastest mile must be corrected to a height of 15 cm in  order
to utilize this equation.  As before, no data analysis was
utilized to justify the proposal of this equation.  An equation
of similar form (and also without data analysis to justify  its
adoption) is published in AP-42.  It is used to predict
industrial wind erosion emissions.

     The latest equation proposed for evaluating wind erosion
emissions was developed by Gillette and Passi  (1988).  Their
equation was utilized to develop wind erosion emissions estimates
for the  1985 National Acid Precipitation Assessment Program
 (NAPAP).  In their work, they proposed the following equation  for
estimating wind erosion emissions:


                      N
                                     G(U)p±(U]dU
                                  Uti
where E is the mass of dust emitted in the time period AT;  C  is  a
constant determined by calibration; i is the index of summation
over N different erodible areas within the region of interest; R±
is the effect of soil roughness; g(LA)  is the effect of field
length, Li;-  A± is the area of land being  considered; G(U)  is the
vertical mass flux of dust as a function of wind speed, U;  p±(U)
is the probability density function of the wind speed during  the
time period of interest, and Uti is the ith threshold wind speed
for dust emission.  Details of how each of the above terms  are
obtained are given in the authors paper  (Gillette and Passi,
1988).  The dust emission data used to calibrate the model  was
that obtained by Gillette et al. ,  1978.
                              2-210

-------
                            SECTION 3

 ACTIVITY DATA AND EMISSION FACTORS USED FOR INVENTORY  PURPOSES


INTRODUCTION

     The following is an evaluation of the source activity data
for agricultural tilling,  wind erosion,  unpaved roads, paved
roads construction/demolition, mining and quarrying, and storage
piles.  The criteria used in this evaluation were completeness
and extent of activity input data,  magnitude of the inventory
area to which the activity data was applicable or available for
(i.e., local, regional,  or national),  accuracy of data  (when
possible),  frequency of updates and cost.

GENERAL INFORMATION

     The first approach use to locate sources of activity data
was to examine existing PM10 fugitive dust emissions inventories.
Individuals listed on the SCRAM bulletin board as the EPA
regional office contacts for PM10 inventories were contacted
first.  After speaking with individuals from each of the 10 EPA
regional offices, the general consensus was that a region only
maintained a fugitive dust PM10 emissions inventory if there were
nonattainment area(s) within the region.  If an inventory was
prepared, it was prepared by the State or local air pollution
agency.  When the State and local air pollution agencies (a least
one or more from half the EPA regions were contacted)  discussed
their inventories many had divided the above source categories
into either point or area source emissions.  Point sources would
be unpaved and paved roads, storage piles,
construction/demolition and mining and quarrying activities.
Area sources would include paved and unpaved roads not covered as
point sources, wind erosion and agricultural tilling.  Many of
the nonattainment areas are urban,  therefore agricultural tilling
is frequently not a factor.  In some states the activity levels
are maintained in permit files; however most inventories are for
SIPs and the activity data is site specific and has to be
obtained by the air agencies.  Only in a few instances  (such as
VMT for area source paved roads) were federal, state, or local
non-air pollution agencies contacted by the states air pollution
agencies to obtain activity data.

     The second approach used to locate activity data sources was
a literature search of previously developed PM10/TSP fugitive
dust emission inventories.  These documents were searched for
federal, state,  or local agencies that were contacted to provide
activity data for these emission inventories.  From the list of
agencies developed (DOT, DOC, USDA, FHWA, DOL, etc.), a phone
survey was conducted starting at the federal level.  The federal
agencies were questioned as to activity availability using

                               3-1

-------
evaluation criteria listed above..  For many of the source
categories, the activity data (when available) was available^at
both the national and state level.  Very few agencies maintain
data at the county or local level and were unable to confirm  if
the states or county agencies gather these statistics.  As a
consequence the next step was to contact the state agencies to
request the availability of activity and emission factor
correction parameters information.  At least one state per EPA
region was contacted for many different state agencies such as
DOT, DOC, SCS, and DOL.

     As a result of the telephone and literature surveys a couple
of common emission factor correction parameter and activity data
sources were located.  For the source categories examined all the
required meteorological data was available from the National
Climatic Data Center (NCDC) in the form of either the monthly or
annual Local Climatological Data summaries (NCDC,1991).  An
example of these summaries can be found in Appendix A.  These
reports are available directly from NCDC at an approximate cost
of $1 per month per station or $180 per diskette.  (The price per
diskette is based on an estimate to receive the fastest mile per
day for six years of data at one location).  The Local
Climatological Data Summaries can also be located in many
libraries.  There are approximately eight to nine thousand
substations in the U.S.  These substations collect at a minimum
the minimum and maximum daily temperatures and daily
precipitation amounts.  Most cities and major airports will
maintain all statistics.  The Climatological Data for each of the
50 states contains an index of the stations in that state.  The
data element, fastest mile (which is used in calculating wind
erosion emissions) is being phased out at many locations and  is
being replaced by peak wind statistics.  The publication of
monthly reports generally, have a three month lag time and the
annual summaries have a lag time of about six months.

     A second source of general information is the County
Statistic Tape File 3  (CO-STAT 3) which is available from the
U.S. Bureau of Census.  The data found in the County Statistic
Tape File 3  (CO-STAT 3) pertains to agricultural, business,
construction, education, housing, labor, population,  and service
industries.  The cost of the tape is $175.  The data is based on
1980 data plus other Census sources and other Census years up to
1987.  The data is national, state, and county or county
equivalent  (Alliance 1990).  This source would probably be most
useful for developing a national PM10 fugitive dust emission
inventory at a county level when activity data is not available
and surrogate activity indicators are needed, i.e., state VMT for
unpaved roads is estimated by the FHWA but a surrogate either
(housing or population) might be necessary to distribute the  VMT
to a county level.
                               3-2

-------
     As will be illustrated in the following source category
subsections, data availability is dependent on several factors.
Air pollution agencies will have very source specific data on
emission factor correction parameters or activity data for a
limited area.  The federal non-air pollution agencies maintain
activity data (VMT, acres of land planted, etc.)  for sources on
either a national and/or state level.  Finally the data is
sometimes collected but is very difficult to extract in the
necessary format or the data is simply not available.  An example
of this is that most states take traffic counts on state roads.
The problems in converting these counts to VMT include not
including counts for all roads systems within a county and
associating these counts to a road segment within the county.
The second example is that the U.S. Bureau of Mines survey all
the nonfuel mines in the U.S. each year to gather several types
of data including production figures.  The problem is that these
mines are promised total confidentiality in exchange for their
cooperation, thus the data, although collected cannot be used.
Individuals at both the U.S. BOM and U.S. Mineral Survey were
rather certain that the U.S. BOM was the only agency to collect
production figures.

     The following subsections detail the activity data
requirements of each of the fugitive dust source categories.  The
activity data and emission factor correction parameters required
to develop emissions estimates are discussed according to the
source of information be it EPA, State, or local air pollution
agencies, federal/State non-air pollution agencies, or PM10
fugitive dust emission inventory documents.  All contacts made
with the Federal, State, and local agencies were personal
telecommunications with individuals at these agencies.

AGRICULTURAL TILLING

     The following is the AP-42 emission factor equation for
determining the quantity of dust emissions from agricultural
tilling per acre of land tilled.

                    E=k(4.80)sA0.6 (Ibs/acre)

where:
     E    =    emission factor
     k    =    particle size multiplier  (dimensionless,
               aerodynamic particle size multiplier are given in
               AP-42 for [total particulate, <30, <15, <10, <5
               and <2.5 microns])
     s    =    silt content of surface soil  (%)

Both studies reviewed in section two,  (Cuscino, et al., 1981 and
Cowherd et al.,1974) for agricultural tilling are cited as
references in AP-42.  The particle size multiplier  (k) values
given in AP-42 are the only source found for this correction

                               3-3

-------
parameter.  The other correction parameter, silt content of
surface soil (s),  along with the acres of land tilled, and how
many tillings per year were surveyed.

Non EPA Federal, State, and Local Agencies

Silt content--

     A national map of soils types is available from the USDA,
Soil Conservation Service (SCS) dated 1988.  Personnel at the
National Agricultural Statistics Service  (NASS) indicated that
state SCS can provide county level maps of soil types.
Apparently a National Soil Survey for each county in the U.S. was
conducted and most have been completed.  One should contact the
state SCS office and talk to the soil scientist regarding the
availability of the data.

Number of tillings and acres of land planted—

     The acres planted for each crop by state for each year can
be obtained from the. USDA.  The delay on the annual state summary
is approximately four months.  NASS currently does not have the
data but are working on a central database.  This data is
available currently by contacting the state Soil Conservation
Service.

EPA regional office or state or local air pollution agencies

     The state air pollution agencies also indicated that
activity data and emission factor correction parameters were
obtained from the county SCS.  Other sources mentioned were
aerial photos to estimate acreage of farm land, site-specific
silt measurements (as in Arizona), and use of state permit
information  (as in Ohio).

Emission inventory documents

Silt content--

     Examples of silt content sources from PM10 fugitive dust
emission inventories are as follows. In developing an old TSP
inventory, Vermont used the U.S. SCS estimate of an average silt
content for Vermont agricultural soils and an average P.E. index
(State of Vermont, 1979).  Other documents suggested using soil
types obtained from soil survey reports published by the SCS,
county or state agricultural departments, and farmers or growers
trade associations (Jutze et al., 1974). Silt content may be
obtained from the state conservationist at the SCS through the
use of Land Use Maps  (Cowherd and Guenther, 1975).  The soil
types for California  (CARB,  1987) were obtained from the
University of California Division of Agricultural Sciences
generalized soil map of California developed in 1980.  CARB

                               3-4

-------
obtained the climatic factor from the USDA SCS Annual Wind
Erosion Climatic factor, 1986 interim map.

Number of tillings and acres of cropland —

     Tillings and acres of cropland sources were also cited in
several PM10/TSP emission inventories.  Vermont's 1979 SIP
indicated that tilled acreage was estimated by planimeter
measurement of cropland area from state Land Use Maps and
multiplied by the county ratio of tilled land to total cropland
obtained from the Vermont Census of Agriculture (State of
Vermont, 1979).  Crop acreage statistics by county can be found
in annual bulletins published jointly by USDA's Statistical
Reporting Service and the State university system except in
California where the data comes from individual county
agricultural reports (Jutze et al., 1974).  The Census of
Agriculture publishes harvested cropland by county (Cowherd and
Guenther, 1976).  For California the acreage of crops can be
obtained from each of the 58 counties' 1987 county crop reports
prepared by the county agricultural commissioner (CARD,  1987).

