United States
  Environment*] Protection
  Agency
Office of Air Quality
Planning and Standards
Research Triangle Park. NC 27711
EPA-454/R-95-008
"October 1991
     DEVELOPMENT OF
       A PLAN FOR A
SURFACE COAL MINE STUDY
      and

-------
                                    EPA-454/R-95-008
       DEVELOPMENT OF
         A PLAN FOR A
SURFACE COAL MINE STUDY
                     U.S. Environmental Protection Agency
                     Region 5, Library (PL-12J)
                     77 West Jackson Boulevard, 12th Floor
                     Chicago, IL 60604-3590
         Emission Factor And Inventory Group
       Emissions, Monitoring, And Analysis Division
       Office Of Air Quality Planning And Standards
         U. S. Environmental Protection Agency
          Research Triangle Park, NC 27711

                October 1991

-------
This report has  been reviewed by the Office Of Air Quality Planning And Standards, U.  S.
Environmental Protection  Agency, and has been approved for publication.  Any mention of trade
names or commercial  products is not intended to constitute endorsement or recommendation for use.
                                   EPA-454/R-95-008

-------
                                   PREFACE
    This interim report was prepared by Midwest Research Institute under
    U.S. Environmental Protection Agency (EPA) Contract No. 68-DO-0137, Work
    Assignment No. 68.  The principal authors of this report are Dr. Greg Muleski of
    MRI and Mr. Clifford Cole and Mr. Steve Vardiman of TRC Environmental
    Consultants. Dr. Muleski was assisted by Dr. Chatten Cowherd and Ms. Karen
    Connery of MRI. Mr. Joe Touma and Mr. Dennis Shipman of the Office of Air
    Quality Planning and Standards served as EPA's technical monitors of the work.

    Approved:
    Charles F. Holt, Ph.D., Director
0   Engineering and Environmental
     Technology Department
    October 29, 1991
    MRI-OTS\R10-31 2nd
                                                                           III

-------
                                CONTENTS
 Preface  	  iii
 Executive Summary  	ES-1

       1.     Overview  	   1
                   1.1    Site selection	   2
                   1.2    Evaluation of emission factor/dispersion model
                         methodology	   2
                   1.3    Pit retention	   3
       2.     Test Mine  Site Selection Criteria  	   7
       3.     Emission Factor Verification and Dispersion Model Evaluation  .   9
                   3.1    Short-term testing for emission factors  	   9
                   3.2    Long-term monitoring for dispersion model
                         evaluation	  17
                   3.3    Data analysis and model evaluation   	  24
       4.     Pipit Retention  	  27
                   4.1    "Checkout" study	  27
                   4.2    Field pit retention measurements  	  28
                   4.3    Data analysis and model-building  	  29
 References	  31

 Appendix

       Synopsis of prior  coal mine dispersion model evaluation studies  .... A-1
MRI-OTS\R10-31 2nd

-------
                          EXECUTIVE SUMMARY
      At present, ambient particulate matter (PM10) impacts from surface coal
mining operations are assessed in the following ways:

1.    The mine's operating plan is reviewed to identify major PM10 emission
      sources, such as blasting, overburden removal, haul trucks, etc.  The plan
      is also reviewed to determine source activity rates—such as tons of coal
      mined or tons of overburden removed per year—over the effective life of
      the mine.

2.    An emission factor (mass emitted per unit source activity) is proposed for
      each major source.  Factors are usually selected from Section 8.24  of
      AP-42.

3.    Values from items 1 and 2  above are combined to estimate annual and
      worst-case-day PM10 emission rates from the mine.

4.    Dispersion models (such as the Industrial Source Complex [ISC] model)
      are then used to simulate the atmospheric transport of the estimated
      emissions. Resulting ambient air concentration estimates are then
      compared against National Ambient Air Quality Standards (NAAQSs) or
      Prevention of Significant Deterioration (PSD) increments.

      The study proposed here addresses  issues involving the modeling process
described above for surface coal mines. Specifically, the Clean Air Act
Amendments (CAAA) requires the Administrator to "analyze the accuracy of. . .
models and emission factors and  make revisions as may be necessary to
eliminate any significant overprediction." The purposes of this study are to:

•     Improve available  emission factors for surface coal mines.

•     Develop a comprehensive data base of source activity levels, on-site
      meteorological conditions, and air quality data.

•     Conduct a model evaluation study to assess how the current methodology
      predicts the ambient air quality impact from mines.
MRI-OTS\R10-31 2nd
                                                                     ES-1

-------
      In general terms, the study combines extensive long-term air quality and
meteorological monitoring with intensive short-term, source-directed testing. The
long-term data collection is necessary to answer the following questions:

•     Does the current methodology result in systematic overprediction of air
      concentrations?

•     If so, what is the degree of overprediction?

•     How well do ISC model results match measurements in time and space?

If the answers to these questions show that problems exist, a different approach
is necessary to find out how and why differences occur.

      No matter how sophisticated, long-term monitoring of ambient "far-field"
concentrations alone cannot answer questions such as the following:

•     What portion or portions of the methodology are most responsible for
      overprediction?

•     Can the identified portions be modified so that systematic overprediction is
      effectively removed?

•     What fraction of material  emitted at the bottom of the pit eventually
      escapes?

      To answer these types of questions, the long-term monitoring program
must be supplemented with the  intensive short-term monitoring and pit retention
tracer programs that incorporate more source-directed  measurements.
Quantitative examination of separate steps in the emission factor/dispersion
model methodology is necessary to answer the "how" and "why" questions.
 Furthermore, the paniculate tracer studies will provide a quantitative basis for
development of a pit retention algorithm.

      The proposed program represents an excellent opportunity for cooperative
 agreements between agency and coal companies/trade groups. A variety of
 participation levels are available.

 •     At a minimum, industry will be asked to provide  historical source activity
      data, such as production rates and maintenance records for mine
      equipment.  This information can be treated as "CBI" — confidential
      business information

 •     Mines may  also be asked to  supply historical ambient air quality
      monitoring and meteorological data
 ES-2

-------
       With agreement from state and regional offices, some mine air quality or
       meteorological equipment could be relocated or loaned out for use at
       another mine during the long-term monitoring program.

       For mines selected as test sites, higher levels of cooperation are possible,
       ranging from arranging for electrical power drops and for field laboratory
       space to providing technician help who, under direct supervision by the
       contractors, would assist in field team support.
MRI-OTS\R10-31 2nd                                                              ES"3

-------
                                SECTION 1

                                OVERVIEW
      This section presents an overview of major components of the proposed
study.  The field sampling program consists of extensive long-term air quality and
meteorological monitoring combined with intensive, short-term source testing. In
addition, pit retention will be studied under a separate set of experiments.  It
should be noted that substantial cost savings are expected if sets of experiments
are conducted and coordinated at the same mines.

      To avoid confusion about what is meant by terms such as "ambient air"
and "near-source," the following defines terms that are used throughout this
document.

      "Source tests" refer to air quality/meteorological measurements made in
the immediate vicinity of an individual emission source (such as  a road) with the
intention of characterizing that source.  Most of these measurements are
required for "exposure profiling," which relies on simultaneous multipoint
concentration and wind speed measurements over the effective cross-section of
the dust plume to determine source strength.  As such, these measurements are
those necessary to develop the improved emission factors.

      "Near-field air monitoring" refers to measurements made further away than
source tests, but still within the vicinity of an individual source. A major
difference is that near-field measurements do not normally span  the plume cross
section.  Rather, these measurements  will generally be made at  a height of 1 to
2 meters and are  usually made for comparison with ISC or other model-
generated concentration estimates.  Unless otherwise specified,  near-field
measurements will be made within the pit.

