United Stales                   EPA-600/9-89-Q85
             Environmental Protection
                                      September 1989
v>EPA     Research and
            Development
            FIFTH SYMPOSIUM ON

            FUGITIVE EMISSIONS:

            MEASUREMENT AND CONTROL

            (May 3-5,  1982, Charleston, South Carolina)
             Prepared for
            Office of Environmental Engineering
             and Technology Demonstration
             Prepared by
            Air and Energy Engineering Research
            Laboratory
            Research Triangle Park NC 27711

-------
                  RESEARCH REPORTING SERIES


Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in  related fields.
The nine series are:

    1. Environmental Health Effects Research

    2. Environmental Protection Technology

    3. Ecological Research

    4. Environmental Monitoring

    5. Socioeconomic Environmental Studies

    6. Scientific and Technical Assessment Reports (STAR)

    7. Interagency Energy-Environment Research and Development

    8. "Special"  Reports

    9. Miscellaneous Reports


This report has been assigned to the MISCELLANEOUS REPORTS series. This
series is reserved for reports whose content does not fit into one of the other specific
series. Conference proceedings, annual reports, and bibliographies  are examples
of miscellaneous reports.
                        EPA REVIEW NOTICE


This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.

This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.

-------
                                         EPA-600/9-89-085
                                         September 1989
               FIFTH SYMPOSIUM ON



FUGITIVE EMISSIONS:  MEASUREMENT AND CONTROL

      (May 3-5, 1982,   Charleston, South Carolina)



               EPA General Chairmen:

      D.  Bruce Harris and William B. Kuykendal
    Air  and Energy Engineering Research Laboratory
    (Industrial Environmental Research Laboratory)
     Research Triangle Park, North Carolina  27711
                    Prepared for:
         U. S.  Environmental Protection Agency
          Office of Research and Development
               Washington,  DC  20460

-------
                      ABSTRACT

    These proceedings document presentations at the Fifth Symposium
on Fugitive Emissions, held May 3-5,  1982, in Charleston, South Carolina.
The symposium was sponsored by the U. S. Environmental Protection
Agency's Industrial Environmental Research Laboratory (now known as
the Air and Energy Engineering Research Laboratory) in Research
Triangle Park, North Carolina, as part of the Agency's continuing effort
to develop methods for the measurement and control of airborne and water-
borne fugitive emissions  from energy and industrial processes.
    The  objective of the  symposium  was to bring together  people from
              *
industrial, academic, research,  and government organizations with exper-
ience or  interest  in fugitive emissions problems to exchange information of
mutual potential benefit.
    The  program included presentations by individuals from a variety of
organizations describing their experience and viewpoints regarding the impact,
measurement, and control of fugitive emissions.  An international flavor
was provided by presentations by authors from Belgium, Canada, and Sweden.
                              11

-------
                         CONTENTS


SESSION I—Session Chairman Alfred B. Craig (EPA/IERL-RTP)

Results of Measurement Programs for Fugitive Hydrocarbons
    Using a Downwind Crossectional Flux Analysis Method—
     P. R. Harrison (Envirosol)                                  1-1

Test Protocol for Evaluating Fugitive Emissions from Active
    Coal Storage Piles—D. Carnes,  P. Catizone, T. Kincaid
    (TRC Environmental Consultants); D. B. Harris (IERL-RTP)   2-1

Characterization of Fine Particulate Emission Factors for Paved
    Roads—C. Cowherd,  P. Englehart (Midwest Research Insti-
    tute)                                                      3-1
SESSION II--Session Chairman William Hague (Julius Koch,
            USA, Inc.)
Micron Droplet Dust Suppression Proves Out in Variety of Fugi-
    tive Dust  Applications—W. Hartshorn (Sonic Development),
     L. Strand (Andeze AB,  Sweden)                              4-1

The Optimization of Wind Screens for Fugitive Emission Control
    Using Wind Tunnel Tests—C. J. Williams (MHTR)*            5-1
Evaluation of Field Test Results on Wind Screen Efficiency—
    A.  Larson (TRC Environmental Consultants)*                 6-1

Effects of Street Sweeping of Fugitive Emissions from Urban
    Roadways—D. F.  Gatz (Illinois State Water Survey)           7-1

Evaluation of Road Carpets and Chemical Road Dust Suppressants—
    A. Larson (TRC Environmental Consultants)*                 8-1
Evaluation of  Weathering  Characteristics  of Dust Suppressant
    Chemical  Additives—W.B. Kuykendal, D. C. Drehmel,
    B. E. Daniel (IERL-RTP)                                    9-1

SESSION in--Session  Chairman John E. Yocom (TRC Environ-
             mental Consultants)

On the Use of SFg Tracer Releases for the Determination of Fugi-
    tive Emissions—B. Vanderborght, J. G. Kretzschmar, T.
    Rymen (SCK/CEN, Belgium); F. Candreva, R. Dams (INW-
    RUG, Belgium)                                             10-1
(*) Alternate paper provided.
                                  iii

-------
Contents (cont.)
An Atmospheric Tracer Investigation of Fugitive Emissions
    Transport in the Colorado Oil Shale Region—G. A.
    Sehmel (Pacific Northwest Laboratory)                    U-l

Laboratory Testing to Improve Rail Car Sealant Spray and
    Loading Techniques for the Abatement of Fugitive Coal
    Dust—C.J.  Williams. W. F.  Waechter (MHTR)**          12-1

Studies of Nontraditional Fugitive Particulate Control Tech-
    nique s--B. M. Nicholson (EPA/OAQPS), M.  Borcherding
    (City  of Portland), G. Ekhardt (State  of Minnesota),
    R.  Mohr  (State of  Colorado)**                             13-1
Evaluation of the Efficiency of a  Charged Fog Generator in
    Controlling Inhalable Particles at a Steel Plant (C. V.
    Mathai, B. M. Muller (AeroVironment); W. B. Kuyken-
    dal (EPA/IERL-RTP)*                                   14-1

SESSION IV--Session Chairman James A. Dorsey (EPA/
             IERL-RTP)
Estimation of Ambient TSP Impacts of Coal Storage and Hand-
    ling Facilities--R. C.  Wells, D. C. Doll (Enviroplan),
    J.  Hattrup (Baltimore Gas and Electric)                   15-1
A Determination of the Impact of  Fugitive VOC Emissions
    from a Municipal Hazardous  Waste Incinerator on the
    Surrounding Community--G. A. Holton (Oak Ridge National
    Laboratory). L. J. Staley (EPA/IERL-Cin)*                16-1
Impact of Fugitive Emissions on  PM-10  Concentrations—
    T.G. Pace  (EPA/OAQPS)**                               17-1
Application of Dispersion Dictated Mass  Balance for Calcula-
    ting Fugitive Dust Emissions—C. F. Cole  (TRC Environ-
    mental Consultants),  J. G.  Moldovan (Anaconda Minerals),
    P. B. Kunasz (Consultant)                                 18-1

Modeling the Emission of Aerosols in and Around a Metallur-
    gical Plant--B. Vanderborght, I. Mertens, J. G. Kretz-
    schmar (SCK/CEN,  Belgium); F. Adams (UIA, Belgium);
    R. Dams (INW, Belgium)**                               19-1

SESSION V--Dennis J. Martin (TRC Environmental Consultants)

New York State Industrial Coal Pile Drainage Regulations and
    Guidelines--C. Hornibrook (NYSERDA)                    20-1
 (*) Alternate paper provided.
 (**) Abstract only.
                                    IV

-------
Contents (cont.)

Calibration and Verification of a Coal Pile Drainage Model--
    J. A. Ripp, G. T. Brookman. P. B.  Katz (TRC Environ-
    mental Consultants); J. G. Holsapple (New York Power
    Pool)                                                    21-1

Coal Pile Simulation Study--A. Schumacher, E.G.  Hanson
    (Environmental Science and Engineering)                   22-1

Control  of Acid Problems in Drainage from Coal Storage
    Piles—H. Olem, T. L. Bell, J. J. Longaker (TVA)         23-1

ATTENDEES                                                 A-l

-------
     RESULTS OF  MEASUREMENT PROGRAMS  FOR FUGITIVE
     HYDROCARBONS USING A DOWNWIND CROSSECTIONAL
                  FLUX ANALYSIS METHOD
                    PAUL R.  HARRISON

                       Envirosol
               Environmental Solutions
                1700 N.  Fiske Avenue
             Pasadena,  California 91104
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
                             1-1

-------
                                 ABSTRACT
     A method to quantify fugitive volatiles organic compounds (hydro-
carbons) has been in the developmental process for the last seven years.
Recent activities have significantly improved the method and have quanti-
fied the precisions and accuracies under both controlled and field
conditions.  This paper is a summary of various applications and their
results.  Special attention is given to the precision (repeatability) of
the data in field conditions.   Facilities,  components, and other area
sources have been selected to  demonstrate the flexibility of the method.
Actual data are given.

     The major conclusions are that the technique is operational to most
applications with precisions of less than 30%,  with 8% nominal.   Ultimate
accuracies are not yet determined  due to lack of an acceptable standard.
The method is the most flexible available and is providing emission rates
hitherto unavailable.   The application to the "bubble rule" is encouraging,
                                    1-2

-------
I.   INTRODUCTION

     This paper is a summary of a developmental program to further de-
velop a method to quantify fugitive emission rates of VOC's under all
reasonable meteorological and physical condition.  The initial impetus
for the work was a problem with applying the U.S.EPA emission rates (AP-
42) for refineries to non-refinery operation, i.e. a small crude oil gas
treatment plant.  In this initial program the measured leak rate for
individual components differed from the published factors by orders of
magnitude.  By use of a downwind plume mapping technique in conjunction
with the first reported "Direct maintenance program", an understanding
of the differences was achieved.1  Since this initial program, performed
in 1976, the technique has been further perfected and applied to total
refineries and petrochemical facilities as well as their sub-units and
individual components.  In addition, Natural Seeps, Impoundments, waste-
water treatment facilities, tankers and storage tank emission rates have
been quantified, usually to much better precision than achieved by more
passive measurements.  Due to the response characteristics of the con-
tinuous VOC monitor preferred for these measurements, and due to the
question of quantitative determinations of precisions and accuracies
the method had not, until now, been advertised as operational.  Since
very few data existed for fugitive emissions and due to the unusually high
variability of most fugitive sources, the task of objective evaluation
of the method was not at all trivial.

     This paper will briefly describe the technique, summarize the past
applications and present data for some emission rates found.  It will also
present the results of precision and accuracy tests performed on recent
data.

     Results will be offered to show that the technique is quite flexible
and will provide adequate, defensible emission factors at a relatively
inexpensive cost (the more precision required, the more work required).
At this point the method is operational.  Although it can be performed
by semi-technical personnel, tests should conducted under the direction
of an experienced person, especially if one desires a cost effective pro-
gram and a data set with optimum accuracy.

II.  THE TECHNIQUE

     The method makes very few assumptions.  The most critical is that
the downwind signature has a Bionominal or "Gaussian" distribution,
especially in the horizontal.  Since we know that on the average, the
turbulence structure will distribute itself to the Gaussian condition
in the micro and mesoscale meteorological volumes, the assumption is
relatively sound.   The key is to choose the distance from the source
and the wind conditions so as to map the plume at the most likely isotropic

                                  1-1
                                  i  \j

-------
turbulent condition.  As a matter of practicability this condition
must be compromised, but one can recognize the condition and compensate
if necessary by early data reduction and a presurvey measurement.

     The formula used is straight out .of Turner's handbook2 (or any
other pertinent resource for that matter) and is for a  centerline con-
centration of a ground level source.

     From diffusion theory and measurements, the following equation
applies for ground-level sources:2

                         Q = IT  oy az  AX max u c K

where

     Q  =  Emission rate of hydrocarbons  (gm/sec)

        =  Peak, net concentration  of  hydrocarbons  in  Gaussian fit
 AY max            .   x
           curve  (ppm)

     K   -  Conversion  constant  from ppm to  yg/m3;  665  x 1CT6  gm/m3  for
           methane  at  20°C

     av  =  Lateral  extent  of Gaussian plume (m)

     az   =   Vertical extent  of  Gaussian plume  (m)

      C  *  Correction  factor from methane equivalent to actual mass
            emission rate due to instrument  response

      •7  -  3.141	

      u  =  Mean wind speed (m/sec)

 All parameters are usually obtained from field measurements.  (In some
 instances,  az is obtained theoretically from  ay.    Care must be taken
 to prevent over or underestimation of  this value in areas severely affected
 by local,mechanically-induced turbulence.)

      The transect method of sampling is used to actually  map  the dimen-
 sions and concentrations of a plume.  From this mapping,  a reasonably
 accurate emission rate can be calculated.

      Two types of plume mapping have  been used:  single level transect
 at an optimum distance and conditions  from source, and a  complete  plume
 crossectional mapping.  The former method is less work (expense) but may
 not be as accurate due to the necessity of calculating the vertical  extent
 of the plume (
-------
made except reasonable isotropic turbulence.  For close-in measurements
of area sources this method is necessary since gz calculated  from  cry
may overestimate the emission rate.  Please see Figure 3.

The actual physical measurement techniques will not be discussed at this
time since, although they are not particularly unique, they do require
specific protocols that must be followed in order to achieve  optimal
accuracies.

III.  EXAMPLES OF APPLICATIONS

      This section will present a selection of the actual applications to
the quantification of fugitive VOC from various facilities and devices.
Please note that fugitive VOC emission rates vary greatly within and from
facility to facility.  The following data will demonstrate the typical
precisions found in steady-state conditions as well as those  in highly
varying conditions.
     REFINERIES, TOTAL FACILITIES

     REFINERY I

     In  this case  the wind  conditions were unusually ideal. A single  level
 set  of transects were used  in the calculations, but the vertical extent
 of the plumes were verified by use of a slow aircraft.  Five data  sets
 were obtained over a three  month period.  Each "signature" was unique
 but,the  data agreed very well, i.e.:

                   Test            Q (tons/day)

                    1                  29.0
                    2                  24.3
                    3                  27.1
                    4                  24.3
                    5                  28.1

                        AVERAGE        26.6 + 2.2, (8%)

     REFINERY II

     In  this case  the conditions were such that the transect line  was very
 close to the edge  of the complex.  Both vertical and a single level mapping
 were taken.  Figure 2 is a  typical signature of the close-in facility fugi-
 tive VOC emissions.  Note that the single level transect would miss part
 of the plumes.  Comparison  of the two types of data resulted in an emission
 rate of  3.2 tons/day but with a precision of twenty three percent.  These
 data demonstrate the loss of precision due to the unfavorable transect or
mapping  conditions as well  as the possible effect of a non-uniform emission
 rate (a  turn-around was underway at Refinery II).
                                 1-5

-------
0    PETROCHEMICAL  (A sub-unit and components)

     Most petrochemical data are proprietary; however, we can use one case
to demonstrate how a single unit can be isolated from the total complex-

     In one case a sub-unit was quantified both upwind and downwind. At
the same time each individual emission point (component) or area was quan-
tified.

     The sum of the most significant of the 19 leaking components was
706 + Ibs per day.  The downwind minus the upwind value was 824 Ib  per
day,~thus there was an agreement within fifteen percent between the
two methods.  (Both single and multilevel transects were made but the
multilevel results were used preferentially in these  calculations.) It
is interesting to note that the precision of these single day comparisons
were better than the comparison in results taken on separate days,  i.e.
the variance of four separate emission rates using the plume mapping
 technique was 21%.  This, of course, was a result of  process varia-
 tions  as well as the data sets themselves.

     Finally, the small components were bagged and timed to find their
 emission rates.  A.  tail gas compressor was two large  to bag, thus both
 a single level and  multilevel transect was used to gain results of  88 and
 81 Ibs/day  respectively.  As one can see, the  agreement is quite good under
 these  conditions.   Other individual areas quantified  were loading racks,
 a flare pipe  leak,  and a spill; each was identifiable within the upwind
 transects.  An  intermittent  process such as "Trap Blowing" was  also  quan-
 tified (148 Ib/hr  of  operation).  Please see Figure 4.

 0    CRUDE  OIL GAS  TREATMENT PLANT

      In this instance  the  fugitive VOC emission rate  for  the  total  facility
 was  initially 74  Ib/day.   After the directed maintenance  program the  rate
 was  measured downwind  as 7  Ib/day.

 0    PROCESSED CRUDE  OIL STORAGE TANK

      This  double  seal  storage  tank contained warm  crude  that  had under-
 gone dewatering and used a heater/treater  process  to  remove  volitiles.
 The emission rate during draw-down was 25  Ib/day.  This measurement used
 a single level transect  and a  plume centerline vertical  profile for de-
 termining  the plume dimensions.

      NATURAL OCEAN SEEPS

      Over the last several years  a large  natural  ocean seep  has been
 quantified by use of  both  the  single  and  multilevel  methods.   The  source
 strength is known to  vary  from visual  observations;  thus precisions also
 include natural variations:-^
                                    1-6

-------
                   NATURAL SEEP EMISSION RATES

                                                Emission Rate
             Year                                  T/day

             1977                                   8.7
             1978                                  10.7
             1979                                  21.0
             1980                                  16
                                                   25
                                                   20
                                                   24
                                                   22
                              AVERAGE 1980         21 + 3.2 (15%)
Since 32% was non-methane, it was determined to be equal to one-half the
county hydrocarbon inventory and that it was commercially feasible to trap
the gas for sale as product.  Finally, a bonus was realized from some of
these measurements in that we were able to document revised dispersion
coefficients over the cold California oceans.

0    REFINERY WASTEWATER TREATMENT

     Currently, under a U.S. EPA contract, an evaluation of the emission
rates of refinery wastewater treatment facilities is taking place.  Sub-
units and components are also being quantified.  These data will be
presented later after review.  They will include preliminary information
such as effect of covers, operating characteristics, age and configura-
tion.  Detailed comparisons of of purely statistical treatment of the
data versus the graphical treatment used are also being determined.
Since these devices are so dynamic in the material received it represents
a severe test for the method.  For example, not only are the locations
of components usually not ideal, the input to them is highly variable
and represents a large mix of VOC constituents.  Also, the liquid surface
conditions vary greatly and can control the fugitive emission rates as
much as the potential VOC in the bulk water.  This study will also dis-
cuss the accuracy of the methods in greater detail.

IV.  ACCURACIES AND/OR PRECISIONS OF THE METHOD

     Although the "plume mapping" or "downwind crossectional analysis"
method of quantifying fugitive emissions has been used in even a more
primitive form for particulates as well as for other gases and tracer
studies, there has been some resistance to its acceptance.  The primary
reason may be that, although it uses off-the-shelf equipment and is
relatively simple in theory, it is tedious and requires judgement as to
the best measurement location to optimize the turbulence conditions so
to approach the classical normal distributed plume.  It marries the real
world with turbulence and statistical theory in a rather unique way that
crosses disciplines, thus making it suspect; i.e. it's a unique combina-
tion of established procedures.

                                    1-7

-------
     A recent obstacle that has been removed is the calibration of the
continuous monitor to the unique hydrocarbon mix encountered in most
areas (if only one VOC is present there is no problem).  By using in-
field analyses for total, NMHC, and MHC, we can determine the calibration
factor C in equation (1) for each condition.  Since the VOC are all con-
verted to methane, we can eliminate the necessity of calibration to each
constituent.  Since the procedure is a step-wise progression of operations
to the data, the variations are easily exposed and not lost in the interven
ing calculations.  Said another way, the process must not only "suit in a
reasonable final data point, each step or parameter must also make physical
sense.

     As part of all these studies we have attempted to focus part of our
efforts toward objective determination of precisions and accuracies (few
of the data sets are complete)i

     Table I is a summary of the precision of much data available to date.
Please note most of these cases are dynamic and contain process variations
as well as variations due to methodologies.

     Recent data  taken over several days verify these results but
also contain standard deviations varying as high as 69% and as low as
18%.  (For data taken at different times during the same day, the preci-
sions improve to 15%.)  Naturally we observe that the quality of the
data sets goes down as the variation among the data goes up.  However,
the realistic test in these data is to compare data taken in close time
sequence on the same device.  These comparisons are those shown in Table
I.  These are a great improvement over the thousand percentage deviations
obtained in other studies on wastewater facilities.

     Accuracy tests were also conducted by comparison of the results of
the downwind method to metered releases and to a temporarily ducted
source.  The ducted sources were less accurate and more variable than
the transect method due to leaks an.* the effects of wind direction orien-
tation to the ducts.  Local turbulence was also a major factor in the
poor showing of the ducted method.  Standard deviations ranged from 220%
to 7%, much larger than the transects.  Ratios of transect  to ducted
resulted in an accuracy of 150% + 68%.  Since we lost much  of the VOC
from the ducts we are not surprised the ratios are usually  in greater
than 100%.   A similar test using a metered rate of release  of propane
resulted in a ratio of transect to metered rate of 1.6 + 19%.  Although
the precision was in the range seen before, the accuracy is not as good.
We have since discovered that as the propane was released,  some of it
was nearly liquid at the rotometer.  Thus, the reported metered release
rate is too low, especially since the metering apparatus is cooled by
the gas phase change.   The resultant ratios confirm this finding.  In
conclusion, we know the precision for these tests is within 30% and the
average accuracy is better than 60%.  A definite test of the system's
ultimate accuracy has not yet been made due to these difficulties in the
accuracy of the calibration methodologies themselves.  It is probably
equal  or better than the precisions.
                                    1-8

-------
                                 TABLE I

                 PRECISION OF DATA SETS AVAILABLE TO DATE


         Test                 a ,  Precision, %        Number of Tests

     Refinery I                       85 days

     Refinery II                     NA

     Petrochemical                   15                     2 methods

     Tail gas compressor              9                     2 methods

     Petrochemical                   21                     4 days

     Ocean Seeps, 1980               15                     5, same day


     Refinery II Wastewater Treatment

       0 Test in succession, 1       21                     3
       0 Test in succession, 2       31                     3
       0 Test in succession, 37                     3
       0 Test in succession, 4       28                     3
       0 All tests, same device       69*                   13



*Conducted over several days and includes process variations.
                                   1-9

-------
     Finally, by using purely statistical manipulations on the raw data
sets, the agreements are also acceptable.  Variations within the various'
statistical methods themselves vary between a standard deviation (a) of
7% to 13%.  The purely statistical analysis and graphical methods agree
within a range of 60% to 120% (+ 40% to 20%) of each other with one of
three comparisons within 3%.

CONCLUSIONS

     From the data presented we can make the following conclusions:

     0  A method exists to quantify fugitive VOC with precisions typically
        within 30% and to 8% for well behaved plumes.

     0  Ultimate accuracies are not yet determined but are known to
        be better than £ 60%.

     0  The method uses a combination of classical theories in a
        unique way.

     0  The method has great utility and can be used on most sources
        accessable downwind (one can also forecast wind shifts).

     0  This method is providing data never before quantified with
        acceptable precisions or accuracy.

     0  The method has application to the bubble rule .and can verify
        the effectiveness of the directed maintenance program.

 REFERENCES

 1.   Harrison, P- R. ,  "Comparison of Component Emission Rates to AP-42  for
     a  Gas Treatment Plant at Ellwood, CA,"  AR.CO Production Co., Envirosol
     Report No.  1461,  1976.

 2.   Turner, D.  B. ,  "Workbook of Atmospheric Dispersion Estimates," U.S.  EPA
     (PHS) Publication 999-AP-26, 1970.  (NTIS PB191482.)

 3.   Harrison, P. R. and Mass, S. T. , "Dispersion Over Water:  A Case Study
     of a  Non-Bouyant  Plume  in the Santa Barbara Channel, California,"  AMS
     Joint  Conference,  November 29, 1977.
                                     1-10

-------
VIRTUAL SOURCE
OF FUGITIVE EMISSIONS
ELEVATED
POINT SOURCES
              Figure  3»  Schematic representation of vertical point profiles (1,  2. 3, 4, 5) and
                         elevated horizontal transect lines (A, B, C, D).

-------
                        (1)
                      SUMMIT
         II HO
mxucT
UMf(J)
 I
I—•
CO
                      Ficuin
                               (k)

                             COMFOMENI
  (0

COUPONEM1

 S1SIEH
                                                           IIINUCT LIMES
                 TMNSECI IINES

                   \
             Figure 4:  Schematic of cross-sectional measurement for subunits,  components,
                         or component system.

-------
                                  Questions

Greg Holt, Oak Ridge National Lab:
     Q:  How far downwind were you measuring for API unit?
     A:  Well, you know, when you are In structures, you have to be careful
          because you have to go far downwind to try to get your turbulence
          distributed; sometimes that 1s not practical.  In this particular
          case, we are approximately 5 meters downwind.
     Q:   What was your time frame for your measurements?  For each transect or
          total?  Well, your averaging time, considering your air movement.
          I'm thinking your Gaussian assumption 1s awfully close.  I would think
          perhaps a box model would be just as good.
     A:   Well, one of the things I can't argue with 1s when you take the data
          you plotted up, 1t sure looks Gaussian and It also satisfied the
          Gaussian tests within reason, there 1s no such thing as a perfect
          Gaussian.  The averaging times, typically, 1t takes a half hour to
          get a data set, which Is good enough.  Sometimes you get wind shifts,
          that 1s why we've altered the technique, rather than take several  at
          one level, and then several at the next level, etc., to work your  way
          up, we take a couple at each level and move up and down - we keep  going
          that way as opposed to doing all the averaging at one level, because
          a shear will come along sometimes and your plume will get bent over.
          It's no real problem because you see that happening and you just
          measure directly up rather than slantwise, so we've altered that.
          We get kind of a data set within 15 minutes or so and just repeat it.
                                     1-13

-------
George Sehmel,  Pacific Northwest Laboratory
     Q:   What 1s downwind distance and how long 1s that wind effective?  Do you
          feel  plume 1s contained 1n lower 2 weters?  Not sure what you did there.
     A:   Well, these are more theoretical questions than practical questions.
          The practical answer to that Is that you take sufficient height until you
          look  at your signature and you take sufficient points until you actually
          see this coming out, so you see the data at Us real time and you map the
          plume, and you don't care what the tall  up there looks like that much
          becasue the real Important part of 1t 1s the center line of the plume and
          you take sufficient height and you clearly define that at least 7Q% of
          the plume and the only problem you get Into 1s that you have to make sure
          you take enough data points discrimination to get close to the maxima.
          If you don't do enough of them within the maxima, then you have a problem
          with what is the maxima.   You can get fairly large error that way.  Now,
          1n some of these techniques you notice that we talk about selected inter-
          perlated data you are actually searching and 1nterperlat1ng a maximum
          concentration at the center line.  And you will find that 1t helps a little
          bit  but 1t doesn't  change 1t that much.  The real question 1s that I don't
          assume anything, I  actually measure It.  There are no assumptions here at
          all,  except  that It's Gaussian distributed and then If you are really con-
          cerned about that (1t being Gaussian) you can actually test 1t for its
          Gaussian  fit, which we  do.  The statistical techniques are purely statist-
          ical  techniques, using  a logrlthmic Gaussian fit and those are your
          results and  you compare that to the graphical, which just assumes a
          straightforward Gaussian.  It is not a box model really, 1t certainly
          looks like  1t doesn't have sharp wedges at all; if anything, 1t 1s a
          Gaussian.
                                       1-14

-------
Q:   You were taking measurements at a horizontal,  were you also taking
     measurements at vertical as well?
A:   We took several measurements at various levels.   You will "find it is
     easier to take the horizontal at various levels  than try to go up and
     down It at various horizontal positions.
Q:   Then you were at more than one plane?
A:   Absolutely, you map both the vertical.  I showed one example of a single
     plane which again was more or less a ground level  source which assumed
     certain vertical dispersion characteristics, but we progressed from that
     a couple of years ago and by various means, either by long  poles or actu-
     ally a tethered sonde balloon system where we actually record the height
     and tow the balloon back and forth 1f 1t 1s a very high type plume.  Yes,
     we get both vertical and horizontal dlsperson parameters.

Q:   In one of your examples, you mentioned you were  as close as 5 meters to a
     unit source.  What was your regular distance when doing entire refineries?
A:   One-quarter of a mile.
Q:   And how high did you go?
A:   In a total refinery, remember we're looking for  fugitives only, you'd
     have to go maybe 500 feet.   At this particular  refinery,  I don't have
     that data here and It's somewhat proprietary,  but we use an aircraft as
     well as simultaneous surface  measurements.  The problem with the total
     refineries 1s that you have cooling towers, you  have flares, you have
     stacks, etc., so the assumption was that 1f you  took your transects close
     enough In, and this happened to be the case, and flares were on the down-
     wind side, so you get close enough so that the flares were not touching
                                  1-15

-------
     down 1n your transect line and took about one transect line, you can use
     your assumption of vertical  dispersion characteristics and come up with
     a good number, using only one transect line of about 3 meters.  On that
     particular refinery, that was what we did and It worked very well.  We
     were looking only for fugitives,  that was the purpose of that particular
     task; although we did measure the total  refinery.  But to separate the
     fugitives from the flares and other materials, we could not go too high.

Q:   In a tethered balloon, what do you do, bring your sample down a small
     diameter tube or how do you do that?
A:   Well, 1t depends on the height.  We can get up to about 500-600 feet
     with tethered lines and bring the sample down.  There's no real problem
     with that, It's just a matter of  lift and some of the other work we've
     done on total chemistry of the detached plumes, we've been able to lift
     quite a bit up Into the plume and get easily a few hundred feet of line
     up there.  If you get much higher than that, you just can't 11ft that
     much line and then we go Into remote control samplers, but you don't
     get a continuous sample, you get  a point source sample.  It's hard to
     map a plume from one point or point data taken 1n sequence over a period
     of time.  We've towed the balloon several  hundred feet high.
Q:   What are your remote control samplers?
A:   I've built these myself.  I  own the balloons, the telemetry and the
     sampler.

Q:   The best method?
A:   It depends on what you're looking for.  Most of the time,  we've been
     after sulfates, sulfur, etc., but you can either use absorbents, you
                                   1-16

-------
can use a bubbler, whatever you want, 1t doesn't matter.   It's just a
matter of turning the pump on and off at precise times,  recording the
meteorological phenomena, time, location, depending on what your posi-
tion 1s, but It's nothing really sexy about It,  you just turn  It on and
off, apply 1t to whatever level you want 1n the  plume, but once you get
above, 1f you want a continuous sample,  say 500  feet,  you almost have
to use a slow flying aircraft.   But, for large  facilities, that 1s not
so bad.  It 1s that point source plume that 1s the  problem with an air-
craft.
                             1-17

-------
   TEST PROTOCOL  FOR EVALUATING FUGITIVE  EMISSIONS

            FROM ACTIVE COAL  STORAGE PILES
                     David Games
                     Pietro Catizone
                     Thomas Ciacaid

          TSC Environmental Consul cants,  Inc.
          800 Connecticut Boulevard
          East Hartford, Connecticut 06108

                          and

                   D.  Bruce Harris

    Industrial Environmental Research Laboratory
    Research Triangle Park, North Carolina 27711
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved lor
presentation and publication.
                         2-1

-------
Introduction

    A sampling program  was developed to collect  fugitive particulate  matter
•missions  data  from an  active  coal  storage  pile  servicing  an  electric
utility.  The  data were used  to  determine the  relationship between  program
costs and relative accuracy for variously sized monitoring networks.

    The  upwind-downwind sampling  method was  selected for the study  because
it  is the  most appropriate  method  for assessing  the  source  strength  of
relatively  large sources that cannot  be sampled  by traditional  techniques.
This method entails  measuring  downwind concentrations, correcting  this value
for  ambient  background  concentrations  (upwind  contribution),   and  using
diffusion equations  to  calculate  apparent source  strength.   These equations
assume  that pollutant  diffusion  has a  normal distribution  about the plume
centerline  in  both  the horizontal and  vertical dimensions.   The equations
also  assume a  uniform emission rate  of pollutant and that  total  reflection
of  the  plume  takes  place at the  Earth's  surface.   The   area  source  is
represented  as  a  virtual point  source  using the  techniques  of Turner.1
Because  the position of the virtual point source  varies with  stability,  the
downwind distance of the samplers changes slightly between tests.

    Since  emissions  from fugitive dust sources are  normally estimated  from
measurements made at remote distances  from  the source, loss of material  from
the  plume  between   the  source and  the measurement  point  will  affect  the
source   strength estimate.    If   the   source  being  measured  generates   a
signficant  number of large particles*  many  of the particles may  be deposited
before the  plume reaches the measurement point.   To  take the effect  of plume
depletion  into  account,  correction  factors   were developed  to  adjust  the
apparent  source  strength.   By  collecting  dustfall samples  at  downwind
distances  where  suspended particle  concentration  is  also  being  sampled,
deposition  velocity  may be calculated by taking  the ratio of  the  deposition
rate  to  the  immediate ground  level  air concentration.   Chamberlain  and
others  describe  a method  for  computing  a  source depletion relationship  by
modifying   the   equations  published  by  Sutton.2   The  depletion  fraction
using  this  method  is  a  function  of  downwind  distance,   Pasquill-Gifford
atmospheric stability  class,  source  elevation  and  particle   deposition
velocity.   The estimates of source  strength, corrected  for dry  deposition,
were  statistically analyzed to determine the  cost effectiveness of variously
sized monitoring networks  for  similar  area sources.