WIND EROSION

Emission factors

     For wind erosion three emission factor equations, the
associated correction factors, and activity data were
investigated.  These emission factors equations are Woodruff and
Siddoway (1965), Gillette and Passi (1988), and AP-42 industrial
wind erosion.

     The modified Woodruff and Siddoway windblown dust equation
cited in Cowherd et al.  (1974) is of the form:

                   ES=AIKCL'V  (tons/acre/year)

where:

     Es   =    suspended particulate fraction of wind erosion
               losses of tilled fields
     A    =    portion of total wind erosion losses that would be
               measured as suspended particulate,  (estimated to
               be 0.025)
     I    =    soil erodibility,  (tons/acre/year)
     K    =    surface roughness factor,  (dimensionless)
     C    =    climatic factor
     L'   =    unsheltered field width,  (dimensionless)
     V   =    vegetative cover factor,  (dimensionless)

Cowherd et al.  (1974) provides soil erodibility  (I) values based
on the predominate soil texture class.  Soil types can be
obtained from the USDA  (see agricultural tilling).  Values of K,

                               3-5

-------
L, and V for common field crops are also presented  in  Cowherd e_t
al. (1974).  The climatic factor  (C) can be calculated as
follows:
where :

     W

     PE
                       C = 0.345WA3/ (PE) A2
               mean annual wind velocity  (mph) corrected to a
               standard height of 30 feet
               Thornthwaite's precipitation-evaporation index
               (sum of 12 monthly ratio's of precipitation to
               actual evap-transpiration) .
The climatic factor can also be obtained from maps  that  use the
Weather Bureau data to calculate monthly climatic factors  (NCDC,
1991) .

     The second emission factor equation  (Gillette  and Passi,
1988) is as follows:
 N


i-l
                                    r
                                jAT \G(lf)
                                   »
                                              dU
where :

     E
     C
     N
     g(Li)
     A±
     G(U)

     P±(U)

     U±
               mass of dust emitted in the  time period delta T
               constant, determined by calibration
               number of different erodible areas within the
               region of interest
               effect of soil roughness
               effect of field length  (L±)
               area of land being considered
               is the vertical mass flux  (mass/unit  area/time)  as
               a function of wind speed  (U)
               probability density function of wind  speed during
               the period of interest
               ith threshold wind speed for  dust  emission.
With knowledge of  soil types and Land  Use  Maps  (see agricultural
tilling) , wind speed  statistics, and several  tables from Gillette
and Passi  (1988) the  above equation  can be solved.
                                3-6

-------
     The third emission factor equation is the industrial wind
erosion equation obtained from AP-42.  It can be used to estimate
wind generated particulate emissions from mixtures of erodible
and nonerodible surface materials subject to disturbances in
units of grams/meter/year as follows:
                                If

                                i-l
where:

     E    =    emission factor
     k    =    particle size multiplier (aerodynamic particle
               size multipliers for 30, <15,  <10 and <2.5 microns
               are presented in AP-42)
     N    =    number of disturbances per year
     PA    =    erosion potential corresponding to the observed
               (or probable) fastest mile of  wind from the ith
               period between disturbances.

Correction parameters are needed to determine P±.   These  include
the type of material, the shape and size of any piles, and
fastest mile of wind.  With this data and the help of other AP-42
tables in the industrial wind erosion section, the above equation
can be solved.

State Agencies

     The state Soil Conservation Service provide data for soil
characteristics (number of disturbed/undisturbed acres, mass
fraction of surface soil particles smaller than 50 microns,
fraction of soil that can be suspended, soil  credibility and
surface roughness).  The soil types can be obtained from the same
sources as for agricultural tilling.  There is county data from
state conservation agencies for mass fractions of surface soil
particles smaller than 50 microns.  The seasonal planting that
could be used to determine when and how often agricultural lands
are disturbed can be obtained from the USDA which provides usual
planting and harvesting dates by crop and state  (some states are
broken down in smaller areas or to county levels).  The most
recent document containing this information was published in
April 1984 and the previous publication date was August of 1972.
The only other source of data found was that the state of Ohio
uses permit information to determine wind erosion estimates of
storage piles.
                               3-7

-------
CONSTRUCTION

     Two sources of emission factors for construction/demolition
activities were investigate.  The first was the PM10 gap  filling
document (MRI, 1988)  which contains a construction site
preparation and a demolition of structures emission factor
equation.  The gap filling document sites the Kinsey et  al,
(1983) report in its references.  The demolition of structures
equation is the same emission factor as that found in AP-42 for
storage pile drop operations and the discussion on it can be
found in the storage piles.  The construction site emission
factor is as follows:

                                E  =    Et + Ee + Eh

where:

     E    =    emission factor
     Et   =    topsoil emission factor  (201bs/VMT)
     Ee   =    earthmoving emission factor  (4.3 Ibs/VMT)
     Eh   =    hauling emission factor  (10 Ib/VMT).

     A source for construction VMT was not- located.  Those places
contacted were the State Department of Motor Vehicles, the State
Information Processing Center, Highway Registration, Highway
Building Department, R.L. Polk, Transportation Data Center,
Construction Equipment and Leasing, Caterpillar Equipment and
Automotive News.

     The other emission factor equation is the AP-42 heavy
construction emission factor of 1.2 tons/acre of
construction/month of activity-  The emissions document  in
support of this emission factor is Cowherd et al. (1974).  A
discussion of the AP-42 heavy construction emission factor
activity data follows.  However, construction dollars were also
investigated since both Cowherd et al.  (1974) and Heisler  (1988)
have utilized construction dollars by SIC construction type as  a
factor to determine the acres of construction.

Non EPA Federal, State, and Local Agencies

     At the federal level there are three agencies that  publish
the value of construction.  F.W. Dodge Division of McGraw Hill
Information Systems Company reports the monthly value of
construction contracts for privately and publicly owned  new and
major alteration projects in the U.S. by their regional  division
(since January 1985 all F.W. Dodge regions conform to Census
regions with the exception of the west region).  They also
publish national monthly value of construction contracts for
privately and publicly owned projects for the following
categories: housekeeping residential, nonhousekeeping
residential, commercial, manufacturing, educational and  science

                               3-8

-------
hospital and other health treatment, public building and
nonbuilding construction.  The federal highway administration
publishes in its annual Highway Statistics (FHWA, 1990) state
construction dollars spent by agency (federal, state, local).
The Census Bureau (U.S. DOC, 1987) publishes every five years
(last published in 1987) construction value by state and by SIC
(1521, 1522, 1531, 1541, 1542, 1611, 1622, 1623, and 1629).  The
building permit section of the Bureau of Census provides county,
monthly construction values through three formats, diskette and
hardcopy which contains the county information and printed
publication.  There is also a annual report that summarizes the
data from the previous 12 months.  The monthly reports are
available with approximately a one month delay and the annual
report is generally available May of the following year.  There
is a cost to the general public but these reports are free to
other federal agencies.

     The NC Department of Labor  (DOL) maintains statistics on
building permits  (from counties/cities) by residential,
nonresidential or commercial building type.  Cost of the building
is estimated at the time the permit is acquired  and published in
the Division of Statistics monthly reports.  We were also
informed by the NC DOL that F.W. Dodge estimates use a ceiling.
The NC report has a few months time lag.

     Several states contacted used the above data sources.
Illinois Commerce Research Department releases F.W. Dodge
bimonthly summaries.  The Texas Real Estate Research facility and
the Kansas Department of Commerce obtain data from the U.S. DOC
monthly construction permit data.

     The Georgia Economic Development Department maintains only
the number of permits.  Both the Pennsylvania and Massachusetts
DOL have historical construction statistics but because of state
budget cuts no longer maintain these statistics.

EPA regional office or  state and local air pollution agencies

     In Indiana building permits contain estimates on dollar
value and acres of construction.  At the county air agencies in
Arizona, the construction permits contain duration information.

Emission inventory documents

     Acres of active construction can be obtained from  Public
Works or building department construction permit  files  (Jutze  et
al.. 1974).  Other places suggested were the  county  or  state
planning departments, building or trade associations,  county APCD
permit files, and bank  published economic revenues for
metropolitan areas.  Amick  et al.  (1974)  suggests the  land
planning council, county engineers  and county assessors for
obtaining the acreage under construction  for  residential,

                               3-9

-------
industrial and commercial construction.  In California the
Statistical Abstract published by the Department of Finance
contains data on residential building  (GARB, 1987).  The
California Department of Finance, Financial and Economic Research
unit maintains data on commercial, industrial and institutional
building valuation.

MINING AND QUARRYING

     For mining and quarrying operations, there is no one single
emission factor equation.  In fact, U.S. EPA (1990) separates
this category by commodity (metal, minerals, coal and tailing
piles [Evans and Cooper, 1980]) and further by processing
activity  (vehicle loading/unloading, blasting,  crushing,
drilling, overburden removal and unpaved road travel).  Most of
these emission factors are single factors i.e., Ibs/unit.  The
complexity is compounded since each mining commodity and process
has a different unit.  Several of the possible units for mining
and quarrying activities are tons of crushed stone processed,
tons of coal, number of blasts, cubic yards of overburden
removal, VMT, number of holes drilled, tons of coal loaded, tons
of overburden, tons of ore processed, or amount of land in use
for disposal of mill and processing wastes.

     Because of the variety in the above emission factors, all
the emission studies reviewed in section two apply (Cook et al,
1980:  Marple et al., 1980:  Axetell, 1978:  Axetell and Cowherd,
1984).

Non EPA Federal, State, and Local Agencies

     From the U.S. Bureau of Mines  (BOM) and the Mineral
Information section of. the U.S. Geological Survey it was
determined that the BOM sends a yearly survey to all mines
 (nonfuel) .  The U.S. BOM collects this data with the help of the
states.  In return the states coauthor the state chapters of the
annual minerals yearbook.  The BOM is the sole collector of
production data surveys.  Data is published with a two year lag
time in the yearbook on a national commodity level.  State and/or
county level data retrievals are made by special request of the
mining and quarrying statistical department.  Production numbers
are confidential therefore, at the state or county level the data
is often unavailable for confidential reasons.  To determine
whether a state would release the confidential information each
state would have to be individually contacted.  The individuals
spoken to at the federal level were unaware of any state that
would provide this information.  The U.S. DOE publishes annual
coal production and number of mines by state and type of mining
 (surface or underground) with a one year lag time.  This
information is obtained from the Energy Information
Administration, Form EIA-7A, "Coal Production Report".  These
production figures excludes mines producing less than 10,000

                               3-10

-------
short tons of coal during the year.  When speaking with the EIA
National energy information center they were unable to locate a
publication which presents county level coal production figures.