      "Far-field air monitoring" will refer to measurements made at  a
considerable downwind distance from the nearest emission source.  As such,
these measurements focus on characterizing the air quality impact of the mine as
a whole rather than on any individual source or source category.  Unless
specifically stated otherwise, far-field measurements will be made beyond  the rim
pit (i.e., at surrounding grade).  In this document, "ambient" and  "far-field"  are
equivalent terms.  These are the measurements that will be compared against
dispersion model predictions to evaluate the current methodology.
MRI-OTSW10-31 2nd

-------
1.1    SITE SELECTION

      An initial step in the proposed study is the identification of candidate test
sites for the field work.  Selection of mine sites at which testing will be conducted
will begin with identification of the minimum site criteria needed to  accomplish
testing objectives.  Candidate sites  will be evaluated with  respect to surrounding
terrain, availability of power, ease of access to mine activities and  to the mine pit,
absence of outside interferences (particularly from adjacent mines), and
willingness of mine owners to participate in the study.  Site selection is discussed
in greater detail in Section 2  of this plan for a surface coal mine study.
1.2   EVALUATION OF EMISSION FACTOR/DISPERSION MODEL
      METHODOLOGY

      This portion of the study provides the basis for an objective evaluation of
the emission factor/dispersion model methodology as currently applied. This
effort will involve a relatively long-term (approximately 4 months) monitoring of air
quality, meteorology and mining activity program. Data will be used to compute
fugitive dust emission rates, and the dispersion/transport of emissions  will be
simulated with the ISC model.  Modeled and measured concentrations will be
compared.  Model evaluations can be highly sensitive to the kind of statistical
comparisons that are used  to judge model performance.  Consequently,
considerable effort will be devoted to defining exactly how the comparison will be
made, and what pairwise and ensemble statistics will be used.

      The long-term data collection is necessary to answer the following
questions:

1.    Does the current emission factor/dispersion  model (EF/DM) methodology
      result in systematic overprediction of air concentrations?

2.    If so, what is the degree  of overprediction?

3.    How wel! do ISC model results match measurements in time and  space?

      Various time averaging periods are possible for PM10 concentrations.  In
the regulatory sense, 24-h  concentrations are most important because the
 NAAQSs and PSD increments are based on that time period.  On the other
 hand, currently available dispersion models typically use hourly meteorological
 data to predict hourly concentrations.  The hourly predictions  are then combined
to  provide longer-period averages.
                                                                 MR CT

-------
      Until recently, ambient PM10 concentrations could only be sampled by
reference methods for periods on the order of 24 h. Within the past year,
several equivalent PM10 measurement methods have become available, providing
far greater time resolution than the older high-volume sampling methods.
Table 1  summarizes the advantages and disadvantages of various field sampling
equipment that could be used.

      Questions such as (1) to (3) above ask only if problems exist; a different
approach is necessary to find out why differences occur.  No matter how
sophisticated the deployment on time resolution, long-term monitoring of "far-
field" concentrations cannot answer questions such as the following:

4.    What portion or portions of the methodology are most  responsible for
      overprediction?

5.    Can  identified portions be modified so that systematic  overprediction is
      effectively removed from the methodology?

      To answer these types of questions, the long-term  monitoring program is
supplemented with short-term monitoring programs. During these periods, the
contractor team will make source-directed measurements. Quantitative
examination of separate steps in the EF/DM methodology is necessary to answer
questions such as (4) and (5).

      Thus, the model evaluation portion of the proposed study combines long-
term monitoring of air quality, meteorology,  and source activity together with
short-term field measurement of emission factors.   The combination permits each
step in the current methodology to be independently evaluated.

      Specific details on methodology evaluation portion  of the study are
presented in Section 3.
1.3   PIT RETENTION

      The previous portion of the proposed field study would evaluate the
emission factor/dispersion model methodology as it is currently used.  A separate
set of field experiments has been proposed to develop the data base needed to
modify the existing methodology to better account for the retention of paniculate
matter within the pit at a surface coal mine.

      Pit retention (or "pit trapping") is the tendency for paniculate matter to
remain inside a mine pit, rather than dispersing downwind and impacting ambient
receptors. Neither the ISC model nor the other EPA dispersion models can
account for the retention of paniculate matter in the mine pit—instead, ISC acts
MRI-OTSXR10-31 2nd

-------
                                         Tabfe 1.  PM10 SAMPLING OPTIONS
         Type
  Representative
     samplers
   Time
 averaging
   period
           Advantages
      Disadvantages
     High volume
Wedding, Anderson
6 to 24 h
EPA Reference Method for PM
                                                                                10
                                                    Averaging period comparable to

                                                    Can operate on portable generator
                                                    power
Requires AC power

Cannot provide fine time
resolution of concentrations
    Continuous
Beta gauge, TEOM
(tapered element
oscillating
microbalance)
Continuous
Provides very fine time resolution
of concentration
Requires "clean" AC power,
and does not run well on
portable generators

Generally requires
temperature-controlled
enclosure for reliable
operation

Most expensive option
     Saturation
«
6
"PRO-2"
6 to 24 h
Battery powered

Least expensive option

Relatively rugged and easily
deployed/moved
Not an equivalent method

Cannot provide fine time
resolution of concentration

-------
as if the emissions occur at grade and disperse downwind, that is, as if the pit
exerts no  influence.  Failure to include pit retention may introduce very large
overpredictions.  The ISC model predicts maximum paniculate concentrations
during low wind speeds under stable atmospheric conditions; however, it is
during these same low wind speed stable conditions that pit retention is most
pronounced.

       The effects of pit retention will be evaluated by releasing tracer agents
(preferably paniculate tracers, although some gaseous tracers  may be used) at
the bottom of a mine pit and measuring their concentrations at the top of the pit.
      The dissemination and detection of paniculate tracer materials present
considerable technical challenges. Furthermore, a variety of potential tracer
materials are available.  In general, durable candidate materials that are easily
detected (such as encapsulated dyes) tend to be expensive.  Less expensive
materials, on the other hand, are  usually harder to detect in low concentrations.
Finally, the "weatherability" of certain materials can extend their applicability in
the field program.

      For these reasons, candidate tracers and dissemination techniques will be
evaluated during a "checkout" study. This program, to be conducted well in
advance of the SCM field tests, will provide practical guidance on how the field
tracer studies should be conducted.

      To the extent practical, field pit retention tests will be coordinated with the
source tests.  This approach provides a means to simultaneously and indepen-
dently evaluate each step in the emission factor/dispersion model methodology.
In so doing, the sources of overprediction may be identified with  confidence.

      Section 4 of this report discusses the "checkout" and surface coal mine pit
retention studies in greater detail.
MRI-OTSNR10-31 2na

-------
                                SECTION 2

                  TEST MINE SITE SELECTION CRITERIA
      Mine sites will be selected by comparing each candidate mine's attributes
against  predetermined criteria of desired location, available data, site
configuration, etc.  It is anticipated that at least one Wyoming surface mine will
be selected.  The criteria given below represent guidelines rather than "hard and
fast" rules. The process of site selection requires many qualitative and
subjective decisions.  Candidate mines will be judged on the following criteria:

•     General Location. Although mines with existing PSD monitoring networks
      provide ready candidates, most PSD networks are not sited to provide the
      data needed in this study.  In addition, mines that are relatively isolated
      would be preferred because interference from neighboring mines may
      influence results.  While it may not be possible to find completely isolated
      mines, those located away  from neighboring mines would be preferred
      over those in proximity to other mines. Finally, the site should be isolated
      from terrain features influencing local airflows. That is, the terrain
      immediately surrounding the mine should be fairly flat.