                              Monitoring Program

    Design  of  the sampling array  to  monitor  source emissions  is illustrated
in  ?igure  1.   The  active area of the  coal pile,  which constitutes the major
source  of  fugitive  coal dust, measured approximately 177 meters  (580 feet)
long, 116 meters (380 feet) wide,  and was oriented with the  long  axis in the
east-west  direction.  Activities within this area included  loading  onto the
active pile through  a telescopic chute  and spreading and reclaiming  coal by
bulldozers  and  scrapers.   The level of  activity  within   the  source  area
varied  from test to test  and  therefore the  generation  rate of  the  fugitive
emissions  also varied.

    The  orientation of the sampling  array  with  respect to  the  source was
based on  5-year  average  seasonal  wind  rose  data  collected at a  nearby

                                   2-2

-------
                  UGEHO
      A  HI6N-WUMC SAMPICI
                          vim sue sfuaivf
         ANO MlfiH VOLUHI CASCADE INMCfM
      O  JO-FOOT TOJtt UIIH NICH WHUff SAMTLHIS
      m  30-FOOf fttJEI UIIN MIGN VOLIMC SAMTIEU
      "  ANO NEIEOMN.OGICAL SfNSOM
      O  OUSI f All
      *  ME1EOMH08ICML SCiSMS Al I FCCT
      Q STATION IB
                     H
MIT
'rtlCII VOIUMf .SANPtfR lOCATf* I401
 APPBOI.  1000 Fill UPUINO OF SOIMCC
               Figure ).   Final  monitoring network  layout  and site topography  (elevations  In feet).

-------
Rational  Weather   Service   station.    The   upwind  station   was   located
approximately 1,000 feet  south of the aoucce  area.   Three downwind  sampling
locations were  established at approximately  150 feet, 1,000  feet and  1,400
feet,  based  on estimates and preliminary analyses  of  source  strength  and
deposition characterization.   The crossvind  dimension of  the  array  spanned
approximately a 45° sector.

    Initiation of a test  occurred only when wind flow was at speeds  greater
than  1 aps  (2.2 mph).   The  length  of  each  test  was  determined  by  the
activity level  within the  source area  and  the time  required  to collect a
measurable sample at  all  stations within the  dust plume.   Experience showed
that  when filters  became  visibly  dark  in  color,   sufficient  particulate
matter  had  been  accumulated  for accurate weighing.   Site  characteristics
during the test periods are described in Table I.

              TABLE I.   Site characteristics during test periods
   Test
  Number
          Wind
         Speed
         Jm/s)
   Wind         Temp.
Direction+t     (°C)
Pasquill-
Gifford
Stability
Category
Coal Pile
Activity
 t

tt
                                                                   2  bulldozers
                                                                   2  scraper(s)

                                                                   2  bulldozers
                                                                   2  scraper(s)
3
4t
5t
6
7
8
9
2.3
3.6
1.5
3.6
3.1
3.6
2.9
146
220
60
147
146
145
143
23.9
23.9
27.8
24.4
24.4
21.1
23.3
B


B
B
B
B
1 bulldozer
1 scraper (s)
1 bulldozer
2 scraper (s)

1 bulldozer
1 scraper (s)
2 bulldozers
2 bulldozers
2 scraper (s)
2 bulldozers
                                                             2  scraper(s)
                                                        -	==^==
Preliminary  tests  to  determine lateral  and  downwind configuration  of
sampling array.
Tests  that  were terminated  because  of shifting  wind flow out  of the
desired sector.
Readings are in degrees relative to true north.
                                 2-4

-------
                              Monitoring Equipment

     The high-volume  samplers  used  to collect  total  suspended  particulate
 matter (TSP)  data were Misco Model 680 samplers.  These  samplers are equipped
 with constant  flow controller*  to maintain  the airflow  through the  filter
 medium at  a relatively  constant  volume  rate  of  40 cfm.* At  the  colinear
 stations 13, 24 and 33,  an additional high-volume  sampler  equipped with  a
 Sierra size  selective  inlet (SSI)  and a Sierra five-stage cascade  impactor
 (CZ)  was rua simultaneously at approximately  20 cfm.   This  lower flow  rate
 was used  to reduce particle bounce  within the  impactor  head.   The  design
 cut-point  of 15 um at 40  cfm for  the  SSI  was altered  to  about  16  um  by
 the change  in   flow rate.  Sierra  glass-fiber filters.  Models  C-230-GF  and
 C-305-GF,  were  used  as collection substrates for  the  cascade  impactor  and
 standard high-volume measurements,  respectively.

     In addition to the  high-volume  data,  particle  deposition measurements
 were made  using  20.3  centimeter  (8-inch)  diameter  dustfall  buckets.   The
 buckets were mounted  on  1.8-meter  tripods  to  prevent contamination  by
 saltating  particles near  the  ground.  Grab  samples  of  coal were  collected
 from the  source  area  during  each test  and  sealed  in glass  jars.  These
 samples were used  to  determine moisture content and  grain size distribution
 of  the  source material.

     Meteorological data were collected routinely during  each  test at station
 24  and included sky cover,  temperature, wind  speed  and wind direction.   Wind
 speed   and  wind direction  were  also monitored near  the source  area.   A
 Climatronics  Mark  III  Wind  System was placed  at  the  10-meter level  on  a
 telephone  pole   located  at station  24.   A  similar  wind  system, Climatronics
 Electronic Weather Station,  was mounted on  a  tripod 1.8  meters  (6 feet)  high
 and placed near station 12.  wind  speed  and direction data were recorded  on
 analog  charts at both  locations.  A standard mercury  thermometer and  a Bendix
 aspirated  psychrometer  system were   used  to  obtain   ambient  temperature,
 dew-point temperature and  relative  humidity  data.

                               Analytical Methods

    Glass-fiber  filters were inspected for  defects,  numbered  and  stored  for
 24  hours in  a desiccator.  Filters  were weighed before and  after testing  in  a
 controlled  atmosphere  where  the  temperature  was  between  20°C  and  25®C,
 and the relative  humidity below 50  percent.   Basults  were verified for  10
 percent of   the  filters,   randomly  selected,  according  to  the   criteria
 presented in  the EPA Quality Assurance Handbook.3

    Each  dustfall  sample  was   filtered  to  determine  the total  settleaole
 particulate  matter.  The  total water  soluble  content  of the samples   was
 determined by the  evaporation of an aliquot of the  sample.  Both values  were
 corrected by  using a  control sample for the fluid used  (water).   The sum of
 the  total  suspended solid matter and  total water soluble matter was defined
 as the  total  dry settleable particulate matter.

    The moisture content of  the bulk samples of coal  was determined  following
 the  American Society  for  Testing  and Materials (ASTM)   procedure  D3173-73,
 "Moisture in  the Analysis  Sample of Coal and  Coke."  Grain size distribution
was determined following ASTM procedure D410-38,  "Sieve Analysis  of Coal."
(*) 1 cfm = 0.0005 m3/s.          2-5

-------
    Hourly  average  meteorological  data   were  determined  from  the   data
collected during  each t«at.   Average  wind  «p««d and wind direction values
were  obtained  by   reducing  the  analog   charts   and   were  converted  to
appropriate  engineering   units.    Atmospheric   stability   conditions   were
characterized  as Pasquili-Gifford  stability  categories  and  estimated  as
suggested by Turner.^

••suits and Discussion

    The  calculated  emission  rates  varied among  sampler*   because   of  a
combination  of random  and  systematic  errors.  Random error in  the   data
results from  factors influencing  data collection,  data  reduction procedures
and  from  the  contamination of  samplers  by  downwind  sources  within  the
sampling  array.   These   factors  were  present  in  all  five   tests,   and
theoretically  are manifest  as  the deviation  among individual estimates  of
the  source  emission rate.   Systematic error, introduced through the use  of
the  Gaussian equation, was not  quantifiable in this  study  because actual
source  emission  rates were  not  known.   The  accuracy  of  the predictive
equations was considered to  be a factor of 2.

                               Surface Sediments

    Surface  sediments within the  source area were  characterized  in  terms  of
moisture  content  and grain  size distribution.   Both of  these  parameters are
important because of their influence on the  emission  rate.  As  the  moisture
content  increases,  particles  bind  together  and  the   erodability  of  the
surface  material decreases.   The  grain  size  distribution   provides  an
indication of  the availability of particles  in  the finer fractions  and also
the  surface  roughness  which  influences   wind  flow   in  the  vicinity  of
individual grains.   The moisture  content on  test days was between 3 percent
and  4  percent by weight.   The surface  sediment  ranged  from  clay to pebble
size fractions.   About 50  percent of the  surface material  was larger than 4
millimeters  and the  quantity of material  finer  than 62.5  vm  ranged between
0.5 percent and 5 percent.

                      Suspended Sediment Characteristics

    Within  the source area, surface  sediment  is  mechanically  broken  into
smaller  size fractions  by  the  activity of bulldozers  and  scrapers.   This
activity occurs routinely  and  prevents the development of a surface  crust by
cementing agents  or  an ablative surface  where fines are selectively  winnowed
and  coarser  materials stabilize  the  surface.    The  finer fractions may  be
eroded solely by  winds, or by a combination of winds and vehicles perturbing
the  surface  and ejecting  particles  upward  into  regions  of  higher  wind
velocity.  Airborne  particles will  remain in suspension  if wind turbulence
can overcome gravity effects.   The modes of transport include  surface creep,
saltation and  suspension.

    TRC's  study  focused  only  on  material  leaving the source  area in  the
suspension mode.   This material was analyzed  to determine the  distribution
of grain  sizes.  (Histogram plots of  samplers located  at  downwind  distances
of 219  meters, 481  meters and 612  meters are presented  in Figure  2.)   The
data  indicated  that  there was  a  paucity  of  material  in   the  1.3-um   to
4.2-wm  size  range.   Also,  a  slight decrease  was  observed  in  the  greater-
than-10-um  size  range  with   increasing   downwind distance.   This  trend,

                                2-6

-------
  100
3
i

I*
i
                 FIRST ROW
  100
   •0

 §
 i
 5
 r
                                               SECOND  ROW
  100
i
i

I
*

l
                                               THIRD ROW
 STAGS 1


10.2 » 1C *
4.2
2


10.2 j*
    3      I
2.1 - «.2 !•  I
                                          1.3 - 2.1 IN
  BAOCUP


0.0 - l.J
                  HIW-WUJHE STASE MO O>U)£5POmiM GRAIN SIZE
     Figure 2.  Grain size distribution for center!ine samplers,


                             2-7

-------
coupled  with  a  slight  increase  in  the  finer   fractions  with  downwind
distance, is believed to reflect deposition of coarser materials.

    Observed  dustfall with  downwind  distance  is  illustrated  in  Figure  3.
The bulk of the material was deposited within 50 meters of the source  where
the  average  deposition  rate  was 11 grana/m* /min.   The  deposition  rate
decreased  to  1.7  grams/rf /ain and  1.0 grams/a* /min at  downwind  distances
of 160 meters and 380 meters, respectively.

                           Source Emission Strength

    Bach TSP measurement was used  to  back-calculate source emission  strength
using Gaussian diffusion equations.   To solve  these equations  for  the source
strength term,  the source  area was  represented by a  virtual point  source
which was  located using Turner's  techniques.   The  results for the  complete
tests (nos. 3, 6, 7,  8 and  9)  are  presented  in Table IX.  TSP  concentrations
were corrected for  background contamination  by subtracting the  upwind sample
concentration   value  from  the   value  observed  downwind.    Background
concentrations  were  low   (approximately  30  ug/m* )  and  did  not  change
significantly  during  the   test  period.   The  downwind   TSP   concentration
decreased  dramatically between samplers located  at line 1 and line  2.   The
apparent  source  strength   values  were  also corrected  to  account  for  the
depletion  of  the dust cloud as a result  of  the dry deposition of  airborne
particulate  matter.   The   calculated correction   factors  for   the  three
downwind  distances   are  1.24,  1.37,  and  1.46   for   lines  1,   2  and  3,
respectively.

                          Estimation  of Program Cost

    Program costs were  estimated by  grouping those tasks required  to attain
a  specific objective and by developing man-hour and other direct  costs for
each  group.   Table III illustrates  how tasks were grouped  to estimate the
costs  associated with using the  upwind-downwind test method.   The  relative
importance of  each component identified in Table III varies among  programs.
For  example,  the  proximity  of   the monitoring  site  to  the organization
completing  the tests affects  travel expenses.   However,  basic assumptions
may  be made  and costs  developed for  variously sized  programs.   Figure  4
compares the  total  program  cost with  the  number of  high-volume  samplers used
to monitor the  fugitive  emission source.   The increase  in cost  is a step
function  owing  mainly  to  the addition  of  field   personnel to  handle  the
increased  work load.
                                   2-8

-------
 400
  SO-
 300
  100-
  M
                     lib-
40     80
  160    200    240    280
OMMIBO OISTMCT (mtun)
Figure 3.   Comparsion of total dustfall collected by sampler with
            downwind distance from source.

M
40
m
i *
1"
10
0
i
•
•

"
•
•









i























j































i
































































1
•

1
•
•

       0   1   2  34   S   6  7   8   9   10  II  12  13  U  IS  IS  17  18
                                  HI-VOC  SMPUXS

  Figure 4.  Comparison of total program  cost and the number of high-
             volume samplers using  the  upwind-downwind method to
             determine fugitive  source  emissions.
                                     2-9

-------
                                                 TABU  II.   BatlMtad Bmlaalon B«taa (or *ba Aiaa Bouroa
DO
 I
T«it fl fast 1C Taat ff Taat IB »•« ft
Dlatanoa Bat IM tad Dlatanca Batlaatad Dlatanoa BatlMtad Dlatanoa Batlaatad Dlatanoa Bit luted
ttom Bnlaalon fro* Bnlaalon titm Bataalon fro* Eal>«lon '«>• Balaalon
Bowroa* R*t« Bouioa* IUt* Bouroa* B»k* Bowroa* Bat* Booroa* Bat*
at. t Ion (») |q/a| |.| (4/a| (•! la/a) !•! |9/«1 !•! <9/«l
11 111,
12 HO.
i> lot.
14 2)5.
IS 2i«.
21 171.
22 421.
2) 4S1.
24
2S 470.
It 47a.
17.
7.
7.
IB.
It.
27.
It.
IB.
—
14.
32.
27
31 9B7.4) IB.
)) CBB.B l».
14 «B7.t IB.
35 4««.l 34.
177.
1»4.
211.
23B.
2J».
—
4M.
41*.
471.
472.
477.
—
•»^
•04.
iO>.
iO«.
SB.
34.
32.
4S.
S3.
—
sc.
47.
33.
34.
SS.
—
—
3*.
44.
2».
1B4.
201.
21*.
24t.
2t7.
3B3.
432.
4a3.
411.
4*1.
St.
37.
24.
4B.
ii.
S3.
St.
St.
42.
S3.
1B3.
201.
21».
247.
2tl.
3B1.
431.
4(2.
401.
403.
4Bt.
4»S.
—
•11.7 7S.4 til.
•1B.S 43.2 t2B.
«l».
13B.
70.
S7.
•4.
1«2.
IDS.
<7.
Si.
3».
57.
Ot.
IBS.
—
St.
t3.
1B4.
1B7.
20t.
134.
253.
175.
303.
434.
4tS.
40t.
40*.
4*B.
sot.
—
~
•27.
• 3B.
**.
71.
40.
•3.
04.
4t.
4B.
31.
3t.
J7.
SO.
•1.
—
—
IS.
St.
















         •Olatanoa (roa »ouroa*  !• tka  41«tanoa !• aiatais  batwaam tk«  •aaylaca and the  victual point
         batwaan taata baoauaa tha poaltlon of tha virtual point aoucoa Is a (unotlon o(  ataoaphado atabllttv
TklB
v«(l««

-------
                    TABLE III.   Components of program cost
    o  Management and program administration
    o  Development of test plan
    o  Calibration and preparation of equipment
    o  Travel to/from the monitoring site; shipment of equipment
    o  Data collection
    o  Checking and cleaning equipment
    o  Data reduction
    o  Data analysis
    o  fteport preparation
                   Determination of Optimum Sampling  Protocol

    To determine  the optimum sampling protocol, estimates  of  source emission
rates were  analyzed using standard statistical  techniques.  A summary of the
results  for each  test is  contained  in  Table IV.   The values  for skewness
range from  a low of 0.066  for  test  7  to a high  of 0.662  for  test 8, which
indicates  a slight to moderate bunching  of  the  smaller  values  about  the
mean.  The  values  for  kurtosis range from a  low  of -1.244 for test  6  to a
high of  1.485 for  test 7.   Hone of the values for kurtosis are indicative of
a significant departure from normality.


          TABLE IV.   Statistical analysis of emission rate estimates

             Number               Standard    Coefficient
  Test         of          Mean   Deviation       of
 Number   Observations	(g/s)    (g/s)	Variation   Skevness  Kurtosis
3
6
7
8
9
14
12
12
15
14
29.9
57.4
68.8
113.0
77.1
13.6
12.6
19.8
41.8
28.1
45.6
21.9
28.7
37.0
36.4
0.382
0.102
0.066
0.662
0.639
-0.096
-1.244
1.485
-0.321
-1.047
    The  wide  range of  values  exhibited  by  the  coefficient of  variation
indicates  that the mean and  standard deviation  do not change together, and
for that reason the data from the five tests could  not  be  combined to obtain
the desired  solution.   Instead each  test was analyzed  independently and the
results were  compared  to determine  the  variability of  relative accuracy for
sampling arrays of  different  size.   The data values from  the five tests were
combined,  however,  to determine  whether  the sample  population  exhibited   a
normal distribution and  whether  any  significant differences  exist among the
data  values  because  of  the  techniques employed   in  estimating  the source
emission rate.

                                 2-11

-------
    Test for normality.    Because the mean  of the  emission rate values wert
different  for  each  test,  the  values were  standardized  separately  for each
test using the following expression:


         Q   ~ Q
    Zt m-H - 1       i - 1, 2, 3r  ... n                                  U>
where:
         Q. . represents the emission rate values

             is the sample mean for the Q values  for test j
          S .  is the sample standard deviation of  Q values in test j
          n.  vas the Dumber of Q values  in  test j


The  standardized values for  the five tests were  combined and grouped in nint
categories.   A  chi-square -goodness-of-fit'  test indicated  that  the  Zj/s
were values  from a normal distribution.

     Analysis of  variance.    A  univariate  analysis  of  variance  was  also
performed  on the  data  values   obtained  for  each  test.   The  array  of
high-volume  samplers  was divided  into three lines and  the samplers  in each
line grouped into  five positions.   This division resulted in  the layout shown
in Table V.


               TABLE V.   Layout of samplers by sampler -ID number
Line
1
2
3

1
11
21
31

2
12
22
32
Position
3
13
23,24,25
33

4
14
26
34

5
15
27
35
 For each test,  the following model for the Q values was employed:

     QijK -     1 +ai +flj * WJij •»• eijk      i - 1, 2, 3
                                                j • 1, 2, 3, 4, 5            (2)
                                                k • 1,

 where:
              7    is the general mean

              ai is the effect due to line i

              8 j is the effect due to position j
                                 2-12

-------
           (aft ) ij is the effect due to the interaction of line i with
                 position j
            *ijk I* *&* error term

                 i* the number of observations in line i and position j
The  e^  values are  assumed  to be  independent  random  variables normally
distributed  with a  aean  of zero  and a  variance  of a1 .   Also,  all three
effects are assumed to be  fixed effects.

    An analysis  of variance  was performed  for the five tests  (3,  6, 7, 8 and
9).  For all  tests except test 9, there was  no  significant effect because of
the  interaction of  line  and  position.   An analysis of  variance  was  also
performed  on  the data for all five  tests  combined.  For this  case there was
not  a  significant effect  because of  the  interaction  of line  and position.
There was  also  no difference among the  lines,  but there was  a  significant
difference among the positions.  Duncan's multiple range test  was performed
on the  position means using an  alpha  level of  0.05   (se« Table  71). 4  This
test showed that position 5 was  significantly different  from positions 2 and
3. 'For all tests combined,  the ranking of the  position means from largest to
smallest was  5,  1, 4,  2  and 3.  This ranking indicates a trend for the outer
positions  in the rows to have a higher mean than  the inner positions.


 TABLE VI.   Duncan's multiple range test for estimated emission rates

 Sampler         Mean Value of Estimated Emission Rates (g/s)        All Tests
 Grouping     Test 3     Test 6  ,  Test 7     Test  8     Test 9     Combined

 All samplers  29.9        57.4       68.8       113. 0       77.1        70.1
Line 1
Line 2
Line 3
Position 1
Position 2
Position 3
Position 4
Position 5
17.5
33.7
40.6
38.4
17.6
24.6
31.7
37.3
53.7
62.4
55.2
62.0
60.0
51.4
65.9
54.9
58.7
71.8
86.6
73.3
62.2
69.8
61.5
82.8
131.9
101.2
109.0
157.8
95.0
72.7
109.2
166.1
95.8
65.9
69.6
93.7
71.8
51.8
74.5
105.9
71.5
69.3
69.1
82.7
61.3
55.4
69.1
98.4
    The difference  between the positions  within the  sampling  array suggests
that  representing  the area of fugitive emissions  by a  virtual point  source
does  introduce a  bias  into   the  estimates  of emission  strength.  Samplers
located  farthest  from the plume centerline  estimate higher  source emission
rates  than  does the population mean.   Conceivably, the  source  area might  b«
better  represented by  a  line source  with  segments  weighted  differently  to
reflect  the variability  in  emissions  within the  source  area.   An  iterative
process may be used to  determine  the proper weighting  factors  for  each  line
segment to  eliminate  the bias in  the data set.  However,  this technique would
require a different description for each  test  of  each source with  the  result


                                2-13

-------
that determining  the  emission  characteristics  of the  line  source would  be
difficult for programs that attempt to use a reduced number  of samplers.

    Relative Accuracy  Os ing  Reduced Humber  of Samplers.    It was  desired  to
determine the number of samplers  that  would be  required to estimate  the  mean
within plus or minus  (100  z  r)  %, where  r  is a number  between 0 and  1.   The
coefficient of variation  was employed  to obtain  the  desired  estimates.   The
coefficient of variation  from each  test was  treated  as  a  datum point.   The
mean of five values was 0.339 and the  standard error  of this mean  was 0.112.
A 95 percent confidence interval  for the mean coefficient of  variation may be
obtained by the following expression:
where:
               x is the mean coefficient of variation

         ta/2, v is the value of Students' t -distribution with a
                 probability level of a/2 and degrees  of freedom,  v

            s/V"n"is the standard error of the mean

         For this case a » 0.05

         v • n-1 « 4, and the value of t is 2.776.

The  lower  bound of the  95  percent confidence  interval on  the mean  is. 0.228
and  the upper bound is 0.451.
                         •j.
     For any number of samplers  (n)  the  following equation may be  solved for a
value of r:


      m ^72, n-1  (c.v.)
               /"n~"                                                         * '

where:
         e.v. is a value of the coefficient of  variation

Using  successive values  of n  from 4  to 25  inclusively,  values  of  r  were
obtained for  the mean coefficient of variation in addition to the  upper and
lower confidence  bounds for the coefficient of variation.  To illustrate the
results in terms of  relative  accuracy,  the number of  high-volume  samplers  (n)
were tabulated  with values of  1-r (see Table  VII) .   The  total  program  cost
for   incremental   increases  in  program  size   were   compared  with  relative
accuracy  using  the  mean  coefficient  of  variation.   These  results   are
presented in Figure 5 and show  that the most cost-effective program would use
10 downwind samplers.

Conclusions

     A study was  conducted to  determine  the effect of  varying  sampling  network
size on the  cost and  relative accuracy  of  the information  obtained.   The
results of the  study indicate that the  most cost-effective program would  use
10 downwind samplers  and would provide, when  deployed in accordance with  the
guidelines of  Kolnsberg, an estimate of  emission strength  accurate to within

                                  2-14

-------
approximately   25  percent   of  the   value  possible   using   30   or  acre
samplers.       These  results,  however,  should  only  be  assumed  valid  for
similar fugitive emission sources.  Additional work  is necessary to determine
the  minimum  size  sampling   network  for  more  complicated  source areas  or
prevailing meteorological conditions.

                    TABLE VII.  Values  of  1-r for «  of 0.05
n
4
5
6
7
8
9
10
11
12
13
14
15
16
17

18
19
20
21
22
23
24
25

3.182
2.776
2.571
2.447
2.365
2.306
2.262
2.228
2.201
2.179
2.160
2.145
2.131
2.120

2.110
2.101
2.093
2.086
2.080
2.074
2.064
2.056

Lower
Limit
C.V. - 0.228
.637
.717
.761
.789
.809
.825
.837
.847
.855
.862
.868
.874
.879
.883

.887
.890
.893
.896
.899
.901
.904
.906
1-r Values
Mean
C.V. • 0.339
.461
.579
.644
.689
.717
.739
.758
.772
.785
.795
.804
.812
.819
.826

.831
.837
.841
.846
.850
.853
.857
.861

Upper
Limit
C.V. • 0.451
.282
.440
.527
.583
.623
.653
.677
.697
.713
.727
.740
.750
.760
.768
i
.776
.783
.789
.794
.800
.805
.810
.815
                                  2-15

-------
to

0>

        Ul
        »—I
        t_l
        •-•
s  v
i
         1.0
        0.0
        0.6
        0.4
          0
                   20
                                                            n
                                                            ''

30                      40

       TOTAL PROGRAM COST X 1(T
                                                                            •  "

                                                                                   50
         Figure 5.  Comparison of relative  accuracy (utilizing mean coefficient of variation)
                    to total program cost.   Numbers denote the quantity of downwind samplers.

-------
                                  References


1.  O.B.  Turner,   Horkaook  of  Atmospheric   Disunion  Estimates.    U.S.
    Environmental Protection Agency  AP-26.   1970. (NTIS PB191482.)

2.  O.B.  Slade,  Ed.,  Meteorology  and  Atomic  Energy.   U.S.  Atomic  Energy
    Cosniaaion, July 1363.  445 pp.

3.  Environa«nt*l  Protection  Agency,  Quality  A««urance  Handbook   for  Air
    Pollution Measurement Systems.  EPA-600/4-77-027A.1S77.(NTIS PB273518.)

4.  O.B.  Duncan,  *t-tests  and interrala  for  comparisons  suggested  by the
    data.'  Biometrics 31: 339-59   (1975).
                                  2-17

-------
CHARACTERIZATION  OF FINE PARTICULATE  EMISSION FACTORS
                      FOR PAVED ROADS
                             by

                   Chatten Cowherd, Jr.
                   Phillip J. Englehart
                Midwest Research Institute
                   425  Volker  Boulevard
               Kansas City,  Missouri  64110
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
                             3-1

-------
                                   ABSTRACT
     This paper  presents  the results  of an  expanded measurement  program to
develop  emission  factors  for  particulate   emissions  generated   by  traffic
entrainment of paved road  surface  particulate matter.  The emission sampling
procedure used  in this program  provided emission  factors for  the following
particle size ranges:   <  15  ym, <  10  pm,  and < 2.5  urn  aerodynamic diameter.
Testing was performed at sites in  the  Kansas  City  and St. Louis areas.  These
sites  were  representative of  significant urban paved  road  emission sources
within the following land  use categories:   commercial/industrial,  commercial/
residential,  expressway, and  rural  town.

     The measured  inhalable  particulate emission factors  ranged from 0.06 to
8.8 g/VKT.  Lowest emissions  were measured for the  "expressway"  road category;
highest  emissions   were   measured  for  the  "rural  town"   road  category.
Approximately 90% of the IP emissions (< 15 urn aerodynamic diameter) consisted
of particles smaller than 10  ym in aerodynamic diameter,  and approximately 50%
of the IP emissions  consisted of particles  smaller  than  2.5 ym  in  aerodynamic
diameter.

     Correlation   analysis   of   particulate   emissions   with   parameters
characterizing  the  source conditions   showed  the   existence  of  a  relatively
strong  positive  relationship   between  intensity   of emissions and  roadway
surface silt loading.   This  relationship was  used  as  the  basis  for derivation
of predictive emission factors  for each particle size range.  The equation for
IP emissions was found to represent measured IP emissions  more accurately over
a much larger range of values than does the AP-42 single-valued  factor.

     To facilitate  the use of these  particle size specific  equations in the
development  of  emission   inventories,  a  classification  system  of mean  or
typical silt  loadings  as a  function  of roadway category  was  derived.   These
mean  silt  loadings  were  then   inserted into  the  respective  emission  factor
equations  to  derive  a  matrix  of emission  factors  for  specific  roadway
categories and particle size  fractions.
                                      3-2

-------
                                 INTRODUCTION


     Traffic-entrained particulate  from paved roads has  been  identified as a
major  cause  of  nonattainment  of  air  quality  standards  for  total  suspended
particulates  (TSP)  in urban  areas.1    Therefore,  the quantification  of this
source  is  necessary  to  the  development  of  effective   strategies   for  the
attainment and  maintenance  of the  TSP  standards, as  well  as  the anticipated
standard  for  particles  smaller  than   10  micrometers   (urn)   in  aerodynamic
diameter.

     Few  data  on  directly measured  dust  emissions  from  paved  streets  are
available in the literature.  An isolated  study of dust emissions from a paved
road in the Seattle area yielded an emission factor of 0.83 Ib/vehicle-mile at
20  mph.2»3   The  test road was  noticeably  dusty  and  had no  curbs  or street
cleaning program; it  was  located adjacent to gravel  roads and unpaved parking
lots  from which  dirt was  tracked.   Dust  emissions  generated  by  vehicular
traffic with  average daily traffic  exceeding 200 vehicles were  estimated to
equal the amount removed by sweeping every 2 weeks.3

     A  single-valued  emission  factor   of 3.7  g/vehicle-kilometer  for  dust
entrainment  from  paved  roads  was  developed  from  another  field  study.1*
Emission measurements were  obtained using the  upwind-downwind technique with
high-volume samplers.   Thirty-five successful  tests were completed.   It was
determined through microscopy  that  78%  (by weight) of the emissions consisted
of  particulate  less  than  30 ym  in  size.  Also  through optical microscopy, it
was found that  59% of the particulate  collected was mineral matter, while 40%
consisted of  combustion  products.   It  was also concluded  in  this study that
particulate emissions from a  street are  proportional  to traffic  volume but
independent of  street  surface dust loading.

     In  a   third  field   study,  quantitative   emission  factors  for  dust
entrainment from paved urban roads were developed using  exposure profiling.s
Field  testing was  conducted at  three representative  sites  in  the Kansas City
area.   At  one location,  controlled amounts  of  pulverized top  soil and gravel
fines  were  applied to the  road  surface.   Eight tests were performed at the
artificially  loaded  site,  and  five tests  were  made  at a  different site under
actual traffic conditions.  Emissions were found to vary directly with traffic
volume  and  surface  loading  of  silt  (fines).   The  dust  emission  factor for
normally  loaded  urban  streets   ranged  from  1  to   15  g/vehicle-kilometer,
depending upon  land  use.  Approximately 90% of  the emissions  (by weight) were
found  to  be  less  than 30  ym  in Stokes diameter and  50%  less than  5  ym in
Stokes diameter, based on a  particle density of 2.5 g/cm3.  Measured  emission
factors  for  street  particulate  reentrainment  added  to  vehicle  exhaust were
found to be an  order of magnitude larger than the factors for vehicle exhaust
alone.6

     This document presents the  results of an expanded measurement program to
develop particulate  emission  factors  for  paved  roads.  The emission  sampling
procedure used  in this program  provided  emission factors for  the  following
particle size ranges.
                                     3-3

-------
        IP  =  Inhalable  particulate  matter consisting of particles smaller than
               15  ym  in aerodynamic diameter

     PM-10  =  Particulate matter consisting of particles  smaller than 10 ym in
               aerodynamic  diameter

        FP  =  Fine  particulate  matter  consisting  of  particles  smaller  than
              .2.5 ym in aerodynamic diameter

Results are  presented  for  winter testing  in the Kansas  City,  Missouri, area
and  spring  testing  in  areas  of  St.   Louis,   Missouri,   and  Granite  City,
Illinois.   These results are used as a basis for the derivation of a matrix of
emission factors for  specific road  categories and particle size ranges.


                           SAMPLING SITE SELECTION


     Eight candidate  sampling areas  in Kansas,  Missouri, and  Illinois were
designated by  the  Environmental  Protection Agency  (EPA) as  representative
sites for  the  field  study.   These areas represented  a range  of typical road,
traffic,  geographical,  and   environmental   conditions  within  residential,
commercial,  and industrial  land  uses.    Each  sampling  area  contained  a TSP
monitoring site providing historical air quality data.

     A  wide  variety  of road  and  traffic  characteristics was found  in the
presurveys of  these  areas.    Equivalent  hourly traffic  volume  ranged  from
36 vehicles  to  2,944 vehicles.  Road width varied from  22 to 216  ft.  Both
asphalt  and  concrete  street  surfaces,  curbed   and  uncurbed,  were  included.
Street surface conditions ranged from smooth to  rough, and surface particulate
loadings varied from  light  to  heavy  in  comparison  with  typically observed
loadings.