     Several state Geological Survey and Department of Labor  (KS,
PA, IL and CO) publish reports on mineral and coal statistics.
However, no state agency questioned released mineral production
figures due to the confidentiality of the data.

EPA regional office or state and local air pollution agencies

     The county air agencies within Arizona maintain permits on
sand and gravel activities.  Other states also use permit
information.

Emission inventory documents

     Jutze et al. (1974)  suggests obtaining total acreage,
surface condition and size of different sections of tailings
piles from state agencies or departments of mining and minerals,
the minerals yearbook, the state mining association or individual
mining companies. •

UNPAVED ROADS

     The following AP-42 equation is use to estimate the quantity
of size specific particulate emissions from unpaved roads per
vehicle mile traveled.

   E=k(5.9)(s/12)(S/30)(W/3A0.7(w/4)"0.5([365-p]7365)   (Ib/VMT)

where:

     E    =    emission factor
     k    =    particle size multiplier (dimensionless, values
               for k are presented in AP-42 for the following
               aerodynamic particle size ranges [<30, <15, <10,
               <5, <2.5 microns])
     s     =    silt content of road surface material (%)
     S     =    mean vehicle speed  (mph)
     W    =    mean vehicle weight  (ton)
     w    =    mean number of wheels
     p     =    number of days with at least 0.01 inch of
               precipitation per year.

It is recommended that site specific values for silt be used but
if not  available AP-42 provides typical silt content values of
surface material on industrial and rural unpaved roads.  As a
conservative approximation, the silt content of the parent soil
in the  area can be used.   Caution should be used when applying
                               3-11

-------
these values since tests show that road silt content is normally
lower than that in the surrounding parent soil.  AP-42 does not
suggest default values for speed, vehicle weight or number of
wheels.

     The emissions studies reviewed in section two that are cited
in AP-42 includes: Cowherd et al. (1974), Dyck and Stukel  (1976),
Bohn et_al. (1978), Cowherd et al. (1979), Cuscino et_aJL-  (1983),
Reider  (1983)  and Axetell and Cowherd (1984).

Non EPA Federal. State, and Local Agencies

Vehicle Speeds—

     An agency that determined vehicle speeds on all road surface
types and functional class systems was not found.  Federal
Highway Statistics does supply some speed data per state by
functional road class for roads with posted 55 mph speed limits.
NC DOT has some published and available speed data through
libraries but the extent of this data is unknown.

Vehicle weight, wheels and distribution—

     Estimates of national vehicle weight, the number of wheels
per vehicle and travel activity by vehicle type can be obtained
from the U.S. Department of Transportation (DOT) .  Some state
DOT'S  (NC, KS, and IL) maintain vehicle distribution and vehicle
weights (usually trucks) in their files.

Vehicle miles travelled (VMT) and road mileage—

     VMT on the state level is obtained from two sources.  The
annual Highway Statistics  (FHWA, 1990) reports the rural and
urban mileage by surface type and functional classification, but
does not include the local functional class unpaved road mileage.
Local functional class rural and urban unpaved road mileage by
average daily traffic volume (ADTV)  ranges in a spreadsheet
format can be obtained from the U.S.DOT.  County level road
mileage by functional system is also available from the U.S.DOT
approximately every two years but no distinction is made as to
surface type.

     All states are required to provide data'to the U.S. DOT in
order to collect federal aid.  The data provided is part of the
Highway Performance Monitoring System (HPMS).  A copy of the data
ele-3nts requested can be found in the Appendix A.  This data
con   .ns traffic counts for various state roads throughout each
sta

      .affic counts are made in most states on state maintained
roads. .  For some states distinction is made by road surface types
(NC, KS, and Texas), while other state Department of

                               3-12

-------
Transportations differentiate by functional class only.  Most
counts are not annual, they maybe updated every two  (NC) to 5
(IL) years.  Most of these counts are available on county maps
for a fee.  The exception is Texas which publishes annual tables
of county VMT by highway system and surface type.

EPA regional office or state and local air pollution agencies

Silt content—

     Average state silt content values were developed as part of
the 1985 National Acid Precipitation Assessment Program (NAPAP).
It was developed by the Illinois State Water Survey.  The
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.

VMT—

     Surveys and questionnaires are sent by states agencies or
the data are available in their permit files on facility unpaved
road VMT or emission factor correction factor parameters.   The
area source estimates for VMT are often obtained from state and
federal highway maps.  Arizona estimates ADTV on unpaved roads
since traffic counters have become buried under dust.

Emission inventory documents

Silt content—

     Vermont's 1979 SIP used an average silt content obtained
from direct testing for the state and used the state highway
departments estimate of average vehicle, speed.

VMT and mileage—

     Vermont's 1979 SIP used the state highway department town
maps to obtain road mileage and coupled with an estimate of 11
miles per vehicle per day.  Exact mileage by county for different
types of unpaved roads can be obtained from state or county
Highway Department's annual reports on the status of highway
statistics (Jutze et al.,  1974).  Traffic flow and road surface
type maps were obtained from the planning department of the
Illinois State DOT and the Mapping Department of the Missouri
State Highway Commission by Cowherd and Guenther  (1976) .
California draft inventory guidance does not use the AP-42
emission factor for unpaved roads  (CARB, 1987).  The windblown
dust equation they used requires acres of land which they
estimate by obtaining the unpaved road mileage from the

                               3-13

-------
California Highway System Engineering Branch and multiplying by
25 feet (the assumed width of the street).

PAVED ROADS

Emission factors

     AP-42 divides paved roads into two subsections: paved urban
roads and industrial paved roads.  Dust emissions from vehicle
traffic on a paved roadway may be estimated using the following
equation:

                       E=k(sL/0.7)Ap  Ib/VMT

where:

     E    =    emission factor
     L    =    total road surface dust loading  (grains/sq ft)
     s    =    surface silt content
     k    =    base emission factor  (Ib/VMT)'
     p    =    exponent (dimensionless)

The base emission factor coefficient  (k) and exponent (p) in  the
above equation for each size fraction  (TSP, <15, <10> and <2.5
microns) are listed in AP-42.  The two terms s and L when
combined are referred to as the silt loading parameter.  Default
silt loading values for a select number of cities by roadway
category are presented in AP-42.  Using all the.defaults from AP-
42 yields the recommended particulate emission factors for
specific roadway categories and particle size fractions.
Therefore only VMT for each roadway category is needed.

     The emission studies reviewed that are references for the
AP-42 paved urban road emission factors are Axetell and Zell
(1977), Cowherd and Englehart (1984), and Cowherd et al.  (1977).

     The quantity of total suspended particulate emissions
generated by vehicle traffic on dry industrial paved roads per
VMT may be estimated by using the following AP-42 equation:

          E=0.077l(4/n)(s/10)(L/1000)(W/3)A0.7  (Ib/VMT)

where:

     E    =    emission factor
     I    =    industrial augmentation factor  (dimensionless)
     n    =    number of traffic  lanes
     s    =    surface material silt content (%)
     L    =    surface dust loading  (Ib/mile)
     W    =    average vehicle weight  (ton)
                               3-14

-------
     The guideline for industrial road augmentation factor  (I)
suggest 1=7.0 for a paved industrial roadway with traffic
entering from unpaved areas.   1=3.5 for an industrial roadway
with unpaved shoulders where  20 percent of the vehicles are
forced to travel temporarily  with one set of wheels on the
shoulder.  1=1.0 for cases where traffic travels only on paved
areas.

     Typical silt content and loading values for paved roads at
industrial facilities are presented in AP-42.

     An alternate industrial  paved road emission factor for
traffic consisting predominately of medium and heavy duty
vehicles follows:

                  E=k(3.5)(sl/0.35)A0.3  Ib/VMT

where:

     sL   =    road surface silt loading (oz/sq yd)

The particle size multiplier  (k) values are given for various
aerodynamic size ranges (<15, <10,  <2.5 microns).  This
alternative equation can be replaced by a single valued emission
factor for light duty vehicles on heavily loaded roads for <15 or
< 10 microns (0.41 or .33 Ib/VMT).

     The emission studies reviewed in section one that are
references for the AP-42 industrial paved road emission factors
are Bohn et al.  (197.8),  Cowherd et al. (1979), Cowherd and
Englehart (1984), Cuscino et  al. (1983) and Reider (1983).

Non EPA Federal, State,  and Local Agencies

VMT—

     VMT on all road surface  types by functional class for each
state can be obtained from Highway Statistics (FHWA, 1990).  By
special request, the VMT for  all surface types by rural and urban
classes for each state for each month may be obtained.

Other-

     Traffic lane mileage by  rural and urban functional
classifications for each state can be obtained from FHWA  (1990).
Information on where to obtain vehicle weights and vehicle
distributions can be found in the unpaved road activity data
section.  A federal, state or local agency that publishes silt
loading values was not found.
                               3-15

-------
EPA regional office or state and local air pollution agencies

Silt—

     The state of Maine collected site specific  silt loading
values for sanding/salting of paved roads for their  1991  PM10
SIP-  Some site-specific silt content and size distribution data
was collected for Arizona's inventory.

Other—

      Locally derived paved road capacity, number of lanes,  and
ADTV values were obtained for Arizona's inventory-   Details on
EPA agencies obtaining VMT data can be found in  the  unpaved road
activity data section.

Emission inventory documents

     In CARB's draft inventory guidance they suggest  obtaining
VMT from California DOT division of Highway and  Programming for
arterial and collector highways based on HPMS data and CALTRANS
which estimates 1987 VMT by county.

STORAGE PILES

     Total dust emissions from aggregate storage piles are
contributed by several distinct source activities within the
storage cycle:

     1.   Loading of aggregate onto storage piles (batch or
          continuous drop operations)
     2.   Equipment traffic in storage area
     3.   Wind erosion of pile surfaces and ground areas around
          piles
     4.   Loadout of aggregate for shipment or for return to the
          process stream (batch or continuous drop operations).