•     Source/Location Orientation.  Near-source measurements of traffic source
      generally require  long stretches of roads that are approximately
      perpendicular to the prevailing wind.  In  addition, sources should be
      relatively isolated from other important PM10  emission sources within the
      mine.

      The following is a list of example criteria to be applied to a site for
roadway source testing.

1.     There should be at least  10 meters of flat, open terrain downwind of road.

2.     There should be at least 30 meters of flat, open terrain upwind of road.

3.     The height of nearest downwind obstruction  should be less than the
      distance from the road to the obstruction.

4.     The height of nearest upwind obstruction should be less than one-third the
      distance from the road to the obstruction.
MRl-OTS\R10-31 2nd

-------
5.    A line drawn perpendicular to the road orientation should form an angle of
      0° to 45° with the mean daytime prevailing wind direction.

6.    The mean daytime wind speed should be greater than 4 mph.

7.    The test road should have an adequate number of vehicle  passes per
      hour enabling completion  of a test in no more than 3 h.

8.    The traffic mix should be representative of the type of vehicles that
      regularly use the road.

Analogous criteria  have been established for nonroad sources, such as material
handling operations. These include isolation from other important sources and
wind obstructions,  accessibility to the emission source, and orientation  of the
source to  prevailing winds.

•     Mine Activity.  Quantification of the mining activity (traffic volumes, amount
      of coal removed, amount of overburden disturbed, etc.) is needed  for the
      mine model validation effort. Candidate mines will be judged on the
      availability of historical data and on the ease with which these data could
      be collected during any field study.

•     Logistics.  There is a need to locate instruments  near dust-producing
      activities, at locations up and downwind of the mine, and at the  top of the
      mine pit highwall. The candidate  mines will be evaluated on the basis of
      being able to accommodate the wide variety of sampling locations
      necessary.

•     Representativeness.  Preference will be given to  mines that are  neither at
      the start nor near the end  of operation according to the mine plan.
      Furthermore, geographic diversity of the test sites is a desired goal.

      It is anticipated that EPA will initially contact the owners or operators of
candidate mine sites.  Once contacted, the contractor team can make written
requests of information that relate to location, mine activity, power, and logistics.
Based upon a qualitative assessment  of each of these criteria and on each
mine's willingness  to participate in a study of this sort, the candidates will be
reduced to a number of sites to be visited by MRI/TRC personnel. The final
selection used will be based on reexamination, using the same general criteria.
8                                                                MCi'OTS a ic 3-

-------
                               SECTION 3

EMISSION FACTOR VERIFICATION AND DISPERSION MODEL EVALUATION
      Evaluation of the emission factor/dispersion model methodology requires
two types of field sampling.  The first examines, over a period of 4 months,
whether the current overall methodology results in systematic overprediction of
the overall air quality impact of surface coal mines. This examination involves
gathering the data needed to compare ISC-generated concentrations with
measured concentrations to determine how well the current methodology allows
one to predict concentrations near a surface coal mine.

      The second type of sampling is directed to answering why overprediction
occurs and  how well each step in the EF/DM methodology predicts measured
values.  This portion of model evaluation involves short-term measurements both
in terms of individual tests (on the order of 1 to 3 h) and in time spent on-site
(estimated as two or three 2- to 3-week periods distributed throughout the
4-month long-term monitoring period).  See Table 2.
3.1    SHORT-TERM TESTING FOR EMISSION FACTORS

      A review of emission factors applicable to surface coal mines has been
recently completed (MRI, 1991).  This review also presented recommendations
for future source testing, which are summarized in Tables 3 and 4. The
proposed source testing program focuses on the following major sources:

•     Coal and overburden
•     Haul roads
•     General traffic (light- and medium-duty)
•     Overburden material handling operations
•     Coal material handling operations

These sources typically account for 70% or more of paniculate emissions at
surface coal mines (Cole et al.,  1985).

      The source-directed field sampling will employ the "exposure profiling"
concept to quantify source emission contributions and near-field concentrations.
MRI-OTS\R10-31 2nd

-------
Table 2. PROPOSED SCHEDULE

Activity (Relevant Section)
1 . Program Planning
General Testing Protocol
!
!
l
Paniculate Tracer "Check-out" Study (4.1)
- Identification of candidate tracers
- Field Exercise
- Data Reduction and Recommendation of
"model-based" or "mass-balance"
tracer methodologies
2. Surface Mine Site Selection/Equipment
Mobilization/Supply Acquisition (2.0)
3. Short-term Testing for Emission Factors (3.1)
Travel-related (line) sources \b
Western Mine #1
- Uncontrolled Emissions
Source testing
Sample analysis
Data reduction
- Controlled Emissions
Source testing
Sample analysis
Data reduction
Western Mine #2
- Uncontrolled Emissions
Source testing
Sample analysis
Estimated
One*
($1000)
\a
135

70

45\c
105\c
55
1QQ1
O N D
* * * *
* *






J F M
* *






A M J
* * *
* * * *


* * *
• * * *
* *
*
1QQO
J A S





* * * *
* * * ft
* * *
* * * *
*****

O N D





* *

J F M






1Q
A M J






QO
J A S







O N







-------
Table 2 (continued)

Activity (Relevant Section)

Sample analysis
Data reduction
Mine #3 *
- Uncontrolled Emissions
Source testing
Sample analysis
Data reduction/
- Controlled Emissions
Source testing
Sample analysis
Data reduction
Material handling sources \b
Western Mine #1
Source testing
Sample analysis
Data reduction
Western Mine #2
Source testing
Sample analysis
Data reduction
Mine #3 *
Source testing
Sample analysis
Data reduction
4. Source Activity Monitoring (3.3)
Travel-related (line) sources
western Mine w i
western Mine #2
Mine ff3
Material handlinq sources
Estimated
Cost
($1000)



55\c,d



120\c,d




70\c



80\c



80\c,d



e
20


20

- 1991
O N D































J F M































A M J













ft * ft
* * *
* * *

*












1 QQO 	
J A S















*

* *
* * * *
* * * *











O N D
* * * * *
• # * * *





























J F M




«


























	 iy
A M J




* *
* * *
* * *
* * * •












* *
*








QO
yj 	
J A S







* ft













ft *
* ft * * *








O N






























-------
Table 2 (continued)

Activity (Relevant Section)
Western Mine #2
Mine #3 * *
Misc. sources (blasting, road
grading, etc.)
Western Mine #1
Western Mine #2
i Mine #3**
5. Long-term Monitoring for
Dispersion Model Evaluation
Acquisition system design
Western Mine #1
- Equipment deployment/shakedown period
- Long-term (4 month)
"far-field" monitoring (3.2)
- Paniculate tracer study of
Pit Retention (4.2)
Second Year Mine * *
- Equipment deployment/shakedown period
- Long-term (4 month)
"lar-field" monitoring (3.2)
Construction of data base

6. Model Evaluation Protocol (3.1)

7. Emission Factor Review/Development
Estimated
Oriet
OOSl
($1000)

20



25
90
\e
\f
95\d
\e
25

30

30
1QQ1
O N D














J F M














A M J
* *

* *


* * *







1OQ9
J A S
* * » *

* * * *



* * *







O N D
* *

* *











J F M






*






1Q
A M J






* *






QO
J A S














O N














-------
                                                                    Table 2 (continued)
JJ
9


9. Program Management and Reporting
Interim Report
Draft Final Report
Final Report
Estimated
Poet
($1000)
75



— 1QQ1
O N D





J F M





A M J




1QQ9
J A S





O N D





J F M

*


1Q(
A M J




3*J
J A S





O N


*
*
     Notes:
      *  Depending upon the results of the first testing season (i.e., spring-fall 1992), the third mine used for emission factor development/verification may be either
        western or eastern.