     Three major  criteria  were used  to  determine  the  suitability  of  each
candidate  site  for sampling  of road dust  emissions  by the exposure profiling
technique.7

     1.   Adequate space for sampling equipment.

     2.   Sufficient traffic and/or surface dust loading so that adequate mass
would  be  captured   on the  lightest  loaded collection   substrate  during  a
reasonable sampling time period.

     3.   A  wide  range  of  acceptable  wind  directions,  taking  into   account
(a) the street orientation relative to the predominant wind directions  for the
locality, and  (b) upwind obstacles  (houses,  buildings, or trees) to free wind
flow.

Although roads with light traffic were excluded  from consideration,  such  roads
probably  do  not  contribute  substantially  to  total  emissions  of   traffic
entrained" dust  in urban areas.
                                    3-4

-------
     Based on  the above criteria,  nine  sites were' selected  for this testing
program:

     Kansas City Area - three sites

     7th Street in Kansas City, Kansas (commercial/industrial)
     Volker  Boulevard/Rockhill  Road   in  Kansas  City,  Missouri  (commercial/
       residential)
     4th Street in Tonganoxie, Kansas  (rural town)

     St. Louis, Missouri - two sites

     1-44 (expressway)
     Kingshighway  (commercial/residential)

     Granite City. Illinois - two sites

     Madison Street (commercial/residential)
     Benton Road  (commercial/residential)


                              SAMPLING EQUIPMENT


     A  variety of sampling  equipment was  utilized in  this  study  to measure
partfculate  emissions,  roadway  surface  particulate  loadings,  and  traffic
characteristics.

     The  primary  tool  for  quantification of  emissions was  the  MRI exposure
profiler,  which  was   developed  under EPA  Contract  No.  68-02-0619.7    The
profiler  consisted of  a  portable tower  (4  m height)  supporting an array of
four sampling  heads.   Each sampling  head  was operated as an isokinetic total
particulate  matter exposure   sampler  directing  passage  of  the flow stream
through  a settling chamber (trapping  particles larger  than about  50  ym in
diameter) and  then upward through a  standard  8- by 10-in. glass fiber filter
positioned  horizontally.    Sampling  intakes were  pointed into the  wind,  and
sampling  velocity of  each  intake was adjusted  to match  the  local  mean wind
speed, as determined prior to each test.  Throughout each  test, wind speed was
monitored by recording  anemometers at  two heights,  and  the vertical profile of
wind speed  was determined by assuming a  logarithmic distribution.   Normally,
the  exposure  profiler was  positioned  at  a distance of 5 m from the downwind
edge of the road.

     The recently  developed EPA  version  of the size selective inlet  (SSI) for
the  high  volume air sampler was used  to  determine the IP concentrations.  To
obtain  the particle   size  distribution  of  IP,  a  high-volume  parallel-slot
cascade  impactor  (CI) with  greased  substrates was   positioned beneath  the
SSI.   This  five  stage cascade impactor has,  at a flow  rate of  40 SCFM, 50%
efficiency  cutpoints   at  7.2,  3.0,  1.5,  0.95,  and  0.49  ym  aerodynamic
diameter.   SSIs fitted with  high-volume  cascade impactors were  placed at 1-
and  3-m  heights to determine  the  respective IP and  FP mass  fractions of the
total particulate  emissions.
                                    3-5

-------
     Standard  high-volume  air  samplers'  were  used  to  measure   TSP matter
consisting of  particles  smaller  than  about  30  ym in  aerodynamic diameter.
These samplers were operated at a height of 2 m.

     The basic upwind  equipment  included SSIs  and  a standard high-volume  air
sampler.   In  the  Kansas City testing,  two  SSIs at  heights  of  2 and  4 m were
used  to obtain the  IP concentration  of upwind  particulate matter.   In  the
St. Louis  testing,  the primary  upwind  equipment included  a high-volume  air
sampler and  an SSI/CI  with greased substrates.   For the  secondary  deployment
array, two SSIs were used to obtain the vertical distribution of IP.

     Samples of the  dust found on  the roadway  surface  were collected during
the source tests.  In  order to collect this surface dust, it was necessary to
close  each  traffic lane  for a  period  of approximately  15 min.   Normally,  an
area  that was  3   m  by  the width  of  a lane  was   sampled.   For  each test,
collection  of  material  from all travel  lanes  and curb  areas  (extending  to
about  25  to  30  cm  from  the  curbing)  was attempted.   A hand-held  portable
vacuum  cleaner was used  to  collect  the roadway dust.   The  attached  brush  on
the collection inlet was  used  to abrade  surface compacted dust and to remove
dust  from  the  crevices of the road  surface.   Vacuuming  was  preceded  by broom
sweeping if  large aggregate was present.

     The  characteristics  of the  vehicular  traffic  during the  source  testing
were   determined   by  both  automatic  and  manual  means.     The  vehicular
characteristics included:   (a) total  traffic  count,  (b)  mean  traffic speed,
and (c) vehicle mix.

     Total vehicle count  was determined by  using pneumatic-tube counters.   In
order  to  convert  the axle counts to total  vehicles, visual  1-min vehicle  mix
summaries  were tabulated  every  15 min  during  the source  testing.  The vehicle
mix  summaries  recorded vehicle type,  number  of vehicle axles,  and number  of
vehicle wheels.   From this information, the total  axle  counts were corrected
to the  total number  of vehicles by type.

     The  speed of the traveling vehicles was  determined by  noting the posted
speed   limits  of  the  roadway  test   section.    As  a   check  against  this
determination  method,  speeds  of  the  vehicles were  determined  through  the
occasional  use  of a  hand-held  radar gun.  The weights  of  the  vehicle types
were  estimated by consulting automobile  literature and distributors  of medium-
duty  and semi-trailer  type  trucks.


                       SAMPLING AND ANALYSIS PROCEDURES


     The sampling  and  analysis  procedures  employed  in  this study were subject
to  the Quality  Control  guidelines  which  met  or  exceeded  the requirements
specified  by EPA.a.9  As  part of the QC program  for  this study, routine audits
of sampling  and analysis  procedures were performed.  The purpose of  the audits
was  to  demonstrate  that measurements  were  made  within acceptable  control
conditions  for  particulate source sampling  and to  assess the  source  testina
data   for   precision  and  accuracy.     Examples  of  items  audited   include


                                   3-6

-------
gravimetric  analysis,  flow  rate calibration,  data processing,  and emission
factor calculation.

     Participate samples were collected on type A slotted glass fiber impactor
substrates and on type AE (8- by 10-in.) glass fiber filters.  To minimize the
problem of particle  bounce,  the glass  fiber  cascade  impactor substrates were
greased.   The grease  solution was prepared  by dissolving  140  g  of stopcock
grease  in  one liter of reagent  grade  toluene.   No grease  was  applied  to the
borders  and   backs   of  the  substrates.    The   substrates  were  handled,
transported,  and  stored  in  specially  designed  frames  which  protected the
greased surfaces.

     Prior to the initial weighing,  the greased substrates  and  filters were
equilibrated  for  24  hr  at  constant  temperature  and  humidity  in  a special
weighing room.  During weighing, the balance was checked at frequent  intervals
with standard weights to assure  accuracy.  The substrates and filters remained
in  the  same  controlled  environment for another 24 hr, after which a  second
analyst reweighed them as a precision check.  Substrates or filters  that  could
not  pass  audit  limits  were  discarded.   Ten percent  of  the  substrates and
filters taken to the field  were used  as  blanks.   Paper bags  for the  vacuum
cleaner were  conditioned and tared  in a similar manner.

     Prior to  equipment deployment, a  number  of decisions  were made  as to the
potential  for acceptable  source  testing  conditions.   These decisions  were
based on  forecast information  obtained from the  local U.S.  Weather Service
office.  A specific  sampling  location  was  identified  based on the anticipated
wind  direction.    Sampling  would   be  initiated only  if the wind  speed was
forecast between  4 and  20 mph.  Sampling  was not  planned  if there was  a high
probability  of  measurable  precipitation  (normally  > 20%)  or  if  the  road
surface was damp.

     Sampling  usually  lasted   4   to   6  hr.     Occasionally,  sampling  was
interrupted  due  to occurrence  of   unacceptable meteorological  conditions and
then  restarted   when  suitable  conditions   returned.     The   unacceptable
meteorological conditions most frequently encountered consisted of light  winds
(below  4 mph)  and  insufficient angle  (< 45  degrees)  between mean (15-min
average) wind direction and road direction.

     To  prevent   p'articulate   losses,  the  exposed   media  were  carefully
transferred  at the end  of  each run  to protective containers within the MRI
instrument van.    Exposed  filters  and  substrates  were placed  in  individual
glassine envelopes and  numbered file  folders  and then  returned to the MRI
laboratory.    Particulate  that  collected  on the   interior  surfaces of  each
exposure probe was rinsed with distilled water into separate glass jars.

     When  exposed substrates  and   filters (and the  associated  blanks)  were
returned from the field,  they were equilibrated under the  same conditions as
the initial weighing.  After reweighing, 10% were audited to check precision.

     The vacuum bags were weighed to determine total net mass collected.  Then
the dust was  removed  from the  bags and was dry sieved.  The screen  sizes used
for the dry  sieving  process  were the  following:  3/8 in., 4, 10, 20, 40, 100,
                                     3-7

-------
140, and 200 mesh.   The material passing a  200  mesh  screen is referred to  as
silt content.

     The vertical  distributions of the product of plume concentration  and mean
wind speed  were numerically  integrated  to  calculate  emission  factors.   The
size selective  inlet/cascade  impactor sampler combinations provided  reliable
point concentrations for  IP  and finer particle  size  fractions.  Plume height
was determined  by  extrapolation of the vertical  profile  of total particulate
concentration as measured by the MRI exposure profiler.


                                 TEST RESULTS


     The winter testing was conducted  during the months of February and March
1980 at three sites in the Kansas City area.  The spring testing was conducted
during  the  month  of May  1980,  at two sites in  St. Louis  and at two closely
spaced  sites in Granite City. Illinois.

     The  source tests  were  evaluated according to established  QA criteria.
Seven  of  the nine  Kansas  City  tests met all  of the  QA  criteria, while only
three  of  the  ten  tests conducted  in the St. Louis,  Granite City area met the
QA  criteria.  The  spring  testing,  in particular, was hampered  by unseasonably
light  wind  conditions.  Wind speed  for  four of  the ten  spring tests did not
meet the minimum wind  speed criterion of 4 mph.

     The  results  of the 10 runs which met  the QA criteria were used  as input
to  Multiple Linear  Regression   (MLR)  analysis   (see below).    These  runs are
subsequently referred  to as the  "MLR" data set.

     During  each   emissions  sampling run and at  other times  when emissions
sampling  was  not  being conducted, samples of  street  surface particulate were
collected to determine total  particulate  loadings and  silt percentages.  Silt
loadings  on active  travel  lanes  ranged  from about 0.022 g/m2 on  a freeway
(1-44)  to more  than 2.5 g/n)2 on a lightly  traveled rural road in Tonganoxie.
As  expected,  loadings  in  curb areas  substantially exceeded loadings in travel
lanes.   The range of  day-to-day variations  in  loadings  at a  given  site was
generally  within   a  factor of  2.    Higher   loadings tended to  occur  after  a
precipitation event.

     The  upwind and downwind particulate mass concentrations  for the various
particle  size  fractions  measured  during the  field program were  analyzed  to
determine  representative  mass fraction ratios.   The IP concentration  measured
downwind  of the test  road  segment was found  to decrease with  height.   At  a
sampling  height of 2 m, the mean ratio of downwind  IP  to TSP concentration was
0.45  (a  = 0.14),  and the  corresponding  mean  upwind  ratio  was  0  54  (0  =
0.18).   This indicates that  background TSP, although  lower in concentration
contains  a  higher percentage of  IP.   Similar differences are  also evident  in
the mean  upwind versus downwind  PM-10  to TSP ratios and FP  to  TSP ratios.

     The  mean  downwind ratio of FP  to  IP was 0.52 (o  = 0.098) while  the mean
upwind  ratio was  0.53 (a =  0.085).  This  finding implies that  there is  no
                                    3-8

-------
significant  enrichment  of  fine  particles  attributable  to  the  paved  road
source.

     Table 1  summarizes, by  land  use category  and test  series  quality,  the
emission  factor  data.   As  can  be  seen, the  smallest emission  factors  were
measured  in   the  freeway  category  which  also  had the  lowest surface  silt
loadings.  The highest emission factor was measured in the rural town category
which showed a correspondingly high surface silt loading.

     Intercomparison  of  emission factors by land-use  category indicates  that
relative to the mean expressway IP emissions:  (a) mean commercial/residential
IP  emissions  were  approximately 10  times  larger;  (b)  commercial/industrial
emissions were approximately  20  times larger;  and  (c)  the rural  town roadway
produced  IP  emissions that were roughly 60 times  larger.  Relative  to  mean
expressway  silt  loading:   (a)  the  silt  loading  for commercial/residential
roadways  was  approximately  25  times  higher;  (b)  the  silt  loading  for
commercial/industrial  roadways  was  roughly  15  times higher;  and   (c)  silt
loading on the rural town roadway was approximately 115 times higher.


                         MULTIPLE REGRESSION ANALYSIS


     Stepwise MLR  was the method used  to evaluate  independent  variables  for
possible  use  as  correction  factors  in a predictive emission factor  equation.
MLR  is  a statistical  technique  available  in the Statistical  Package for the
Social  Sciences  (SPSS).10  Because  it was  desirable to  have multiplicative
rather than additive  correction  factors  in  the  emission.factor equations,  all
independent and dependent variable data were transformed to natural logarithms
before being entered  in the MLR program.

     The  independent  variables  evaluated  initially  as   possible  correction
factors were  silt  loading (g/m2), total  loading  (g/m2),  average vehicle speed
(Kph),  and average  vehicle  weight  (tonnes).    The  rationale for  including
measures  of roadway  particulate  loading  stems from  findings of an earlier MRI
programs which indicated that the magnitude  of  roadway emissions was directly
related  to variations in  surface   loadings.    The vehicle  parameters—mean
weight  and speed—were  included largely  by analogy  to  MRI's unpaved  road
equation,11 although  it was recognized that  the  dust generation mechanism for
paved roads may  differ from that for unpaved roads.   The moisture content of
the  road  surface  particulate  was   not  included  as  a  correction  parameter
because of the difficulty of collecting a sample without altering its moisture
content.

     Preliminary MLR  analysis of the entire  data  set indicated  that all  the
independent  variables  except  vehicle   weight   were   highly  intercorrelated.
Although  the  stepwise  algorithm  would  include  vehicle  speed   first  in  a
predictive equation,  silt  loading  and total  loading showed  essentially  the
same correlation with  IP emissions  (r =  0.60).   In other words, the variables
represent  a  common set of  source  conditions—either  low  vehicle  speed,  high
surface  loading  and  emissions; or high  vehicle speed,  low   loading  and
emissions.
                                     3-9

-------
TABLE 1.  SUMMARY OF PAVED ROAD EMISSION FACTORS
TQCI. IP emission factor
Land use
category
Commercial/
industrial
Commercial/
residential
Expressway
Rural town
a RSD (relative
DH Iv*a1 at i wo
series
quality
All tests
MLR tests
All tests
MLR tests
All tests
MLR tests
All tests
MLR tests
standard
rlowi at- 
-------
     The decision  was  made to  use  silt loading rather  than  total  loading or
vehicle  speed  in  the  development  of  the  emission  factor equation  from the
"MLR"  data  set.   This  decision  was  based on  the  perception that  (a)  silt
loading  is  the  most  physically  plausible  indicator of the magnitude  of IP
emissions,  and  (b) it  will yield  more  reproducible  results  in  independent
applications  than  total  loading,  a  parameter  which  can be  biased by the
presence of large particles (i.e., gravel).

     Including  silt  loading as the primary  predictor  effectively  precluded
total  loading or vehicle speed from entering the  equation for the  "MLR" data
set.    This  follows  from  the  high  intercorrelations  (multicollinearity)
mentioned  above.    Examination of  the  regression statistics  indicated  that
inclusion  of  vehicle weight  as  a  second  correction  parameter could not be
justified.
     The raw MLR equation for the "MLR" data set was as follows:


                               •IP
elp * 4.37 (sL)0'9                           (1)
where:
             eTP = IP emission factor expressed in grams per vehicle kilometer
                    traveled (g/VKT)

             sL  =  Silt loading of road  surface  particulate matter expressed
                    in grams per square meter (g/m2)

This  equation  explained  73%  of the  variation  in  the  emission factors.   As
noted  earlier,  the  "MLR"  data  set did  contain  data  from  all the  land use
categories sampled during the field program.

     The  comparable  predictive  IP  emission  factor equation normalized  to a
typical value for silt loading was as follows:

                            eiP  = 2'54 < -07T )°'8                          (2)


     The  emission factor  equation  was found  to predict the  "MLR"  series test
data with a precision factor of 2.0.  The precision factor (f) for an emission
factor is defined such  that the 68% confidence  interval for a predicted value
(P)  extends  from  P/f  to  Pf.    The  precision  factor  is  determined  by
exponentiating  the  standard deviation of the differences (standard  error of
the  estimate)  between the  natural  logarithms of the  predicted  and  observed
emission factors.

     The  precision factor  may  be interpreted as  a  measure of "average" error
in predicting  IP emissions from the  regression   equation.   Assuming  that the
actual IP emission factors are normally distributed about the regression line,
it can be stated that approximately 68% of the predictions are within a factor
of  2.    The  effective  outer  bounds  of  predictability are  determined  by
exponentiating  twice the  standard  error  of the  estimate.   The  resultant
estimate  of   predictive   accuracy,  in  this  case   4.0,  then  encompasses
approximately 95% of the predictions.


                                      3-11

-------
     To put the precision factor of the IP predictive emission factor equation
emission factor  into  perspective,  two  comparisons were  undertaken utilizing
the  single-valued  emission  factor  found  in  the  current  AP-42  manual.s
However, before valid comparisons  could be made,  it  was  necessary to convert
the  AP-42  single-valued  factor,  which  represents  TSP  emissions,  to  an
approximate  IP  emission factor.    This  was accomplished  by  multiplying the
AP-42 value  by  0.4  which is the mean ratio of  net IP (downwind minus upwind)
to net TSP concentrations as determined from the data collected in this study.

     The first  comparison  involved  the calculation of a  precision factor for
the AP-42 data set.   The resulting value of 2.1 is a measure of the ability of
the  single-valued  factor   to  represent  the  40  pieces  of  data which  were
averaged  originally to  produce  the AP-42  factor.   The second  comparison
involved the calculation of a precision factor using  the  single-valued AP-42
factor  to  represent the "MLR"  data set,  as  collected  in this  study.   This
comparison yielded a precision factor of 4.4.

     The precision factors and the range of the data values (emission factors)
upon which  they are based  are  presented  graphically in Figure  1.   The ideal
model  has  a precision  factor of 1.0,  implying  that each predicted  value is
identical  to the  corresponding observed  value,  over an  infinite  range  of
emission  factors.   The  most  important  conclusion  that  can  be  drawn  from
Figure  1  is  that the emission  factor  equation,  though  far  from ideal,  does
predict  IP emissions more  accurately over  a much  greater range of values than
does the AP-42  single-valued  factor over  a considerably  smaller range of data
values  corresponding to  the AP-42 data set.   Furthermore, application of the
single-valued AP-42  factor to represent the wide range of  IP emissions from
paved  roads, as measured during  this  program, yields a precision factor which
is  more than  double (4.4 versus  2.0)  that associated  with  the predictive
equation.    This  ability   of  the   predictive  equation   to  more  accurately
represent  variations   in   IP  emissions   is   directly  attributable  to  the
relatively  strong relationship  between roadway  surface  silt  loading  and IP
emissions.

     Though  not the primary focus of the  program,  it was  possible to develop
predictive  emission factor equations  for the   PM-10  and  FP particle  size
fractions using the  same procedure  as  that applied in developing the equation
for  IP.  Derivation of TSP  emission factors for use in developing a predictive
equation  required different initial  calculations, since  only two TSP samplers
(one  upwind, one  downwind) were operated  during  the  measurement phase of the
program.   In essence,  the  initial  calculation  involved  multiplication of the
IP  emission  factor  for  each   run  in the  "MLR"  series  data  set  by  the
corresponding net ratio  of TSP  to IP concentration as measured by appropriate
samplers.   This procedure  assumes that the TSP/IP ratio is  constant over the
vertical extent of the plume.

     The  general  form  of   the  emission equation  factors, applicable  to all
particle size fractions, is as follows:

                      p

           e =  k  (§75)              (metric)
                      p
                                        (3)


(nonmetric)                              M\


  3-12

-------
       -  3
       o

       1
        2
       Q_

           2
                                  AP-42 Applied to Present Data
\
                                   Present Regression Equation
                                                             AP-42 Emission.Factor
                                                Ideal Model
                     1    I  I  J I I III      I    I  I  I  I I  I ll       I    I  1  I  I I  I I
            0.01                   0.1                     1                     10

                                    Emission Factor (g/VKT)


                     Figure 1.  Comparison of emission factor precision
                                           3-13

-------
The  base  emission factor  coefficients  (k,  K),  exponent  (P),  and precision
factor for each size  fraction are listed in Table 2.  For the metric equation,
silt  loading  is  expressed  as  grams   per  square  meter;   for  the nonmetric
equation, silt loading is expressed as  grains per square foot.


           TABLE 2.   PAVED ROAD EMISSION FACTOR EQUATION PARAMETERS
                          (by particle  size fraction)


Particle size fraction    k (g/VKT)    K (Ib/VMT)      P     Precision factor3
TSP
IP
< 10 um
FP
5.87
2.54
2.28
1.02
0.0208
0.0090
0.0081
0.0036
0.9
0.8
0.8
0.6
2.4
2.0
2.2
2.2
a  Represents the interval  encompassing 68% of the predicted values.


     It should be  noted  that  the tendency for the power  term  in the equation
to increase  with  larger  particle size  fraction  is generally  consistent with
MRI's previous paved road equation in which  silt  loading  to the 1.0 power was
employed to account for variations in TSP emissions.


                       EMISSIONS INVENTORY APPLICATIONS


     For  the majority  of  emissions  inventory applications   involving  urban
paved roads,  actual  measurements of silt  loading will probably  not be made.
Therefore,  in  order  to facilitate  the  use  of  the  previously  described
equations,  it is  necessary  to characterize  silt  loadings   according  to  a
parameter(s)  more  readily   available  to   persons  developing   emissions
inventories.  After examination  and analysis of silt  loading  and traffic data
collected during  relevant  MRI sampling programs,  as  well as  surface loading
data gathered in connection with an extensive  study of urban  water pollution
the decision was made to characterize variations  in silt  loading based upon a
roadway  classification  system.    This  roadway classification  system  is
presented  in  Table 3.   This  system  generally corresponds to  the  functional
classification systems employed  by  transportation agency  personnel-  and thus
the data  necessary  for  an  emissions  inventory—number of  road miles oer road
category and traffic counts—should  be easily obtainable
                                  3-14

-------
                    TABLE 3.   PAVED ROADWAY CLASSIFICATION
                                  Average daily traffic
          Roadway type                    (ADT)               No.  of lanes
Freeways/expressways
Major streets/highways
Collector streets
Local streets
> 10,000
> 10,000
500-10,000
< 500
> 4
> 4
2a
2b
J Total roadway width > 32 ft.
D Total roadway width < 32 ft.
     It  should  be  recalled  that  traffic  volume  is   not  the  only  factor
affecting roadway  silt  loadings.   For  all roadways that  provide  access  to
immediately  adjacent  areas,  land  use,  particularly  as  it relates  to  the
potential for mud  and  dirt "tracking," is  important.  Silt  loadings may also
be affected  by  street  surface type and condition,  the presence  or absence of
curb, as  well  as  public works practices  and  season of the year.   However,
given the present  data base,  it is not possible  to incorporate  relationships
between these factors and silt loadings in a manner applicable to the majority
of emissions inventories.

     The data base, made up of  44  samples  collected and  analyzed according to
the procedures outlined  above,  may be used to characterize  the  silt loadings
for each roadway category.  These  samples,  obtained during  MRI  field sampling
programs over the past 3 years, represent  a broad range  of urban land use and
roadway  conditions.    Geometric  means for  this   data  set are  broken out  by
sampling location (i.e., city) and roadway category in Table 4.

     The  sampling  locations  can  be considered  representative  of  most  large
urban areas  in the United  States with  the  possible exception of those located
in the Southwest.  Except for the collector roadway category, the overall mean
silt loadings do not vary greatly from city to city, though the St. Louis mean
for major roads is somewhat lower than the other  four cities.  The substantial
variation within  the collector roadway  category is probably  attributable to
the  deposition  effects  of   land  use  associated  with  the  specific  sample
locations.   It  should  also be noted that  an examination of  data collected at
three cites  in Montana during early spring  indicates that winter road sanding
may produce  loadings five  to  six times higher than the  means of the loadings
given in Table 4 for the respective road categories.
                                   3-15

-------
 TABLE 4.  SUMMARY OF SILT LOADINGS  (g/raz)  FOR  URBAN  PAVED ROADWAYS3 BY CITY


                          	Roadway category	
                           Local         Collector       Major         Overall

        City              Xgb    n      Xg     n       Xg     n       Xg     n


Baltimore0                1.42   2      0.72    4     0.39    3      0.68    9

Buffalod                  1.41   5      0.29    2     0.24    4      0.56    11

Granite City (111.)6         -   -         -          0.82    3      0.82    3

Kansas City6                 -   -      2.11    4     0.41   13      0.60    17

St. Louis                    -   -         -          0.16    3      0.16    3

Overall                   1.41   7      0.92   10     0.36   26         -
a  Freeway/expressway data not included; only one value (0.022 g/m2) obtained.
   Xg's are geometric means based on the corresponding n sample size.
~.  Reference 12.
d  Reference 13.
e  From this study.


     Table 5 presents the emission factors  broken  out  by  roadway category and
particle size.  These were obtained  by  inserting  the above mean silt loadings
into the emission factor equations with parameters  defined in Table 2.  These
emission factors can be utilized directly for emission inventory purposes.  It
is important to  note  that the current AP-42 paved  road emission factors* for
TSP agree quite  well  with those developed  in this  study.   For example, those
cited in connection with MRI's previous  testings were conducted at two roadway
sites in  the  major street and  highway  category.   Those tests  yielded  a mean
TSP emission factor of 4.3 g/VKT versus 4.4 g/VKT as determined from the data
presented here.


                           SUMMARY AND CONCLUSIONS


     The purpose of this study was  to quantify inhalable particulate emissions
generated by  traffic entrapment  of paved  road  surface  particulate  matter
Paved road source testing was performed  at sites representative of significant
emission sources within a broad range of urban land-use categories.

o -,-, Th«,,,;easured  1"halable  Particulate  emission factors  ranged  from  0.06 to
8.77 g/VKT.   Lowest  mean  emissions were  measured  for the  "expressway" road
category;  highest  mean   emissions   were  measured  for   the   "rural   town"
category    Approximately  90%  of  the  IP  emissions consisted of  particles
smaller than  10  mn in aerodynamic  diameter,  and approximately 50%  of  the IP
emissions consisted of particles smaller than 2.5 ym in aerodynamic diameter.
                                    3-16

-------
    TABLE 5.   RECOMMENDED EMISSION FACTORS FOR SPECIFIC ROADWAY CATEGORIES
                         AND PARTICLE SIZE FRACTIONS
                           Emission factor by particle size fraction
Roadway
category
Local
Collector
Major street
and highway
Expressway
TSP
g/VKT
15
10
4.4

0.35
Ib/VMT
0.053
0.035
0.016

0.0012
< 15 um
g/VKT
5.8
4.1
2.0

0.21
Ib/VMT
0.021
0.015
0.0071

0.00074
< 10 um
g/VKT
5.2
3.7
1.8

0.19
Ib/VMT
0.018
0.013
0.0064

0.00067
< 2.5 u
g/VKT
1.9
1.5
0.84

0.16
Ib/VMT
0.0067
0.0053
0.0030

0.00057
     Correlation analysis  of  IP emissions with parameters  characterizing  the
source  conditions   showed  the   existence  of  a  relatively  strong  positive
relationship  between   intensity  of  emissions  and  roadway   surface   silt
loading.  This confirms the findings of earlier testing.5  Based on regression
analysis of a subset of acceptable ("MLR")  test runs,  the following predictive
IP emission factor  equation was  developed:
                                             0.8
                                   2.54
                                                            (5)
where:
elp = Inhalable particulate emission factor (g/VKT)

sL  = Road surface silt loading (g/m2)
     This predictive  equation has  an  associated precision  factor of  2.0  in
relation to the "MLR" data set.  By way of comparison, the AP-42 single-valued
factor (corrected to represent IP emissions)  has a precision factor of 2.1 for
its data set and a precision factor of 4.4 for the "MLR" data set,  which spans
a  much   larger  range of  values  than  the  AP-42 data  set.   Therefore,  the
predictive equation, though  far  from ideal,  does represent  IP  emissions  more
accurately over  a much  larger range  of  values than  does the AP-42  single-
valued factor.   This  fact is directly attributable to  the relationship of  IP
emissions to silt loading.

     Extension of the regression analysis to include emission factor equations
for  other particle  size  fractions—FP,  PM-10,  and  TSP—yielded  a  set  of
equations  in  which  the   power  term for  silt  loading  increased  with  larger
particle  size fraction.    This   result  is  generally  consistent  with  MRI's
previous  paved  road  equation in  which  silt  loading  to the  1.0  power  was
employed to account for variations in TSP emissions.
                                   3-17

-------
     To facilitate the  use  of these particle  size specific equations  in  the
development  of  emission  inventories,   a  classification  system  of  mean  or
typical silt  loadings  as  a function of  roadway category was derived.   These
mean silt  loadings were  then inserted  into  the  respective emission  factor
equations.   The resultant emission  factors for specific roadway  category  and
particle size fractions  can  be  utilized  directly for emissions  inventory
purposes.    By  accounting  for variations  in  silt loading,  these  emission
factors are   significantly  more  reliable  than  an overall  average  emission
factor in  developing  components of  an urban paved  road  emission  inventory.
                                ACKNOWLEDGMENT
     The work  upon which this  paper  is based was  performed pursuant  to  EPA
Contract No. 68-02-2814, Assignment No.  32,  and  EPA Contract No.  68-02-3158,
Technical  Directive  No. 19.        Dennis  Drehmel  and       William  Kuykendal
served as EPA project officer for the  study.


                                  REFERENCES

  1.  Lynn,  D.  L., 6.  Deane, R.  Galkiewicz,  R.  M. Bradway,  and  F.  Record.
     National Assessment of Urban Particulate Problem.  Volume  I -  Summary of
     National Assessment.   U.S. Environmental Protection  Agency.  Publication
     No.  EPA 450/3-76-024,  NTIS PB2636.65, July 1976.

  2.  Roberts, J.  W.,  A. T.  Rossano, P.  T. Bosserman,  G.  C. Hofer,  arid  H.  A.
     Watters.   The Measurement, Cost and  Control  of Traffic Dust  and  Gravel
     Roads  in Seattle's Duwamish Valley.   Paper  No.  AP-72-5, Presented  at the
     Annual Meeting of  the  Pacific Northwest  International  Section  of  the Air
     Pollution Control Association,  Eugene, Oregon,  November 1972.

  3.  Roberts, J.  W.,  H. A.  Watters, C.  A. Margold, and A.  T.  Rossano.   Cost
     and Benefits  of  Road Dust  Control  in  Seattle's Industrial  Valley.   Paper
     No.  74-83,  Presented  at  the 67th  Annual Meeting of  the  Air  Pollution
     Control Association, Denver, Colorado, June  9 to 13,  1974.

  4.  Axetell,  K.,  and J.  Zell.    Control of   Reentrained Dust  from  Paved
     Streets.  EPA Publication No. EPA-907/9-77-007, NTIS PB288325, August  1977.

  5.  Cowherd,  C.,  Jr., C.  M. Maxwell,  and D. W.  Nelson.    Quantification  of
     Dust  Entrapment  from Paved  Roadways.   U.S. Environmental  Protection
     Agency, Publication No. EPA-450/3-77-027, NTIS PB272613,  July 1977.

  6.  Compilation  of  Air  Pollutant  Emission Factors,  Third  Edition,  U.S.
     EPA, Publication No.  AP-42, NTIS PB275525, August 1977.

  7.  Cowherd,  C.,   Jr.,  K. Axetell,   Jr.,   C.  M.   Guenther,   and  G. Jutze.
     Development of Emission Factors for Fugitive Dust Sources.   Final  Report
     Midwest  Research  Institute for  U.S. Environmental   Protection  Agency'
     Publication No.  EPA-450/3-74-037, NTIS PB238262, June 1974
                                  3-18

-------
 8.   Quality   Assurance  Handbook  for  Air  Pollution  Measurement  Systems.
     Volume  II  -  Ambient Air Specific Methods.   U.S. Environmental Protection
     Agency,  Publication No.  EPA  600/4-77-027a, NTIS PB273518,  May 1977.

 9.   Ambient  Monitoring  Guidelines  for Prevention  of  Significant Deteriora-
     tion.     U.S.  Environmental  Protection  Agency,  Publication  No.  EPA
     450/2-78-019, NTIS PB283696, May 1978.