For equipment traffic in storage areas the recommended emission
factor equation is that for vehicle traffic on unpaved roads.
The emissions from wind erosion of pile surfaces and ground areas
around piles are best estimated using the industrial wind erosion
equation (see wind erosion section).   The storage pile drop
operation (loading or loadout)  emission factor equation obtained
from AP-42 is as follows:

          E    =    k(0.0032) t(U/5)A1.3/(M/2)A1.4]

where:

     E    =    emission factor
                               3-16

-------
     k    =    particle size multiplier  (dimensionless)
               [aerodynamic particle size multipliers for <30,
               <15,  <10,  <5 and <2.5 microns found in AP-42]
     U    =    mean wind speed (mph)
     M    =    material moisture content (%)

AP-42 provides a table of typical silt and moisture content
values of materials at various industries.

     The emission reports reviewed earlier that resulted in the
above emission factor equations are Bohn et al. (1978), Cowherd
et al. (1974), and Cowherd et al. (1979).

     The state air pollution agencies gathered storage pile
statistics by three methods.  These methods were aerial photos,
(Illinois in 1988) to estimate the size of stock piles, permit
information, and surveys (Wayne County, Michigan).

     Amick et al. (1974)  suggested contacting the plant operators
to obtain storage pile activity data.  A federal,  state or local
non-air pollution agency that gathers statistics on storage piles
was not found.
                               3-17

-------
                            SECTION 4

                 EVALUATION AND RECOMMENDATIONS
     Field studies in which data was collected to develop
emission factors were reviewed in Section 2.   In Section 3,  the
activity data required to develop emissions estimates from these
emission factors were identified,  and sources from which this
data can be obtained were noted.  This section presents an
evaluation of the field studies, the resulting data, and the
availability of the requisite activity data.   Recommendations for
future research and development efforts based on this evaluation
are then presented.

EVALUATION

     Several broad areas have been identified for evaluation:

          1.   Methodology
          2.   Sampling equipment
          3.   Quality assurance
          4.   Emission factor development
          5.   Documentation
          6.   Activity data requirements

     Evaluations of each of the above areas are detailed below.

Methodology

     In general, two main methodologies have been utilized to
develop fugitive emissions data:  exposure profiling and upwind-
downwind modeling.  As documented in Section 2, other methods
have been used, but not nearly to the extent of these two.
Therefore, most comments presented here will concern one or both
of these two methods.  Several important problems with their
implementation have been identified.

Lack of Statistical Design—

     Very few of the reviewed field study reports documented a
statistically based experimental design.   In only one study
(Cook et al.,  1980) was a thorough statistical analysis utilized
to determine the number of samples necessary to assure adequate
data collection prior to the field sampling phase.

Lack of Procedural Standardization—

     Despite the number of experiments performed to determine  the
emissions from fugitive dust sources, few experiments were
conducted identically even within individual studies designed  to
sample the same source.  Some examples will serve to illustrate

                               4-1

-------
this point.  In the initial study conducted by MRI in 1974
(Cowherd et al.) in which exposure profiling was first utilized,
the filters in the profiler sampling heads were oriented
vertically.  In subsequent studies by MRI, the filters were
mounted horizontally-  No reason or justification was ever given
for this.

     To further illustrate this point of non-standardized
procedures, in the 1986 comparability study by Pyle and McCain,
the heights and numbers of exposure profiler heads for each group
differed, even though the source being measured was the same and
the samplers were located side-by-side.

     Other non-standard items noted in the reviewed studies
include the distance downwind from the source that the exposure
profiler is located, the number and types of ancillary samplers
(i.e. particle sizing devices) located with the profiler tower,
and the methods used to maintain the orientation of the samplers
into the wind and to correct the collected samples to isokinetic
conditions after sampling.

     No detailed procedural method has been developed for
fugitive dust sampling similar to that for Method 5 stack testing
or for locating and operating ambient particulate monitors.

Potential Background Sampling Problems—
     Typically, when exposure profiling has been utilized to
determine emissions from fugitive dust sources, the background
correction has been made using the data from one sampler, usually
located at a lower height than most of the exposure profiling
sampler heads.  Since the majority of the data collected and
utilized has been for TSP rather than PM10, the possibility
exists that there is a change in the particle concentration with
height above the ground.  Thus, correcting an exposure profiler
head located at 12 feet above the ground with the background
levels determined using a sampler with an inlet height of 2.5
feet above the ground could lead to errors in the determination
of exposures.

     Furthermore, the background samples were not collected
isokinetically.  Because the efficiency of the hi-vol depends on
ambient particle size and meteorological conditions, measured
background levels may be biased downward.

Potential Particle Size Distribution Sampling Problems—

     Most determinations of the particle size distribution have
been made using either a single instrument or two particle size
sampling devices located at two heights.  Typically, the device
used to determine the particle size distribution has been a high
volume cascade impactor.  The assumption in this approach is that
the particle size distribution does not vary with height and that

                               4-2

-------
estimation of particle-size specific emissions can be made from
the integrated exposure sample coupled with information on the
particle size distribution.  However, if the particle size
distribution varies with height, this method could result in
erroneous emission factor development.  Indeed, the study by
Axetell and Cowherd (1984) indicated that there was a difference
in the particle size distribution with height.

Unisolated or Incompletely Measured Dust Sources—

     Isolating a particular dust source from others and
completely measuring it are critical steps in developing an
accurate emission factor.  Difficulty with this issue was alluded
to or detected in several reports.  Most noteworthy are Cowherd
et al.,  1974 and Axetell, 1978.  Cowherd et al. estimated
emissions from building construction sites using the upwind-
downwind method.  Wind roses presented in the report indicated
that the source was not isolated from neighboring dust sources.
Thus, the modeled emission rate may not be reliable.  Axetell
noted problems with wind direction reversals which effectively
placed downwind dust mass on the nominal upwind filter.  Although
most of the other reviewed field study reports did not indicate
any problems of this nature, few presented evidence that this
issue was not a problem.  The researchers who use upwind-downwind
modeling are more likely to encounter this problem than those
using profiling because the former is the most appropriate method
of estimating emissions from widely dispersed, less easily
isolated sources.

Incorrect Solution of the Dispersion Model Equation—

     As indicated in the reviews presented in Section 2,
especially those for unpaved roads, the investigators have
frequently rearranged the dispersion equation incorrectly and
have then proceeded to solve the equation in its incorrect form.
Three of the six unpaved road studies employing upwind-downwind
dispersion modeling did not correctly apply the model.  Jutze and
Axetell (1974) claimed to use the equation for a continuously
emitting,  infinite line source in Turner's Workbook of
Atmospheric Dispersion Estimates  (1970).  However, as .is
explained in the summary of that study, they misplaced the term
which accounts for the angle between the wind direction and the
road, sin <|>.  Furthermore, attempts to reproduce their
calculations for source strength, q, were unsuccessful.  Three
examples of these discrepancies are shown in Table 4.1.
                               4-3

-------
TABLE 4.1.  EVALUATION OF COMPUTATIONS BY JUTZE AND AXETELL,  1974
height

-------
emissions.  In those cases, a puff model would probably be more
appropriate.  Incidentally, this is yet another example of
incomplete measurement of the source.

     Other studies that have utilized dispersion models have
ignored terms that must be included in the equation when the
sampler is not at ground level.  One study (Blackwood and
Chalekode, 1978)  utilized the point source dispersion model in
the form that the equation takes for ground level monitors
(z = 0),  despite the fact that it was clearly indicated in the
data presented that the monitor was located well above ground
level.

Sampling Equipment

Use of Outmoded Equipment—

     Although some new measurement technologies have been field
tested (i.e. Pinnick et al., 1985), only rarely have proven,
state-of-the-art samplers been utilized in developing emission
factors for fugitive dust sources.  Typically, the only EPA-
verified sampler that has been utilized in these studies is the
standard high volume sampler.  In those cases where dichotomous
samplers have been used, they were the early versions with the 15
|j.m inlet heads.  However, when the dichotomous samplers have been
utilized, the results have been consistent, especially with
regard to particle size information.

Indirect Measurement of PM10 Concentration—

     No studies examined in this report determined PM10 emissions
directly.  Estimates of PM10 emissions were made by either
interpolating or extrapolating from data on particle size
distributions without direct measurements of 10 |j.m particles.
Sampling equipment which does directly sample PM10 has been wind
tunnel and field tested and found to be reliable in 1982 (Wedding
et al.).

Particle Bounce in Measurement of Particle Size Distribution—

     The majority of the information collected on particle size
distributions was collected using high volume cascade impactors.
These samplers are known to have particle bounce problems
resulting in an overestimation of the fine particle contribution.
Additionally, even though attempts were made to correct for these
particle bounce problems by using greased substrates, the
particle bounce problems still existed.  Early studies which did
not employ greased substrates, utilized a calculational method to
"redistribute" the particle mass by assuming a lognormal particle
size distribution.  However, even some later studies that
utilized greased substrates found it necessary to utilize this
calculational method to redistribute this mass.  This could have

                               4-5

-------
been caused by high filter loadings on the filters.  Aerosol
scientists have known for a long period of time that a heavily
loaded filter (even if treated to avoid particle bounce) can
still exhibit particle bounce problems, since once the filter
becomes loaded,  the additional particles bounce off of the
previously collected particles, despite the treatment to prevent
particle bounce.

     Despite the knowledge that particle bounce problems existed
with these samplers and that the problem could be attenuated by
treating the filters, some studies still continued to use the
samplers without treating the filters.

     Perhaps the most damaging evidence of the particle size
distribution problem associated with using high volume cascade
impactors was found in the study by Axetell and Cowherd (1984)
where dichotomous samplers and cascade impactors with greased
substrates were operated side-by-side.  The results showed that
the dichotomous samplers yielded the most reliable results.

Poorly Characterized Samplers—

     Although EPA has been scrupulous in its development and
characterization of samplers designed for the determination of
ambient pollutant concentrations and for stack testing purposes,
no rigorous testing results have ever been published for exposure
profiler heads.  Although these samplers were typically based on
high volume sampler motors and filter holders, several firms
utilizing this technology have added components to the system
that potentially modify the flow and capture characteristics of
the samplers.  For instance, MRI utilized a roof-shaped cover
with a square inlet fitted with an isokinetic probe tip in
several of its exposure profile experiments.  The purpose of this
cover was to serve as a settling chamber in order to eliminate or
reduce the number of particles larger than 50 ^.m in diameter.
However, the attributes of this sampler concerning flow
characteristics and whether or not this configuration actually
served to remove particles larger than 50 |u.m has never been
tested rigorously, at least not so far as can be ascertained in
the publications examined for this report.