     * * The long-term air quality and source activity monitoring program during the second testing season (i.e., 1993) will be at a western mine (which may be the
        same as the second mine used for emission factor work during the 1992 testing season).

     \a  Cost of preparing protocol included in current OTS work assignments.

     \b  Emission sources to be considered are summarized in Table 3.

     \c  Estimated cost assumes that emission factor experiments are interspersed with long-term monitoring program.

     \d Cost dependent upon location of test mines during the second test season.  Estimate given assumes that long-term and short-term monitoring are
        interspersed at the same location. See notes * and * * above.

     \e  Source activity levels will be recorded during all periods that field crews are present at mines. Consequently, the estimated cost assumes certain savings
        because this activity is concurrent with others. Additional source activity data may be collected by survey of mining companies and trade groups.

-*   \e  Cost estimates assume that sampling Option A is selected (i.e., two or three primary monitoring locations combined with 20 to 30 saturation samplers.
        Should Option B (exclusive use of continuous samplers) be selected, it is estimated that additional cost will be approximately $500,000 for the first
        test season and $250,000 for the second test season.
                                               s from "check-out* study.  Preliminary cost estimate is about $325,000, assuming that the pit retention study

-------
       Table 3.  EMISSION SOURCES RECOMMENDED FOR EVALUATION



Lines sources (travel-related) to be considered6
1 , Haul trucks
2. General mine traffic
3. Haul trucks (overburden)
"Grouped" approach to material handling emissions0
1 . Overburden removal and placement in trucks"
2. Overburden replacement by draglined
3. Misc. overburden handling operations^
4. Misc. coal handling operations6
Recommended
number of
tests per mine3

3-6U, 6-1 2C
3-6U, 6-12C
3-6U, 6-1 2C

3-6
3-6
3-6
3-6
Overall
priority


1
3
2

4T
4T
4T
81
 Notes:
       U = uncontrolled, C = controlled.  Uncontrolled if not indicated.
       Depending on the road network at the mine, it may not be possible to
       separate various travel-related emission.  In that case, because of the
       importance of travel-related emissions, additional toads or other line sources
       (such as scraper travel [see note f below] or road grading) will be tested.
       These tests are expected to be of uncontrolled emissions.
       Emission factor development will  group operations with like materials
       regardless of the equipment involved. This approach is expected to be
       particularly beneficial for overburden-related emissions from draglines, power
       shovels, truck dumps,  etc. It is further expected that test results will be
       combined with the generic materials  handling emission factor data base.
       Dependent upon the type of mine operation (i.e., dragline or shovel/truck;.
       Including truck dumps, loadout for transit, etc  Again, emission factor
       development will group across operations with the same material.
       Scrapers in travel mode rated as 7th overall priority
14

-------
i
                                 Table 4. SUMMARY OF THE PROPOSED FIELD STUDY

Testing phase
Tracer "Check-out" Study
(Section 4.1)
Long-term monitoring for
dispersion model
evaluation (Sections 3.2
and 3.3)
Short-term testing for
emission factors
(Section 3.1)
Pit retention
(Section 4.2)

Objectives of testing phase
1 . Identify suitable materials and
release techniques
2. Compare model-based and
mass-balance approaches
1 . Collect source activity,
ambient air and
meteorological data
2. Compare "whole-mine" impact
against ISC model
1 . Focus on individual sources
rather than whole mine impact
2. Determine how well individual
"steps" in EF/DM approach
match with measurements,
(e.g., source terms,
deposition, dispersion, etc.)
1 . Measurement method
depends on result of checkout
study
Sampler Deployment
Far-
field
X
X

Xa
Source
X

X

Near-
field


X


Testing duration
30 min to 2 h
Mostly 24 h with
some continuous or
- 1 h running
average samples
30 min to 3 h
30 min to 3 h

Testing
time frame
Spring 92
Spring-Summer 92
Spring-Summer 93
Spring-Summer 92
Spring-Summer 93
Summer 92
* Far-field measurements directed toward source characterization in a mass-balance approach.
en

-------
This established and widely accepted test methodology for fugitive emission
sources is briefly discussed below.

      The "exposure profiling" technique for source testing of open paniculate
matter sources is based on the isokinetic profiling concept that is used in
conventional (stack) testing.  The passage of airborne pollutant immediately
downwind of the source is measured directly by means of simultaneous
multipoint sampling over the  cross section of the open dust source plume.  This
technique uses a mass flux measurement scheme similar to EPA Method 5 stack
testing rather than requiring indirect emission rate calculation through the
application of generalized atmospheric dispersion model.

      For measurement of particulate  emissions from roads, a vertical network
of samplers is positioned just downwind and upwind from the edge of the road.
The downwind distance of 5  m is far enough that interference with sampling due
to traffic-generated turbulence  is minimal but close enough to the source that the
vertical plume extent can be adequately characterized with a maximum sampling
height of 5 to 7 m.  In a similar manner, the 10-m distance upwind from the
road's edge is far enough from the source that (a) source turbulence does  not
affect sampling, and  (b) a brief reversal would not substantially impact the
upwind samplers. The 10-m distance is, however, close enough to the road to
provide the representative background concentration values needed to determine
the net (i.e., due to the source) mass flux.

      In addition to the source testing measurements described above, this
portion of the field work will also consider near-field  concentrations and
deposition.  That is, during source testing periods, additional air samplers and
dustfall samplers will be deployed at several distance downwind of the source.
To the extent practical, the deposition portion of the proposed study will be
designed to augment on-going EPA studies on the subject.

      Although the general testing methodology for open dust sources is
reasonably well defined, the  proposed  study will use a "flexible" test schedule
described below to best make  use of available resources.

      The emission factor study will emphasize source testing at mines in  the
western United States. During the first testing season (spring/summer 1992), the
major sources at two mines in  the West will be characterized in terms of
emission factors. During the next testing season  (spring/summer 1993), a  third
mine will be characterized.  This mine  may be either an eastern or a western
mine. The final decision of the location of "Mine 3" for source testing will be
made in conjunction with the EPA technical monitor during the winter of 1992-
1993. Two example factors  contributing to this decision will be:

-------
 •      How well did the new emission measurements match with existing
       emission factors?  That is, were some of the current emission factors
       "validated," or did the need for further study become apparent?

 •      How well did the first testing  season perform in meeting the program's
       goals of geographic, seasonal, and meteorological diversity? Were data
       collected under a fairly broad range of operating practices?

 (Note: Should an eastern mine be  selected for emission factor validation studies
 during the second testing season, long-term monitoring for dispersion model
 evaluation during the second year will still be conducted at a western mine.)

       In addition to questions involving the third test  site location,  the proposed
 "flexible" source field testing approach  recognizes that no "hard and fast"
 decisions on the number of tests  or other particulars of the program are possible
 at this time.  For example, if it is found that currently available emission factors
 appear adequate for Western mines, then the second test season's resources
 can be redirected to a validation of  the factors  at eastern surface coal mines
 (SCMs).  On the other hand, the first year's testing results may indicate that the
 second year's resources would be best directed to a Western mine.  This might
 be true if a thorough reexamination  of western  SCM emission factors is found
 necessary, because the  second test season provides the third test site
 recommended in AP-42 emission factor quality rating schemes.
3.2   LONG-TERM MONITORING FOR DISPERSION MODEL EVALUATION

      In contrast to emission source testing, there is no well-established protocol
for assessing the ambient air quality impact from fugitive sources.  For this
reason, a considerable effort will be needed to establish and implement  an
appropriate protocol.