10.   Nie,  N.  H., et al.   Statistical  Package for the Social Sciences, Second
     Edition.  McGraw-Hill,  Inc., New  York,  1975.

11.   Cowherd, C., Jr.,  R. Bohn,  and T.  Cuscino, Jr.   Iron and Steel Plant Open
     Source   Fugitive  Emission  Evaluation.   Final  Report,  Midwest  Research
     Institute   for   U.S.   Environmental    Protection   Agency,   Publication
     No.  EPA-600/2-79-103, NTIS PB299385, May 1979.

12.   Cuscino,  T.,   Jr.     Total   Suspended   Particulate  Matter  Analysis  in
     Baltimore,  Maryland.   State of  Maryland,  Baltimore,  Maryland,  October
     1981.

13.   Bohn,  R.    Evaluation of Open  Dust  Sources in  the  Vicinity  of Buffalo,
     New York.    EPA  Contract   No.  68-02-2545,  Assignment  1,  Environmental
     Protection Agency, New York, New  York,  March 1979.
                                3-19

-------
                  MICRON DROPLET DUST SUPPRESSION  PROVES OUT
                    IN VARIETY OF FUGITIVE DUST APPLICATIONS
             The work described in this paper was not funded by the U.S. Environmental
             Protection Agency. The contents do not necessarily reflect the views of the
             Agency and no official endorsement should be inferred.
                      By:  Wayne Hartshorn
                           Executive Vice President
                           Sonic Development Corporation
                           Mahwah, New  Jersey  07430*

                           Lennart Strand
                           President
                           Andeze AB
                           Helsinborg,  Sweden
*  305 Island Road
                                         4-1

-------
ABSTRACT

     In recent years, a little known dust suppression technology called
Dry Fog"* has proven successful in troublesome applications.  Normally a
baghouse or conventional wet-spray system is assigned these problem
applications.  This paper will detail this new technology and examine four
individual applications.  Specifically, applications where it is successfully
controlling fugitive dust emissions in-situ — at 40 percent lower cost vs
baghouses.

     Initially, the technology was applied almost exclusively to rock
product plants, which pose a dusting problem characterized by multiple,
widely dispersed dust generation points.  The individual points are
located throughout the entire materials handling system.  More recently
many quarries, chemical plants and coal handling operations have also
reported success.  These will be discussed in specific terms.

     In light of recent interest in coal conversions in many industries,
operating experience in coal handling operations at Dayton Power &
Light and Westvaco Pulp & Paper will be detailed fully.

     Since Dry Fog Dust Suppression is a relative newcomer, the paper
will examine and discuss its basic principle of operation.  Essentially,
small, micron-size water droplets — the size of airborne dust particles
themselves — are generated to blanket the dust and trigger in-situ
agglomeration.  It is accomplished without discernible wetting and
with no chemical additives.  It is the tiny droplet size that
differentiates this technology from any other.

     The paper will also discuss the technological and economic factors
pertaining to matching the Dry Fog technology to other applications.
Included will be the pros and cons of Dry Fog vs other technologies
commonly considered for fugitive dust control.  Feasibility, capital
costs, operating costs, space and ease of installation, additives and
wetting will be explored.
*Dry Fog™ is a registered trademark of Sonic Development Corporation,
305 Island Road, Mahwah, NJ  07430
                                      4-2

-------
INTRODUCTION

     As you know, wet spray dust suppression has been around for decades.
The Dry Fog concept itself is similar to the scrubbing process where a
spray of water droplets remove dust particles out of an air or gas
stream.  However, in dust suppression, rather than pumping the mixture
from a scrubbing vessel, liquid drops agglomerate with dust particles
and knock them down in-situ.

     Many of these wet spray dust suppression systems — commonly
termed chemical or conventional type systems — are saddled with a
major problem.  While they do an adequate job of treating the material
being handled, often times the droplets sprayed do not evaporate quickly
or completely, thus contaminating the product.  Without quick and
complete evaporation, product wetting is inevitable.  In many materials
handling operations, such as cement, aggregate crushing, and coal
preparation operations, it is essential to keep the product dry.  In
coal handling, for example, moisture directly impairs the heating
value of the coal.  With wet coal, too much heat is dissipated in
evaporating that moisture.  Therefore, less BTCJ's of heat per ton of
coal are available to generate steam.
DROPLET SIZE HINDERS EVAPORATION

     The reason why these water droplets produced by conventional
wet-spray systems cannot evaporate quickly or completely is their
size.  They are too large.  The second problem with large droplets is
that when droplet size greatly exceeds that of the dust particles,
there is very little chance of particle-to-droplet contact to trigger
the desired agglomerating action.  Instead, very little particle-to-
droplet contact actually occurs so the dust particles simply move
around the water droplets (Figure l).l

     At Sonic, however, we have engineered a system that is capable of
producing a superfine atomization of water droplets that greatly
enhances particle-to-droplet contact.  And it evaporates before wetting
anything but the dust.  These atomized water droplets are best described
as fog.  Since it doesn't wet product, call it Dry Fog.

     The system operates on one underlying principle:  By producing
water droplets of approximately the same size as the dust particles,
                                    4-3

-------
Figure 1.  Airflow around large water droplet (top)  prevents coal dust
particles from contacting the  droplet.   The dust particle,  however,
easily impacts a smaller droplet (bottom).
                                 4-4

-------
the probability of collision between the two will be extremely high.
The superfine spray of water droplets rapidly agglomerates with the
individual dust particles and  immediately knocks them out of the
atmosphere, causing them to fall down in-situ.  Also, the droplets
equal the diameters of the dusts over a wide range of operating
conditions and particle sizes.

     As mentioned, the tiny droplet size differentiates this technology
from any other wet-spray dust  suppression system.  The superfine fog
rapidly agglomerates with airborne particulate and immediately knocks
it down, in place.  Knockdown  is quick and it is accomplished without
product wetting.

     The Dry Fog system is capable of consistently producing the uniformly
small droplets over all operating conditions.  This is attributable to
the manner in which they are created.  The Sonicore® spray nozzle is
the key to the Dry Fog system  (Figure 2).  It's actually an air-driven
acoustic oscillator that creates a sonic shock wave which shatters the
liquid, producing very fine droplets.

     In addition to suppressing dust while insuring minimal moisture
addition to the product, the Dry Fog system has other advantages over
conventional methods of dust suppression — including baghouses.  For
example, a Dry Fog system can  be installed for as little as 40 percent
of the capital cost of installing a baghouse and in less than 20 percent
of the time.  Since particle knockdown is achieved in-situ, there's no
need for long duct runs to convey the dust to a central collection
point.  In most cases, it can  also be installed while the plant operates
at 100 percent capacity.  A typical baghouse installation can last up
to 3-6 weeks, and sometimes forces production to come to a standstill.

     The system also offers these advantages over conventional wet-spray
dust suppression alternatives:

     — low water consumption  averaging only one to three gallons
        per hour (per nozzle).  Conventional hydraulic systems consume
        anywhere from five to  15 gallons per hour to perform the
        same task.
                                  4-5

-------
                      RESONATOR CHAMBER
LIQUID
                               SONIC  ENERGY
                                  CORE
                AIR OR GAS
Figure 2.  Schematic of the Sonicore® spray nozzle.
                   4-6

-------
        Low water pressure eliminates the need for costly pumping
        systems.  Usually 30-50 psi supply pressures are adequate.
        The Sonicore nozzle will typically operate at 20 psi.
        Air consumption is only seven scfm per nozzle at 65 psi.
        Most plants have sufficient air available.  Larger systems
        will require supplemental air.
        Almost no water  is added to the process — less than 0.1
        percent; this percentage runs as high as 1-10 percent
        with wet spray systems.
        Costly wetting agents and their associated controls are
        eliminated with Dry Fog.
        Dry Fog will not freeze.  Independent research has proven
        that water droplets less than 30 microns will not freeze
        above minus (-) 40°F below zero.
     — Self-cleaning nozzles.  The Sonicore nozzle has no internal
        screens or filters to plug.  The liquid ports are large
        enough to pass large particles and the sonic action of the
        nozzle keeps it clean.
     — Lowered maintenance costs.  The system requires no handling
        of chemicals and/or surface active agents.
     — No major plant modifications required for installing the system.
        Enclosures, skirting/ and shrouds, similar to baghouse
        requirements, are all that are necessary.
     In addition, many engineers overlook additional factors — such
as cost control, space management, and maintenance and operating costs —
when they search for a fugitive dust control system.  The Dry Fog
system cuts capital costs over a baghouse and some conventional spray
                                  4-7

-------
systems by as much as 20-60 percent.  It also takes up less space and
can be installed in many applications where there isn't room for a
baghouse.  For example, the entire system's space requirement is less
than 50 percent of conventional hydraulic systems, and less than
95 percent of the space of a baghouse.  Only electricity and water
constitute annual operating costs.  Also, because there are fewer
parts, equipment, and no chemical handling, maintenance is minijnal.
CONTROL STANDARDS

     The principle objective of the Dry Fog dust suppression system
is to meet fugitive emission control regulations.  Systems are designed
to control practically all types of dusts that measure one to 10
microns in diameter, as well as larger fugitive dusts that measure
up to 600 microns.  Figure 3 shows average emission concentrations
at a typical aggregate crushing plant located in Sweden, with
and without the system in service.
IN-SITU, ON-SITE

     As mentioned earlier, the main feature of this new technology
involves its ability to ensure that almost no moisture is added to
the product.  Coal handling operations for example, require dry coal
at all times for maximum heating efficiency.  Two very recent
applications, one a midwestern utility plant, the other a southern
pulp and paper mill, both solved major dust problems in their coal
prep operations.  The systems have been successfully controlling
fugitive dust for close to two years and without a trace of wet or
frozen coal.

     Dayton lower & Light Company's Longworth Station needed to comply
with tight EPA regulations.  They had a problem suppressing fugitive
dust in their coal prep operation.  Each time they ran the conveyor
that leads to its coal bunkers dust was generated.

     Conventional wet-spray type dust suppression systems controlled
the dust, but left a residual moisture on the coal.  This left them
with the dual problem of excessive energy costs for coal heating  and
frozen coal during the winter months.  With the wet-coal problem,
some of the heat generated inside the boiler was dissipated in
                                     4-8

-------
                    Figure 3






AVERAGE RESULTS AT SWEDISH STONE CRUSHING PLANT






 Points Treated              Dust Level
                          System Off     System On



 Reclaim Tunnel               6.0           1.2



 Crusher Discharge          190.0          16.0



 Crusher Feed                73.0           2.6



 Screen #1                    2.7           0.4



 Transfer Point              15. '0           1.0



 Screen #2                   11.0           0.9
                    4-9

-------
evaporating that moisture.  Consequently, less BTU's of heat per ton
of coal were available to generate steam.

     Dry Fog eliminated both problems at half the capital investment
compared to another alternate method — a baghouse.

     The Dry Fog system was installed at several points in DP&L's
coal handling gallery.  These points include a screen, a crusher,
three transfer points, and a 20' x 40' open bunker area.  The
installation time required for the Dry Fog system was minimal and no
major modifications of the existing plant facilities were necessary.
Also, Dry Fog virtually eliminated maintenance costs.

     For each gallon of water used in the process. Dry Fog yields a
much greater distribution of micron-sized droplets.  This, of course,
means greater knockdown power since the size of the droplets is
equal or close to the size of the airborne dust particles.

     The resulting dust suppressiqn satisfied the regulations
and brought a cleaner air environment for plant workers.

     For DP&L's essentially intermittent operation, the Dry Fog system
functions only once or twice a day.  The application of Dry Fog has
been so successful that they are considering applying it to other
facilities.

     Westvaco Corporation of Charleston, SC holds the distinction of
being the first full scale coal prep installation of a Dry Fog system.

     During its conversion from oil to coal, plant engineers found
a great amount of dust being generated over the crusher and transfer
towers.  To control this problem, several wet-spray dust suppression
systems were examined.  Finally Westvaco installed Dry Fog spray
bars at the inlets and outlets of its coal crusher and transfer points.

     The system proved superior to a baghouse — a baghouse would
have required duct runs ranging from 150-200 feet between its coal
crusher and transfer tower.  Also, lack of adequate space in its
coal prep room, and the high cost that accompanies a baghouse
installation, steered them toward Dry Fog.
                                     4-10

-------
     Dry Fog's tiny droplets triggered agglomeration so that the coal
dust was knocked down in-situ.  This eliminated the need for long
duct runs to carry the dust to a central pickup point.  And, since
Dry Fog's droplets evaporated quickly, Westvaco experienced no
moisture addition to the coal.
SCIENTIFICALLY PRCVEN

     The emission concentrations at the stone crushing plant located
in Vilhelmina, referenced earlier, are fully supported by a series
of tests conducted by the Swedish National Board of Occupational
Safety and Health.

     The test site was a fluidized stone crushing plant that operates
only six months out of the year due to harsh winter months.
Figure 4 gives a plot plan of the crushing operation showing
the location of the primary crusher, screens and hoppers.

     In this particular crushing plant, raw, materials are crushed in
two stages — primary crushing and secondary crushing.  These two
crushers are also the main dust generation points.  After the raw
materials are crushed they are assorted into two fractional sizes —
0-8 mm and 8-16 mm.  Daily output is about 200 to 250 cubic meters.
Crushing rate is close to 50 percent.

     The Dry Fog system was tested over a three day period to determine
its efficiency as a dust control method.  Samples were taken from the
test site with Dry Fog equipment in service for two days.  The third
day of testing was conducted without the Dry Fog system in operation.
Temperature and wind velocity varied over the three days.  Temperatures
ranged from 8°C to 13°C and the wind velocity varied between slight
westerly to northwesterly winds.  Production output each day was
approximately 25-30m3/hr.  Water usage was about 120 liters per
hour.  The system consisted of a total of nine Sonicore nozzles.

     Samples were taken at several points.  The stone
crushing operation's dust levels were measured at the following
points: the reclaim tunnel, where coarse sand is loaded onto a
conveyor; at the jaw crusher; the main bell crusher; the primary and
secondary screens; and a transfer point.
                                   4-11

-------
to
           Finished
           Product
          8-16 mm
                                  Finished
                                  Product
                                  0 - 8 mm
               .^Transfer Point
Screen 2
                                Reclaim
                                Tunnel
           Screen 1
IpTfop Crusher (Bell)

      Bottom Crusher (Jaw)
                                                                            Test Site
                    Raw Material In
                           Figure 4.  Schematic of Swedish stone crushing plant.

-------
     Figure 5 is an overall summary of test results with the Dry
Fog Dust Suppression system in service and without the system
in service.  Dust concentrations were measured in mg/m3.

     These government sponsored tests show that airborne particulate
levels are significantly reduced by the Dry Fog system.  At the same
time, they also show that it accomplished the necessary reduction
without any discernible product wetting.
FURTHER TESTING

     A second, more detailed experiment on the Dry Fog system was
performed at another Swedish crushing site in early 1981.  This was
conducted at an underground mining operation.  Figure 6 is a
plot plan of the mining operation.  Over a four-day testing
period, the system's performance was measured for its ability to
suppress fugitive dusts on a section of this underground mine.
During normal mining operations, the Dry Fog system was activated.
At the same time,'six sampling units measured atmospheric particulate
levels.  Figure 7 is an overall summary of test results with the
Dry Fog system in service.

     Three tests were run during a four-day period.  On the first
day, this particular section of the mine was shut down.  However, the
fugitive dust samplers recorded atmospheric readings for particulate
content for comparison purposes vs particulate levels during normal
operation.  The significance of the overall findings is uncovered in
the three days the mine was in actual operation vs the one day it
was not in service.  The comparison shows that there was no appreciable
difference in airborne particulate levels when the section was
operating with the Dry Fog System in service and the day the entire
mining section along with the Dry Fog system was shut down.

     In the final report, test engineers stated that "there was no
excess of applicable limit values for the total dust volume detected
during the testing period."  This means they didn't exceed the total
dust volume limit allowable according to Swedish local and federal
regulations for the area.  The allowable limit is 10 mg/m3.

REFERENCE
   Schowengerdt, F.D., Brown, J.T.,  "Colorado  School of
   Mines Tackles Control of Respirable Coal  Dust,"
   COAL AGE, April 1976, Copyright 1976 McGraw Hill, Inc.

                                  4-13

-------
                                 Figure 5
            INDIVIDUAL RESULTS AT SWEDISH STONE  CRUSHING PLANT
with Dry Fog
Test:  1
       2
       3

Average
Reclaim
Tunnel

  1.3
  1.0
  1.3

  1.2
                         Crusher    Crusher
                         Discharge   Feed
                          (jaw)     (bell)
 9.3
16.1
22.9

16.0
1.5
3.0
3.2

2.6
Screen
 (#1)

  0.6
  0.3
  0.4

  0.4
Transfer
 Point

   1.4
   1.0
   0.4

   1.0
Screen
  #2

  1.3
  0.6
  0.7

  0.9
without Dry Fog
Test:  1
       2
       3

Average
  5.6
  3.2
  9.1

  6.0
210
190
170

190
84
68
61

73
  3.7
  1.9
  1.9

  2.7
   23
   12
    6.8

   15
  19
   6.4
   2.3

  11
                                   4-14

-------
             Belt Conveyor
                                                       Screen
i
(—•
01
Sampling Unit
                       CT
                                                                 Cone

                                                                 Crusher
                                        Belt Conveyor
                     Figure 6. Schematic of Swedish underground mine.

-------
            Figure 7
TEST AT SWEDISH UNDERGROUND MINE
Points
Treated
System Off
Belt
Conveyor
System On

Screen fl

Screen #2

Transfer
Point 11

Transfer
Point #2

Transfer
Point #3
Total Dust Average
Measured mg/m3 Per Day
0.57
1.51
0.70
1.39
0.65
1.44
0.53
0.59
1.26
1.11
1.03
2.03
1.18
0.76
1.00
2.49
0.68
0.77
1.67
0.91
2.68
3.54
5.13
6.71
1.10
1.29
1.06
1.01
0.78
0.84
0.34
0.35
1.67
0.97
0.55
1.41
1.70
0.72
0.46
2.63
0.75
0.66
0.32
1.39
4.74
3.98
3.38
8.16
1.02
0.66
1.01
1.52
0.78
0.30
0.52
0.29
1.66
0.68
0.88
1.33
1.62
0.69
0.81
2.29
0.89
0.46
0.60
1.22
4.68
2.70
4.18
5.21
0.90
1.15
0.92
1.31
0.74
0.86
0.46
0.41
1.53
1.48
0.82
1.59
1.50
0.72
0.76
2.47
0.77
0.63
0.86
1.17
4.03
3.41
4.23
6.69
Total
Average
1.13

0.58

1.30

1.32

0.89

4.78
                 4-16

-------
           ,~  df CTibcd in this paper was not funded by the U.S. Environmental
          !!!^ ™H SrS£ -^C c?ntents do «* necessarily reflect the views of the
         Agency and no official endorsement should be inferred.
                           FUGITIVE DUST CONTROL STUDIES

                               USING SCALE MODELS AND

                                 MASS  LOSS ESTIMATES*
                           by:   Michael F. Lepage,  Anton E.  Davies
                                 Colin  J. Williams
                                 ROWAN  WILLIAMS DAVIES  & IRWIN Inc.
                                 650 Woodlawn  Road West
                                 Guelph, Ontario, Canada
                                 NIK IBS
(*) Although this is not the actual presentation made at the May 1S82
   meeting,  it closely resembles that presentation, which was
   entitled, "The Optimization of Wind Screens for Fugitive Emission
   Control Using Wind Tunnel Tests, " C. J. Williams (then  of
   MHTR. Ltd.).
                                5-1

-------
INTRODUCTION

     Fluid  model  studies of  bulk storage  facilities provide  an
inexpensive  means  of  testing  dust   control   strategies  before
implementing  them in  the  field.  The potentially  high cost  of
implementing  a  successful  dust  control  program requires  careful
planning to ensure that the best result  is obtained.


     There  has  been  growing pressure  in recent years to  improve
air  quality  around  bulk  storage sites (e.g.  coal  stockpiles,
sawdust  piles and  fly ash  lagoons)  by  reducing  dust  emissions
(see  Figure 1).   Improved  air  quality,  however,  is only  one  of
the benefits  of  a dust control  program, since dust  losses often
represent  losses  of  valuable inventory.   In many cases,  the cost
of  implementing  dust  controls  is more  than offset  in  the  long
run by the  reduction in lost inventory.


     There  are  two  principal approaches to  dust control  at bulk
storage  facilities.   The first  is to  clean  up the  site,  compact
loose surfaces  and  encrust  exposed dust which  would  be  available
for   emission.    The  second  approach   is   to  reorganize  the
stockpiles  and  erect  wind  breaks in  order  to reduce local wind
speeds   below  the   threshold   of  dust  emissions.    The  most
efficient  dust  control  program is  typically a  combination  of
both approaches.


     Fluid model  studies   provide  a  means  of  optimizing  dust
control  while avoiding costly trial  and error in the  field.   By
examining   local  wind  conditions on  a  scale  model in   a  wind
tunnel  or  water  flume,  one can  identify  the  windiest  and hence
most  sensitive  areas  of  the site and can evaluate  the relative
merits  of  different schemes  for localized wind speed  reduction.
To  give   further  meaning  to   the  results,  actual  samples  of
material   from   the   site   are   tested  in   the  wind  tunnel  to
determine  the relationship  between  wind speed  and  the  level  of
dust  emission  (which  can  vary   significantly  from  one  site  to
another).   By feeding  this information together  with  data from
the   scale  model  tests  into   a computer  simulation   of  dust
emission,  wind  reduction schemes  can then be  evaluated in terms
of  their  impact  on  overall  emissions  at  the  site.


      The  present paper  describes the fluid modelling  techniques
currently   being  practiced  at   Rowan  Williams  Davies  &  Irwin
Inc.'s   fluid  dynamics  laboratory   at  Guelph,   Ontario.   The
techniques have  undergone   a  number  of -refinements since they
were   first  presented .   The   objective,  which   has   remained
unchanged,  is   to   determine  the  most  effective  configuration
                             5-2

-------
windbreaks  and  stockpile  geometry  for  minimizing  local  wind
speeds.   The   techniques   in  question  have  been   applied  to
numerous  sites,  such as:   the  coal handling  facilities of  Kerr
McGee  Inc.  at  Trona,  California;  the   Pennsylvania   Electric
Company at  Homer City,  Pennsylvania;  the  Dofasco  steel mill  at
Hamilton,    Ontario;   and  Ontario   Hydro's  thermal  generating
stations at Nanticoke and Atikokan,  Ontario, Canada.
EXPERIMENTAL METHODS

Scale Model Tests

     The  fluid  modelling techniques  for  dust control studies  at
Rowan Williams  Davies  & Irwin Inc. include  testing  a  scale  model
of  the  study  site  in  the   simulated   wind  flow  of  both  the
boundary  layer   wind  tunnel  and  the  open  channel  water  flume.
Photographs of both pieces of  equipment are  shown in Figure  2.


     The  boundary  layer wind tunnel is an open  circulation  type,
27m  long, 2.4m  wide  and  1.8m high.  The  test  section,  located
near  the  downwind  end  of  the  tunnel,  has  a  2.4m  diameter
automatic turntable built into  the  floor.   Spires  and  roughness
elements  can  be placed on the  floor  upwind of  the  test  section
to  simulate  the characteristics  of  the  mean wind and turbulence
approaching   the   study  site  .    Wind  speed  is   controlled
automatically and  can achieve  values  in excess  of 25 m/s.


     The  open channel  water  flume is a  12m long,  1.2m  wide  and
0.5m  deep apparatus that  operates on the same  principles as  the
wind  tunnel  but  uses  flowing  water to   simulate  wind.   Flow
patterns  on  a  scale  model are  visualized  by  injecting  coloured
dye  into  the  stream.    Drifting of  relatively  large  particles,
such as  snow grains, is simulated using  silica  sand.


     Scale models of   bulk   storage   facilities  are  typically
constructed  at   scales  between  1:400  and  1:600.   The  principal
modelling material  is  plexiglas,  with  stockpiles  modelled  in
clay  to   facilitate reshaping.   The  models  are  mounted on a 2.4m
diameter  base that fits on the  turntable of  the wind tunnel test
section.   Various  wind  directions  are simulated by rotating  the
turntable.  The base of  the  model is divided  into  two  sections:
an  inner  disk  and an  outer  annulus.  The  1. 2ra diameter  inner
disk fits into  the test section of the water flume.
     Local  wind speeds on  the scale model are  measured  using an
 array  of  surface-mounted  omnidirectional  sensors   as  shown in
                              5-3

-------
Figure 3.   Speeds  are measured  at  the equivalent height  of 2.0m
above grade full scale and  are sampled at a rate  of  300 Hz.  The
data  are  low-pass  filtered at  200 Hz  which  is  well  above  the
dominant  frequency  of atmospheric  turbulence.  The  local speeds
are   normalized  by  the   free   stream  (gradient)   speed,   U ,
measured  above the  model using a pitot-static  tube.   The results
are   recorded  in   the   form  U/U   and  s/U    where  U  is  the
one-hour  mean  speed and  s is the RMS  (standard deviation) of  the
wind   at   each  sensor   location.    Using  previously  measured
profiles,  the  ratios  are  then  converted  to   the  form U/UQ  and
s/U   where U    is  the  mean  speed at  2.0m  above  grade  in  an
unoSstructed  area.   The  resulting  ratios  can  be thought  of  as
local  wind speed magnification  or  reduction  factors  due  to  the
presence  of terrain features,  buildings  or  windbreaks.


Dust  Sample Tests

      A schematic of  the apparatus for  testing dust  samples  in
the  wind  tunnel is  shown  in Figure  4.   The  quantity  of  dust
lifted into the  atmosphere  is evaluated by weighing portions  of
the   samples   before  and  after   exposure   to  the   wind.   The
apparatus is  designed to distinguish  between  dust particles that
are  sufficiently small to  remain suspended  in the atmosphere and
undergo  turbulent  diffusion (airborne particulates), and heavier
particulates   that  are   subjected  to  the   saltation  process
 (saltating particulates).


      Three adjacent  trays  embedded  in the  floor  of  the  wind
tunnel test section  are filled  with samples  of the  dust.   The
middle tray is  fixed  to the floor  while the  upwind  and downwind
trays are removable  for  the purpose of weighing.  When the wind
 is turned on,  the  saltation process quickly reaches  equilibrium.
The   middle  tray  is  sufficiently   long  that  the downwind  tray
experiences a flux of saltating  particles  leaving the tray equal
 to the  flux  entering  from upwind.   Hence,  any  measured weight
 loss   after  a  time t is attributable to  airborne   particulates.
The   upwind  tray,  on  the   other  hand,  experiences  no  flux  of
 saltating material from  upwind  so  that  the  measured weight loss
 is a  combination   of  both   saltating  and  airborne   particulates.
The   difference  in weight  loss  between  the upwind  and  downwind
 trays,  therefore,  gives  the  estimate  of  the  outward  flux  of
 saltating particulates.


      In  summary,   if  w   is   the  upwind  weight  loss  and  W,  is
 the   downwind  weight  loss   after  a time  t,   and if  A^  and  A,
 are  the surface  areas of the upwind and downwind  trays, \hen:
                              5-4

-------
W2/A2 = emissi°n of" airborne particulates  per  unit  area
W,/A, - W0/A0 = outward flux of  saltating  particulates
 .L  J.    £  £


     Wind  speed is  monitored using  a  pitot-static tube,  which
can  be set  at  any  desired  height  above the  samples,  and  the
tests  are  repeated   for  a  range  of  wind  speeds.   Spires  and
roughness  elements are  used  to generate turbulence.   Although
full scale  turbulence  cannot be generated in  the wind  tunnel  due
to  the  limitation  of having  walls  and a  ceiling,  the  turbulence
produced  is  thought  to be  sufficiently  close to  full  scale  for
the  purpose  of the dust  sample  tests.  The impact of  turbulence
intensity   on   these   tests   is   the   subject   of   ongoing
investigations.


Computer Simulation

     The  computer  simulation  combines  the  results  of  the  dust
sample  tests  and the  wind speed ratios  described  in the previous
sections  with  wind  measurements  from  a  meteorological  station
near  the  study site   in  order  to  predict   dust  emissions  for
individual  wind events.   Dust losses  are  calculated for the area
around  each model  sensor location  and summed to obtain a global
dust  loss  for  the entire  site.   The  simulation has  the capacity
of  integrating  the   results  for  several  years'   worth of  wind
events  to predict  average annual dust losses.


     The  dust loss from  the  area around  each sensor  location is
calculated  hour by hour  as follows.   The reference hourly mean
wind   speed,   U ,  and  wind   direction  are   obtained  from  the
nearby meteorological  station.   The  local  mean  speed  at  2.0m
above  grade  is  calculated  by  taking the product  of  U   and  the
local   wind  speed  ratio,  U/U , that  was  measured   during  the
scale  model  tests.   The resufts  of  the dust  sample  tests  are
then  used  to determine the local dust  losses  for  the given hour.
The   local   RMS ratio,   s/U  ,   is   not  currently  used  in  the
analysis  but  will  be  adde%  in  future  studies   as  more  data
becomes   available   on   the   relationship   between   turbulence
intensity  and the  level  of  dust emissions.   For  the  present,  it
is  assumed  that  the  level  of  turbulence  throughout  the  study
site  is uniformly the  same as  was  generated in the  wind tunnel
during the  dust sample  tests.


     The  dust  sample  results  allow  the  local  dust   loss  to  be
divided  into   its   airborne   and   saltating  portions.    When
calculating  the saltating portion,   consideration  has to be given
to  the fact  that  a  partial  balance  exists  between the  flux of
material  away  from  the  sensor  location  and the  flux  into  the
area   from  upwind.    The   net  flux  is  determined  using  a
                              5-5

-------
two-dimensional   centred-difference   technique   in   which   the
difference in  dust  loss  between  surrounding  sensor locations  is
examined.


     Before  the   local  dust  losses  can  be  calculated,  initial
conditions have   to  be  established  -  i.e.  the  amount  of  dust
available  for  erosion  at each  sensor  location prior to  the  wind
event.  The  initial  conditions  are highly  site-dependent  and can
vary  from one  event  to another.   The  calculated  hourly  dust
losses  are   subtracted   from  the   initial   amount  until   the
remaining  amount  reaches zero.   When the  remaining  amount  at  a
particular  sensor  location  has  reached  zero,  no  further  dust
loss  is  permitted  to  occur  at  that  location.   If  the area  is
continually  disturbed  by  vehicle  traffic,   then  the   material
available  for erosion  may  be  unlimited.   To  account  for  such
cases the  initial condition  is  set sufficiently high so  that the
remaining amount  does not reach zero  during the  event.


Optimization and  Evaluation of  Dust Control Strategies

     The   fluid   model   studies   are   designed  to  evaluate   the
following  types of dust  control strategies:

(i)    development  of  an optimum stockpile  geometry to  achieve
       maximum wind  speed reductions  in critical areas;

(ii)   development  of  haul  roads and operating  procedures  that
       minimize disturbance  of the surface of the storage  piles
       in  critical areas;

(iii)  development   of  an  optimum  arrangement  of  windscreens,
       berms  and  other windbreaks upwind  of critical  areas.


      Semi-permeable  windbreaks   may  also  be  used  downwind  of
critical  areas to trap saltating particles before they  leave the
property,  but  are   generally  ineffective  at  trapping  airborne
dust.   Typically,  it  is not  feasible to  design  a dust  control
program  that works  for  all  wind  directions  but  the program can
be  designed  for  the directions  most frequently  responsible for
high  wind  events.


      The   following  are  some  of  the  practical  limitations  by
which dust control  programs  tend  to be constrained:

(a)   the  cost and  availability of material  for  berms  and  other
      windbreaks ;
                              5-6

-------
(b)   the  need   for   efficient  haul   routes,   unobstructed  by
     windbreaks;

(c)   the need  for sufficiently  large working  areas uncluttered
     by windbreaks;

(d)   temporary   windscreens   must   be   practical   to  assemble,
     disassemble or move about.
     By and  large,  flow visualization on  the  scale model in  the
water  flume  is  adequate  for  determining  which  set  of control
strategies works best  .   A qualitative   judgement,  however,   is
less convincing  than  a numerical  assessment,  and  the  pros  and
cons of  various  possible  strategies  are  best  weighed  by  using
predicted dust emissions from the computer  simulation.


EXAMPLE OF RESULTS

     Figure  5  shows  a  plan  of  the  fly  ash  lagoon  at Ontario
Hydro's  Generating  Station  at  Nanticoke,  Ontario.   During  the
course  of some  thirteen  years  of ash  storage  operations  the
level  of deposited  ash exceeded  the  waterline  in some areas.
Eventually, dry  ash became exposed resulting in significant dust
emissions during strong  winds.