Quality Assurance

     General quality assurance  (QA) practices varied widely
between the studies reviewed.  Certain aspects of the sampling
programs had good to excellent QA.  The aspects of the sampling
programs that were most frequently subject to some form of QA
were maintenance and calibration of samplers; attention to
isokinetic sampling; preparation, handling, and weighing of
filters; and standard procedures for the sampling, handling, and
analysis of surface material samples.


                               4-6

-------
     Reports from the initial and other early profiling studies
provided substantial detail on the procedures followed in the
methodology.  However,  quality assurance measures such as
processing of blank profiler filters, collocation of profile
sampler intakes,  or auditing of profile filter weights were
generally not documented.

     The studies  which utilized dispersion modeling were
generally of lower documented quality than those using exposure
profiling.  For the modeling studies, documentation of quality
assurance procedures and even detailed methodological procedures
was generally lacking.

     In those studies that demonstrated the best quality
assurance programs, additional quality assurance protocols were
delimited for blank evaluations, auditing of filter weights,
statistical analysis to determine the number of samples necessary
to assure an adequate sampling of the source and collocation of
samplers.  Few studies utilized these additional quality
assurance procedures, however, and those that did have higher
quality assurance were typically the later studies.

     Several aspects concerning quality assurance were never
addressed in any of the studies.  Collocation of exposure
profiling arrays  to try and determine the reproducibility of the
methodology was never assessed.  Variability and reproducibility
of methods used to ascertain the silt content of various
materials by obtaining duplicate samples was never performed.

     Audits of data reduction calculations were rarely
documented.  As is indicated in Tables 4.1 and 4.2 in the
methodology evaluation above, calculational errors do occur.  The
need for more audits in data reduction procedures is also
demonstrated in Table 4.3.  Note that Cowherd et al., 1974 was a
report written under contract for EPA.  The investigators'
calculation of particulate concentrations from exposure
measurements could not be duplicated.  To derive concentration
from exposure, the requisite additional data are the area of the
intake, the rate  at which air is sampled, and the duration of
sampling.  Table  4.3 shows comparisons of the results of attempts
to reproduce the  published data for several profile samplers.
                               4-7

-------
  TABLE 4.3.   EVALUATION OF COMPUTATIONS BY COWHERD ET AL.,  1974.

ht
(ft)


10.5
8
5.5
3

sample
rate
(raVsec)

0.0137
0.0130
0.0123
0.0114

plume
concentr.
(mg/mj)

0.9
3.33
7.2
8.13
back-
ground
concentr .
(mg/ra!)

0.0469
0.0469
0.0469
0.0469

corrected
plume
concentr.
(mg/mj)
0.8531
3.2831
7.1531
8.0831

# of
passes


168
168
168
168

inlet
area
(in2)

4
4
4
4

sample
time
(sec)

3600
3600
3600
3600
MRI's
exposure
(rag/in2/
veh.)

0.082
0.289
0.591
0.629
Pechan'
exposure
(mg/inV
veh)

0.06
0.23
0.47
0.49
 a) Pechan formula: unit exposure - concentration * sampling, rate * sampling time / intake area / H passes

      Sampling isokineticity is  also an issue of quality
 assurance.   Studies  varied widely in their attention this issue
 and  in their methods of  addressing it.  Cowherd et al., 1974,
 noted that  changes in wind speed during sampling were
 insignificant.   However,  some of the exposures were
 isokinetically corrected.   The  field comparison study supervised
 by Pyle  and McCain  (1986)  is noteworthy for its application  (by
 some of  the participants)  of servo systems for continuous
 monitoring  and adjustment  of sampling velocity to match wind
 velocity-   Pyle and  McCain found that the servo system can
 maintain isokineticity at  each profiler head,  thus eliminating
 the  need to correct  the  measured exposures.

 Emission Factor Development

      Several major problems exist in the development of emission
 factors  from the data reviewed above.   These problems are listed
 and  discussed in the paragraphs  below.

 Lack of  Data--

      Perhaps the biggest problem is the paucity of data utilized
 to develop  the  emission  factors  currently used for several
 sources.  Emissions  from cattle  feedlots,  landfills,  and unpaved
 parking  lots have not been field measured at all.   Very limited
 testing  has  been conducted for agricultural tilling,  construction
 of buildings and roads, and wind erosion.

 Lack  of  Thorough Statistical Analysis—

      Many of  the studies  (especially the earlier ones)  provided
 little statistical analysis of the data collected.  Means and
 standard deviations  of the results were frequently not
 calculated.   Even those studies  that did provide detailed data
 analysis, either in  the form of  analysis of variance or multiple
 linear regression, often omitted crucial details.   For instance
 several  studies  which indicated  that they used multiple linear '
regression analysis  to determine the coefficients  associated with
correction parameters in the emission  factor equations provided
                               4-8

-------
no theoretical or empirical basis for the selection of those
correction parameters over other potential correction parameters.
In some of the studies,  the statistical analysis was presented,
but the underlying data was not presented.

Capricious Selection of Correction Parameters—

     In many instances,  a correction parameter was chosen without
empirical data indicating a relationship between it and the
emission rate.  Correction parameters were added on the basis of
theory with no data at all, either supporting or contradicting.
This was sometimes done on the basis of analogy to other sources
which were believed to be similar in some regard.  For example,
because the average number of wheels per vehicle was found to be
a useful predictor of emissions from unpaved roads, it was
assumed that the same correction parameter should be used for
paved road emissions.  Without any statistical basis, the wheels
correction parameter was added to the proposed predictive
emission factor equation for paved roads  (Cowherd et al.,  1979).
In fact. Cowherd and Englehart, in their 1985 source category
report on industrial and rural roads, found that wheels was not a
useful predictor of emissions from industrial paved roads.

     In that same report, Cowherd and Englehart recommended the
use of silt content  (i.e. percent silt) in place of silt loading
(i.e. mass of silt per unit area) as a correction parameter in
the unpaved road emission factor equation, despite data analysis
showing that silt loading explained more of the variance in the
emissions data.  Their recommendation was based on two
considerations.  First,  limited tests indicated that silt loading
was a less reproducible parameter than silt content, a finding
which was attributed to those relatively few cases in which no
well defined hard pan had formed beneath the loose surface
material.  Second, existing field data was found to be more
abundant for silt content of industrial roads than for silt
loading.  In response to the first consideration, after careful
comparison of the methodologies used to collect data on both
parameters, no substantial difference in difficulty of sample
collection was noted.  The improved predictive ability of the
equation utilizing silt loading over the one using silt content
would be well worth any possible slight increase in sampling
difficulty-  Regarding the second consideration, the survey of
available correction parameter and activity data in Section 3
found significant sources for neither one of the correction
parameters in question.

     Empirically unjustified modifications were also made to the
wind erosion emission factor.  As was noted in the review of Bohn
et al.,  1978 in the wind erosion section, several  correction
parameters were added without reference to any supporting data.
                               4-9

-------
Documentation

     Many of the problems noted above may be attributed at least
in part to poor documentation.  This includes documentation of
both sampling and analytical procedures, as well as raw field
data.  A lack of documentation frequently prohibited attempts to
reproduce the results.  The problem of undocumented raw data
speaks for itself, but the issue of procedural documentation
warrants some elaboration here.  Three key areas were often
poorly documented.

Method of Correcting Plume Exposure for Background Levels—

     Prior to the 1986 study by Pyle and McCain, the method of
correcting plume exposure for background dust was poorly defined.
In the study by Cowherd et al. (1974), an upwind hi-vol sampler
was used to measure the background dust level.  It is unclear
from the report if the adjustment to measured plume dust level
was made via concentration or exposure.  If the samples were
collected isokinetically, it would not matter; but, since they
often were not, the exact procedure for correcting for background
dust levels was ambiguous.

     All of the other profiling studies prior to 1986 employ
either the standard hi-vol sampler or the hi-vol with a size-
selective inlet (SSI) to measure background dust levels.  None of
these other studies clearly document the manner in which
background dust is accounted for.  All of the participants in the
side-by-side field study supervised and documented by Pyle and
McCain  (1986) used isokinetically sampling hi-vols to measure
background dust (most participants used a profile tower of two or
more sampling heads).

Method of Calculating Integrated Exposure From Sampler Data—

     There were several reports in which this process was not
clearly documented.  In cases in which two different types of
sampling heads were used on a single profiler (e.g. Bohn et al. ,
1978), no indication was given that differences in the sampling
characteristics of the heads were accounted for in the
integration process.  In many instances, the method of
integration was not even discussed.  Again, omissions such as
these make it impossible to reproduce the results.

Method of Deriving Emission Factor Equations—

     Particularly for storage pile emissions, but also to a
lesser extent for other source categories, the method or logic
used to develop emission factor equations was not fully
documented.  This problem was discussed above in the section on
emission factor development.


                               4-10

-------
Activity Data Requirements

     In evaluating and determining the types and availability of
activity data and/or correction parameters for PM10 fugitive dust
emission factors,  it became clear that the availability of data
necessary is clearly influenced by the scale of the emission
inventory being developed.  Conversations with air pollution
agency personnel responsible for developing inventories for
nonattainment areas indicated that either default parameters were
utilized or the agency personnel either went to the facility and
collected the information or a survey was sent to the facility
requesting that they provide the information.  While this
approach is adequate for small scale inventories covering a
limited area or number of facilities, for an inventory the scale
of the 1985 NAPAP inventory, such an approach would be
impossible.  The reason that these air pollution agencies must
utilize this approach however, is that the information needed
either as activity data or as a correction parameter in the
emission factor equation is not information that the facility
typically reports to or is collected by a federal agency or to a
trade association.  Several examples will serve to illustrate
this problem.

     The original TSP emission factor developed for estimating
construction- emissions required activity information on the acres
of construction and the number of months of activity.  Typically,
the information reported on a routine basis for construction
activities has been the number of permits issued or construction
dollars spent.  In some areas, the acres of construction are
available, but there was no instance where the months of activity
were maintained as a data element found in our examination of
sources of activity data.