      Comparison of measured and modeled concentrations has been a subject
of interest to modelers for several years.  What is clear in all of the model
evaluation studies is that the means by which the models are evaluated  is
crucial.  How does one quantitatively judge the performance of an  air quality
model? For isolated stack studies, the method endorsed by the EPA compares
the maximum measured and modeled concentrations without regard to time or
space to allow comparison of one model against  another (Cox, 1988). For
surface coal mines, however, the location of modeled maxima is crucial  because
concentrations decrease dramatically with distance from the mine pit.  The
choice of model receptor locations and measured concentration locations
strongly influence whether a  model appears to 'work* well for surface coal mines
or not.
MRI-OTSVmO-31 2nd                                                              17

-------
      The issues of model performance and model "verification" have been
studied extensively for isolated stack sources but have received little attention for
fugitive dust sources.  (The few prior model evaluations at surface coal mines
are discussed in the Appendix.) There is virtually no uniformity between the
evaluation approaches used previously. The previous studies have all had to
"make do" with the data that were available, and little thought was given to
"designing" a study for the purposes of evaluating model performance.

      A review of these previous studies, and an understanding of how the
Gaussian models function, suggests that the following are desirable features in
designing a mine dispersion model evaluation study:

•     Time Resolution. The time period over which ambient measurements are
      made, over which emission rates are computed, and over which source
      locations are idealized should be as short as practicable.  The
      performance of an air quality model should improve when the input data
      have better ("finer") resolution.  Nevertheless, any analysis of a model's
      performance in predicting PM10 concentrations must consider 24-h time
      intervals because many mines have more difficulty demonstrating
      compliance with 24-h rather than annual increments or standards.

•     Evaluation Protocol. It  is clear that the means by which models are
      evaluated is crucial. The approaches and statistics used to judge a
      model's performance are as important (and, possibly, more  important)
      than the input data themselves.  The Cox approach (Cox, 1988) is widely
      used to compare performance of one  model to another for isolated stack
      sources.  Use of this same Cox protocol (with a variation to accommodate
      fugitive dust sources and modeled versus measured concentrations) is
      desirable.

•     Random Error.  Quantification of fugitive dust emission rates, use of on-
      site meteorological  data, and model simulations are all subject to random
      error. The error in any one measurement propagates throughout the
      process and can adversely affect the comparison of modeled
      concentrations (Cole et al., 1985). To the extent possible, these random
      errors must be minimized by employing  as many samples and as large a
      data base as possible.  In addition, the use of the robust-highest-
      concentration (RHC) statistic inherent in the Cox approach helps to
      overcome random error.

      Unlike the studies described in the Appendix, the dispersion model
evaluation study proposed here is being designed "from the ground up" and can
incorporate all of these features.  In updating meteorological data every hour and
by tracking source activity and emissions, this model evaluation will achieve far
better time resolution than any previous fugitive emissions mode! evaluation
18

-------
 study. By using a variant of the Cox evaluation approach, this study will conform
 to accepted statistical practice.  Four months of sampling at each of two separate
 mines overcomes random errors to yield statistically significant results.

       In brief, the emission factor and dispersion model methodology will be
 evaluated as follows.  A meteorological station will be installed at the surface
 coal mine, and a monitoring  network will be sited to measure maximum
 concentrations due both to individual emission sources and to combined sources
 at the mine.  For a period of approximately 4 months, PM10 will be measured at
 each of the monitors in the sampling grid.  For the purpose of discussion, there
 are two basic sampling options available.

       "Option A" seeks to combine the best features of the available PM10
 sampling methods by combining two or three continuous monitors with 6 to 10
 high-volume samples and 20 to 30 "saturation" samplers.  In this way,  a high
 degree of temporal resolution is accomplished at a few locations within a
 generally well-monitored geographic area.

       Under "Option B," a network of continuous PM10 monitors would be
 established, with the concentration at each monitoring location averaged on a 1-h
 basis. Although, these monitors provide the finest time resolution possible, they
 typically require temperature-controlled enclosures and "clean" electric power.
 The  enclosure and power requirements accentuate the high purchase cost of
 continuous samplers.  In summary, continuous monitors provide a very high
 degree of temporal resolution but can provide adequate spatial coverage only at
 a very high level of cost.

       Throughout the sampling program, field crews will record (a) activity
 parameters (number of blasts, amount of coal and overburden removed, vehicle
 miles travelled by mobile equipment, etc.); (b) independent variables related to
 emission potential (silt, moisture, etc.); and (c) the exact location of emission
 sources.

       The "dynamic" emission inventory maintained over the 4-month  period will
 serve as input to the ISC model.  Activity parameters and independent variables
 will be used in conjunction with available emission factors to compute hourly
 average paniculate emission rates.

       A comparison of measured  and modeled PM10 concentrations, made in
 accordance with a well-defined evaluation protocol, will determine how well the
 emission factor (EF) and dispersion model (DM) combination performs. The
following subsections describe the various components of the model evaluation
 study in greater detail.
MRI-OTS\R10-31 2nd                                                              ^ Q

-------
3.2.1  Model Evaluation Protocol

      As explained previously, it is anticipated that the Cox approach will be
used as a basis for evaluating model performance.  In the Cox approach, the
individual fractional bias and the absolute fractional bias will be used to
determine whether the ISC model and the emission factors overpredict measured
concentrations. Bias is a measure of the overall tendency for the models to over
or underpredict the measured values.

      Whatever the final sampler selection under Options A and B, a key factor
in the dispersion model evaluation process must be the so-called "operational"
component.  That is, fractional bias statistics will be computed for 24-h averaging
periods because that is the time basis for the applicable standards and
increments.  The operational component thus addresses regulatory concerns.

      The "scientific" component, on the other hand, is concerned with a
dispersion model's ability to perform accurately throughout a range of
meteorological conditions and throughout the geographic area surrounding the
emission source.  Clearly, the degree to which the scientific component may be
assessed depends upon the final selection of Option A or B.  Both options afford
the opportunity to compute approximately 1-h fractional bias statistics.

      Thus, with respect to the model evaluation procedures, the principal
difference between Options A and B appears to  be the relative importance paid
to assessing the scientific component in addition to the operational component.

       In virtually all other respects, the rest of the model evaluation protocol will
follow the Cox approach. Background PM10 concentrations determined from
upwind samplers will be subtracted from downwind measured concentrations so
that the analysis looks only at mine paniculate contributions.  Measured and
modeled PM10 concentrations will be used to determine the robust highest
concentrations (RHC), thereby eliminating a great deal of random error.
Fractional bias will be used as a basis for comparison because of its attractive
attributes (symmetrical,  bounded, and dimensionless), and because increasingly
it is becoming the  standard for comparison for model evaluations.  Each of the
above is analogous to the Cox approach for single stack sources.

       It is expected that the model evaluation protocol will parallel the Cox
approach in that concentrations will be  compared unpaired in time and space.
However, "unpaired in space" takes on a slightly different meaning for surface
mines than for isolated stack sources.  At a surface mine, the measured and
modeled concentrations decrease drastically with distance from the mine
(PEDCo 1982) so  that the robust highest concentration (RHC) will almost always
be at the location  closest to the mine in the predominant downwind direction.
 20
                                                                 Mli-OTS R<'-3* 2-.?