     In  order   to  continue  using the   lagoon,   Ontario   Hydro
drained  off  most  of  the  water  and  compacted the  ash surface
sufficiently  to  allow  the traffic of heavy  vehicles.   Several
settling  cells  were created  in the western  half  of the lagoon.
At  present,  the  ash,   which  is  a  by-product  of  the coal-fired
boilers  at  the  generating  station,  comes to  the lagoon  in  a
liquid  slurry poured   from  pipelines  into the  settling cells.
The  ash  is allowed to  settle  and  is  then  removed by drag  lines
and  transported  by  truck to the permanent  storage  area.  A  fluid
model  study  was  undertaken to  determine   the  optimum  method  of
continuing  this  procedure  over   the  next  twenty  years   while
minimizing dust  emissions  at the  site.


     Since  most observed  incidents  of  high   dust   emissions
occurred  during  winds  from  WSW,   the  design  of  dust control
strategies focused  on  that  direction.  Figure 6   shows  contours
of  wind  speed ratios  derived  from the scale  model tests in  the
wind  tunnel.   The  highest  values  occur   in   the  most   elevated
areas which are  exposed to the highest wind speeds.
                              5-7

-------
     Ash samples  were  collected from  the  site and an  example of
the  data   from  wind  tunnel  tests  on  the  samples  is shown  in
Figure  7.   Weight  losses  of  airborne particulates  are  plotted
for  wind  speeds  of  6.0,  8.0 and  10.0 m/s  (measured  at  15  cm
height) and  for  various  durations  of wind exposure.   Wind speeds
of  2.0 and  4.0  m/s  were  also tested  but  no  weight loss  was
observed,   from  which  it was  concluded that  the  threshold  speed
of  dust emissions  is  between  4.0 and  6.0  m/s.   For  durations
greater than  0.2 hours,  the  data  shown in  the figure  fit  to an
expression of the  form  W = at  + b where  W is  the  weight  loss,  t
is  the duration  of wind  exposure,  and   a  and  b are  constants
which  vary with  wind speed.   The  constants a and  b were found to
be approximately related to the square of  the wind speed.


     A scheme for  storing  ash was developed,  taking  into  account
all  of the economical  and  practical  limitations  relating to the
site,  and  is shown in Figure 3.   The plan  shows  the  entire ash
lagoon as  it would appear  after  approximately  twenty years  of
operation.   The  sections at  the  bottom show the  progression of
the  permanent storage  area over  the  same period.  Berms  mounted
with  windscreens  are placed west-southwest of  areas  where ash is
being  dumped.   The  dumping  begins  just  east of the  settling
cells  (see Figure  5)  and  eventually progresses   eastward  until
the  permanent storage area  is filled  to  a  height of  15m  above
the  level  of  the surrounding dyke.  A 15m high porous  windscreen
along  the  western perimeter  of the lagoon protects  the settling
cells  area where much of  the  vehicle  traffic occurs.   Solidity
ratios between   0.5   and  0.8   were  recommended  for  various
windscreens  in   order  to obtain  the  optimum compromise  between
maximum wind  speed reduction  and maximum extent of protection.


      The scheme  shown  in Figure 8 was modelled  and tested in the
wind   tunnel.   To  assess  the  impact  of  the  scheme  on  dust
emissions,   the   computer   simulation   was   run  using  the  wind
conditions  that  occurred  during   a  particular   dust  event,  on
January 13 and  14,  1985,   which  was  chosen  as a  representative
case.   Wind  speed and  direction  during the  event were recorded
at  a  tower  located a  few kilometres  north  of the lagoon.   Wind
directions   were  observed   to   range   between   southwest  and
northwest  and the highest hourly mean wind  speed  recorded was 35
km/hr   (at  10m  above  grade).   As  an initial  condition  for the
event,  it  was   assumed  that  the  entire ash  storage area was
covered in a uniform  thin  layer  of  loose  dust.  Although this
was   an oversimplification  of  the  true  situation,   it  was  an
adequate   assumption   to  assess    the  relative   impact  of  the
recommended berms and wind  screens.
                            5-8

-------
     Figures  9  and 10  show  contours of  airborne  dust emissions
as  predicted by  the  computer  simulation  for  the  present  ash
lagoon  configuration   (Figure  9)  and  for  the  ash  lagoon after
twenty  years  of  operation.   The  size  of  the  shaded  area,
indicating  emissions  of  0.2  kg/m  or greater,  is significantly
reduced  in the  future configuration despite  the  fact  that  the
ash  is  piled considerably higher  than  in  the  present case-    In
the  future configuration, the  areas downwind  of  the protective
berms  and  windscreens  experience  minimal  dust  emissions.   The
upwind  slope of  the  westernmost  berm  in  the  permanent  storage
area  experiences  accelerated winds  and  consequently  high values
of dust  emission,  but in practice this  area would be permanently
sealed and  seeded with grass.


      Integration   of   the  dust  losses   over  the  entire  area
indicated  that  the  future  configuration would  experience about
half  the  total  dust  emissions  of  the present  configuration for
the  same  initial  conditions.   If  permanently  sealed  and  grassed
areas  are  accounted   for  then  the emissions  would  be   further
reduced.
CONCLUSIONS

     Fluid   modelling  studies  are   an'  excellent  visual   and,
quantitative  tool  in the  design  of  dust control  programs  for
bulk   storage  facilities.   Areas   where  refinements  to   the
physical  model approach  are  anticipated  in  the  future  include:
the  establishment of  more realistic  initial  conditions for  the
computer   dust   simulation   that   account   for    the   rate   of
disturbance  by  vehicles,  the  moisture  content  of  the surface
material  and other  factors;  an indepth  study  of  dust  emissions
resulting directly from vehicle activity; a further study of  the
relationship    between   airborne    emissions   and    turbulence
intensity.


     A  further addition currently being  implemented consists  of
applying  a  numerical dispersion  model  to  predict the  rate  of
dispersion of the dust emissions downwind of the study  area,  the
results  of   the  numerical model  can  then be  compared  to  field
measurements  with  high volume samplers.
                             5-2

-------
REFERENCES

1.   A.E. Davies, "A physical  model  approach to  the  solution of
     fugitive emission  problems",  Proceedings of  the 73rd Annual
     Meeting  and   Exhibition  of   the  Air  Pollution  Control
     Association, Montreal, June 22-27,  1980.

2.   H.P.A.  Irwin,  "Design  and use  of  spires  for  natural  wind
     simulation",  National  Research  Council of  Canada,  N.A.E.
     Report LTR-LA-233, 1979.

3.   H.P.A. Irwin "A   simple   omnidirectional   sensor  for  wind
     tunnel studies  of pedestrian  level winds",  Journal  of  Wind
     Engineering  and  Industrial  Aerodynamics,   7,  pp.  219-239
     (1981).
           Figure 1.  Dust emissions at a fly ash storage area.
                               5-10

-------
Figure 2.  The boundary layer wind tunnel (top) and open channel water flume.
                                   5-11

-------
Figure 3.  Omnidirectional wind sensors on a model (top) and a side view of
an individual sensor.

                                   5-12

-------
 CBL1NG
 N^y^s^^^^y^y^^^^
                             PITOT-STATIC  TUBE
 WIND
RAMPv*

AIRBORNE
DUST
xx''*
SALTAT1NG
1 	 ^DUST

^•~^
/I.3CM DEEP /
^7 DUST SAMPLE /


FLOOR/  55
             cm
240 cm
                                                       30cm
Figure 4.  Schematic of the dust sample test apparatus,
                              5-IS

-------
                           12m HIGH
                           DYKE
           SECONDARY
           SETTLING
           CELLS
AREA  BETWEEN
CELLS  IS  LEVEL
WITH  DYKE
                                                  TERTIARY
                                                  SETTLING  POND
                                                    PERMANENT
                                                    STORAGE
                                                    AREA
        PRIMARY
        SETTLING
        CELLS
                  ACCESS  ROAD
                                                      PERIMETER^
                                                      ROAD
'ROADS CONSTRUCTED
-IN  SOFT  LOW LYING
 AREA
                 NOTE- SHADED AREAS  ARE IMMERSED IN WATER
                       OTHER AREAS  CONSIST OF ASH
                                                         I	
                                                        400
    400
            800m
   Figure 5.  Plan of  the  ash lagoon at Nanticoke, Ontario,  in its present state.
                                 5-14

-------
             PREVAILING  O.
             WIND
Figure 6.  Wind speed ratios, U/U ,  as measured in the wind tunnel for winds

from WSW.

                               5-15

-------
  2800
  2400  .
  2000 _
cvj

*
CO
CO
3
   1600 .
   1200 -
O
u2
   800 _
   400 .
                                  LOSS OF AIRBORNE PART1CULATES
                              DURATION (hrs.)
        o
        «
        A
WIND SPEED: 6  m/s AT 15cm  HEIGHT
            8  m/s
            10 m/s
  Figure 7.  Example of dust sample test results showing loss of airborne

  material versus duration of wind exposure.
                                 5-16

-------
            PREVAILING
            WIND
                                         HAUL  ROAD
   15m HIGH
   PERMANENT
   80%  SOLID
   WINDSCREEN
                                                                SEE
                                                                BELOW
                                      -PROTECTIVE  BERMS
           ,3m HIGH  WINDSCREEN
          /80% SOLID
   A—T
                   IOO
200
                                             300
                           400  METERS
                                                                          1st
                                                                         .YEAR
                                                                          10
                                                                         »YEARS
                                                                           20
                                                                          a YEARS
Figure 8.   Twenty-year  storage plan.


                                  5-17

-------
                 PREVAILING
                 WIND
                                                   2
Figure 9.  Contours of airborne dust emission (kg/m )  during a typical dust

event for the ash lagoon in its present configuration.

                                 5-18

-------
 PREVAILING
 WIND
     20.2 kg /m
Figure 10.  Same as Figure 9, but for the recommended scheme after 20 years,

                                   5-19

-------
        EVALUATION OF FIELD TEST RESULTS

            ON WIND SCREEN EFFICIENCY*
                    Alan G.  Larson**

          TRC Environmental Consultants, Inc.
                  Englewood, Colorado
  This paper has been reviewed in accordance with the U.5. Environmental
  Protection Agency's peer and administrative review policies and approved lor
  presentation and publication.
(*)Although this is not the actual presentation made at the
   May 1982 meeting, it describes the same work and covers
   the same general time period. It represents work funded
   under EPA contract 68-02-3115, Technical Directive  117.
(**)Current address:  The Kentwood  Moore Co.,  5690 Denver
    Tech Center Blvd.,  Englewood, CO  80111.
                             6-1

-------
                                 Abstract

      A field measurement  project  was  conducted to  generate data  which
would lead  to  a preliminary  determination of  the  effectiveness of  wind
screen material  as  an inhibitor of  fugitive windblown dust  emissions  from
a  stockpile.   A stockpile  comprised of  fly  ash was  used  as  the  test
material and a wind screen was erected  upwind  of it to comprise the  test
configuration.   Standard hivolume  samplers were  used  to collect  samples
of Total  Suspended  Particulates  (TSP) and  hivolume  samplers fitted  with
size  selective  heads   were  used   to  collect  samples   of   Inhalable
Particulates  (IP)   (less  than  15  /jn).          Particulate  concentration
measurements due to wind erosion from the stockpile both  with and  without
the  wind  screen  were  made.   Low  threshold  anemometery  was  used  to
document wind conditions.   It  was  found  that  the wind screen did  reduce
the  fugitive  emissions  of   TSP  but  the  reduced emissions of  IP  were
insignificant.
                                Introduction

       The  emission  of  windblown fugitive  dust from stockpiles can be  a
 significant problem  at mines,  power  plants,  steel  mills,  rock-handling
 facilities,  and  many  more  industrial   concerns  which   stockpile  raw
 materials, fuel,  and waste materials.  Stockpiles which  are inactive (not
 drawn  upon)  can  be effectively  treated  for  fugitive  dust  control  by
 several  methods   including  chemical  crusting  agents,   coverings,   etc.
 Active stockpiles,  however, present a much  more  difficult dust  control
 problem  because   of   the   need  for  access  and  because  of   surface
 disturbances  caused during access.  One approach which has been suggested
 as a means of  reducing windblown  fugitive emissions from  both  active  and
 inactive stockpiles employs wind screen materials.

       Wind screen material  is comprised  of a  fabric-like  construction of
 high tenacity  polyester  fiber,  woven  to  varying degrees  of porosity  and
 prepared in different widths.   The  manufacturer holds that the material,
                                      6-2

-------
when  erected  vertically  like  a  fence, will  reduce the  wind speeds  at
levels directly  adjacent  to the  ground surface  in  the lee  of the  wind
screen.   Further,  the  manufacturer  contends,  when the  wind  speed  is
reduced due to the  presence of the wind screen, windblown  dust  in  the lee
of the screen will also be  reduced.   It is therefore expected that use of
the wind screen material as a  wind  break  will result in reduced  fugitive
windblown particulate matter  in  a limited  region downwind from the  wind
screen.   This report  describes  a  preliminary   investigation  which  was
conceived  to  determine  the actual   effectiveness  of  one type  of  wind
screen material toward control  of fugitive dust emission.

      TRC Environmental  Consultants,  Inc.  (TRC), under  contract  to  the
EPA  Industrial  Environmental  Research  Lab  (IERL)  in  Research  Triangle
Park,  North  Carolina,  has  performed  a   preliminary  field  study  which
investigated  the  effectiveness  of  wind  screens  to   control  fugitive
windblown dust  from stockpiles.  The study was  conducted around a  test
stockpile of  fly ash material.   The stockpile  was  located  in a  remote
portion  of  property  owned  by Colorado Public Service Company's  Valoont
Power  Plant  in Boulder,  Colorado.   The site  was chosen  because  of  the
high  incidence of  consistent  wind direction  (westerly)   in  several  wind
speed  categories.

       Fly  ash material  was  chosen to  comprise the test  stockpile for this
field  investigation  because it is a  common waste material where  coal is
used  for fuel,  it  is comprised of relatively small particulates, and thus
is susceptible  to  be  windblown  at  moderate wind  speeds,   and  it  was
readily  available  at no  cost.   The fly ash  being susceptible  to  wind
erosion  simplified  the  experimental  work  because measurement of material
removed  from the  test  stockpile was possible over a  wide range  of wind
conditions.   All of  this  was judged as being  optimum to  lead to the most
expeditious   determination   of  the  effectiveness  of  the  wind  screen
material as an inhibitor  of windblown particulate matter from stockpiles.

       The  study  was  designed to determine the total suspended particulate
matter  (TSP)  and   the  inhalable   particulate  matter   (IP,  less   than
15 fjfn.)      due  to wind action.   The  emissions of particulate matter  were
                                    6-3

-------
measured from a simulated  active  stockpile before and  after  installation
of a  wind  screen.  Hivolume  samplers (hi-vols)  and  hi-vols fitted  with
Sierra  size  selective heads  were installed  upwind  and  downwind of  the
test stockpile to determine the TSP and  IP concentrations under different
wind conditions.  The upwind sampler  measured background  concentrations
and  the downwind  samplers  measured  the  particulates  emanating  from  the
test stockpile along with the background  concentrations.   Anemometers  and
wind vanes were located  upwind  and downwind of the wind  screen to record
wind  speeds  and  directions  that   resulted in  the  windblown  particulate
emissions.

                        Test Methodology

       The  test  site was located on a native grass covered mesa.  The mesa
 is  elevated above  the  surrounding   terrain  by  about  50  feet* and  is
 oriented in an east-west direction which coincides with  the direction of
 the most frequent and highest  winds.  The test site was situated on the
 eastern portion of  the mesa in order to  allow the west winds to  stabilize
 prior  to reaching the test site.  The test  stockpile  of  fly ash  material
 was placed  in  a  rectangular geometric  shape.  The fly  ash material was
 taken  from  the  waste  dump  where the  Valmont  Plant  waste material is
 disposed.   The elevated  landform provided an isolated test  site with an
 unobstructed fetch  of about  3/4  mile*               (for west winds) and
 was  isolated  from  interfering  particulate  matter  sources   by  about
 3/4 mile to the west, about 1/2 mile to   the northwest, and about  1/3  mile
 to  the east where  the Valmont  waste dump  is  located.    Tests  were not
 conducted  during easterly  wind   conditions,  so  the  only  interferences
 experienced were  those located to the northwest and west.   It  is believed
 that during most conditions when interferences were  experienced,  those
 interferences  were  adequately   documented  by  upwind   sampling.   The
 stockpile was shaped to a size  of approximately 8 feet wide, 12  feet  long
 (N-S),  and  A feet high.   The  leading  edge  of  the  stockpile  was located
 approximately 12 feet downwind of the wind screen.   The wind  screen  used
 in  the study  was  commercially  available  high  tenacity polyester  fabric
 wind screen manufactured by Julius Koch USA,  Inc.   The  screen was  6  feet
 tall,  165   feet  long,  and  had  approximately 50 percent  porosity.   The
  (*) For nonmetric units used in this paper, please use: 1 ft = 0. 3 m,  and
     1 mi = 1.61 km.
                                    6-4

-------
screen  material  was  mounted on  steel  poles  set  in  the  ground.   Wind
screen   material   of   50   percent   porosity  was   selected  based   on
recommendations of the manufacturer which were directed  to  configurations
most  likely to  result  in  maximum  wind speed  reduction  at the  levels
adjacent to the ground surface.  Material six  feet tall was  chosen  so  as
to enable  employment  of  a stockpile being a significant  height  above  the
flat mesa  test surface  (A feet) and  having  a stockpile height to  wind
screen  height  ratio  of  2 to 3—also  as recommended by  the wind  screen
manufacturer.  A section  of  wind screen material  165 feet  in length  was
chosen  so  as  to  avoid end  effects  at  the  test  stockpile   (wind  eddying
effects) and  to  allow  for a reasonable  wind  direction  variation  of  as
much as + 30°.

       After the  test  stockpile  was  placed,  it was allowed  to dry  before
 tests  were conducted.   Following  the  drying period,  the   stockpile  was
 covered between tests with  a  tarpaulin to protect  It from exposure  to
 precipitation.  Samples of  fly  ash material  were collected during  each
 field   experimental   day   and  were  subsequently  analyzed  for  moisture
 content.

       The  instrument array  used to generate  test data was  comprised of
 three  pairs of hi-vol samplers  and  three anemometer/wind vane  sets.  The
 upwind hi-vol pair  measuring  TSP and  IP were located 20  feet  upwind of
 the  screen   on   the  east-west  centerline   of  the   stockpile.    Two
 Climatronics  Mark 3  anemometer/wind  vane  sets  were located adjacent to
 the  upwind-  hi-vol s  at  heights  of   A   feet   and   12  feet.   Another
 anemometer/wind  vane  set was   located  adjacent   to  the  stockpile  at  a
 height  of  A   feet,  level with  the  top of  the  stockpile.   One  pair of
 hi-vols measuring TSP and IP  were located on  the E-W centerline 20  feet
 downwind and  another pair  was  located on  the  same centerline  50  feet
 downwind of the stockpile.   Two hi-vols measuring only  TSP were located
 50 feet downwind  of  the  stockpile,  one 30  feet  north  and one  30  feet
 south   of   the  east-west  centerline.   The  wind   screen  was 165  feet in
 length  with   approximately  half of   the  screen   on  either  side  of  the
 east-west  centerline.  A diagram of  the study  site is shown in Figure  1.
                                   6-5

-------
         .20'
                                             N
                       Wind Screen
                       165' Long
                                                           30'
    O
    A
121*4-8!*

 t
         	
-------
     The plan for each test consisted of 6 separate steps:

     1.  Remove tarpaulin from stockpile.
     2.  Set up anemometer/wind vane equipment*
     3.  Shift  hi-vols to  be on  the  centerline of  the wind  and
         stockpile.
     4.  Rake stockpile surface to loosen any packed ash.
     5.  Install filter paper in hi-vols.
     6.  Measure  dust  concentrations  for   approximately one  hour
         total sampling time.

      Steps 1 and 2 were  only required at the beginning  of each test day
as usually two or three tests were conducted  on  a given  day.  During the
testing period,  the  hi-vols were moved,  as necessary between individual
tests,  to be on the centerline of the  wind  and the stockpile.  Each  test
period lasted approximately 60 minutes.   At  the  end of  the test day, the
equipment was removed  to a storage area and  the tarpaulin was replaced  on
the  stockpile.   When  more  than  one  test was conducted during  a  single
day, Step  4  was repeated just prior  to each  additional test.   This was
done  to  assure  that  each  test   dealt with  the   same  particle  size
distribution of  fly ash as much as  possible.

      A total of  7 tests were run  without   the wind  screen and  12  tests
were run with the wind screen  installed.

                           Data Analyses
Assay Methods

       The  data  collected  during  each hour-long  test  consisted  of hi-vol
 filters  for both  TSP and  IP  samplers,  wind  direction, wind speed,  and
 moisture  content  of  the surface  fly ash material.  Particle size analysis
 of the material  on the stockpile  surface was done from  a single composite
 sample obtained  from  all  the  faces of  the stockpile.
                                6-7

-------
      The  hi-vol filters were assayed  gravimetrically in accordance  with
40 CFR  50 Appendix  B  Reference  Method  for Determination  of  Suspended
Particulates  in  the Atmosphere  (Hivolume  Method).

      Wind data were  compiled  from  visual strip  chart  recordings  to
represent  hourly average directions  and  speeds,  peak gusts, and  an  index
consisting of a gust-duration product.   The compilation of averages  and
peak  gusts  is  straightforward  resulting in  conventional  values.    The
gust-duration product  index  was devised  in  an attempt  to derive a  wind
representation  that  would  directly  relate  to the  physical  processes  of
the wind  erosion phenomena.   It was  reasoned that,  since wind  erosion is
a direct  function  of both the  wind  speed and the  duration of high  wind
speeds,  an index incorporating  both  wind speed  gusts and  the  duration of
those   gusts   should   relate    directly   to    the   measured   sampler
concentrations.   In  practice  it is  difficult to  determine  the  level  at
which a wind  gust  becomes important.  Therefore  the gust-duration  index
was  derived  simply  by  integrating  the  total area  under  the  wind  speed
curve for each  experimental period.   In this way a set  of indices  was
developed which  relate to one another  but not to  other  field experimental
work.   The  indices  are  in arbitrary  units   and  can only  be  used  as  a
measure of  the  relative  wind  erosion  potential  for each  of the  tests
represented in this analysis.

      Moisture  content  of  the fly  ash  material  was  determined  from
samples of  the  material  collected  during   each  experimental   day-   Each
sample  of fly ash material consisted  of about  60 grams of  ash.   Samples
were  taken  from each face of the  stockpile,  placed  in  a sealed bag, then
transported   to the  assay  laboratory.   Samples  were  placed   in  dried
crucibles,  and  promptly  weighed   to determine the  mass  of fly  ash and
moisture  prior  to drying.   Samples  were baked at  105°C for one hour to
drive  off  all  moisture,   then placed  in  a dessicating   chamber  until
cooled.  After  cooling, the samples were weighed again to  determine the
moisture  loss.   Moisture  content was represented  as  percent reduction (by
mass) due to drying.   The  mean moisture content of  the stockpile for an
experimental  day was  determined as a mean of the  values from all  faces of
the  stockpile.
                                  6-8

-------
      Particle  size  analysis  of  a  composite  ash  sample,   which   was
comprised  of material  from all  the  faces  of  the  stockpile,  was  done
manually with  eight sieves  from  number 10  to 325.   The  sieves are  USA
Standard  Testing  Sieves  which  conform  to  ASTM  specifications.    The
sieving  of  the  composite  sample was  done  manually  in  order that  the
mechanical   action   of  the   sieving   process  which   might   break   up
agglomerated  particles would  be  held  to  a  minimum  and  the  resulting
analysis would as nearly  as possible represent the  true surface  particle
size distribution.
Selection of Data To Be Used for Analysis

       A total of 19 tests  were conducted  to  represent  wind  erosion  both
 with and  without the  presence of  the  wind screen.  The  data collected
 were assayed as  described above and a  series  of  comparisons  were made to
 assist in  selecting those  tests  that  would  best  represent the effects of
 the wind screen  as  an  inhibitor of  fugitive  windblown  dust.

       For   several  reasons  not all  the  test data  collected  were used in
 the analysis.  As a way  of selecting  the  most  representative  TSP and IP
 concentration data, all  of the concentration  and  wind  data  were plotted
 on  scale   diagrams so  as  to  graphically  illustrate  the   tests   which
 exhibited  consistent wind and dust erosion conditions.  It was  found  that
 the wind  direction  variability (especially during  peak gust  conditions)
 during  some tests   caused  a portion  of the  eroded dust  to  pass by  the
 sampling array during some  portions of  some tests.  In addition  to  this,
 sampler  failure resulted   In inadequate   representation  of  upwind  or
 background concentration levels  on  two tests.  As  a  result,  six of  the
 nineteen tests were only partially successful leaving  eleven  to be  used
 for analysis.   The TSP and  IP concentrations along  with the  supporting
 wind data  and measurements  of other  test  conditions are listed In
 Tables 1,  2, and 3.
                                     6-9

-------
TABLE 1:  TSP & IP CONCENTRATIONS
  FOR TESTS WITHOUT WIND SCREEN
TEST
NUMBER
3
5
6
7
Means
AX 20 =
AX 50 =
TSP
UPWIND
225
118
70
82
124
X (at 20 feet) ~x
X (at 50 feet) -x
(yg/m3)
Ax 20
151
97
195

148
Upwind
Upwind
Ax 50
95
6
16
7
31


TSP & IP CONCENTRATIONS FOR
8
9
10
11
12
13
17
Means
300
72
80
88
—
44
201
131
138
53
16
59
—
4
189
76
72
86
7
—
—
24
88
55
IP (yg/o3)
UPWIND AY 20
158
35
57
70
80


TESTS WITH
126
35
72
53
20
6
86
57
26
40
22
—
29


WIND SCREEN
81
23
—
4
7
21
2
23
AY 50
34
21
19
—
25

•>

78
25
—
—
33
15
42
39
              6-10

-------
                   TABLE 2:  WIND DIRECTION, SPEED,  PEAK GUST (nph)
                               AND  GUST DURATION  INDEX

                           DURING TESTS WITHOUT WIND SCREEN
TEST
NUMBER
3
5
6
7
Means
U-12
DIR*
285
225
240
220

U-12
SPEED
10.5
11.0
13.5
8.5
10.9
U-12
PK GUST
20.6
22.5
24.2
14.4
20.4
U-4
DIR
295
215
215
210

U-4
SPEED
8.0
10.0
11.5
7.0
9.1
U-4
PK GUST
15.5
• 19.8
19.5
16.6
17.8
D-4
DIR
290
205
210
220

D-4 D-4 GUST DURATION
SPEED PK GUST INDEX**
7.5
10.0
12.5
7.0
9.2
15.5
20.3
19.2
17.4
18.1
52
52
63
40
52
            WIND DIRECTION, SPEED, PEAK GUST (mph) AND GUST DURATION INDEX

                            DURING  TESTS  WITH  WIND  SCREEN
8 290
9 280
10 250
11 230
12 285
13 300
17 210
Means
13.0
10.0
13.2
12.2
12.0
11.4
16.0
12.5
25.0
18.0
21.4
20.8
20.6
20.2
19.2
20.7
__
—
240
275
285
255

__
—
10.5
11.0
10.5
13.0
11.2
__
—
16.4
17.4
16.5
19.8
17.5
290
280
25.5
230
250
265
235

3.8
3.0
4.5
4.8
4.0
3.7
4.5
4.0
12.5
8.5
7.5
5.6
7.5
6.9
10.1
8.4
68
49
58
60
61
54
71
60
 *U-12 » Upwind,  12 feet  height.  Direction (DIR) = bearing (true) in degrees.

**Gust Duration Index is  in arbitrary units.

                  MOISTURE CONTENT OF SURFACE FLY ASH MATERIAL
   DATA FROM FIVE FACES OF STOCKPILE ON DAYS  HAVING WIDEST VARIATION OBSERVED
                               (PERCENT BY MASS)

          NORTH FACE   EAST FACE    SOUTH FACE    WEST FACE    TOP   AVERAGE

Day 1         2.7          1.6          1.1           1.9       1.7     1.8
Day 2         1.4          1.0          1.3           2.6       0.8     1.4

                                    OVERALL AVERAGE hJOISTURE CONTENT    1.6
                                     6-11

-------
                   TABLE 3:   PARTICLE SIZE DISTRIBUTION
                       OF SURFACE FLY ASH MATERIAL
SIEVE
NUMBER
10
50
100
140
180
250
325
Pan
SIZE OF SIEVE
OPENINGS (ym)
2,000
300
150
106
83
58
45

RANGE OF PARTICLE
SIZES COLLECTED (ym)
> 2
300
151
107
84 -
59 -
46 -
< 45
,000
- 2,000
- 300
- 150
106
83
58

PERCENT OF
TOTAL SAMPLE
22.88
24.18
13.26
12.82
7.45
14.74
4.61
.07
100.01
                            Results
Wind Speed Reduction

      A  measure  of  the effectiveness  of the  wind screen  can  be  shown
 through  an  analysis  of the  reduction  of wind  speeds  resulting  frotn the
 presence  of  the  wind  screen  material.   The  data  sample   that can  be
 employed  for  this  purpose  is  somewhat larger  than is  shown in  Table  2
 above  since it does  not  matter  about the  success of  the  concentration
measurements.   For  this   efficiency,   the  tests  are  used  which  were
 conducted when  the  wind screen  was in  place  and wind  measurements  were
made at the same heights above the ground both  upwind  and  downwind of the
wind screen.

      The average wind  speeds  and  peak gust wind  speeds as  summarized in
 Table  4  show  that   the  average  downwind  wind  speeds  are  reduced  to
 36 percent  of  the   upwind   value   and   the  peak  gusts   are   reduced  to
 41 percent  of  the  upwind  value.  These  reductions should be  expected  to
vary somewhat  by location  for  both distance downwind  of  the  wind screen
and  for  height above  the  ground.   The values  presented  here  should  be
 representative of  the  fugitive  windblown dust measured  in  these  tests
because  the  downwind  wind   speed was measured  adjacent  to  the stockpile
and at  the height  of  the stockpile.
                                  6-12

-------
                TABLE 4:  WIND SPEEDS UPWIND AND DOWNWIND
                  OF WIND SCREEN AT 4 FEET HEIGHT (mph)
TEST
NUMBER
11
12
13
14
15
16
17
18
19

UPWIND
SPEED
10.5
11.0
10.5
10.0
12.5
12.7
13.0
11.5
10.5

DOWNWIND
SPEED
4.8
4.0
3.7
3.5
4.0
4.0
4.5
4.0
4.0

UPWIND
PEAK GUST
16.4
17.4
16.5
16.0
19.6
18.4
19.8
18.8
16.0

DOWNWIND
PEAK GUST
5.6
7.5
6.9
7.0
7.9
6.2
10.1
6.5
7.5
Means
% REDUCTION
WIND SPEED
54
64
65
65
68
69
65
65
62
64
Z REDUCTION
PEAK GUST
66
57
58
56
60
66
49
65
53
59
Airborne Concentration Reductions
      The most  graphic and  direct  measure  of  the  effectiveness  of the
wind  screen  as an  inhibitor  of windblown fugitive  dust can  be  seen by
close examination  of  the  TSP and  IP  concentration  values  as  shown in
Table  1.   Several  relations  of  significance  are  described  in  the
following paragraphs.

      Without the  wind screen,  an  average  of 148 y g/m   increase  in TSP
concentration is  seen  at  20  feet  downwind.  By  the  time  the windblown
plume  has  traveled   to  50  feet   downwind  the  average  increase  in
concentration  value  has  dropped  to  31 yg/ffl  above  the upwind  value.
With  the  wind  screen  in  place,   an  average  of 76  yg/nr  increase in
concentration  is  seen  at  20 feet  downwind.  However,  an  increase of
55 yg/m3 is  still seen at  50  feet downwind.  These  differences  are of
significance.  Without  the wind  screen, the increase in  TSP  concentration
above background  levels  is high, but  drops  off  rapidly probably  due to
deposition.   With  the wind  screen,  the   increase  in  TSP  concentration
isn't as high  above  background  levels as  in  the  case without the screen
but  the  elevated  concentrations  remain  elevated for  a  longer  period.
This  latter  case is  probably due  to  increased  turbulent  suspension of
particulates,  where  the  increase  in  turbulence results  from  the air
flowing over the wind screen.      fi ,_

-------
      It can  be  seen  from  the  wind  speed  averages  and  the  wind gust
duration indices (as shown in Table  2)  that the winds  were higher  during
the tests with  the  wind  screen  than during  the  tests without the screen.
Therefore,  there was a greater potential  for windblown fugitive emissions
during the tests with the wind screen.   Some adjustment should  be made so
as  to  allow   a   more   direct  comparison   between   the   two   sets  of
concentration  values.    Probably  the  best  manner   in  which  such  an
adjustment can  be  is to  normalize   the  mean concentrations  to  the  Gust
Duration Index  (I), since  this  Index  incorporates both wind  speed  and
duration.   Such a  normalization  results  in the following:
      Normalized Concentration Without Screen -  —=—   -   —rr  • 2.85

      and,

      Normalized Concentration With Screen *  -•;—  «  777   • 1.27
                                               I       oU
then  the   ratio   between  concentrations  with   the   wind   screen   and
concentrations without the wind screen is simply,

                 1.27
This says that when TSP concentration  differences  at  20 feet downwind are
normalized to  a  common base  (wind conditions),  the  wind  screen reduces
the windblown TSP  dust  concentration  to approximately  45  percent of what
it is without the wind screen.