     The EPA recommended PM10 emission factor for construction
activities requires activity data on the VMT for various types of
construction vehicles.  This information is not available at any
level and can only be obtained by making site-specific
measurements.  Thus, use of this emission factor to develop a
state or national level inventory would be impossible.

     The paved road emission factor requires knowledge of the
silt loading for use as a correction parameter in determining
emissions estimates from paved roads.  The database readily
available for use by most personnel involved in emission
inventory development for silt loading is AP-42, which contains  a
very limited number of measurements from an even more limited
number of sites.  An equation has been suggested for estimating
the silt loading on public paved roads based on ADTV.  However,
since the emission factor is given in Ibs/VMT and ADTV is
directly related to VMT, why not develop an emission factor  that
only requires ADTV to begin with.
                               4-11

-------
     The unpaved road emission factor requires correction
parameter information on silt content,  number of wheels, vehicle
weight and vehicle speed.   While there is limited information at
the national and state level on the distribution of vehicles by
weight and number of wheels, it is not broken down by type of
surface and cannot be obtained at the county level.  Detailed
information on silt content (especially at the county level),
like the silt loading for  paved roads,  is virtually non-existent.

     Further examples of this problem of availability of activity
data necessary for development of emissions inventories at
anything other than a very localized inventory scale include
industrial wind erosion (the fastest mile statistic necessary to
calculate emissions is being phased out and replaced in NCDC
summaries) and surface mining operations (15 different types of
activity data including VMT, holes drilled,  blasts, VMT by
scrapers, hours of bulldozer use and tons of coal mined are
required).  Other fugitive dust sources have similar problems.

RECOMMENDATIONS

     The following ten recommendations are based on the
evaluations of the studies in which fugitive emissions were field
measured, the activity data and correction parameters required to
utilize emission factors to develop emission inventories,  and the
availability of that data.

1)  Additional field testing should be conducted on the source
categories with the following order of priority:  construction,
wind erosion, paved roads, storage piles, agricultural tilling,
mining and quarrying (especially quarrying), and unpaved roads.

          Testing of construction emissions is needed because
     only two field studies of that source have been conducted
     and because of the expected relative contribution of that
     source, particularly  in urban areas.  Many of the PM10 Group
     I and II areas are urban areas where construction activities
     are likely to be found and to be contributors to PM10.

          Wind erosion is  high on the list because the empirical
     data on dust emissions due to wind erosion is very limited
     and because wind influences all of the other sources.
     Additionally, work on wind erosion may be required to
     determine how to evaluate wind erosion in urban areas.
     "Street canyon" effects for mobile source emissions have
     been recognized for some time.  Do these same "street
     canyon" effects influence wind erosion from exposed areas
     within cities because of a channeling effect?  Questions
     such as these need research and investigation.

          The other categories listed above were placed in the
     priority list according to the extent and quality of field

                               4-12

-------
     data currently available,  and the potential magnitude of the
     source in inventories..   The other source categories,
     landfills,  feedlots,  and unpaved parking lots,  are of lower
     priority regarding development.   However,  in certain areas
     and at certain inventory scales,  these sources  may be
     important and will require a significant development effort
     to determine adequate emission factors.

2) Standardized procedures should be developed for conducting and
documenting fugitive emissions field tests, resulting data
analysis, and emission factor development.

          There are currently no such standard procedures
     available for the measurement of fugitive dust  emissions.   A
     procedures manual would eliminate much of the often
     unnecessary variation in sampling equipment configuration,
     analytical procedures,  and quality assurance.  An emissions
     database generated from generally consistent field testing
     projects will have less variability and will permit more
     meaningful analysis than a database built on many dissimilar
     or non-comparable projects.

          The procedures manual should include guidelines for the
     following areas:

     •     equipment configuration, including the distance of
          downwind samplers from the source,  the number of
          sampling heads as a function of estimated  plume height
          (for profiling), and the number and type of ancillary
          samplers.

          method of correcting for background dust levels

     •     necessary testing conditions, particularly regarding
          meteorological parameters

     •     method of acquiring isokinetic exposures (for
          profiling)

     •     selection of the appropriate dispersion model (for
          modeling)

          In addition, or as an alternative, to the last item,
     investigators could be required to justify their choice of a
     dispersion model.

          Many of the remaining recommendations could also be
     incorporated into the procedures manual.

3) Emphasis should be placed on utilization of state-of-the-art,
well characterized sampling equipment.
                               4-13

-------
          A primary example of this is the dichotomous sampler.
     The great advantage of the dichotomous sampler concerning
     particle size distribution is that it operates as a virtual
     impactor.  As a consequence,  it does not suffer from the
     particle bounce problems associated with the high volume
     cascade impactors,  since there is no physical surface on
     which the particles can bounce.  There are a few drawbacks
     to the use of dichotomous samplers in determining emissions
     from fugitive dust  sources.  The main drawback is the low
     sampling rate associated with this sampler.  This is
     especially true for the fine particle (< 2.5 |J.m) flow, which
     is only 0.1 that of the total sampler flow.  As a
     consequence,  it can be difficult to obtain enough mass on
     the fine fraction filter to measure above the noise level
     associated with the balance.   An additional problem area is
     that the cut-point  for the inlet is wind speed sensitive.
     Thus the upper limit for the particle size cut-point varies
     with the wind speed.  However, this does not apply to the
     fine fraction, since typically the material in the < 2.5 (am
     fraction is not affected by the sampler cut-point.  Despite
     these limitations,  the dichotomous sampler represents a well
     classified, state-of-the-art particle sampler.

          Perhaps EPA should give some thought to development of
     a "trichotomous" sampler having two virtual impactor stages,
     one at the current  fine particle cut point and one at the 10
     Urn cut point.  The  upper limit could then be wind speed
     sensititve, and it  would not matter, since the internal flow
     dynamics would then handle the PM10 fraction.

          Optical particle sizing devices should also be
     carefully considered for characterizing the particle size
     distribution of the dust plume.  In some circumstances,  the
     relatively short run time needed to collect a sample permits
     multiple samples to be taken at several heights in the plume
     before the character of the plume changes substantially -

          The use of hot wire anemometers and automatically
     adjusting flow controllers should be encouraged, if not
     required, when conducting exposure profiling.  This will
     help to insure that samples are collected isokinetically.
     This may not be feasible with dichotomous samplers,  however.

4}  Statistical methods should always be employed in both the
experimental design and  the field data analysis.

          Field experiments should .be designed such that the
     results will generate relatively certain (e.g. 95%
     confidence) conclusions regarding the relationships between
     the monitored variables over the ranges experienced in the
     test.  Similarly, specific statistical criteria should be
     used in determining the utility of a particular variable in

                               4-14

-------
     explaining variation in emissions.   Statistics should at
     least play a key role in determining which variables are
     used'in a predictive equation.   Furthermore,  summary
     statistics that  reflect the amount  of variability in the
     measured emissions  which is explained by the  predictor
     variables should be reported with each revision of the
     equation.

5)  Insure the use and documentation  of quality assurance
practices in emissions measuring projects.

          Development of a Quality Assurance Project Plan (QAPP)
     for each environmental monitoring and measurement effort has
     been a requirement  of the Office of Research  and Development
     (ORD)  since May  30,  1979.  Interim  guidelines for
     preparation of QAPPs were published in 1980 (U.S.  EPA).
     These guidelines specify that the following items must be
     addressed in the QAPP:

     •     QA procedures  for measurement  data in terms of
          precision,  accuracy, completeness,  representativeness,
          and comparability

     •     sampling procedures

     •     sample custody

     •     calibration procedures and frequency

     •     analytical  procedures

     •     data reduction,  validation,  and reporting

     •     internal quality control checks and frequency

     •     performance and systems audits and frequency

     •     preventive  maintenance procedures and schedules

     •     specific routine procedures to be used to assess data
          precision,  accuracy, and completeness of specific
          measurement parameters involved

     •     corrective  action

     •     quality assurance reports  to management


          Clearly,  many  of the reviewed  field studies, including
     those done after 1980, have not addressed all of the above
     items.   If every future field measurement project conforms
     to  these guidelines,  the collected  data will  be of a much

                               4-15

-------
     greater quality and value.   For very large emissions testing
     projects,  it may be advisable to have an organization other
     than the one performing the sampling conduct the QA audit.

6)  Whenever exposure profiling is chosen as the method of choice,
require that dispersion modeling of the emission rate be
conducted also.

          In other words,  measurements taken for exposure
     profiling could also be utilized in dispersion modeling.
     Since exposure profilers also could be utilized to determine
     concentration, the dispersion model could be solved for the
     various heights of the profiler heads to determine the
     source strength of the source.  This would serve as a check
     on the results of exposure  profiling.  When the emission
     rates calculated by the two different methods are in
     significant disagreement,  the investigators would be alerted
     to possible problems in their equipment configuration.
     Additional data requirements would be minimal.

7)  Explore emissions modeling alternatives to the framework
described in the Workbook of Atmospheric Dispersion Estimates
(1970) .

          The models described and discussed by Turner are, with
     few exceptions, intended for use with continuously emitting
     sources or sources from which pollutant releases last longer
     than the travel time between the source and the receptor.
     Many of the fugitive dust sources discussed here are
     sometimes, if not always,  intermittent.  For example,
     depending on the traffic density, paved and unpaved roads
     are sometimes more aptly described as intermittent rather
     than continuous.  Models designed specifically to estimate
     pollutant releases which occur in short intervals have been
     developed in recent years.   These are typically referred to
     as puff models.  Adaptation of these models to line-type
     sources may be required.

8)  The availability of activity data and data on correction
parameters should be considered when developing predictive
emission factors.

          The issue of inventory scale is critical here.  As was
     pointed out in the evaluation of activity data above, many
     of the correction parameters required for use in the'
     emission factors available now require data that have not,
     are not, and probably will not ever be collected by a state
     regional,  or national agency or trade association.  The only
     way of obtaining much of this information is to perform a
     site visit and collect it or to send out surveys to
                              4-16

-------
     individual facilities to collect it.  If EPA still wants to
     pursue development of activity-specific emission factors, it
     should also consider one of the following options.