-------
3.2.2 Recommended Field Sampling Approach

3.2.2.1     Sampling Equipment

      The dispersion model evaluation portion of the proposed study requires
monitored air quality values. Earlier, two sampling options were presented:

      Option A combines a few continuous samplers with numerous so-called
"saturation" samplers.  The latter samplers require on the order of 6 to 24 h to
collect a sample on a filter. Thus, Option  A uses a network of samplers to
directly measure the 24-h concentration of regulatory interest and to provide
PM10 concentrations referenced to other averaging periods ranging from 1 h and
up.

      Option B uses continuous  PM10 monitors exclusively to collect far-field
samples.  In this sampling network, concentrations are referenced to 1-h
averaging periods, which are combined to form the 24-h averages needed for
regulatory interpretations.

      It is  recommended that Option  A be employed for the proposed field
study.  This option is viewed as providing  the greatest spatial coverage possible
at a  reasonable cost while still providing the type of data needed to assess the
scientific component of the  dispersion model. Specifically, the following types of
equipment  are proposed for the study.

1.    Two or three primary monitoring stations,  each station with:

      a.   A continuous PM10 monitor such as a beta gauge or tapered element
           oscillating microbalance.

      b.   At least one each Andersen size-selective-inlet (SSI) and Wedding
           PM10 inlet mounted on a high-volume sampler.

      c.   At least one saturation sampler.

      d.   A recording meteorological station.

      The  different samplers will be colocated.   Furthermore, the "whole-air"
      (i.e., noncontinuous)  samplers will be operated over the same time
      periods.  If only one set of high-volume samplers is available, then they
      will be operated on a 24-h schedule.  If more than one of each type is
      available, then shorter periods  (such as 8 and  12  h) are recommended.

      Each of the primary sampling locations will be equipped with a trailer and
      either a power drop or a heavy-duty diesel generator to provide the
MRI-OTS\R10-31 2nd                                                              21

-------
      "clean" electrical power needed to operate the continuous monitors.  (If
      generators are used, special precautions will be taken to avoid sample
      contamination by exhaust.)

2.    Approximately 20 to 30 saturation samplers deployed throughout the
      geographic area downwind of the mine.

This recommended sampling approach permits evaluation of both the operational
and the scientific components of a dispersion model evaluation and also provides
a spatially thorough monitoring network at a reasonable cost.

      The next section discusses how the recommended sampling equipment
would be sited to meet the requirements of the proposed study.  The stylized
mine used as an example in the siting discussion is based on TRC's and MRI's
experience at surface coal mines.
3.2.2.2    Sampler Deployment

      To adequately test ISC, the model predictions and measured
concentrations should be made at the location of the maximum predicted
concentration because this is the value of regulatory interest.

      For illustration purposes, a stylized power shovel mine is shown in
Figure 1. At the example mine, the major sources are found to be coal haul
trucks, overburden haul trucks, overburden handling, and overburden dumping.
General traffic is also important but is confined to the two roads shown in the
figure.

      Mining activity rates and locations  at the example mine have been
evaluated along with the site-specific meteorological data to determine prevailing
wind direction and the sampler locations.  It is proposed that at least three
portable samplers be placed in an arc immediately downwind of the major
sources.

      Significant mining emission sources are rarely grouped together; rather the
sources may be spread over distances on the order of miles. To measure the
cumulative effect of separate sources, the model should be tested at not only the
points of maximum concentration from individual sources but also at points
where the combined effect is greatest.  A second set of samplers is placed at the
points where emissions from individual sources combine to produce a  secondary
maximum (Figure  1). To determine what  the background contribution  is, three
additional locations are  proposed, placed  in upwind and crosswind directions
22
                                                                MR> CTS

-------
           TOPSCIL  REMOVAL
     COAL BLASTING
     AND LOADING
      PREVAILING WIND

      DIRECTION
OVERBURDEN
BLASTING  AND LOADING
               OVERBURDEN DUMPING
          TOPSOIL REPLACEMENT
 COAL HAUL ROAD
  COAL
PROCESSING
                                                  OVERBURDEN
                                                  HAUL ROAD
                                               O
                                               o
                          Maxinxin concentration samplers (assunes four largest PM-10 sources are the two row
                          overburden blasting/loading, and overburden dumping


                          Cumulative concentration sanplers
             ^B      Q   Upuind/bsclcground concentration samplers


 Figure 1.   Proposed sampling arrays at a stylized truck-shovel mine.  Filled
             symbols represent primary sampling locations, hollow symbols
             represent saturation samples.
 MRI-OTS\R 10-31 2nd
                                                      23

-------
from the mining activities.  It is proposed that roughly half of the saturation PM10
samplers be sited to sample combined concentrations.

      The exact placement of samplers at the actual mines tested will be a
function of both the configuration of the mine and the local meteorology.  Until
the time that an actual test site decided upon, exact placement of samplers
cannot be made.
3.2.3 Source Activity Monitoring

      A "dynamic" emissions inventory will be necessary to accomplish the goals
of the model evaluation. Unlike most regulatory modeling exercises, which are
interested in "annual average" or "worst-case" conditions, this program needs to
track actual source activity over time. Source activity levels will be recorded
during the entire period that field test crews are present at surface coal mines.

      Source extent activity data will be collected with a variety of tools. For
example, in addition to visual observation and note taking, video cameras and/or
time-lapse photography may be used to determine activity on roads or at other
mining sources.  For roads and other travel-related sources (such as coal or
overburden truck dumps), pneumatic axle counters will be used to supplement
other approaches.

      The use of a variety of data acquisition methods aids in resolving source
activity levels on an appropriate temporal basis.  Source activity (both rate and
physical location) will be resolved to 4-h or shorter time periods. The 4-h period
corresponds to one-half work shift at a mine.  Based on MRI's and TRC's
experience at coal mines, most emission sources can be  viewed as relatively
constant in time and space over a work shift.  The use of a half-shift as the basic
time averaging unit provides a margin of safety.  In addition, at the start of each
day, the field crew chief will ask the  mine superintendent  or his designee to
describe any unusual events (equipment downtime,  etc.) during the past 24 h.
3.3   DATA ANALYSIS AND MODEL EVALUATION

      The final task will be to tabulate measured and modeled values and to
determine the performance of the ISC model in accordance with the model
evaluation  protocol discussed earlier to present a 'finished product." That is to
24

-------
say, the interim and final reports will present all data analysis and interpretation
of the results, including:

•     Comparison of measured and estimated emission factors.

•     Any modifications to emission factor models.

•     Comparison of measurements and individual steps in the EF/DM
      methodology.

•     Performance (as will be defined in the evaluation protocol) of ISC in
      assessing the "whole mine" impacts.
MRl-OTS\R1D-3l2nO                                                               25

-------
                                SECTION 4

                              PIT RETENTION
       The result of prior pit retention work funded by EPA was a first-cut pit
retention algorithm as a function of meteorological and pit dimension data. The
available pit retention work suffers from two deficiencies; however.  First, the field
data upon which the analytical work was based consisted of videotapes of smoke
puffs released from inside mine pits, and from these videotapes, opacity (and
paniculate concentrations) was inferred. Second, the majority of the analytical
means of quantifying the effect of pit retention that were derived were very
involved, and did not lend themselves to inclusion into the ISC model.

       Unlike the studies discussed in  Section 3, there is very little technical
basis upon which a pit retention field study can  be built.  Therefore, far more
initial effort is required to develop appropriate techniques.  Candidate materials
and techniques will be evaluated during a "check-out" study.
4.1   "CHECKOUT" STUDY

      Dissemination and detection of particulate tracers in appropriate particle
size ranges present a formidable technical challenge.  For this reason, a
preliminary "checkout" study is proposed.  This program would be carried out
during the fall of 1991 in the general Kansas City area.