      It  is   also  seen   from   Table  1  (as  noted   above)  that  the
concentration differences  stay  higher  with distance with  the wind  screen
than without  it.   Therefore,  to  obtain a more  representative measure of
the total effectiveness of the  wind  screen,  this difference with  distance
should also be incorporated.   This incorporation can be made  as follows.
                                   6-14

-------
                                                    (Ax 20 + Ax 50)
                                                   A     2    y      .
                                                          T        * 1 • .
Normalized Mean Concentration Without Wind Screen  -^	1	*   « 1.73

      and,
Normalized Mean Concentration With Wind Screen  -\     *     /  « 1.10

The ratio between  mean conditions with the wind  screen  and  without it is
simply,

                 1.10
 This  then says  that  the  total effectiveness  of the  wind screen  is  to
 reduce  the downwind fugitive  windblown  TSP emissions  to 64  percent  of
 what  they  were without  the wind screen.

      The  situation is  considerably  different   for   IP   concentrations.
 Also  from  Table   1   it  is   seen  that  the  mean  differences  in  IP
 concentrations  without  the  wind screen  are  29  and  25 u g/m   at  20 feet
 and  50  feet,  respectively.    With  the  wind   screen  in  place  the  mean
 differences  in  IP  concentrations  is  23  and  29  ug/m  .   Considering the
 accuracy  and  resolution of  particulate sampling  with  the hi-vol sampler
 and   gravametric  assay,   these concentration  differences are   for  all
 intents  and  purposes  the  same.   If,  however,  the  same normalization
 exercise  as was  used above is  carried  out,  it  is  seen that,
                                                      x 20 + Ax50
                                                               \
                                                               J
                                                               >
       Normalized IP Concentration Without  Screen  -^	?	f    • .52

       and,

                                                 (Ax 20
                                                 /A
       Normalized  IP Concentration With Screen  -^	     /    - .52
 then,  the  ratio between conditions with the wind screen and  without it  is,
                   .52  _
                   751     1
                                    6-15

-------
To reiterate, the  wind  screen,  although it  reduces  the  wind speed in its
lee  area  where  the windblown  emissions  originate,  appears  to  make  no
difference  in the  fugitive  emission  of  windblown  IP  (small  particle)
material.

       It  must  be  said  that  concentration data  of  the  type  collected
 during this set  of experiments  are inherently variable.  This variability
 is due in  large part to variations  among  the atmospheric processes which
 result in  the windblown activity.  Also,  various  aspects of the  sampling
 system  used  to  make  the  concentration  measurements  lead   to  data
 variabilities.   In  this  analysis,  a great  deal of  care  was   taken  to
 screen the  data  collected  in  order  to assure  that  as  many  of  the
 irregular  occurrences as  possible were  eliminated  from the data base.
 This was done  in an attempt to  remove those factors  that  might  obscure
 the  true  behaviors.  It  is  felt  that   this  data   screening  was quite
 successful  and  the  resulting  data base  is of  comparative  high  quality.
 However,  by  screening  the  data  base  and  eliminating  those cases during
 which atypical  conditions  existed,   the  data base  was  made  to be quite
 small. For this reason  it  is  recommended  that the results  presented here
 be  respected as  preliminary.

                          Conclusions

       Wind screen  material  will reduce both  the  wind  speed  and fugitive
  windblown  dust  emissions  from  stockpiled material.  The  amount  of  wind
  speed reduction  was  found  to  be  64  percent  of  the  mean  wind  and
  59 percent of the peak gust values.   That  is  the  mean  wind  was  reduced to
  36  percent  of  the upwind  value  and  the peak  gusts were  reduced  to
  41 percent  of   the upwind  value   in  the  conflguation  tested.   This
  configuration  was  with a  six  feet  high  wind  screen  where  the  downwind
  wind speed was  measured at a point  four feet  above  the ground  and sixteen
  feet downwind from  the wind screen.

        The  amount  of  reduction  of  windblown fugitive  TSP  is  55 percent
  lower at a location 20 feet  downwind of  the  stockpile.  That  is, the TSP
  was 45 percent  of  what it was  without the wind screen.  When  considering
                                    6-16

-------
both the data from 20 feet and 50 feet downwind  of  the  stockpile,  the TSP
reduction was 36-percent, or 64 percent  of what it was without  the  wind
screen.

      The analysis  of reduction of  IP (snail particulate  emissions) did
reveal some  small  differences  between the  case  with the wind  screen and
without  it.  However,  the differences are  at  an  insignificant  level.  It
is concluded  that  the presence  of  the wind  screen makes  no significant
difference in the amount of windblown fugitive IP emissions.

      The  overall  conclusion  is that  the  wind  screen both  reduces the
wind  speed  in  its  aerodynamic wake and  also  reduces the  particulate
emissions.    There   remains   a   question   about    the   wind   screen's
effectiveness at  reducing the  emissions  of  inhalable  particulate matter
( < 15 urn)  and  additional measurements  will be necessary to resolve  that
issue.
                                   6-17

-------
                EFFECTS  OF STREET SWEEPING ON  FUGITIVE EMISSIONS
                                FROM URBAN ROADWAYS
            The work described in this paper was not funded by the U.S. Environmental
            Protection Agency. The contents do not necessarily reflect the views of the
             Agency and no official endorsement should be inferred.
                                  Donald F. Gatz
                           Atmospheric Chemistry Section
                            Illinois State Water Survey
                             P.O.   Box 5050, Station A
                             Champaign, Illinois 61820
                                    ABSTRACT


    To  determine   the  effects  of   street    sweeping   on   urban    particle
concentrations,  three  standard high volume  samplers were operated near a four
lane road in  a  commercial  area,  and   another  sampler  was  operated  in  a
residential  area,   in  Champaign,  Illinois.  Results show that street-related
sources increased  the total airborne  particle  mass  about  20 yg/m3  at  7  m
downwind  when   the wind had a component  perpendicular to the street.   Analysis
of variance, comparing periods with and without sweeping in spring  and  summer
sampling  periods,   showed  higher downwind-upwind concentration differences in
summer than in spring for east and west   winds,  apparently  caused   by  higher
street dust loadings in summer, but no effect from street sweeping.
                                      7-1

-------
1   INTRODUCTION
    Provisions of the Clean Air Act enacted in  1977 required  states   to  revise
their  State  Implementation  Plans  (SIPs) for all areas  that  had not  attained
National Ambient Air Quality Standards (NAAQS).  The  Illinois   SIP   for  Total
Suspended   Particulates   (TSP)   was   conditionally  approved by  the  U.S.
Environmental Protection Agency (USEPA) with certain minor deficiencies.   One
of  the  reasons  for  the  conditional  approval  of  the Illinois TSP SIP was
inadequate documentation of the impacts and  effects  of   various  controls  on
non-traditional  fugitive  sources  of  TSP.   These  sources include paved and
unpaved roads, parking lots,  and agricultural lands.  This and  similar   studies
have  been  designed to correct those deficiencies and will be  submitted to the
USEPA as part of the SIP for that purpose.  The results of these studies  will
be  used  by  the  Illinois  Environmental  Protection  Agency  (IEPA) to define
further the estimated impacts of non-traditional fugitive  dust  sources   on  TSP
non-attainment  areas  throughout  the state.  They will also be used to refine
the control strategies which may need to be applied to various   non-traditional
fugitive dust sources.


    The purpose of this study was to assess the effects of street  sweeping  on
air  quality  in  Champaign,   Illinois,  by  comparing  aerosol  concentrations
measured by high volume (hivol) samplers in the presence and absence of regular
street sweeping.


    This study was planned so as to  utilize  the  program of   regular  street
sweeping  being  conducted  in  Champaign  by  a  research group from the Water
Survey's Surface Water Section and funded by USEPA as part of   the  Nationwide
Urban  Runoff  Program  (NURP);  the  Principal  Investigator   of that  study is
Michael L.  Terstriep, Head of the Surface Water Section.
2  EXPERIMENTAL METHODS
2.1  AEROSOL SAMPLER LOCATIONS
    A map of a portion of Champaign, showing the aerosol sampling  sites   within
the  NURP  study  basins,  is  shown  in  Figure  1.  Three hivol  samplers  were
operated in the commercial "Mattis South" basin along Mattis Avenue.   As  shown
in the figure, two samplers were located on the west side of Mattis Avenue, and
one on  the  east  side.   An  additional  hivol  sampler  was   operated  in  a
residential  Champaign neighborhood; this is designated the John Street  site in
Figure 1.
                                     7-2

-------
 i
CO
                                                                                       JOHN
                                                                                       STREET
                                                                                       NORTH
                                                                                       BASIN
                                                                                       NO. 5
MATTIS
SOUTH
BASIN
NO. 2
                     MATTIS, EAST
                                                                 BASIN ^
                                                                 NO. 3
                                                                                                 JOHN
                                                                                                 STREET
                                                                                                 SOUTH
                                                                                                - BASIN ^
                                                                                                 NO. 4
 MATTIS.
 WEST (N)
       MATTIS. WEST (S)
                                                                            CHAMPAIGN. ILLINOIS STUDY AREA
                                                                          IEPA-ISWS URBAN STORM RUNOFF STUDY
                                                                            NATIONWIDE URBAN RUNOFF PROGRAM

                                                                          Aerosol Sampler Site       H Automatic Sampler Sites
                                                                             SCALE               9 Wet and Dry Fallout Sampler
                                                                                                  Raingage
                 Figure 1.   Map of a  portion of Champaign,  Illinois, showing  aerosol sampling sites,
                      drainage basins,  and sampling sites  of the Water Survey's  project in the
                        Nationwide Urban  Runoff Program  (Modified from Terstriep et al, 1981).

-------
    The hivol samplers near Mattis Avenue were all located 7.0  m   (  t  0.4   m)
from  the  curb  to  minimize  any  effects  of distance from the  street on  the
concentrations measured by the various samplers.  At the John Street   site   the
hivol was 17.6 m north of the curb.
2.2  AEROSOL SAMPLING METHODS
    All the aerosol samplers were installed so that their inlets were  2  m  above
ground level.  Thus, they conformed to the standard height, i.e.,  between  2  and
15 m, at which hivol inlets must  be  placed.   The  hivol  samplers   were  all
positioned  so  that  their  inlets  faced  the street.  It was important  to be
consistent on this detail since there is  evidence  (Ortiz,   1978)  that  hivol
sampling  effectiveness  varies  with  sampler  orientation to the air flow, at
least in strong winds.


    The standard hivol samplers were provided and calibrated  by the IEPA,  which
also  provided  replacement  pumps  when  airflow fell below  minimum acceptable
rates.  The IEPA also provided glass fiber  filters  for  the hivol   samplers,
weighed   the   filters   before   and   after  exposure,  and  calculated  TSP
concentrations.
    The hivol filters were changed  at  approximately  9:00  a.m.   each  Monday-
through Friday, from 16 April to 20 October, 1981.  This schedule  provided  five
24-hr duration filters each week from each sampler-  The first  filter   of   the
week began Monday morning and the last ended Saturday morning.
2.3  STREET SWEEPING OPERATIONS
    The street sweeper made single passes along both east and west  curbs  of the
Mattis South basin once per week, normally on Tuesday morning,  between  15 April
and 26 May 1981.  The sweeper used was a  1973  Elgin,  Model   Pelican   "S,"  a
three-wheeled  mechanical  sweeper with right and left side gutter  brooms and  a
main rotary broom.  Its sweeping path using one outside broom was 2.4 m (8  ft)
wide.
    Measurements of collection efficiency by Bender  (1982)  indicated   that  the
sweeper  collected  about  67/&  of  the  total  dust  mass  on  the  street.   This
approximate efficiency was maintained for all particle sizes down   to   125 ym,
below which efficiency decreased markedly.  Other measurements  by  Bender (1982)
indicated that 46$ of the total dust mass was in particles  greater than
in diameter.
                                      7-4

-------
    In six sweepings of the Mattis South basin, the median mass  collected  was
28.8  g/curb m (102 Ib/curb mi).  The maximum collection was 44.5 g/curb m  (158
Ib/curb mi) and the minimum 13.2 g/curb m (47 Ib/curb mi).
2.4  SUPPORTING DATA
    Several  kinds  of  supporting  data  were  also  collected   to   aid   in
interpretation  of  the  air  quality  data.  Precipitation was measured at the
Mattis Avenue and John Street  raingage  sites,  as  shown  in  Figure   1,  and
provided  to  us,  along  with  street  sweeping data, by the Water Survey NURP
project mentioned earlier.  Wind measurements were made  at  the  Water  Survey
Headquarters in Champaign, about 4 km east of the Mattis Avenue sites.
3  RESULTS AND DISCUSSION
    An illustration of a small portion of the available data,  plotted  against
time,  is  given in Figure 2.  The data include TSP measurements at four sites,
along with daily rainfall amount, dates of street sweeping,  and  winds.   Wind
directions  are  shown by direction category.  The westerly wind category (W in
Figure 2) includes days when the hourly mean wind direction was from the 202 to
338  degree  sector  for at least 75% of the hours between 6 a.m. and midnight.
Wind direction was not considered  relevant  in  the  remaining  hours  of  the
sampling  period because only a very small fraction of the traffic occurs then.
Similarly, the easterly wind category (E) includes days when  the  hourly  mean
wind  direction  was from the 22 to 158 degree sector at least 75% of the hours
in the same time interval.  The "other" wind direction  category  (0)  includes
all  other  days.   Mean  wind speed during sample collection and daily maximum
1-minute gust speed are also shown.


    Figure 2 shows that rainfall was frequent and relatively heavy  during  the
period  shown, with seven days having 1.27 cm (0.5 in.)  or more.  Streets were
swept at approximately weekly intervals during the period.  Figure  2  includes
only  a  small  fraction  of  the  data  collected,  but it illustrates several
relationships that occurred throughout the sampling period.  One  of  these  is
the  generally  high  correlation  between  TSP  observations  from the several
samplers, including the one located in a residential area.  Generally, on  days
when  TSP  concentrations  were unusually high at one site, they were unusually
high at all sites.  There  were  still  differences  in  concentration  between
sites,  of  course.   A  very  obvious  difference  in  Figure  2  was that the
residential  (John)  site  had  consistently  lower  concentrations  than   the
commercial  sites.   Differences  between  collectors  on opposite sides of the
commercial roadway are also apparent in Figure 2.
                                     7-5

-------
  280
   200
   100
    20

   1.5

   1.0
 •£0.5
       TSP
                    MATTIS, EAST SIDE
                    MATTIS, WEST SIDE (N)
                    MATTIS, WEST SIDE (S)
                    JOHN STREET
RAINFALL
i
I
_ p
fin ,

I

n
n


i
n
I ' '
STREET SWEEPING
I I
I I

I


I


I


I
w
E
0
WIND DIRECTION CATEGORY
- O OO OO
O
-O OO O
II I 1
1 i
0 O
o
O OO O
o
O 0
OO OO OO O
1 1 1
-
    18
    15
      - AVERAGE WIND SPEED
1 - 1
60
40
0
i 	 1 	 i 	 1 	 1 	 1 	 r 	 1 	 1 	
~ MAXIMUM WIND GUST
il 1 1 1 1 1 !
16 20 25 28 1 5 10 15 20
Apr May
                              STARTING DATE FOR FILTERS


Figure 2.  Illustration of data collected  in 1981 for  evaluation of effects
          of  street  sweeping on air  quality.
                                    7-6

-------
    In assessing possible effects of street sweeping on  air  quality,  it  was
first  necessary  to  establish whether the test roadway was indeed a source of
aerosols.  To examine this question, I divided the data according to  the  wind
direction categories mentioned earlier-  This was done to assess which samplers
were upwind, and which downwind, when the wind direction was either easterly or
westerly (i.e., having a large cross-wind component, relative to Mattis Avenue)
and to identify those days when the wind (1) was from either north or south, or
(2)  blew  from both east and west at different times during sample collection,
so that it was difficult to assign upwind and downwind.  Results for  the  east
and  west  wind cases are shown in Figure 3, and those for the "other" winds in
Figure 4.


    Figure 3 includes only measurements made in easterly or westerly winds, and
shows  a  clear  tendency  for the downwind samplers to have higher TSP values.
The upwind samplers experienced higher TSP concentrations with westerly  winds.
For  example,  the upwind means were 71 ug/m3for 26 days with west winds and 60
ug/m^for 30 days with east winds, a difference of 18J,  relative  to  the  east
wind  value.   For both east and west winds, however, the downwind samplers had
higher TSP values, clearly  showing  the  road  to  be  a  source  of  airborne
particles.   As  before,  the  solid  line  in  the  figure  represents perfect
agreement between upwind and downwind samplers.  The dashed line is  the   least
squares  regression line for estimating the downwind concentration, D, from the
upwind value, U: D = 0.94(U) + 21.7.  Since the slope of the regression line is
close  to  1.0,  the  intercept, 21.7 ug/nr gives a good indication of the  extra
airborne dust concentration that may be attributed to the roadway.   The   slope
near  1.0  shows  that  the  added  TSP  concentration caused by the roadway is
approximately constant for all upwind concentrations.


    To consider the  possibility  that  the  differences  in  mean  upwind  TSP
concentrations were the result of differences in mean wind speed, the following
wind speed data are presented.  The mean value of the daily  mean  wind  speed,
for  26  west  wind  samples,  was  6.4  m/sec  (S.D.   =2.0 m/sec), while the
corresponding value for 30 east wind samples was 5.0 m/sec (S.D.  = 2.1 m/sec).
Analogous  means  for the maximum gusts were 24.0 m/sec (S.D.  = 8.6 m/sec) for
west wind samples and 18.2 m/sec (S.D.  = 5.4 m/sec)  for  east  wind  samples.
Thus,  the  relative differences in mean wind speed (28$) and mean maximum gust
speed (32?) between easterly and westerly winds were somewhat greater than  the
relative  differences  in TSP concentrations, but wind speed differences cannot
be ruled out as a possible cause.  Another possible cause is that the site  was
closer  to  agricultural sources with west winds, since it was located near the
western edge of the Champaign-Urbana metropolitan area.


    Figure 4 shows the relationship between TSP concentrations on the east  and
west sides of Mattis Avenue for the "other" wind category.  There was no strong
tendency for either side of the street to  have  higher  values,  as  would  be
expected in the absence of a strong wind component perpendicular to the street.


    The evidence presented thus far makes it clear that the street  was  indeed
the  source  of  airborne  particles,  since  the  downwind  sampler had higher
                                    7-7

-------
                                              °  WEST WINDS

                                                EAST WINDS
              25
50
75   100   125   150   175   200   225   250

   DOWNWIND TSP, M9 m'3
Figure  3«    Comparison of TSP concentrations upwind and downwind  of  Mattis
         Avenue on days when winds were predominantly  from  the   east   or
         west.
                                   7-8

-------
   a.
   CO


   111
   o
   C/5
   LU

   S
180




162




144





126





108





 90




 72




 54





 36





 18
               i      i
            MATTIS AVENUE
              18    36    54    72    90    108   126   144   162   180


                             EAST SIDE TSP,M9 m~3
Figure 4.  Comparison of TSP concentrations  on  the  east and west  sides  of

          Mattis  Avenue on days  when winds  were not  predominantly from the

          east or west.
                                   7-9

-------
concentrations in both east and west winds, but neither side had-  predominantly
higher concentrations with "other" winds.  Next  we  address  the   question   of
whether street sweeping had any effect on reducing the strength of this  source.
This is done by comparing downwind-upwind differences during periods  with   and
without  regular  weekly  sweeping.   A day was considered to be in the  "swept"
period if the street was swept on that day or any of the previous   seven days,
and   in   the  "unswept"  period  otherwise.   A  second  comparison  examines
commercial-residential differences with and without  weekly  sweeping   (by   the
same criteria) at the commercial site.  (The residential site was  not in one of
the experimental  or  control  basins,  and  thus  was  swept  irregularly,   at
intervals of 2-4 weeks, during the sampling period.)


    Table 1 presents results of a two-way analysis of variance (Brown,  1977)  on
downwind-upwind  differences  ( TSP1).   It  examines both sweeping and  time  as
sources of variance.  The upper part of the table  gives  mean  downwind-upwind
differences  in  TSP  concentrations  for  swept  and  unswept  periods  in both
"spring" (up to 30 June) and "summer" (after 30  June).   There  appear   to   be
differences  between  seasons  (with  higher  concentrations  in summer  than  in
spring), but not between periods of sweeping and not sweeping.  The analysis  of
variance  table  in  the  lower  part  of the table confirms that  there  were  no
significant differences between periods of sweeping and no sweeping,  but  also
indicates  that  the  differences  between seasons were significant at the 0.10
level, but not at the 0.05 level.   Table  2  presents  results  of a   similar
analysis of variance on commercial-residential TSP differences ( TSP2).   Again,
differences between seasons appear large (with concentration differences  higher
in  spring  than  in  summer  in  this case), but differences between swept  and
unswept periods look small.  The analysis of variance table in the lower part
of Table 2 confirms that the differences between swept and unswept periods were
not significant, but also shows that seasonal differences were  significant   at
the  0.0001 level.  Interactions between sweeping and time were not significant
in either comparison.


    The main question addressed by this paper has now been answered.   Although
the  test  roadway  was  shown  to be a source of TSP, no evidence was found  to
indicate that street sweeping had any effect, either beneficial or detrimental,
on air quality at 7 m downwind.


    Table 3 summarizes TSP concentrations and the two TSP differences discussed
above.   Mean  TSP  values  are  shown  for  three different data  sets:  (1)  all
samples, (2) the samples used in the downwind-upwind comparison,   and   (3)   the
samples used in the commercial-residential comparison.  It is apparent that  the
mean TSP concentrations in the commercial  (Mattis  )  and  residential   (John)
sites do not vary much between sample sets.  For example, the Mattis mean shows
a maximum of 87 Pg/m3 and a minimum of 84 ug/m3 in the  three  data   sets   shown.
Similarly,  at  the  John  site  the highest mean was 65ug/m3and  the lowest 62
Pg/m3  Thus, regardless of the data set, mean TSP values were  higher  at both
sites  in  spring than in summer.  The Mattis site had higher concentrations by
about 20ug/nH in the spring and about 8 pg/m3 in the summer -
                                     7-10

-------
 Table 1.  Results of 2-way analysis of variance
on upwind-downwind TSP differences (ug/tn ) across
                the test roadway
                 Spring
Summer
No. of Cases
Mean
S.E.M.
Source of
variance
Time
Sweeping
Interaction
Error




Sum of
squares
603.4
20.7
23. A
8369.3
Table 2.
Swept
12
12.9
4.2
Degrees
freedom
1
1
1
41
Unswept Swept
11 4
12.8 19.8
5.2 1.8
of Mean
square F
603.4 2.96
20.7 0.10
23.4 0.11
. 204.1
Unswept
18
23.0
3.1
Tail
probability
0.0931
0.7518
0.7368

Results of 2-way analysis of variance
of commercial-residential TSP differences



(ug/m )

Spring Summer
No. of cases
Mean
S.E.M.
Source of
variance
Time
Sweeping
Interaction
Error



Sum of
Squares
2251.4
8.1
159.1
12166.8
Swept
27
22.0
3.0
Degrees
freedom
1
1
1
86
Unswept Swept
19 8
19.6 7.2
2.0 3.5
of Mean
square F
2251.3 15.91
8.1 0.06
159.1 1.12
141.5
Unswep t
36
11.0
1.7
Tail
probability
0.0001
0.8116
0.2919

                    7-11

-------
              Table  3.   Summary  of  TSP  concentrations  and differences
                  by  season for three  different sample sets
                                           Spring
                                           Summer
Sample set

All samples

Site
*
Mattis
John
N Mean -

53 86 -
52 65 -
S.E.

6
5
N Mean

77 62
74 54 -
S.E.

2
2
A TSP1
Mattis         23   87   -   7
Downwind-upwind          ,
                      difference
               23   13
                      33   65
                      22   22
A TSP2
      ***
Mattis
John
Mattis-John
difference
46   84   -   5
46   62   -   4
                                     46    21
68   63
68   55
                      44   10
  *Includes days  with at  least  one  sample  from both  sides of Mattis Ave;
     both sides weighted  equally.
 **Includes days  with measurements  on both sides  of  Mattis Ave.,  and east
     or west winds  (as defined  in the text).
***Includes days  with measurements  on both sides  of  Mattis Ave.  and at
     John St.,  with no restriction  on wind direction.
                                7-12

-------
    TSP concentration differences, however, display different seasonal behavior
depending  on  the  data  set used.  The mean commercial-residential difference
(&TSP2) was higher in spring than summer, as were TSP  concentrations  at  both
sites.   The  mean  downwind-upwind difference (ATSP1 ), on the other hand, was
larger in the summer, when the Mattis site had lower concentrations.


    The higher TSP concentrations in spring are consistent with an agricultural
soil  dust  source,  which  would be expected to be stronger in the spring from
tilling operations, lack of ground  cover  by  crops,  and  generally  stronger
winds.   It is reasonable for the Mattis site to have higher TSP concentrations
than the John St. site because it is closer  to  the  sources  with  prevailing
winds,  and  because  the  John St. site is located in a somewhat older part of
town, where the more mature  trees  may  help  to  remove  particles  from  the
atmosphere.


    The  downwind-upwind  differences  were  greatest  in  summer,   when   TSP
concentrations  were  lower  at  both the John and Mattis sites.  Others (e.g.,
Cowherd and Englehart, this Symposium) have found that roadway emissions depend
on  both  traffic  volume  and  roadway surface silt loading.  We could find no
evidence for significantly greater traffic volumes in summer  than  in  spring,
but  according  to data provided by Bender (1982) , the mean and standard error
of  11 spring measurements of total street loadings were 108 ± 5 g/curb-m, while
the  same  values for 10 summer measurements were 189  111 g/curb-m.  Thus, the
spring/summer ratio of loadings, 1.75, is very close to that  of  ATSP1,  1.69,
and  it  appears  that  differences  in street surface loadings between seasons
could well have caused the observed seasonal differences in ATSP1.
4  SUMMARY AND CONCLUSIONS
    To determine the effects of street sweeping on  urban  TSP  concentrations,
three  standard  hivol  samplers  were  operated  near  a  four lane commercial
roadway, and an additional sampler  was  operated  in  a  residential  area  of
Champaign,  Illinois, between April and October, 1981.  When winds were blowing
across the test roadway in either direction, downwind  concentrations  at  7  m
from the curb exceeded upwind by an average of about 20yg/m^ .  in a comparison
of downwind-upwind TSP concentration differences during  a  period  of  regular
weekly  street  sweeping  with  those  from  another  period  when sweeping was
discontinued, analysis of variance was unable to detect a significant effect of
sweeping  on the amount of TSP produced by the roadway.  Thus, for a commercial
urban roadway clearly shown to  be  a  source  of  airborne  particles,  street
sweeping  could  not be shown to have any effect in reducing the amount of dust
produced.
                                    7-13

-------
5  ACKNOWLEDGEMENTS
    This research was carried out under support from the Illinois Department of
Energy  and Natural Resources, Project 10.093.  Mr.  William P.  Murphy was the
Project Manager.


    The results were achieved only  through  the  dedicated  efforts  of  Susan
Wiley,  who  collected most of the samples and performed the data analysis, and
Bruce Komadina, who devised a convenient  means  of  mounting  the  hivols  and
supervised  their  installation.   Thanks are also due Randall K.  Stahlhut, for
programming assistance, Eberhard  Brieschke, who helped with the  installations,
and  Mr-   Kenneth  Porter,  Mr.    Lawrence  Boastick, and Dr.  and Mrs.  Glenn
Stout, who allowed us to operate  samplers on their property-  The Illinois EPA,
through  the  assistance  of  Mr.  Arden Ahnell, Mr-  Bob Button, and Mr-  John
Shrock, provided, calibrated,  and maintained the standard hivols, and  provided
and weighed the filters.  We also thank Mr.  Michael Terstriep and Mr.  Michael
Bender for their cooperation in supplying data.
6  REFERENCES

    Bender,  G.  M.,  1982.  Personal  communication.   Mr.  Bender is associated with
         the Water  Survey's  Nationwide  Urban Runoff Program.
    Brown, M..B., Editor,  1977:  BMDP-77 Biomedical Computer Programs,  P-Series,
         University of California  Press, Berkeley, 880 pp.
    Ortiz, C. A.,  1978:  Aerosol  collection characteristics  of  ambient  aerosol
         samplers.  M.S.  Thesis,  Texas A & M University, College Station, Texas.
    Terstriep,  M. L.,  G.  M.  Bender,  and D. C.  Noel, 1981: National Urban Runoff
         Project,   Champaign,  Illinois.   Evaluation   of   the Effectiveness of
         Municipal  Street Sweeping  in  the  Control   of   Urban  Storm  Runoff
         Pollution.    First  Annual    Report,    Prepared    for  the   Illinois
         Environmental Protection  Agency and the  U. S. Environmental Protection
         Agency,  Region  V.   Illinois State Water  Survey, Champaign.
                                       7-14

-------
                  EVALUATION OF THE EFFECTIVENESS
                 OF CIVIL ENGINEERING  FABRICS  AND
               CHEMICAL STABILIZERS IN THE REDUCTION
             OF FUGITIVE EMISSIONS  FROM  UNPAVED ROADS*
                                 by

                          Alan  G. Larson**
                         Donald L. Shearer

                TRC ENVIRONMENTAL CONSULTANTS, INC.
                       Englewood, CO   80111
                         Dennis C. Drehmel

      U.S. EPA, INDUSTRIAL ENVIRONMENTAL  RESEARCH LABORATORY
                 RESEARCH TRIANGLE PARK, NC  27711

                         Gary  W.  Schanche

                   U.S.  ARMY CORPS OF ENGINEERS,
            CONSTRUCTION ENGINEERING RESEARCH LABORATORY
                        CHAMPAIGN,  IL  61820
        This paper has been reviewed in accordance with the U.5. Environmental
        Protection Agency's peer and administrative review policies and approved ior
        presentation and publication.
(*)Although this is not the actual presentation made at the May 1982
   meeting, it describes the same work and covers the same general
   time period. It represents work funded,  in part, under EPA con-
   tract 68-02-3115.   The title of A.G. Larson's presentation was
   "Evaluation of Road Carpets and Chemical Road Dust Suppressants.
(**) Current address: The Kentwood Moore  Co.,  5690 Denver Tech
   Center Blvd.,  Englewood,  CO 80111.
                              8-1

-------
Introduction

    Emissions  of fugitive dust from unpaved  roads  constitute a significant
portion of  the  total  suspended  particulate  matter  (TSP)  emitted  in  the
United States.  These  roads may  be managed  by  local governments  for  the
use  of the  public  or  managed  privately  by an   industrial  user  (e.g.,
mining, lumber,  oil  and gas,  manufacturing).   Emissions of  fugitive  dust
from  unpaved  roads may  exacerbate an  air  quality  problem,  and  possibly
cause  violations  of  the  National  Ambient  Air  Quality  Standards  (NAAQS).
Therefore,  many  state  and  local   governments  have  required  dust  control
measures to be used on  unpaved roads.

    The  U.S.   EPA's   Region   8,   its   Industrial   Environmental  Research
Laboratory-RTF (IERL),  and the U.S. Army  Construction Engineering Research
Laboratory  agreed  to conduct  a  field  study at  Ft.  Carson,  Colorado,  to
evaluate the  effectiveness  of  two  potential  control  techniques—civil
engineering fabrics  (road  carpets)  and  chemical  dust  suppressants.   TRC
was contracted by the EPA  to  design and manage a  short-term study program
to evaluate these materials and to  coordinate with the Army on a long-term
study- The  Army  was enthusiastic about  conducting  the tests at Fort Carson
because of the  significant number  of  unpaved road miles on  the facility.
The Army wanted  to find an  effective,  but  inexpensive,  control technique
for   fugitive  road  dust   control.    EPA  Region  8  was   interested   in
participating  in  the study because an  effective  solution  to  the  fugitive
road  dust  problem  is  of concern  to Region 8.  EPA-IERL  was  interested in
evaluating  the  overall  effectiveness  of  these two  specific  dust  control
techniques.

    The short-term study was designed  to  measure  the  effectiveness of each
fugitive dust  control  application  at   two  points  in  time—immediately
following  application  of  the   road  carpet  or  chemical  suppressants
(measuring  maximum control efficiency)  and  after  a few months of road  use
by  local  vehicles  (measuring  average  control   efficiency).   Also,   by
limiting  the  measurements   to   a  short   time   period   (in  this  case,
approximately 1 hour),  the influence  of  weather  (i.e.  precipitation)  and
abnormally  high winds  was minimized  and  any  bias in  the  results  due  to
these  elements  reduced.   The   long-term study  was  designed  to measure  the
overall  efficiency of  each  dust  control   application.    The  effects  of
weather, abnormal  traffic patterns,  and   inconsistent  wind  patterns  were
measured.  By reclaiming  the  high-volume (hi-vol)  filters every  3 days,
the decay in  the effectiveness of  the  chemical  test sections  could also be
observed.  In   these   studies,  TSP   and   inhalable   particulate  matter
originating from each  section  of  road carpet or  chemical dust suppressant
were measured.