     •    Undertake a parallel emission factor development effort
          oriented towards both larger scale inventories in which
          activity-specific emission factors are still developed,
          and emission factors for the facility requiring
          activity data based upon routinely reported and
          published information.  An example would be surface
          coal mining.  Develop activity-specific emission
          factors that require VMT for hauling vehicles, but also
          develop an emission factor for total mining activity
          based on tons of coal produced for utilization in
          developing state, regional or national level emissions
          estimates.
     •    Perform an evaluation using robust statistical analyses
          to determine appropriate surrogate indicators that can
          be used as a basis for developing estimates of the
          activity-specific correction parameters currently
          utilized in emission factors.  For example,  determine
          if there is a significant statistical correlation
          between construction dollars spent and construction
          vehicle VMT upon which estimates of construction VMT
          can be made using reported information on construction
          dollars.

     •    Try to establish inter-agency agreements between EPA
          and other agencies such as DOT,  DOL,  etc.,  to make the
          necessary information a required piece of information
          in their standard reports.

9)  Encourage individuals,  organizations,  and government agencies
which have collected data on correction parameters or activity
levels to submit it to the Clearing House for Inventories and
Emission Factors (CHIEF).

          Having adequate emission factors and activity data
     developed is a major first step towards improvements in
     estimating emissions from fugitive dust sources.  However,
     if the information is unavailable or it takes a long time to
     distribute it, its value'is greatly diminished.  Having the
     data available on the CHIEF bulletin board would greatly
     facilitate the dissemination of the most recent information.

10)  Determine the relative importance of temporal activity data
to  PM10 emission inventory development.

          Temporal information concerning activity data for PM10
     fugitive dust sources is virtually non-existent, especially

                               4-17

-------
     at the hourly or daily level.  For a very limited number of
     activity data types,  monthly information is available, but
     typically, only annual data is available.  If hourly or
     daily information is of a high priority  (as it is for ozone
     inventories),  then a significant research and development
     effort will be required to develop the information.

     Finally, EPA must realize that institution of these
recommendations will not be inexpensive.  In many respects,
development of emission factors for these sources will be almost
like starting at the beginning.  The majority of the emissions
testing for these sources was performed from the late 1970s to
the mid 1980s.  No direct measurements of PM10 for these sources
have been made.  Although adequate funding will go a long way
towards improving and developing emission factors for fugitive
dust sources, without a well thought out program of research and
development, the results will not be adequate to meet the
requirements of fugitive PM10 emission inventory development
efforts.
                              4-18

-------
                            SECTION 5

                            REFERENCES
Alliance.  Improved Emission Inventory Methodology for  Other
Stationary Source Fuel Combustion Categories Draft Summary
Report.   EPA contract number 68-D9-0173, September 1990.

American Society of Testing and Materials.  "Standard Method  .for
Collection and Analysis of Dustfall."  Method D 1739-62.

Amick, R.S., K. Axetell,  D. Wells.  Fugitive Dust Emission
Inventory Techniques. 67th Annual AWMA Meeting, Denver, CO.   June
9-13, 1974.

Axetell, Kenneth, Jr. and Chatten Cowherd, Jr.  Improved Emission
Factors for Fugitive Dust from Western Surface Coal Mining
Sources.  EPA-600/7-84-048.  U.S. Environmental Protection
Agency,  Cincinnati, Ohio.  March  1984.

Axetell, Kenneth, Jr.  Survey of Fugitive Dust from Coal Mines.
EPA-908/1-78-003.  U.S. Environmental Protection Agency, Denver,
Colorado.  February 1978.

Axetell, Kenneth, and Joan Zell.  Control of Reentrained Dust
From Paved Streets.  EPA-907/9-77-007.  U.S. Environmental
Protection Agency.  July 1977.

Blackwood, T. R., P. K. Chalekode, and R. A. Wachter.   Source
Assessment:  Crushed Stone.  EPA-600/2-78-004L.  U.S.
Environmental Protection Agency,  Cincinnati, Ohio.  May 1978.

Bohn, Russel, Thomas Cuscino, Jr., and Chatten Cowherd, Jr.
Fugitive Emissions from Integrated Iron and Steel Plants.   EPA-
600/2-78-050, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. March  1978.

Chalekode, P. K., J. A. Peters, T. R. Blackwood, and  S. R.
Archer.   Emissions from the Crushed Granite Industry:   State  of
the Art.  EPA-600/2-78-021.  U.S. Environmental Protection
Agency,  Cincinnati, Ohio.  February  1978.

Connor,  Andrew D., Thomas E. McGuire, and Murray S.  Greenfield.
Wind Erosion Testing by Portable  Wind Tunnel at an  Iron and Steel
Plant.  79th Annual Meeting of the Air Pollution Control
Association, Minneapolis, Minnesota.  Paper No.  86-22.1.   June
22-27,  1986.
                                5-1

-------
Cook, Frank, Arlo Hendrikson, L. Daniel Maxim, and Paul  R.
Saunders.  Fugitive Dust From Western Surface Coal Mines.   EPA-
600/7-80-158.  U.S. Environmental Protection Agency,  Cincinnati,
Ohio.  August 1980.

Cowherd, Chatten, Jr., Russel Bohn, and Thomas Cuscino,  Jr.   Iron
and Steel Plant Open Source Fugitive Emission Evaluation.   EPA-
600/2-79-103, Environmental Protection Agency, Research  Triangle
Park, North Carolina.  May 1979.

Cowherd, C.C. Jr., C. Guenther.  Development of a Methodology and
Emission Inventory for Fugitive Dust for Regional Air Pollution
Study.  EPA-450/3-76-003.  Prepared for U.S. EPA, Office of Air
and Waste Management, Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina.  1976.

Cowherd, Chatten, Jr., Kenneth Axetell, Jr., Christine M.
Guenther, and George A. Jutze.  Development of Emission  Factors
for Fugitive Dust Sources.  EPA-450/3-74-037.  U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.  June
1974.

Cowherd, Chatten, Jr., Christine M. Maxwell, and Daniel  W.
Nelson.  Quantification of Dust Entrainment from Paved Roadways.
EPA-450/3-77-027.  U.S. Environmental Protection Agency.  July
1977.

Cowherd, Chatten, Jr. and Phillip J. Englehart.  Paved Road
Participate Emissions—Source Category Report.  EPA-600/7-84-077.
U.S. Environmental Protection Agency.  July 1984.

Cuscino, Thomas, Jr., Gregory E. Muleski, and Chatten
Cowherd, Jr.  Iron and Steel Plant Oven Source Fugitive  Emission
Control Evaluation.  EPA-600/2-83-110, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.
October  1983.

Cuscino, Thomas A., John S. Kinsey, Richard Hackney, Russel Bohn,
and R. Michael Roberts.  The Role of Agricultural Practices in
Fugitive Dust Emissions.  Contract report for the California  Air
Resources Board.  June 1981.

Cuscino, Thomas A., Robert Jennings Heinsohn, and Clotworthy
Birnie, Jr.  Fugitive Dust from Vehicles Travelling on Unoaved
Roads.  General Technical Report NE-25, ODC 907.3; 187:  111:  273-
425.1, USDA Forest Service.  1977.

Dyck, Rodney I. J. and James J- Stukel.  "Fugitive Dust  Emissions
from Trucks on Unpaved Roads."  Environmental Science and
Technology. Vol.  W. No. 10. October  1976. pp.1046-48.
                               5-2

-------
Elam, C. J., J. W. Algeo, T. Westing, and A. Martinez.
Bulletin D.  California Cattle Feeders Association.   Bakersfield,
California.  June 1972.

Elam, C. J., T. Westing, J. W. Algeo, and L. Hokit.   Measurement
and Control of Feedlot Particulate Matter.  California Cattle
Feeders Association.Bakersfield, California.  February  1971.

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.A., and R. Passi.  Modeling Dust Emission Caused by
Wind Erosion..  Journal of Geophysical Research, Vol. 93,  NO.
D11, pp. 14,233-14,242.  November, 1988.

Handy, R. L., J- M. Hoover, K. L. Bergeson, and Darwin E.  Fox.
"Unpaved Roads as Sources for Fugitive Dust."  Transportation
Research News.   1975.  pp. 6-9.

Heisler, S.L.  Interim Emissions Inventory for Regional Air
Quality Studies.  Electric Power Research Institute Report, EPRI
EA-6070.  November, 1988.

Jutze, George, and Kenneth Axetell.  Investigation of Fugitive
Dust  Volume I - Sources, Emissions, and Control.  EPA-450/3-74-
036-a, U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. June 1974.

Kinsey, John Scott, Phillip Englehart, and Alan L. Jirik.  Study
of Construction Related Dust Control.  Contract No. 32200-07976-
01.  Minnesota Pollution Control Agency, Roseville, Minnesota.
April 1983.

Lemon, Richard W., Richard W. Bloomingdale, and Karen J. Heidel.
Derivation of Suspended Particulate Emission Factor for Motor
Vehicle Use of Unpaved Roadways,  unpublished study by the Pima
County Air Quality Control District.  February 1975.

Marple, Virgil, Kenneth Rubow, and Orville Lantto.  Fugitive Dust
Study of an Open Pit Coal Mine.  Contract No. USDI/BOM J0295071.
U.S. Department of Interior, Bureau of Mines, Washington,  D.C.
September 1980.

McCaldin, Roy 0.  Fugitive Dust Study for Pima County Air Quality
Control District, Tucson, Arizona.  AQ-91A.  Air  Quality  Control
District of Pima County.  October  1977.

Midwest Research Institute.  Gap Filling PM,n Emission Factors
for Selected Open Area Dust Sources.  EPA-450/4-88-003.   U.S.
Environmental Protection Agency.  Research Triangle Park,  North
Carolina.  February 1988.

                               5-3

-------
Muleski, G. E., C. Cowherd, Jr. and T. Cuscino, Jr.   Current
Procedures for Open Source Particulate Emission Measures.
Midwest Research Institute.  December 1983.

Muleski, G., and D. Hecht.  PMTr Emission  Inventory of Landfills
in the Lake Calumet Area.  Contract No. 68-02-3891.   U.S.
Environmental Protection Agency.  September  1987.

National Climatic Data Center.  Local Climatological  Data.   1991.

Pasquill, F.  Meteorology and Atomic Energy  1968.  U.S. Atomic
Energy Commission.  1968.

Peters, J. A. and T. R. Blackwood.  Source Assessment:  Beef
Cattle Feedlots.  EPA-600/2-77-107.  U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.   June
1977.

Pinnick, R. G., G. Fernandez, B. D. Hinds, C. W. Bruce, R.  W.
Schaefer, and J. D. Pendleton.  "Dust Generated by Vehicular
Traffic on Unpaved Roadways: Sizes and Infrared Extinction
Characteristics."  Aerosol Science and Technology.  Vol. 4.  1985.
pp.  99-121.