      The first step in the checkout study involves identification of several
candidate tracer materials (such as encapsulated Rhodamine B or other dyes
and other less expensive materials) and associated analysis methods. Concur-
rently, a suitable test site in the Kansas City area will be identified. The site
should consist of a parallel road and ridge which are roughly perpendicular to
prevailing winds.  The ridge should be largely unvegetated;  an abandoned or
inactive quarry  site may be ideal.

      Next, methods of releasing the tracer material into the wake of a moving
vehicle will be examined.  At  present, It is anticipated that a simple auger feed in
an induced draft may be suitable. In a separate  set of experiments to be
conducted at MRI's main laboratories, weathering characteristics of the tracer
MRl-OTS\R10-31 2nd                                                               27

-------
material will be examined to determine if the material would be acceptable as an
"on/off indicator" of wind erosion during the dispersion model evaluation study.

      The final step in the checkout study involves near- and far-field air
measurements to determine what release rates are necessary and how escape
fractions calculated by different measurement approaches compare.  Note that a
line source arrangement should eliminate difficulties in locating samplers near the
plume centerline. To the extent practical, different distances from the ridge to
the release point will be examined to determine how the necessary release rate
and relative accuracy between the measurement approaches  vary as a function
of distance to the samplers.
4.2   FIELD PIT RETENTION MEASUREMENTS

      In this task the tracer releases will be made and concentrations at the top
of the pit will be measured. Meteorological data will be measured and recorded
at the surface for subsequent correlation with measured escape fractions. A
10-m meteorological tower and wind speed/direction and radiometer (or AT)
instrumentation will be positioned in the prevailing upwind location from  the pit at
surrounding grade. Only one major meteorological instrument tower has been
proposed because ISC allows only one set of met data to be input.  Recall that
one of the site selection criteria requires that the area immediately surrounding a
source test site be relatively free of factors influencing wind flow. Recall too that
by coordinating tracer studies with the  short-term, near-field experiments,
concurrent  in-pit  wind measurements will be obtained.

      At present, it is not possible to provide a definitive measurement
approach.   A major outcome of the "checkout" study will be determination of
what techniques  provide acceptable accuracy.
28
                                                                MR -C~S =•:-?• 2'

-------
4.3   DATA ANALYSIS AND MODEL-BUILDING

      The final task in this portion of the work deals with data analysis and
development of an algorithm for use with the ISC dispersion model. As was
discussed in the preceding section, the interim and final reports will present all
data analyses and interpretations,  including:

•     Comparison of measured escape fractions with estimates from available
      models (such as the Fabrick and Winges equations).

•     Interpretations of measured escape velocity as a function of
      meteorological conditions and pit/source geometry.

•     Recommendation of an algorithm for use with ISC to account for pit
      retention at SCMs.
MRI-OTS\R10-31 2nd                                                              29

-------
                              REFERENCES
Cole, C. F., and A. J. Fabrick, "Surface Mine Pit Retention," Journal of Air
      Pollution Control Association, V 34, No. 6, p. 674 (June 1984).

Cole, C. F., B. L Murphy, J. S. Evans, A. Garsd, 1985, "Quantification of
      Uncertainties in EPA's Fugitive Emissions and Modeling Methodology at
      Surface Coal Mines," for National Coal Assoc., by TRC Environmental
      Consultants Inc., Englewood, Colorado (February 18, 1985).

Cole, C. F., J. S. Touma, J. L Dicke, and K. D. Winges, 1989, "Pit Retention of
      Paniculate Matter at Surface Coal Mines," Paper  No. 89-114.5, presented
      at the 82nd Annual Meeting of the Air & Waste Management Assoc.,
      Anaheim, California  (June 25-30, 1989).

Cox, W. M.,  1988, "Protocol for Determining the Best Performing Model," U.S.
      EPA, Office of  Air Quality Planning and Standards, Research Triangle
      Park,  North Carolina (June 1982).EPA, 1983, "Studies Related to
      Retention of Airborne Particulates in Coal Mine Pits—Data Collection
      Phase," Contract No. 68-03-3037, U.S. EPA, IERL, Cincinnati, Ohio
      (August 1983).

EPA, 1985, "Dispersion of Airborne Particulates  in Surface Coal Mines—Data
      Analysis," EPA-450/4-85-001, U.S. EPA, Research Triangle Park,
      North Carolina (January 1985).

EPA, 1986, "Continued Analysis and Derivation of a Method to Model Pit
      Retention," EPA-450/4-86-003,  U.S. EPA, Research Triangle Park,
      North Carolina (January 1986).

EPA, 1990, Draft Supplement B to the Guideline on Air Quality Models
      (Revised), U.S. EPA, Office of Air Quality Planning and Standards,
      Research Triangle Park, North Carolina (September 1990).

Holzworth, G.C.,  1971, 'Mixing Heights, Wind Speeds, and Potential for Urban
      Air Pollution Throughout the Contiguous United States," EPA AP-101, U.S.
      EPA (May 10,  1971).
MRi-OTS\R10-31 2nd                                                             31

-------
MRI, 1991.  "Review of Surface Coal Mining Emission Factors," Interim Report,
      Contract No. 68-DO-0137, Assignment No. 10, U.S. EPA, OAQPS,
      Research Triangle Park, North Carolina (July 1991).
PEDCo, 1982.  "Characterization of PM10 and TSP Air Quality Around Western
      Surface Coal Mines," by PEDCo Environmental, Inc., Kansas City,
      missouri and TRC Environmental Consultants,  Inc., Englewood, Colorado,
      for U.S. EPA, Air Management Technology Branch, Research Triangle
      Park, North Carolina (June 1982).

-------
                          APPENDIX
           SYNOPSIS OF PRIOR COAL MINE DISPERSION
                  MODEL EVALUATION STUDIES
MHI-OTS\R10-31 2nd                                                 A"1

-------
               SYNOPSIS OF PRIOR COAL MINE DISPERSION
                      MODEL EVALUATION STUDIES
      Although model evaluations of isolated point sources are relatively common,
model evaluations of surface coal mines are rare. It is instructive to examine the
few coal mine model evaluations in order to discover what work has been done, and
to determine how mine model evaluations can better be conducted. This appendix
discusses the methodology, findings, and shortcomings of the surface mine model
evaluations known to the authors.
1. KOMPetal.

SYNOPSIS:

      Mark J. Komp and others (Komp et al., 1984) report the findings of a study
comparing the performance of the COM and the ISCLT.model used to predict
annual average TSP concentrations in the Powder River Basin of Wyoming. The
scale of the output was 25 x 40 km (regional scale). Using identical source
emission rates and locations, the ISCLT and  COM models were run, and resulting
isopleths of concentration were plotted. TSP emission rates were based on
permitted production rates at each of six mines in the basin, and for this reason it is
likely that modeled concentrations overpredicted actual TSP concentrations because
these mines are all permitted for greater production than is actually achieved. The
authors found that the ISCLT model and COM model isopleths were nearly identical,
judged in a qualitative sense.  Another run of the ISCLT model was made with
deposition, and the resulting concentrations were found to be significantly smaller.
Comparisons of modeled and measured TSP were made for calendar years 1980,
1981, and 1982 at 11  hi-vols located near the mines. However, it appears that the
same meteorological data were used for all three years.  Modeled concentrations
invariably overpredicted measured concentrations.