    Several   manufacturers   of   the   road   carpets   and   chemical  dust
suppressants   were contacted.    It was  the objective   of  the   tests  to
evaluate as many of the generically different  products within each  control
technology  as possible.   Space constraints limited the  test to three road
                                  8-2

-------
carpet  and   three   chemical   dust  suppressant  test  sections.   The  road
carpets consisted  of spunbonded  woven polypropylene,  needlepunched  woven
polypropylene,   and  nonwoven  polypropylene  sections.   A  cross-sectional
drawing of  the  road carpet  application  is  presented   in  Figure  1.   The
chemical  dust   suppressants   tested  were   an   oil-latex   emulsion,   an
elastomeric. asphalt  emulsion,  and a calcium  1ignosulfonate.   An untreated
road section was also maintained as a control.

Methodology

    The test  road   is an  existing gravel  road  located  on Ft.  Carson Army
Base,  just  south of Colorado  Springs,  Colorado.   Traffic  on  the  road  is
generally light  duty trucks  and  automobiles.   Occasionally,  a heavy duty
truck  or  troop  transport uses  the  road.   The test  road was  divided into
six test sections  and one control  section.  Each  section was approximately
200 yards (183 meters) long and 30 feet (9 meters) wide.

    The  three   road carpets  were each   installed  over  the   gravel  road,
employing  standard  road  constructor!  methods   with  special  techniques
applied  when recommended  by  the manufacturers.   The  road  carpets  were
covered with  approximately  6  inches  (15  cm)  or  2-inch  (5 cm)  or smaller
aggregate.

    For application  of  the  oil-latex emulsion and  the  elastomeric asphalt
emulsion,  the   gravel  road  was  first  scarified  to a  depth  of  4   to  6
inches.   The  latex  emulsion  was  applied as   1  part emulsion to  5  parts
water  solution  and  blade-mixed by several passes with  a road- grader.  The
scarified  section  was  then  graded  smooth  and packed.    After scarifying,
the  asphalt  emulsion section was windrowed  with  a road  grader  and  the
emulsion  applied  over   the   scraped  section.    After   several  windrowing
passes  with  the grader  to mix the  road base and  emulsion,  and further
application  of  the  emulsion,  the section  was smoothed and  packed.   The
calcium lignosulfonate  section was  graded smooth and the  chemical applied
to the gravel road surface.   No scarifying was recommended by the supplier.

    Hi-vol samplers  were  installed  on platforms  along  the  test road.  The
platforms each  supported a pair of hi-vols—one  for  collecting TSP and one
for collecting  particles  in  six  size  fractions  smaller than  30  ym.  The
hi-vol pairs were  located 14  feet  (4 meters)  above  ground downwind of each
test section at  a  distance of 50  yards (46 meters)  from the downwind edge
of  the  section.   One pair was located approximately 75 yards (69 meters)
upwind  of  the  test road (Figure  2).  The  operation  of the  hi-vols was
regulated by  wind  direction  controllers.   The hi-vols  operated  only when
the wind was  out of a direction ^60°  either  side of a  line  normal to the
road.  This operating mode helped  reduce  cross contamination from one test
section to  another.  An  additional  hi-vol to  measure   background  TSP was
located in an undisturbed  area of the  fort  several  miles  downwind of the
experimental area.
                                8-3

-------
                                     COARSE
                                  AGGREGATE
 « 1  •*   *   •••  • - • — •.•••••«.. ••»*•*    •••  ••»•  •  • ^   • w •  •  •
-V:::v^;-v;:.Vr.:^^-:v:V.;.v://>::^^^^/:»;v>v>VV-y:^
 • •: •-'• :.V'. v>''-/.-.'-."'-.:: • • ::-V-^; ;•".•:•: •>••: ••::: •:'•• •/•:-. -\V •/.;•:• :•:;.-: •• V: v; v.: f
:>:.« {•-•: .":••::;:.••;:.•;.• ;.\<.;::- >:-.v.v.^'.:-0': V >:' !>•:•: :VVA..^v:v;.; y.:j
••#vy:.;:^-;^
hTs*^l'laS^^^S^3a=22S^ "^*^
r»-Tra i-1'.*s>*^C*rSjy^j.iJLji^'^g^Vi^SirmrJS

                                                                      ROAD
                                                                    CARPET

                                    EXISTING
                                     GRAVEL
                                       ROAD
FIGURE  1:    ILLUSTRATION  OF  ROAD
                     CARPET  APPLICATION
                             8-4

-------
                                                  DBACKGnOUND HI-VOL
                RESIDENTIAL  AREA
     D
200YDSH
00
I
en
                       D
                       [
50 YDS
       D
                                                                DHI-VOLS
D
D
a
DMETEOROLOGY

     STATION

  D
PAVED
ROAD
FABRIC
FABRIC FABRIC CONTROL
*2 *3 SEC
CHEMICAL
iM
>\ o^^ D UPWIND HI-V

ELEMENTARY
SCHOOL

IHAf-FlU -• »-«-w
COUNTER METER

CHEMICAL
*,2
CHEMICAL
OL V.
TRAFFIC
COUNTER
PAVED ROAD


                                  NOT TO SCALE
                                                                    N
                           FIGURE 2:  FIELD TEST DESIGN

-------
    Wind speed  and direction  were measured  at  a meteorological   station
located approximately  150  yards  (137  meters) downwind  of Chemical  Section
3.   The  test  road was  equipped  with   traffic  counters  at  each  end   to
measure the  traffic  volume and patterns  on  the  road and  a  speed meter  in
the control section to measure vehicle speeds along the road.

    Following the long-term tests, each  chemical will  be  evaluated  for the
following criteria:

    •  Effectiveness in reducing TSP  and  inhalable  dust particles from the
       road surface.

    •  Frequency of chemical retreatment of test  road sections.

    •  Cost of chemicals for each  application and  costs over the course  of
       the 1-year test period.

    From  these  criteria,  the  cost   effectiveness  of  each  chemical  dust
suppressant will be determined.

Results

    At  the  time the  road  carpets were  installed,  only  2-inch or  smaller
sized  aggregate  was  available.   This   aggregate   had  a  relatively  high
percentage (8 percent,  by  weight,  less  than 97 y m)  of fine materials.   At
the request of local residents, the road  section was  watered to reduce the
dust.  Consequently, a hardpan road surface  developed  and any dust  control
potential  whi-ch  the  road  carpets may  have  exhibited  was  negated.    In
retrospect,  a washed  aggregate  appears  essential  for  testing  the  dust
control capabilities of  road  carpets.  Plans are  being made  to treat the
road  carpet  sections  with   chemical   stabilizers   and   to  measure  the
effectiveness of the chemicals in reducing fugitive road dust.

    Raw data have been  compiled  for total  suspended  particulate matter and
for  particles  less  than  30 ym  in diameter  for  the   three  chemical  test
sections,  the  control  section,  and  the  upwind  and  background stations.
The  sub-30 ym  particles   were  collected  in  six  stages,  with   average
particles sizes of:

    Stage 1     6.8 ym
    Stage 2     3.3 ym
    Stage 3     1.8 ym
    Stage 4     0.9 ym
    Stage 5     0.5 ym
    Stage 6     Remainder of particles
                                  8-6

-------
The  respirable  range  of  particles  is  a  combination  of  Stages  2-6.   The
total respirable particle concentration can be expressed as^':

   X  = (0.19 x S2)  + (0.70 x S3) + S4 + S5 + S6

    Where: X  " respirable particle concentration (ug/m^)

           S2 * concentration of particulate matter on Stage 2
           S3 = concentration of particulate matter on Stage 3
           S4 = concentration of particulate matter on Stage 4
           S5 = concentration of particulate matter on Stage 5
           S6 - concentration of particulate matter on Stage 6

    Particulate matter  concentrations  have  been  tabulated for  the first
nine  long-term  sampling  periods.   These  concentrations  are  presented  in
Table  I.   The  particulate matter concentrations measured  from the  calcium
lignosulfonate,  Chemical  Section   1,   are   consistently  higher   than  the
concentrations  measured  from  Chemical  Sections  2 and  3   and  the  control
section.

    The   manufacturers   suggested   application   rate   for   the   calcium
lignosulfonate  was  too  low  to  provide  sufficient  covering  of  the   road
surface  or to  allow for  absorption  of the  chemical  into   the  road  and
therefore did not result  in  measurable  dust  control.   The manufacturer has
since   modified  the   recommended   application   procedures   to   include
scarifying' of the road  surface  prior  to chemical  application  and  increased
dosage  of  the  chemical.    The manufacturer  has  al.so  required  ambient
temperature   consistently   above   45-50°F   before   application   of  -the
chemicals  can  begin.    Therefore,  the  collection  of  any reliable   dust
concentrations  will   have   to wait  until  the  calcium  lignosulfonate  is
reapplied in spring 1981.   One  thing  that appears to  apply to the  calcium
lignosulfonate  test  section  is that  grading  the  road prior to chemical
application loosened  enough  surface particles  to  cause particulate matter
concentrations  which  are  greater  than  those  measured  from  the  control
section.  The  control  section had  been  packed  by the  combined effects  of
precipitation and traffic.

    The  concentrations  from  the  road sections treated  with   the  oil-latex
emulsion  and  the elastomeric asphalt  emulsion are  generally  much lower
than  the  control section.   When  compared  with the  control  section,  control
efficiencies  of the  oil-latex emulsion  range   from  approximately   6-70
percent  for  respirable  particles and approximately  13-80  percent for  TSP.
Control  efficiencies   for   the   elastomeric   asphalt   emulsion  range   from
approximately  30-96   percent  for  respirable  particles  and   approximately
56-86 percent for TSP.
                                  8-7

-------
                                                                                      iAim: i
                                            HuiiHiirutl  ttonuviit rut loiiu   (im/1"  )  "I inii|ilrnlilu  mul lul«l mmi'didi'd  p«rllciil«t* matttr
oo
oo
SAHI'LUIC
puiuon
12-14-00
Rttiplriiblf
TSP.
17-17-80
Re up 1 cob In
TSP
12-20-BO
RcMplmble
TSP
12-21-80
RtfNplrnble
TSP
13-26-BO
Rcupl ri'blu
TSP
11-29-80
Rniipl rub In
TSP
1 - 2-81
Rcxplrublc
TSI1
1 - 5-81
Hnn|il rnblu
TSP
1 - 8-81
Rciplmblf
TSP
CIICMICAL f)
OIL-UlliX
rnui.sioii

111
282

<>
91

m
54

15
51

27
5)

4 ft
102

}i
82



94
25)

tt
57)

it
909
cimiRoi.
SI:CTIOH

116
)26

20;
454

24
44

51
102

62
105

26
269

76
162

«
274

244
434
uruniD
STATIOII

«•
)65

90
it

26
116

76
66

)8
85

>t
149

)4
)9

1 1
167

<4
102
BACKCROUXO
STATION

9
24

> >
71

1
12

t
14

II
12

n
15

5
1)

* *
24

t«
• t
                                  i   The"  conctrnt r«t loin  irn niu  dntii
                                 "   At tlm  tlimi  tlili r«purt  UUM ptL-pinml,  tliu  ilnln fur  lIu'Hi!  pulnM  lirnl IIVHII nuimurrd «nd
                                       uflgliud but  li«d not  ln'ril prUCCIIIMM!  llmnix') I'1" '|i»illly  MIUIIIHIIII » |i(ci||r»i«ii.

-------
    Because  of  the  impending  inclusion  of  three  more  chemical  test
sections,  the short-term tests have been  postponed  until  all  test sections
are  operational.   Two  short-term  tests  have  been  completed  for  the
existing test  sections.   The limited data  from these two  tests  show that
for the oil-latex  emulsion,  TSP  has been  reduced more  than  50 percent.
For the elastomeric  asphalt  emulsion,  TSP has been  reduced  by 90 percent.
Additional  short-term  tests  are planned  immediately after  application  of
the new chemicals and again in the summer. (Tables II and III.)

    The influence  of wind  speed, wind  direction,  and  vehicle  use  rates
have  not   been  compiled  for  correlation -with  the  particulate  matter
concentrations.  These correlations will be made in spring and summer 1981.

Conclusions

    The effectiveness of road carpets as  a  fugitive dust  control  technique
on  unpaved  roads is  uncertain.   The results  of  field  tests  to evaluate
this  effectiveness  are  inconclusive.   In  order  to  adequately  study  the
emission control capability of  the road carpets,  a clean,  washed aggregate
must be used over the carpet.

    The application  of  chemical dust suppressants  to unpaved  roads  is  an
effective  method of controlling  fugitive dust.  A wide  range  of  control
efficiencies can be realized.   For an oil-latex  emulsion,  up to  70 percent
control of  respirable  particles and up  to 80 percent  control of  TSP  has
been  measured.   For an  elastomeric  asphalt emulsion, up  to approximately
96 percent  control  of res.pirable particles  and up  to  86  percent of TSP has
been measured.   The  dependence  of the control efficiencies  on wind speed,
vehicle speed, and  weathering of the chemicals is still being evaluated.

    The cost:benefit  ratio  for each  chemical  application  is  also  being
evaluated.    This evaluation  will  include  the  initial  cost   of chemical
application,  costs  of successive  reapplication, and  control  efficiencies
achieved.   From  this  costrbenefit  ratio,  chemical  dust  suppressants which
can be  tailored  to  the economic and dust control  requirements of the user
will be presented.
 References

 (i'V.A.  Marble,  "A  Fundamental  Study  of Inertial  Impactors" (Ph.D.
      theses,  University  of Minnesota,  1970).
                                  8-9

-------
                  TABLE II
         LATEX  EMULSION CONTROL EFFICIENCY

DATE
12-14-80
12-17
12-18
12-20
12-21
12-23
12-24
12-26
12-27
12-29
12-30
12-31
1-02-81
1-03
1-05
1-06
1-08
1-09
3-24
3-26
3-27
3-29
3-30
4-01
4-02
CONTROL
SECTION*
325.8
400.0
453.9
215.1
44 .4
80.4
102.0
103.6
105.0
105.0
214.6
269.3
206.9
162.4
204.1
273.6
333.7
433.8
118-7
68.0
37.5
42.3
45.7
58.7
68.0
LATZX
EMULSION*
282.4
146.6
90.8
67.8
54.0
51.8
50.6
52.1
53.1
71.6
102.3
102.3
88.8
82.1
103.6
133.6
137.3
143.5
64.9
60.0
54.5
43.5
37.9
54.4
64.3
CONTROL
/.
13.3
63.4
80.0
68.5
—
35.6
50.4
49.7
49.4
31. S
52.3
62.0
57.0
49.4
49.2
51.2
58.9
66.9
54.7
11.8
—
—
17.1
7.3
5.4
(**) Below this line represents respirable particles;
    above this line represents TSP
                        8-10

-------
                      TABLE III

         EMULSIFIED ASPHALT CONTROL EFFICIENCY

DATE
12-14-80
12-17
12-18
12-20
12-21
12-23
12-24
12-26
12-27
12-29
12-30
12-31
1-02-81
1-03
1-05
1-06
1-08
1-09
3-24
3-26
3-27
3-29
3-30
4-01
4-02
CONTROL
SECTION*
325-8
400.6
453.9
215.1
44.4
80.4
102.0
103.6
105.0
105.0
214.6
269.3
206.9
162.4
204.1
273.6
333.7
433.8
118.7
68.0
37.5
42.3
45.7
58.7
68.0
EMULSIFIED
ASPHALT*
64.1
64.3
64.3
36.1
19.2
25.5
29.2
24.5
20.4
31.9
51.1
51.1
40.1
33.5
42.7
58.1
70.3
90.6
30.6
40.1
44.8
26.1
12.8
21.5
26.8
REDUCTION
Z
80.3
83.9
85.8
83.2
56.8
68. 3
71.4
76.4
80.6
69.6
69.6
81. 0 **
80.6
79.4
79.1
78.8
78.9
79.1
74.2
41.0
—
38.3
72.0
63.3
60.5
    i
(**) Below this line represents respirable particles;
    above this line represents TSP
                        8-11

-------
                EVALUATION OF WEATHERING CHARACTERISTICS OF DUST
                         SUPPRESSANT CHEMICAL ADDITIVES

          by:  William B. Kuykendal, Dennis C. Drehmel, Bobby E.  Daniel
               U. S. Environmental  Protection Agency
               Industrial Environmental Research Laboratory
               Research Triangle  Park,  North Carolina 27711
                                  ABSTRACT
      This paper presents  the  results of an experimental program to evaluate
 the effect of natural weathering on selected dust suppressant  chemical addi-
 tives.   A previous study  had  evaluated the performance of eight commercially
_available additives.  Four  of these were selected for evaluation of their
'weathering characteristics.   The four additives were:  Coherex, Lignosulfonate,
 SP 301, and Polyco 2151.  Three  concentrations of each additive were sprayed
 on a test panel of pulverized coal and then exposed to the weather for periods
 of 30,  60, and 90 days.   The  test panels were then evaluated in a wind tunnel
 to determine the performance  of  the additives after weather exposure.  The
 results show that no significant performance degradation occurred for any of
 the additives tested.
           This paper has be«n reviewed in accordance with the U.S. Environmental
           Protection Agency's peer and administrative review policies and approved for
           presentation and publication.
                                     9-1

-------
INTRODUCTION

     The Environmental Protection Agency has sponsored an active program
to develop technology to suppress fugitive particulate emissions.  A
facet of this program has been aimed at the reduction of fugitive particulate
emissions through the application of dust suppressant chemical additives.
In addition to the work described by this paper, two other projects have
been conducted by EPA to evaluate the effectiveness of chemical dust
suppressants.  In an earlier in-house study, Drehmel and Daniel1 evaluated
the relative effectiveness of 10 different dust suppressants using a
wind tunnel.  Their results will be briefly reviewed in the following
section.  Under a contract study, Larson^ evaluated four dust suppressants
in a field study conducted over several months on an unpaved road in
Colorado.

     Before chemical dust suppressants can gain widespread use, they
must demonstrate the ability to effectively suppress fugitive emissions
in a manner that is cost effective when compared to other fugitive
control options.  A major factor in determining cost effectiveness is
how often the dust suppressant needs to be reapplied.  Two factors
determine the durability of a dust suppressant:  wear and weathering.
In applications such as inactive storage piles where there is
little or no wear, the only factor affecting durability is weathering.
The purpose of this work is to evaluate the effectiveness of several
dust suppressants after exposure to natural weathering.

PRIOR WORK

     The initial wind tunnel evaluation on the relative effectiveness of
dust suppressants by Drehmel and Daniel tested 10 commercially available
products.  These products were evaluated in the same wind tunnel using
the same technique described in the experimental section of this paper.
Of the 10 dust suppressants evaluated, 8 gave consistent results and are
summarized in Table 1.  Each dust suppressant was evaluated over a range
of application rates to determine the effectiveness of each as the
application rate varied.  In order to compare the performance of each
dust suppressant on a common basis, the effectiveness for each, expressed
as entrainment velocity, was determined for an application rate calculated
at a cost of $300 per acre.  Entrainment velocity is defined later in
the paper, but the higher the value, the more effective the additive.
It should be emphasized that the dilution ratios determined by the above
method were not necessarily the ones recommended by the manufacturer.

     From these results four dust suppressants were selected for the
weathering study:  Coherex, Lignosulfonate, SP-301 and Polyco 2151.  The
selection was based on having a variety of chemical makeups and manufacturers
rather than on performance alone.

     A brief literature review of weathering test procedures was con-
ducted to identify the key weather parameters that should be included in
the experimental plan.  Since the purpose of the study was to evaluate
                                   9-2

-------
    TABLE  1.  ADDITIVE  PERFORMANCE RESULTS  FROM DREHMEL AND DANIEL
Additive
Coal Dyne
Coherex
CPB-12
Lignosulfonate
Oil & Water
SP-301
Pentron DC-3
Polyco 2151
Chemical
Unknown
Petroleum Resin
in Water
Acrylic Latex
Lignosulfonate
Mineral Oil
Synthetic Polymer
Acrylic in Water
Synthetic '
Copolymer
Additive Supplier Performance
Aquadyne
Wit co
Wen Don
Wen Don
Wen Don
Johnson & March
Apollo
Borden
16.8
13.4
9.8
18.6
18.3
6.4
9.1
15.2
^ntrainment velocity (m/sec)  for an application cost  of  $300/acre.
                                     9-3

-------
weathering effects on dust suppressants, which are usually sprayed on  coatings,
the available literature on paint and building materials was reviewed.   Three
publications3*4*5 "were located along with one ASTM procedure.6  Each reference
identified the key weather related parameters as:  solar radiation (particularly
ultraviolet), relative humidity, rainfall, and temperature.  In addition,
it was decided to vary the weathering exposure time so that a degradation  rate
could be determined.

EXPERIMENTAL

     Since an evaluation of the weathering characteristics was the primary
objective of the study, the selection of the weathering method was important.
Natural weathering has the advantage of being "real world" testing and
requires a minimum amount of experimental apparatus.  The disadvantage is, of
course, that the weather related variables cannot be controlled.  Simulated
weathering tests conducted indoors do offer a controlled weather environment,
but there are no standard accepted test procedures (see References 3 and 4).
In addition, they are experimentally complex.  For these reasons the natural
weathering method was selected.

     Having selected the dust suppressants to be evaluated and identified  the
key weather related parameters, it was then necessary to develop a test  matrix.
The variables selected were dust suppressant, concentration of the dust
suppressant, and weather exposure time.

     The selection of the dust suppressant concentrations required balancing
several factors.  It was decided to select three concentrations for each of
the four dust suppressants.  Furthermore, it was desirable from the stand-
point of the statistical design that the concentrations for each additive  be
selected such that the medium concentration was half way between the high
and low concentrations.  Care had to be taken in the selection of the high
and low concentrations to ensure that the experimental procedure would be
able to detect a result.  If too high a concentration were used, the maximum
speed available in the wind tunnel (30 m/sec) would not erode any of the
test material.  Conversely, too low a concentration would show no effect when
compared to an untreated test material.  Based on Drehmel and Daniel's results,
it was possible to select dust suppressant concentrations that would yield
results within the range of the experimental capability.  The values of  the
dust suppressant concentrations selected are given in Table 2.  It should  be
emphasized that these concentrations are not the concentrations recommended
by the manufacturer.  Rather they were selected to accommodate the experimental
procedure used so that weathering effects could be determined over a range of
concentrations.

     From a practical standpoint it was decided to limit the outside exposure
time to a maximum of 90 days.  It was felt that if a dust suppressant  were
still effective after this length of time that 'it would be useful as a fugitive
control technique.  Intermediate intervals of 30 days and 60 days were
selected in addition to zero days (no explosure).  This variable of exposure
time posed a problem in that the weather occurring during any 30 day period
would not be the same as for any other 30 day period.  The test matrix accommo-
dated this by including three replicates of the zero day tests, two replicates
                                       9-4

-------
      TABLE 2.  ADDITIVE CONCENTRATIONS TESTED
Additive
   Concentration,  1/m
Low      Medium      High
Coherex
Lignosulfonate
SP-301
Polyco 2151
0.022
0.005
0.005
0.0018
0.044
0.011
0.022
0.0034
0.066
0.016
0.038
0.0052
                           9-5

-------
of the 30 day tests, two replicates of the 60 day test, and one set of  tests
for 90 days.  All of the weathering tests were conducted between mid-August
and mid-November 1981 at EPA's Research Triangle Park, North Carolina,  facility.

     These variables were developed into a partial factorial design which
consisted of 96 tests.  These tests are summarized in Table 3.  One of  the test
samples was destroyed during the course of the experiment so that 95 tests
were actually conducted.

     The material selected for use in evaluating the dust suppressants  was
pulverized coal.  A sieve analysis gave an average size of about 200 pm.
The size distribution is shown in Figure 1.  The coal was then placed in a
test panel made of plexiglass which had a test volume 305 mm x 305 mm x 19 mm
deep. The test panel was always loosely filled with the coal, even with the
top of the panel, and any excess coal was brushed off.  The dust suppressant
was then applied with a spray bottle and allowed to dry overnight.  After dry-
ing, the test panels were placed in outside racks and exposed to the weather.

     Shortly after initiating the exposure tests a problem became apparent with
the planned exposure test procedure.  A severe thunderstorm occurred with the
resulting large rain drops impacting on the treated coal causing the surface
to be destroyed.  It was decided that the best solution would be to place
a cover over the test panels.  The selection of the cover material became
quite important.  Although the test panels would no longer be exposed directly
to rainfall it was necessary to maintain as many of the weather related par-
ameters as possible, particularly solar radiation.  Several candidate cover
materials were selected and evaluated on an ultraviolet spectrophotometer.
Ultraviolet was selected since it had been identified as a major factor in
prior weathering studies.  Two of the materials screened were judged to
be satisfactory.  These were polyethylene (0.0762 mm thickness) and styrene
(0.1905 mm thickness), and the results are shown in Figure 2.  Also shown in
Figure 2 is a plot of the relative energy of the sun as a function of wave
length.  As can be seen from Figure 2 both polyethylene and styrene transmit
about 70% of the incident solar radiation.  Polyethylene was selected as the
cover material because it was readily available.  The polyethylene was
fashioned into an open tent and placed over racks each of which held eight
test panels as shown in Figure 3.  This permitted free air circulation  under
the cover which allowed the test panel to "see" ambient temperature and
relative humidity.

     After exposure to the weather the test panels were brought inside  for
evaluation in a wind tunnel.  The wind tunnel had a 0.61 m square cross
section with an active length of 9.76 m.  The first 4.88 m was used for
establishing the desired velocity profile prior to the location of the  test
panel.  The test panel was located in the wind tunnel such that the coal
surface was flush with the bottom of the wind tunnel.  Past the test panel, the
remainder of the active section was used for disengagement.  Beside the
active section of the tunnel is an air return loop.  In the air return  section
is a baghouse to clean recycled air to the active section of the tunnel.
Design of the active section was intended to produce a turbulent boundary
layer similar to that found for wind passing over open ground or a storage
pile.  Details of this design are given by Viner et al.'
                                     9-6

-------
                         TABLE. 3  TEST MATRIX
Additive
Coherex
Lignosulfonate
SP-301
Polyco 2151

Concentration
Levels
3
3
3
3

Exposure
Days
0
30
60
90

Replicates
3
2
2
1
TOTAL
Planned
Tests
36
24
24
J.2
96
Actual
Tests
36
24
23
ii
95b
aA total of 12 tests.
 One test sample was destroyed during the experiment.
                                     9-7

-------
                                                      1000
         Figure  1.   Sieve analysis  of  pulverized coal.
           • AIR VS AIR
           • AIR VS POLYETHYLENE
           AAIR VSSTYRENE
           • AVERAGE SUNLIGHT
      200
225
250      275
  WAVELENGTH, nm
Figure 2.  Optical transmittance of candidate cover materials.
                           9-8

-------
                                                                               Cover Frame
CO
 i
CO
                                                                                       Polyethylene Cover
                                                             Test Panels
                            Figure 3.  Exposure  racks with  test panels and polyethylene  cover.

-------
     Each test panel was initially weighed prior to being tested in the
wind tunnel.  The wind tunnel was operated at several velocities, begin-
ning at the lowest and advancing monotonically.  After maintenance of a
velocity for 3 minutes, the tunnel was shut down to remove and reweigh the
test panel.  With the test panel put back, the next velocity was tested.
The process was repeated until the maximum test velocity was reached or until
the emission rate exceeded 13.04 g/min.  The emission rate of 13.04 g/min
was selected arbitrarily as a very high emission rate which would indicate
that the dust suppressant had failed and the test should be terminated.

RESULTS AND DISCUSSION

     In general each test conducted in the wind tunnel was similar in that the
dust suppressant nearly completely prevented erosion until a critical velocity
was reached.  At this critical velocity the dust suppressant failed quickly
resulting in a large emission rate.  This critical velocity, termed entrainment
velocity, was determined by measuring the velocity which produced an entrain-
ment rate of 4.89 g/min selected arbitrarily as indicating failure of the dust
suppressant.  Figure 3 illustrates the rapid failure of the dust suppressant
as the velocity is increased to the entrainment velocity.  For the case of
Figure 3 the entrainment velocity for 30 days outside is seen to be approxi-
mately 17 m/sec.

     The results for each of the four dust suppressants evaluated are given in
Figures 4 through 7.  These plots show entrainment velocity as a function of
weather exposure time for each of three concentrations.  In general each plot
shows that the entrainment velocity increased with increased dust suppressant
concentration as expected.  What was not expected was that there was no
detectable degradation in the performance of any dust suppressant over the
period tested.

     Several factors should be kept In mind when examining these data.  First,
as mentioned earlier, the concentrations tested were not intended to duplicate
concentrations suitable for field use.  The selection of pulverized coal in a
shallow test panel is likewise a departure from a practical application.  The
use of a polyethylene cover compromised the results by limiting the solar
radiation and eliminating direct contact with rain.  The exposure procedure
used deviated from ASTM D 1435 in that the exposure racks were oriented flat
instead of at the prescribed 45° facing south.  Finally, these were static
tests with no activity on the test panels.  Nevertheless, these results are
believed to be valid for the intended purpose of the study; namely, to evalu-
ate the weathering characteristics of the four dust suppressants tested.

CONCLUSIONS

     For the four dust suppressants evaluated it can be concluded that there
was no significant degradation in performance over the 90 day test period.
In addition, the experimental test procedures used are believed to be valid
and would have detected any significant degradation had it occurred.  These
are welcome findings because they mean that dust suppressant chemical additives
do offer an effective means of suppressing fugitive particulate emissions.
                                     9-10

-------
    32.5
    26.0
    19.5
    13.0
     6.5
             i     i     I    r
            DAYS OUTSJOC
                   0
                — 30
                mm. fiO
                                I    I     I     I
            J	I
                           12  15   18   21   24   27  30
                        VELOCITY, m/sec
                                                           2
   Figure 3.  Wind tunnel data:  Lignosulfonate  (0.005 1/m )
o
0
CO
Z
30
27
24
21
IS
15
12
 9
 6
 3
 0
            \
                  LOW CONCENTRATION
               -- MEDIUM CONCENTRATION
               -— HIGH CONCENTRATION
                  \    \    \     \     1
\
\
            10    20   30   40   50   60   70   80   90   100
                         TIME OUTSIDE, days
            Figure 4.  Weathering tests:   Coherex.
                           y-n

-------
B
§

>
g
w
M
3
12'!_

9

6
3
C

_
, LOW CONCENTRATION
- _ ^^ MEDIUM CONCENTRATION
	 ^. . HIGH CONCENTRATION
1 1 1 1 1 1 1


«•

—
_
1 1
 Ed
          10
      20
         80   90   100
             30   40   50   60   70


               TIME OUTSIDE, days


Figure 5.  Weathering tests:  Lignosulfonate,
u

-------
u

1
g
H
i

j • • * • ' ' •
24 fc_._ ._^ — — -11"*"*' *V»
J^« ^ •• ^* ^""^"^^••^^»^^. A ^J^
21 f ^"^
1
18
15
12
9
6

3
0
L * ^^"
•
-
_
_ » LOW CONCENTRATION
B MEDIUM CONCENTRATION
- ,-A.^ HIGH CONCENTRATION
1 1 1 1 1 1 1

1 1
	 ^A.
r*-" "
^x
"^4 I
_
—
—
_

—
1 i
10   20   30   40   SO   60   70   80   90
              TIME OUTSIDE, day*
Figure 7.   Weathering tests:   Polyco 2151
100
                  9-13

-------
                               REFERENCES
1.   Drehmel, D. C.,  and Daniel, B. E.  Relative Effectiveness of Chemical
     Additives and Wind Screens for Fugitive Dust Control.  In Proceedings:
     Third Symposium on the Transfer and Utilization of Particulate Control
     Technology:  Vol. IV.  Atypical Applications.  EPA-600/9-82-005d,
     July 1982.

2.   Larson, A.  Evaluation of Road Carpets and Chemical Road Dust
     Suppressants.  Presented at Fifth Symposium on Fugitive Emissions:
     Measurement and Control, Charleston, S. C., May 1982.

3.   Mitton, P.B., and Richards, D. P.  A Model for Objective Development
     of Accelerated Weathering Tests.  Journal of Paint Technology, Vol. 43,
     No. 563, December 1971.

4.   Hoffmann, E.  Weathering of Paint Films and Development of Accelerated
     Tests.  Journal of Paint Technology, Vol. 43, No. 563, December 1971.

5.   Masters, L. W.,  Wolfe, W. C., Rossiter, W. J., Jr., and Shaver, J. R.
     State of the Art on Durability Testing of Building Components and
     Materials.  NBS Report No. NBS1R73-132, March 1973.