Pyle, Bobby E. and Joseph D. McCain.  Critical Review of Oven
Source Particulate Emission Measurements.  Part II -  Field
Comparison.  EPA-600/2-86-072, U.S. Environmental  Protection
Agency, Research Triangle Park, North Carolina. August, 1986.

"Reference Method for the Determination of Suspended  Particulates
in the Atmosphere  (High Volume Method),"  Federal Register,  36,
28,  Appendix B, 22388-22390, 25 November  1971.

Reider, J. Patrick.  Size Specific Particulate Emission Factors
for  Uncontrolled Industrial and Rural Roads.  Draft Final Report.
EPA  Contract No. 68-02-3158.  January 1983.

Roberts, John  Warren.  The Measurement, Cost and Control of Air
Pollution From Unvaved Roads and Parking-  Lots in Seattle's
Duwamish Valley-  Thesis  for Master of Science in  Engineering,
University of  Washington.  1973.

Roberts, John W., Harry A. Watters, Carl  A.  Mangold,  and August
T. Rossano.  "Costs and Benefits of Road  Dust Control in
Seattle's Industrial Valley."  Journal of the Air  Pollution
Control Association.  Vol. 25.  No. 9~i  September  1975  nn  948-
52.

Sartor, J. D. and G. B. Boyd.  "Water Pollution Aspects of  Street
Surface Contaminants."  U.S. Environmental Protection Agency
Publication No. EPA-R2-72-081.  November  1972.
                                5-4

-------
Sehmel, G  A   "Particle Resuspension from an Asphalt Road  Caused
by Car and Truck Traffic."  Atmospheric Environment.  Vol.  7.
pp.291-309.  July  1973.

State of California Air Resource Board.  GARB Draft Inventory
Report.  Sacramento, CA.  1987.

State of Vermont,  Agency of Natural Resources, Department of
Environmental Conservation.  State of Vermont Implementation Plan
Revision.  Appendix A.  January 1979.

Struss, S.  R. .and W. J. Mikucki.  Fugitive Dust Emissions From
Construction Haul Roads.  Special Report N-17.  U.S. Army
Construction Engineering Research Laboratory.  February  1977.

Turner, D.  Bruce.   Workbook of Atmospheric Dispersion Estimates.
Publication Number AP-26.  U.S. Environmental Protection Agency.
Research Triangle Park, North Carolina.  Revised  1970.

U.S. DOE.   Coal Production (Annual), DOE/EIA-0118.  November,
1990.

U.S. DOC,  Bureau of Census.  Industrial Series Census of
Construction.  Table 10, Value of Construction Work for
Establishments with Payroll by Location of Construction Work.
1987.

U.S. DOT,  FHWA.  Highway Statistics, 1989 (Annual).  October,
1990.

U.S. Environmental Protection Agency.  "Reference Method for the
Determination of Suspended Particulates in the Atmosphere  (High
Volume Method)."  Federal Register.  36, 28.  Appendix B.  pp.
22388-22390.  November 1971.

U.S. Environmental Protection Agency.  Quality Assurance Handbook
for Air Pollution Measurement Systems.  Volume II:  Ambient Air
Specific Methods.   Research Triangle Park, North Carolina.
November 1976.

U.S. Environmental Protection Agency-  Cascade Impactor Data
Reduction with SR-52 and TI-59 Programmable Calculators.
Research Triangle Park, North Carolina.  Publication No.  EPA-
600/7-78-226.  November 1978.

U.S. Environmental Protection Agency.  Quality Assurance Handbook
for Air Pollution Measurement Systems Volume II: Ambient Air
Specific Methods.  EPA-600/4-77-027a.  May 1977.

U.S. Environmental Protection Agency.  Ambient Monitoring
Guidelines for Prevention of Significant Deterioration Research
Triangle Park, North Carolina.  1978.

                                5-5

-------
U.S. Environmental Protection Agency-  Interim Guidelines and
Specifications for Preparing Quality Assurance Project Plans.
QAMS-005/80.  Washington, D.C.  December 1980.

USDA Statistical Reporting Service, Agricultural Handbook Number
628, Usual Planting and Harvesting Dates for U.S. Field Crops,
April, 1984.

Vekris, S. L.  "Dispersion of Coal Particles from Storage Piles."
Hydro Research Quarterly.  Vol. 23.  No. 2. 1971.

Wedding, J. B., A. W. Weigand, and T. C. Carney.  "A 10 [am
Cutpoint Inlet for the Dichotomous Sampler."  Environmental
Science and Technology.  Vol. 16,  pp. 602-06.

Woodruff, N.P- and F. H. Siddoway-  "A Wind Erosion Equation."
Soil Science Society of America Proceedings.  Vol.  29.  No. 5,
September-October 1965.
                               5-6

-------
                           APPENDIX A

     NOTES AND INFORMATION CONCERNING ACTIVITY DATA  SOURCES


     The state of Virginia does not have a PM10 nonattainment
area and therefore has never  developed a PM10 fugitive dust
inventory.

     Region I  said that the region collects emission estimates
from the states with nonattainment areas.   For Region I these
states  are  Maine  and Connecticut.

     The PM10 nonattainment area in Maine is Presque Isle.  They
prepared a  1991 SIPs on sanding/salting of paved  roads.  They
collected site specific counts,  mileage and silt  loading values.
Some of their  traffic counts  were  obtained from Maine's DOT.

     In Region VII there are  not any nonattainment areas so no
information on PM10 fugitive dust is collected nor is a PM10  SIP
developed.   The states in region VII are Iowa,  Kansas,  Missouri
and Nebraska.

     Bob Judge of Region I stated  that  attainment states in his
region  do a point source inventory only and therefore do not
collect fugitive  dust activity  data.

     Region V  stated only nonattainment  areas collected fugitive
dust inventories  and they are done by  the  state or local
agencies.

     In 1988 MRI  performed a  PM10 study on Illinois's two
nonattainment  areas,  Chicago  and Granite city.  MRI sent
companies surveys for VMT, used monitoring data for background
emission.  They took aerial photo  to estimate sizes of stock
piles and parking lots.

     In Indiana they use AP-42  emission factors for paved and
unpaved roads, storage piles, open sources and wind erosion.  To
obtain  VMT  they sent questionnaires to facilities and asked"for
surface loading and if not provided they use the AP-42 defaults.
For agricultural  tilling they obtained county-level data from the
county  soil conservation services.  For area source paved and
unpaved roads, they obtained state and federal highway maps of
ADTV, and from county/metro roads they obtained traffic counts.
They do not do a  mining and quarrying PM10  emission estimate.
For construction  they used building permits which contain the
estimated acres and cost of the construction.
                               A-1

-------
     Region IX stated that the state and local air agencies did
the inventory work.  In Arizona the state performed the inventory
and in California and Nevada local agencies estimated the
emissions.

     In Arizona,  Engineering Science developed a base inventory
for modeling work.  They performed paved road sampling, size
distribution, and silt content.  The traffic data was locally
generated data from paved road traffic tapes which contain
capacity, number of lanes, and ADTV.  For unpaved roads the ADTV
was estimated.  For Yuma county,  the use of a traffic counter is
impossible since the counters become buried within minutes in
dust.  For determining activity levels for construction, storage
piles and sand/gravel they obtained county air pollution permit
files.  For agricultural tilling they took aerial photos to
estimate acreage and took silt measurements.

     Within Nevada the permits call for all fugitive dust to be
suppressed or a facility is in violation.  VMT is developed from
hydrographic basis or some actual Nevada DOT data.

     From U.S. EPA, National Air Data Branch (NADB)  it was
learned that to determine PM10 fugitive emissions in AMS the
truck VMT is obtained from the Truck Inventory and Use Survey and
county level registration is obtained from R.L.  Polk.   In their
files are county level paved road VMT for 1985 for the following
states:  California, Delaware, District of Columbia,  Georgia,
Illinois, Iowa, Kansas, Maryland, Massachusetts,  Michigan,
Minnesota, Mississippi, Nebraska, Nevada, New Mexico,  North
Dakota, South Dakota, Texas, Utah, West Virginia, and Wyoming.
AMS also states that at present,  no methodology to estimate
activity levels on a county basis nor emission factors are
available for miscellaneous wind erosion.  They also state that
at present, no methodology is available to estimate activity
level at the county level.

     Ohio uses their permits to obtain activity data,they
inventory paved and unpaved roads, mining, storage piles and wind
erosion.  They use AP-42 emission factors with the exception of
wind erosion where they use an older AP-42 emission factor
equation.  The original area source inventory was developed by
PEI for Ohio SIP during the late 1970's and is currently updated
by population.

     Region I stated that Connecticut adjusted street data to fit
monitoring data.  New Haven is the only nonattainment area and
the modeling group is presently investigating the problems.  In
1979 an attempt at a construction inventory was made.

     In 1982 a fugitive dust inventory for Massachusetts was
estimated.  Presently there is no methodology.


                               A-2

-------
     There are no nonattainment areas in New Hampshire, but  they
do inventory PM10 for all their S02 sources.

     There are also no nonattainment areas in Rhode Island,  but
they do some monitoring.

     In Michigan,  the Wayne County Pollution Agency sent out
surveys to collect VMT and storage pile data.  For an area source
paved road inventory,  the VMT projections were made by
Southeastern Michigan Council of Government (SEMCG).  The SEMCG
also periodically does traffic counts for Detroit.

     West Virginia has presently sent surveys to industries  for
the November 15,  1991  PM10 SIP.

     The Allegheny County,  Pennsylvania Air Agency is presently
sending surveys to industries.  Their last fugitive dust emission
report was prepared by TRC.

     The state of Vermont did a 1979 SIP inventory.  They sent a
copy of the Appendix A:   Non-Attainment Area Emission Inventory.

     New Jersey uses ambient monitoring data to derive PM10
emissions for the State.

     New York goes to a specific  area that has been identified as
having potential fugitive dust problems.   New York then evaluates
the potential emissions and then  proceeds with the permitting
process (a site specific  basis)  if necessary.   Examples of these
specific sources include  quarries,  sanding operations, and road
dust.  New York goes to DOT for actual highway information.  New
York monitors especially  dusty areas to determine the
significance of the problem.  No  estimation of fugitive PM10  for
construction/demolition,  wind erosion or agricultural tilling is
done.
                                A-3

-------