COMMENTS:

      This study shows the "equivalence" of two long-term Gaussian models.  The
use of permitted rather than actual production rates likely accounts for the models'
overprediction of measured concentrations.  Consequently, the findings are of little
interest in judging performance of emission factors and models.

      TIME RESOLUTION:
            EMISSION  RATES: Annual average
            MEASURED CONCENTRATIONS:  Annual average
            MET DATA: Annual STAR  data
      POLLUTANT: TSP
      METHOD: Paired in time and space
      STATISTICS: None

MRl-OTS\R10-31 2nd                                                            A*3

-------
2. DAILEY, B.

SYNOPSIS:

      Bernie Dailey of Wyoming DEQ used actual production data, combined with
Wyoming emission factors, and the CDMW and ISCLT models to predict geometric
mean annual average TSP concentrations in the Powder River Basin of Wyoming
(Dailey, 1984).  Various combinations of central wind speed categories, wind speed
exponents, available meteorological data sets, and several model refinements were
used to compare measured and modeled TSP values at 15 regional hi-vols, for
calendar years  1980 through 1983.

      In general, the  CDMW model tended to underpredict concentrations at
receptors far from the mines. Dailey also compared ISCLT and CDMW, finding  no
appreciable difference in model performance.  Despite high correlation  coefficients
(r = 0.92 in some comparisons), Dailey concluded that simple model correction
factors would not suffice to improve performance. Instead, he recommended a full
examination of  emission factors, meteorological data sets, and deposition
phenomenon as future investigations.  The work in this unpublished review is
thorough and well planned—an ambitious undertaking for  1984.

      Measured and modeled TSP annual means in all instances were compared
with regression plots of geometric mean  values, with least square fits and  correlation
coefficients.

COMMENTS:

      Again, this study shows the "equivalence" of two long-term Gaussian models
but is restricted to annual average  TSP concentrations at hi-vols some  distance  from
the mines. The findings are of limited interest in answering the  question "How well
do the emission factors/models predict short-term concentrations that would be used
to permit a mine?"

      TIME RESOLUTION:
            EMISSION RATES:  Annual average
            MEASURED CONCENTRATIONS:  Annual average
            MET DATA:  Annual  STAR data
      POLLUTANT: TSP
      METHOD: Paired in time and space
      STATISTICS: Regression analyses; correlation coefficient
A-4

-------
 3. TRC

 SYNOPSIS:

       As part of its Emission Factor Development Study (EDS), TRC Environmental
 Consultants compared long-term (annual average) and short-term (6-h) TSP
 concentrations with ISCST and  ISCLT modeled concentrations near the Belle Ayr
 Mine in Wyoming (TRC,  1981).  The emission rates were computed using TRC's
 EDS emission factors—in fact, to the extent that the measured and modeled
 concentrations agreed, the model comparison was viewed as a verification of the
 emission factors.

       Comparison was generally good, although limited data pairs were available.
 For short-term (6-h) periods, examination of 48 data pairs indicated that agreement
 of measured/modeled concentrations was "within a factor of two" 41% of the time;
 and "within a factor of three" for 56% of the pairs. For annual average time
 intervals, comparison of annual average TSP measured at 4 hi-vols near the Belle
, Ayr Mine is extremely good—all annual average measured and modeled
 concentrations agreed within a few micrograms per cubic meter over a range of 24
 to 68 |ag/m3.

       The major drawback to the study is that it used the EDS emission factors,
 which have not gained widespread use and which have not been subject to review.
 TRC had the definite advantage of having detailed on-site activity parameters and
 independent variables for the study. In fact, the use of activity data and emission
 factors which were measured at the very same mine from which the verification
 study was performed certainly helps the model/emission factor performance, but
 raises doubts about the representativeness of the model comparison findings.

 COMMENTS:

       The fact that the emission factors and the model test data were derived from
 the same mine casts doubt on the findings.  Use of both short-term and annual
 average concentrations is welcome. The fact that the EDS emission factors have
 not been widely used detracts from the usefulness of this model performance study.

       TIME RESOLUTION:
             EMISSION RATES: 6-h values
             MEASURED CONCENTRATIONS:  Annual average and 6-h
             MET DATA:  On-site hour-by-hour; on-site STAR data
       POLLUTANT: TSP
       METHOD: Paired in time and space
       STATISTICS:  Regression analyses; correlation coefficient
 MRl-OTS\R10-31 2nd                                                              A"5

-------
4. TRC (confidential)

SYNOPSIS:

      A comparison of measured and modeled PM10 and TSP concentrations at a
Midwestern surface coal mine over a 5-month period was performed in an
unpublished, confidential study (TRC, 1986). The ISC model and the AP-42
emission factors were used as a basis, and 120 individual days were examined at
five different  sampling locations.  PM10 concentrations at the mine averaged about
one-third the magnitude of the TSP values. Great care was taken in locating source
locations individually for every day of the test period, and PM10 and TSP emission
rates were similarly computed on a daily basis.

      TSP concentrations were overpredicted slightly on average; PM10 values were
severely overpredicted on average. Individual 24-h values showed a large degree
of scatter.  The overpredictions were attributed to ISC's failure to properly account
for particle deposition (as opposed to settling) in the ISC  model. However, the
locations of all of the samplers were sufficiently close to the mine sources that the
study was  not a good test of deposition—insufficient travel distance between
sources and  samplers would not permit deposition to reach equilibrium.

COMMENTS:

      This study is an improvement over prior comparisons in that it strives to
examine short- and long-term concentrations of both TSP and PM10, and it uses
24-h resolution of source emission rates and source locations.  However, the large
scatter in individual short-term results is discouraging; and the fact that the study
has not been released minimizes its usefulness to the modeling community.

      TIME  RESOLUTION:
            EMISSION RATES:  24-h values
            MEASURED CONCENTRATIONS:  5-month  average and 24-h
            MET DATA:  On-site hour-by-hour
      POLLUTANT: TSP and PM10
      METHOD:  Paired in time and space
      STATISTICS:  Regression analyses; correlation coefficient;
             scatter plots
A-6

-------
TECHNICAL REPORT DATA
ff'/i'Osc rtad In^lruclKin; on /"<• reverse bifon lamp
1 RF.PORT NO
EPA-454/R95-008
2
4 TITLE AND SUB1 ITLE
Development of a Plan For Surface Coal Mine Study
7. AUTHOR(S)
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Midwest Research Institute
Kansas City, Missouri

12JSPONSORING AGENCY NAME AND ADDRESS
Emission Factor and Inventory Group (MD-14)
Emissions Monitoring and Analysis Division
Office of Air Quality Planning and Standards
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
15. SUPPLEMENTARY NOTES

in in.i!' j
3 RtCIP"- NT s ACCESSION NO
1
5 REPORT DATE
October 28, 1991
6 PERFORMING ORGANIZATION CODE
8 PERFORMING ORGANIZATION REPORT NO
10 PROGRAM ELEMENT NO.
11 CONTRACT/GRANT NO
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE

16. ABSTRACT
This report outlines the steps and considerations involved in developing test plans
and protocols for improving emission factors for western surface co-:! mines in
response to requirements of Section 234 of the Clean Air Act of 1990.
17.
a. DESCRIPTORS
KEY WORDS AND DOCUMENT ANALYSIS
b.lDENTIFIERS/OPi

18. DISTRIBUTION STATEMENT
19. SECURITY CLA
20. SECURITY CLA
EN ENDED TERMS C. COSATI Field/Group

5S (This Report I 21. NO. OF PAGES
43
BS (Tins page I 22 PRICE
EPA Form 2220—1  (R«v. 4-77)   PREVIOUS  EDITION is OBSOLETE

-------