6.   ASTM D 1435, ASTM Book of Standards, Part. 27, 1971.

7.   Viner, A. S., Ranade, M. B., Shaughnessy, E. J., Drehmel, D. C., and
     Daniel, B. E.  A Wind Tunnel for Dust Entrainment Studies.  In
     Proceedings:  Third Symposium on the Transfer and Utilization of
     Particulate Control Technology:  Vol. IV.  Atypical Applications.
     EPA-600/9-82-005d, July 1982.
                  TABLE FOR CONVERSION TO ENGLISH UNITS



          To Convert                 to                Multiply By

            m/sec                  ft/sec                  3.33

                                  gal./acre             1069.73

                                    acre                4047

                                     in                    0.0394

                                     ft                    3.28

                                   gr/min                 15.34
                                     9-14

-------
   ON THE USE OF SF^-TRACER RELEASES FOR  THE DETERMINATION OF FUGITIVE
                                EMISSIONS

               B.  VANDERBORGHT, J. KRETZSCHMAR and T. RYMEN
           Nuclear Energy Research Center,  B-2400 Mol, Belgium
                           F. CANDREVA, R.  DAMS
           Institute for Nuclear  Sciences,  B-9000 Gent, Belgium
     Summary

     The principle  and applicability  of  tracer releases  for the accurate
     measurement  of  fugitive emissions  is discussed  and  illustrated with
     some  practical examples.  At  the emission  source SFg-tracer  gas  is
     released at  a  well-known constant  rate.  With proper  adaption of  the
     tracer  emission  device  to  the fugitive emission   source thorough
     mixing  and  homogeneous  dispersion  of  tracer and  pollutant  can  be
     obtained. In this way the tracer to  pollutant  concentration ratio  is
     the same in  the  dispersed plume  as in the emission and the pollutant
     emission rate  can be  calculated  by measuring  the tracer to pollutant
     ratio at any place in the plume.

           The work described in this paper was not funded by the U.S. Environmental
           Protection Agency. The contents do not necessarily reflect the views of the
           Agency and no official endorsement should be inferred.

1. INTRODUCTION

     Although dispersion calculation  for  emissions from  high stacks pre-
dict a  maximum  ambient  air  concentration  at  a certain  distance from  the
source,  the  highest concentrations are  very  often measured  very close  to
the  factory.   Control  of point-source  emissions  has in  certain  problem
areas not produced  the anticipated  improvement in ambient air quality.  The
reason appears.to be fugitive emissions  at low- or ground-level  altitude.
Fugitive emissions  seem to  be  especially important for  aerosol emissions
from metallurgical  plants. For the  complete understanding of air pollution
situations - through emission-inventories,  mathematical  modelling etc.  -
it is necessary to  be able to quantify the fugitive emissions.
     Since the emission area  is  usually  not well defined, and since volume
flow rates are usually unknown,  the normal emission measurement  techniques
cannot be applied.  The measuring methods  actually described  in the  litera-
ture can  be   classified  into "quasistack"  sampling or roof  monitoring  in
which the  emission  area  is  by   some  means  physically restricted  or  into
"reversed modelling"  where emission is  calculated from immission measure-
ments and mathematical dispersion models.(1)
In  this  paper the  principle and possibilities of a  tracer technique  for
the  quantification   of  fugitive  emissions from  a metallurgical plant  is
described.
A common feature  of all  mentioned procedures  is  the discontinuous  charac-
ter preventing routine continuous emission measurements.

2. PRINCIPLE

     The principle  of  the  procedure  is as  follows :
A tracer  component  is discharged  at a  constant, well-known  rate at  the
emission site of  the  pollutant.  When  the tracer and the pollutant  are  well
mixed in  the emission plume  and when they are dispersed In the same  way,

                                   10-1

-------
then the  concentration ratio of  both  components will  remain constant all
over the plume and equal to  the emission rate ratio :
or
where x * concentration        [M.L~3
      Q - emission  rate        [M.T'1
      p - pollutant
      t - tracer

Knowing  the tracer  emission  rate and  measuring the  concentration of the
tracer  and  the pollutant  in the  plume give  the  possibility to  calculate
the  unknown pollutant  emission  rate  Q . It  is obvious  that the  quality
of  the  results will depend on  the  validity  of equation  (1)  for which a
good  homogenization and  equal  dispersion  of  tracer  and  pollutant  is
necessary.  The characteristics  of the tracer  release (place, time,  flow-
rate, temperature,  shape  of  nozzle)  and the choice of the  measuring points
for  the  simultaneous determination of  pollutant and tracer  concentrations
must be optimized as a function of this homogeneity.
     When  the  fugitive  emissions  out  of  a  factory  workplace  are  to be
measured,  different  sampling  strategies are  possible.  Suppose a building
in  which  two  furnaces are  the sources  of  uncontrolled emissions  of  metal
fumes  (figure  1).  Aerosol  is  evacuated from  the  inside atmosphere of  the
workplace  by means  of  natural draft  through  rooftop  ventilation openings
and  by  means of a ventilator with the hood a few meters above one furnace
and  the exhaust opening  in the  wall  or  the  roof of  the building.  These
emissions are  easily entrained in the  wake  on the lee  of  the building  and
the  source  cannot be considered as a point source.
For  the  determination of  these  fugitive  emissions  - by  the   method of
tracer  release and  measuring  tracer  and pollutant  concentration  -  three
different sampling  locations  are possible.

1)  Tracer and  pollutant are sampled outside at  distances  from a  few meters
   up to a  few hundreds meters downwind the building. With this  configura-
    tion the  total emission  of  the building is measured.
2)  Sampling units are placed along the ventilation opening in the rooftop.
3)  Isokinetic  sampling  is performed in the exhaust tube of the ventilator.

In  the  two  latter   cases  homogenization  of  tracer and pollutant  becomes
very critical. By  proper  adjustment   of  the  tracer  release and sampling
places  the  emissions  of  the  different  sources  are  discerned  and  the
emissions  through  the rooftop and through the  ventilator  can be  calculated
separately.  In  this paper  the  properties  of  the three  methods  will  be
compared.

3.  EXPERIMENTAL

     In  an  antimony  (Sb)  metallurgical plant  the  technique  has  been used
to  quantify the fugitive emissions  emanating from a  18m  high 30mx20m long
workplace   in  which  a  converter  and  a  refinery  furnace were  installed


                                   10-2

-------
(figure 1). One  of the experiments,  in this workplace,  will be described
here as an example. Uncontrolled Sb dust emissions  originated mainly from
the filling and  the  slag  removal from  the  convertor as well  as  from the
same  actions  on  and  the  emptying  of  the  refinery  furnace.  Minor leak
emissions  on  filter  installations  did also occur.  Sb aerosol  had  a mean
aerodynamic diameter  of 1 to 2 urn.
     Sulfurhexafluoride  (SFg)  gas was  used  as  a tracer. SFfi  is an inert,
non-toxic  gas,  stable up  to about  500°C,  routinely detectable up  to 50
ng.ra"3 and normal background concentrations  are below the detection limit.
SF6 was  discharged at  the convertor  (main emission  place)  at a  rate of
about 27 g.min"*.
     SFg and Sb were  simultaneously  sampled  at  three places in the roof of
the workplace  (W^  to W3  in figure  1), at  one place (WQ) in the exhaust
tube  of  the hood  above  the  convertor and  8  places  outside  at distances
between 15 and 180m from the hall  (Rj  to Rg  in  figure 2).
Up  to  this distance  sedimentation and  deposition  of Sb  aerosol is neglec-
tible,  and there  is  no  basic  difference between  the   dispersion  of Sb-
aerosol and SFfi gas.
The sampling  places  outside were  chosen in'such  a  way  that  interferences
from  other parts  of  the  factory were  minimal.  Four sampling periods of
each  30  minutes  were performed.  During sampling  the  Local wind direction
at  the  30m-level was between  200°  en  205° with a windspeed  between 7 and
8,5 m.s"1  and neutral atmospheric stability.  For the sampling points in
the roof  (W1  to W3)  the aerosol was sampled on the  same filter for the  4
periods as the access to these  places  was  difficult.
     Aerosol  samples  were  taken using  LIB-type low  volume  samplers with
Whatman  41 filters.  Sb  was  determined  by neutron activation  and X-ray
fluorescence. SF6  samples  were collected  by filling plastic  bags and were
gas  chroraatographically  analysed. For  further  details on the sampling and
analysis procedures see  reference (2).

4. Results and discussion

     The results of the  emission calculation are summarised in table I.

TABLE I  :  Sb emission from a workplace as  determined by  the concentration
           ratio  Sb/SF6 in  :  ambient air downwind of the  building (R1 -> R7)
                             ventilator exhaust  tubo (WQ)
                             rooftop of the building (Wj   - W3)
sampling
period
1
2
3
4
Average
Sb emission in kg/h
based on sampling points
Rj -» R7 r.s.
1.6 ± 0.4 23 Z
1.1 ± 0.4 33 Z
2.6 ± 0.5 19 Z
1.0 ± 0.2 22 Z
1.6 24 Z
W0
2.5
1.3
2.7
0.7
1.8
Activities of
Convertor
8' filling
6' filling 4' slag
12' filling
9' filling
Refinery furnace
17' slag
short emissions
3' slag

     r.s.  relative  standard deviation in procent on 7 measurements

                                 10-3

-------
                               30  m
    Ventilator
t
V/9////)

Con
tor

ver-
%.
/HO
'#,.
^Sb
% SF6 ^
00
'//
y/~ "
-f---j---T 	 "
W, W2 W3

Refine-
ry
furnace


> _oE51Jn3 	
o
3
Sb203  aerosol
                                             1-2
Fig. 1  - Layout of emission and sampling points in the  workplace
Figure 2  - Layout of emission  and  sampling points



                            10-4

-------
sampling place
W0
Wl
w2
W3
average
Sb emission in kg/h
averaged over 4 periods
1.8
1.4
1.4
0.9
1.4 ± 0.37
     The relative  standard  deviation  on  the  emission  determination for
each sampling period  through the 7 sampling points  (R1  +  R7) outside the
building and  on  a distance  larger than 30 m  ranges between 19  and  33 %
with an  average  of 24  %,  and no  sampling place gave a systematic devi-
ation from  the  average.    In sampling  point  Rg,   15 m downwind  of the
roofcenter,  on the other  hand the Sb/SFg  ratio  was  systematically higher
than in  the  other points.  At this  short  distance,  the dispersed plumes
from the  two  Sb  sources  (converter and furnace)  were  not yet  well  hom-
ogenized, resulting in a higher  Sb/SFg  ratio  in the plume  of  the refinery
furnace   than  in  the  plume  of the  convertor,  where  the SFg  emission was
performed.  At larger distance the  plumes of  the two Installations seem to
be  well mixed  since  the  Sb/SFg  remains   from place  to   place  constant
between  acceptable limits.  The considerable  aerodynamic  turbulence in the
wake at  the   leeside  of the  building,  is  responsible  for this  good re-
sult.
     The calculated  Sb emission  for  the consecutive sampling  periods is
compatible with  the  emission causing  activities on  the plant  apparatus,
except   that  the  emission during  period  4  is  lower  than expected in com-
parison  with  the  three former periods.
     The sampling  points W1  to W3  and the hood  of the  ventilator, and
consequently  also point WQ ,  are  placed  above  the convertor. Visual obser-
vation   of the emission plumes of  the  convertor and of  the refinery fur-
nace indicates that  emission results  obtained  by  these  points  can  in a
first approximation  be equalized  to the emission from  the convertor and
that the contribution  of  the  refinery furnace to the  measurements  in
those points  is neglectlble.
     With the measurements  Inside  the  workplace  - respectively  in the
rooftop   (W1   to W3) and in  the  exhaust tube (WQ) -  two  different methods
to calculate  the emissions  can  be used. With  the  first  method the global
pollutant emission is  obtained,  with the  second it  is  possible to deter-
mine what fraction goes through  the hood,  respectively the roof.
a. By using  the  total 'SFg  emission in  equation (1)  and   the  SFg  and Sb
   concentration respectively in the points WQ, W^,  W2  and W3, the  total
   Sb-emission from   the  convertor  is  obtained. Averaged over  the  four
   sampling periods,  the emission  results  obtained  by the points W«  to W3
   are   respectively  1,8 -  1,4 - 1,4  and 0,9  kg Sb/h (table  I) and  aver-
   aged   over  the four points :  1,38 ± 0,37  kg/h or 1,38 kg/h ± 27 %.  If
   SF6  and Sb were well homogenized in the  plume  above  the convertor the
   four  results should be the same.
   Considering the relatively long sampling time (4x30'), the  27 %  rela-
   tive  standard deviation  on the  emission calculation  indicates that the
   homogenization  of  Sb  and SFg  right  above  the  convertor is  not  com-
   pletely attained.
                                    10-5

-------
b. The measurement  of  the flow  rate  in the  exhaust  tube enables to cal-
   culate the mass flow of SFg and  Sb  through the hood and the ventilator
   (Qfl)-                                    ,  T  ,
   Subtraction from the total SF6 emission (Qgp J gives the SFfi emission
   through the roof (Qgp  ) by draft ventilation.  Substitution of this
   value in equation (1)  gives the  Sb  emission through the roof (Qst,)
   instead of the total Sb emission,  this  of  course with the condition of
   sufficient homogenlzation of  tracer and pollutant. The results of this
   exercise are summarized in table II.

TABLE II : Sb and SF6 emission through the ventilator' and the roof of the
           workplace
sampling
period
1
2
3
4
average
emission in kg/h

Q?F6 (i)
0.24
0.39
0.41
0.39
0.36
QSF6 (2>
1.48
1.18
1.20
1.23
1.27
Qsb CD
0.34
0.32
0.68
0.17
0.38
sampling
place
Wl
w2
W3
average
emission in
qR f'
kg/h
»)
1.08
1.07
0.70
0.95
(1) through the volume flow rate measurment  in the exhaust tube


<3) Q?>, - Q*    .
The  sum  of the  Sb-emission through  the  roof and  through the ventilator
should be  the same as  the  total  Sb  emission from the convertor determined
by  the  previous  procedure (sub  a).  Averaged  over  the  four  sampling
periods this gives :

                     h  Qsb  " °'95 "*"  °'38 * **3 k8 Sb/h

which  is  only  a 'few procents  less  than the  1.4 kg/Sb  found by the  other
procedure  (table I).
It  can easely  been  shown  mathematically  that  both  results  should  be
Identical  with complete mixing of tracer and pollutant, I.e. when
                  -  rxSb
     The  total Sb emission  averaged over  the  four  sampling periods,  is
slightly higher when determined  through  places  outside the building  (Qsh
•1.6  kg/h)  than through  the  sampling points in  the roof and the venti-
lator  (Qsb  -  1.4  kg/h).   This was  to  be expected  since  the places
outside  are  influenced by  the emissions  of  the  complete building  while
the points inside were  focussed  on  the  emissions of the convertor only.
Moreover  measurements  outside can  be  positively  interfered  by the  back-
ground caused  by  other parts of the factory. Measurements of  the Sb con-
centration upwind the building showed  that this  was a minor problem.
                                  10-6

-------
5.  Conclusions and epiloge

     Fugitive emissions can be quantified by  the  method  of  tracer release
and measuring the  tracer to pollutant  ratio  in the dispersed  plume.  The
critical factor  of the method  is the  homogenization  of the  tracer  with
the pollutant.  This  can  be optimized by proper  adjustment  of the tracer
emission device,   the  sampling time  and the selection  of  the  sampling
places•
With concentration measurements a few  tens  or hundreds meter away  from
the source, in the example of the paper this  is outside  the  building,  the
tracer  to  pollutant  concentration  ratio remains  constant  from  place  to
place within  acceptable  limits.  The  total  pollutant  emission  from  the
building can be determined.
With concentration measurements close  to the source,  for example  in  the
rooftop  and  the  ventilation of  a  workplace,  sufficient homogeneity  is
•ore difficult  to obtain.   Nevertheless can the emission  be  determined
with reasonable  precision,  even  in  presence  of  another source  close  to
the main source. The efficiency of an exhaust hood can be evaluated.
Within the sampling time, the emission  of the pollutant  can  fluctuate.  If
during this time the tracer emission and the  dispersion  conditions remain
constant, the time averaged pollutant emission can nevertheless be deter-
mined*
     The  procedure  has  been  used   to determine emission  factors  for
fugitive  emissions.  These  data  have   subsequently  been   used  for  the
mathematical modelling  of the Sb  concentration in the environment around
the  plant  (2,3,4).  The  procedure  has  also  been  used  to  determine  the
fugitive  emissions  of  a  metal  refinery  installation  before and  after
construction  of  a  system for  reduction of  the  fugitive emission.  With
this information the reduction  of the pollution level in the environment,
due  to  this  investment,  has been estimated using  a bi-gaussian disperion
model.

     This research was carried  out in  the "National R &  D Programma,
Leefmilieu-Lucht"  of  the  DPWB  (Ministery of Science Policy).

References

(1)  "Wanted : fugitive emissions"
     S.  Budiansky. Environmental Science  and  Technology V14, N8  (1980)
     904-905.
(2)  "Luchtverontreinlglng  door een  metallurgisch bedrijf.  Eindrapport
     Natlonaal R-D Programma Leefmilieu-Lucht
     Diensten  voor de  Programmatie  van het  Wetenschapsbeleid,  Brussel,
     1981.
(3)  "Influence  of  the  meteorological  input  data  on  the  comparison
     between  calculated  and measured aerosol ground  level  concentrations
     and depositions"
     I. Mertens, J. Kretzschmar and B.  Vanderborght
     Proceedings  of  the  "12th ITM  on  air  pollution  modelling  and  its
     application"
     NATO/CCMS Palo Alto  August 1981
(4)  "Depositie rondom een non  ferro  bedrijf"
     B. Vanderborght, I.  Mertens, J. Kretzschmar
     Extern N4 (1981) in  press

                                 10-7

-------
          AN ATMOSPHERIC  TRACER  INVESTIGATION OF FUGITIVE EMISSIONS
                 TRANSPORT  IN  THE  COLORADO OIL SHALE REGION

                               George A. Sehmel

                         Pacific Northwest  Laboratory
                          Richland, Washington  99352

            The work described in this paper was not funded by the U.S. Environmental
            Protection Agency. The contents do not necessarily reflect the views of the
             Agency and no official endorsement should be inferred.

  This paper is based  on work  performed  under the U.S.  Department of Energy
  Contract DE-AC06-76RLO 1830..  This work  was done at the Pacific Northwest
  Laboratory (PNL), operated for the Department  of Energy by Battelle
  Memorial Institute.

                                    ABSTRACT


      It is anticipated that spent  -shale-disposal-areas  could become  prime
fugitive  emissions sources.  Atmospheric tracer experiments  were conducted
to investigate transport  and  dispersion of fugitive  emissions  in the complex
terrain near a proposed spent  shale disposal  area.   The atmospheric  tracer
gas SFs was released from the  Federal  oil  shale lease Tract  C-a  in Colorado
in July 1981.  The SFs tracer  gas  release rate was 26 kg/hr  at a tracer
release height of  1.7  m.

     Airborne SF$  tracer  gas  concentrations were sampled  at  32 radio-
controlled bag-sampling stations (at  15-min sampling times)  and  along roads
with syringe grab-samplers.  Sampling  sites were 0.6 to 13 km  from the
tracer-gas release site.

     There were four experimental  time  periods for releasing and sampling
tracer gas:  two evening  and two morning time periods.  Results  include
information on the following:   a time-history of pollutant concentration
changes upwind and in  gulches  before,  during, and after sunrise;  identifi-
cation of major transport flow paths  in the evening  and morning; and the
behavior of pollutant  mixing  and channeling in valleys  and downwind  of valley
confluences.


                                      11-1

-------
                                 INTRODUCTION
     Spatial  variations in the airborne transport of pollutant plumes from
fugitive emissions from the developing shale oil industry are complex for
multiple local  flows in the oil  shale regions of Piceance Creek Basin of
northwestern Colorado.   Significant local  spatial variations in pollutant
plume transport are expected because of the complex terrain.  For instance,
overall drainage flows  affecting downwind  airborne pollutant plume concen-
trations are a combination of multiplicity of drainage flows from gully trib-
utary systems.   Each tributary has a drainage plume which may rapidly lose
its plume identity when mixing with other  drainage flow plumes.  The mixing
rates of these plumes cannot be adequately predicted.  When the tributary is
large, i.e., a gulch, cross-gulch mixing of the two converging plume flows
may be slow.  For instance, as was shown by Sehmel (1981) for August 1980,
cross-gulch mixing was  usually incomplete  at the confluence of the tracer
plume from Corral Gulch and the ambient air plume from Box Elder Gulch.  A
mixing distance was defined, for nocturnal drainage flow, as the distance
downwind of the confluence for which the two plumes were uniformly mixed
across the gulch.  The  mixing distance is  also compared to gulch widths
upwind of the confluence of the two gulches.  For the site and times inves-
tigated, the mixing distance was from 7 to 20 times the gulch widths, 0.3
and 0.1 km widths upstream from the confluence of the two gulches.

      In order to investigate downwind pollutant plume transport from poten-
tial  fugitive emissions, atmospheric transport investigations were conducted
with  tracers (Clements  et al. 1981, Sehmel 1981, and Whiteman et al. 1981)
                                              10 14 10 « M

                                              ——  *—f— «
                                              VIlOHITCMS         I
          FIGURE  1.   Map  Showing  Tract  C-a Location in Piceance Basin

                                     11-2

-------
 in  an  area  in  an  adjacent  to  Federal lease"  Tract  C-a  in  Colorado."  The site
 is  shown  on  the map  in  Figure 1.   The  site  is  in  the  Piceance Creek basin,
 -32  km southeast  of  Rangely,  Colorado.   Tract  C-a is  along  Corral  Gulch which
 drains  into  Yellow Creek,  a tributary  of Whvte River.  Previously,  both non-
 depositing  SF5 tracer gas  and depositing tracer particles were released
 from the  bottom of the  Corral  Gulch  in August  1980 experiments (Sehmel  1981).

 	Air  pollution control regulations are  based  partially  upon 3-hr maximum
"concentr'atfons, 24-hr concentrations,  and annual  average concentrations.   In
 contrast, gas-bag sampling time periods  were usually  15  min in this investi-
 gation.   This  short  time period,  compared to longer time periods for compli-
 ance with some air pollution  regulations, was  selected to investigate the
 physics of  air pollution transport.  The rationale for using  the 15-min time
 period  was  that estimates  of  air  pollution  concentrations can be made for
 longer  time  periods  by  integrating tracer concentration  data  as a function
 of  time,  possibly a  time period representative of a 3-hr air  pollution  regu-
 lation.   In  contrast, transport and  dispersion information  is lost  if time-
 averaged  tracer-gas  concentrations are experimentally determined only for
 the  longer  time periods specified in regulations  for maximum  air pollution
 concentrations.

     Even shorter sampling durations are important for investigating physics
 required  to  develop  either site-specific or more  generic airborne pollutant
 transport and  air pollution budget models in complex terrain.   The  sampling
 time requirement  will be dependent upon  the multiplicity and  complexity of
 pollutant plume flows and  the required evaluated  accuracy of  predictive mod-
 els  for different distance scales.   In this investigation,  more spatial  and
 temporal  details  in  identifying major  and secondary plume transport flows  and
 interactions were obtained from syringe  samples collected nearly instantane-
 ously.  An  advantage of using syringe  samplers was a dense  sampling grid at
 low  cost.  A disadvantage was that syringe  samples were  collected in a  time
 sequence, rather  than simultaneously throughout the sampling  grid.   Neverthe-
 less,  concentrations investigated with syringe samplers  were  essential  for
 interpretation of concentrations  investigated  with the widely spaced time-
 integrated  gas-bag sampling sites and  for identifying multiple plume flows
 and  interactions, both drainage and  upslope flows.

     The  objective of the  July 1981  experiments reported here were  to simu-
 late pollutant transport and  dispersion  from a proposed  spent-shale disposal
 site during  both  day and night time conditions.   The tracer simulant was SF5
 gas, a  non-depositing and non-reactive gas.  The  July 1981  SF6 tracer-gas
 experiments were  conducted with the  tracer  release site  at  the proposed spent
 shale  disposal site  on terrain just  north of Corral Gulch.   The terrain was
 a sloping, undulating area between the hilltops along Corral  Gulch  and  Dead
 Horse Ridge.
                            EXPERIMENTAL PROCEDURES
     Airborne tracer concentrations were  investigated  using  one  tracer
 release site and multiple  airborne concentration  sampling  sites.   Details
 follow which describe tracer release, site  locations,  gas-bag  sampling,
 syringe sampling, tracer analysis and meteorological data.


                                     11-3

-------
     Tracer release times raifged fro~m~3 to 9 hours.  There were four experi-
mental time periods:   two evening and two morning periods.  During the
evening time periods,  tracer release rates were constant from 1500 to
2100 MDT (July 15) and 1800 to 2100 MDT (July 17).  During the morning time
periods, tracer release rates were constant from 0330 to 0930 (July 19) and
0300 to 1217 MDT (July 21).	-	

     Tracer release times included airflow transition times around sunset and
sunrise.  The evening  experiments were conducted first.  The results for eve-
ning experiments were  less successful than those for morning releases.  On
the first night, the radio control system malfunctioned for automatically
collecting gas-bag samples as a function of selected time periods at remote
sites.  A power lead in the battery supply shorted.  On the second night the
system was operational, but the experiment was stopped early due to potential
safety hazards from lightning in a severe local storm.

                                Tracer Release

     During each experiment, the SF5 tracer gas was released at a constant
rate of 26 kg/hr at a  height of 1.7 m.  The tracer gas release rate was con-
trolled with rotameters.   The reported release rate was calculated from pre-
and post-weighting tracer release cylinders.   The S?s was released through
a 1.0 cm inside-diameter  tube.

     The SF5 was released from the gas phase in storage cylinders.  Buoyance
effects on the airborne tracer plume were minimized because tracer-gas cool-
ing is minimized during this release procedure, as compared to the greater
cooling during tracer  release and SFs plume subsidence that occurs if SFs
is released from the liquid phase in storage tanks.

                                Site Locations

     Tracer release and downwind sampling locations are shown on the map in
Figure 2.   Important locations are the tracer release site, the tracer-gas  .
sampling sites and the meteorological measurement sites.

     Airborne tracer-gas  concentrations were investigated by examining gas
samples collected in either gas bags or syringes.  Gas-bag sampling stations
were placed in four radio-controlled signal zones for controlling gas sampl-
ing pumps.   The zones  were chosen to selectively sample major plume flow
directions.  Sampling  pumps in each zone were controlled with tone-specific
radio-receivers placed in each zone.

     Locations of gas-bag sampling stations 1 through 32 are shown by num-
bered locations in Figure 2.  These thirty-two radio-controlled sampling sta-
tions encompassed the  release site.  Distances from the tracer release site
to these sampling stations range from 0.6 to 6 km.  Sampling stations were
located at sites with  road access.

     The principal  gas-bag sampling zones were zones 2 and 3, which encompass
regions of drainage flow  from the tracer release site.  Air pollution concen-
trations for upslope flows were also of interest.  Consequently, gas-bag
sampling stations in all  four zones were simultaneously activated (or deacti-
vated) for selected gas-bag sampling time periods.


                                      11-4

-------
   FIGURE 2.  Topographic Map Showing Locations of the Tracer Release Site,
              Sampling Sites 1 thru 32 in Four Radio-Control led Gas Bag
              Sampling Zones and Roads Driven During Collection of
              Grab Samples
                               Gas-Bag Sampling

     Five electrical-control outlets at each gas-bag sampling station were
used for activating (or deactivating) air pumps for filling gas sample bags.
These 5 L, double-walled, sample bags* were prepared by encasing each sample
bag with an outer polyethylene bag as a diffusion barrier to prevent contami-
nation from SFs in both ambient and adjacent sample bags.   An air pump, or
pumps, was attached to each electrical-control outlet, which was activated
(or deactivated) by a specific radio-signal tone.  The pump sampling times
and non-operating times were controlled by operation of a tone-signal radio-
repeater, which was in a line-of-sight of antenna for each sampling station.

     Each gas-bag sampling station had the following components.  An antenna
was attached to a bamboo pole, 5.2 m in height, which was taped to a steel
fence post.  The antenna was attached to a battery-powered radio-signal
decoder.  The decoder was a radio receiver that responded to one of the four
transmitted tone signals.  Each tone signal activated (or deactivated) sam-
pling pumps in only one sampling zone; the zones shown in Figure 2.  Electri-
cal wires extended upward from the five electrical-control outlets on the
decoder.  For most gas-bag sampling stations, each electrical-control outlet
was attached to only one sampling pump.  In this case, sampling tube inlets
for all five pumps were located at a sampling height of 1.5 m.  In other
cases, ten gas-sampling pumps were located at each gas-bag sampling station.
* Industry bag, 10 by 15 in., 2.5 mil, flow meter fully inserted, no tubes.
  B Bar B, Suite 34, 121 West Whittier Blvd. Lahabra, CA  90631.
                                      11-5

-------
Two sampling pumps were simultaneously operated at these stations by con-
necting each pump-signal outlet to two air sampling pumps.  Sampling tube
inlets were located at heights of 0.3 and 5 m, respectively.  Sample outlet
tubes from the sampling pumps extended into a double garbage bag within which
were five gas sampling bags.   One dark garbage bag was inserted into another
dark garbage bag to reduce light transmission into_gas sampling bags as well
as to physically protect the  bags.
                               Syringe Sampling

     Relative locations of syringe-sampling sites are shown by dashed  lines
in Figure 2.  The 50 cm3 syringe* samples were collected at distances  of 3
to 13 km from the tracer release site.  Most syringe samples were collected
along the sides of a triangle formed by roads along Duck Creek, Yellow Creek,
and the road adjacent to gas-bag sampling stations 17,  18 and 20, hereafter
called road 18.  Syringe samples were rapidly collected by driving along
roads, stopping, flushing a new syringe with ambient air by reaching out
through the driver's window,  collecting a syringe sample, stoppering the
syringe inlet, marking the time and odometer reading on the syringe, and
immediatey driving to the next sampling location.

                                Tracer Analysis

     Collected gas-bag and syringe samples were analyzed in a field labora-
tory for SFg tracer gas concentrations.   The laboratory was in the area
labeled MOP in Figure 2.  This laboratory was always outside the plume of
high SFg concentration.  Neverthless to further minimize the possibility of
sample contamination, all samples were analyzed for SF5 concentrations
before tracer was released for the subsequent experiment.
     The SFg concentrations were determined with an electron-capture gas
chromatograph** (GC).  The GC was field-calibrated with SFg calibration gas
contained in cylinders at concentrations from 20 to 2000 pptv (parts per
trillion by volume/volume) .  However,  concentrations in some samples were
greater than the 2000 pptv upper limit for the calibration gas.   In these
cases, a gas sample was diluted by using a 50 cm3 syringe.  Dilutions were
made by partially expelling high-concentration gas from a syringe and sub-
sequently drawing ambient air into the syringe.

                             Meteorological  Data

     Meteorological data were recovered at the tracer release site and at
sites labeled MET. 1, MET. 2, and MET. 3 in Figure 2.  Data are for 15-min.
averages.  Meteorological data included wind speed, wind direction and ver-
tical temperature gradients (T60 m - T10 m).  Winds coming from the north
are described as aoO  wind direction while winds coming from the west are
described as a 270  wind direction.   Meteorological data collected at the
MET. 1 site included wind speed and direction for both 10- and 60-m heights

 * Single-use syringe, Plastipak, BD-5663, Becton-Dickinson, Rutherford,
   NJ  07070.
** Field-Portable Tracer Gas Monitor,  Model 215UP, Systems, Science and
   Software, JaJolla, CA  92038.
                                      11-6

-------
However at other sites, only wind speed and direction were collected at one
height.  At the MET. 2 and MET. 3 sites, wind speed and direction were
measured at 10 m.  At the tracer release site, wind direction and speed were
measured at 2.1 m.
                           	-RESULTS "
     Meteorological data and airborne SF§ concentrations are summarized:
one figure is used for each experiment.  General features of these figures
are described here.  Subsequently, data for each figure (experiment) are
discussed.

     The order of presenting results for each experiment is the meteorologi-
cal data, the SFs concentrations determined from radio-controlled gas-bag
sampling stations, and SFs concentrations determined from syringe grab
samples.  Concentrations for SFs tracer are reported as pptv, i.e., 1 pptv
equals 10-12 parts by volume.  In the figures, airborne SFs concentrations
are superimposed on topographic maps.  Each map is for successive time per-
iods for which radio-controlled gas-bag samples were collected or for which
a series of syringe grab-samples were collected or for which a series of
syringe grab-samples were collected.  For both sample collection types, the
time period is in the figure title.  In addition, to indicate the collection
sequence for syringe samples, selected sampling times are also printed
adjacent to selected concentrations.

     Sulfur hexafluoride concentrations are shown in different graphical rep-
resentations.  For gas-bag samp.les, concentrations are printed adjacent to
each sampling station location (station locations are indicated by solid cir-
cles).  In some cases upper and lower concentrations are printed adjacent to
a solid circle.  In these cases, the concentration represented by the upper
printing is for the 15-m height while the lower printing is for the 0.3-m
height.  In most cases, however, only one concentration is printed adjacent
to a solid circle.  These concentrations are for a 1.5-m sampling height,
i.e., near-respiration height.

     Some SFs concentrations for gas-bag samples have special meanings.  In
these cases, either a zero or blank is printed adjacent to the solid circle.
A zero indicates that a sample was analyzed, but the concentration was below
a detection limit of 1 pptv.  If no concentration is printed, this means no
samples were collected because either an air pump malfunctioned or the
sampling station was not operated.

     Concentrations from syringe grab-samples are shown graphically.  The
graphical representations are oriented with the abscissa approximately par-
allel to roads used for collecting the respective syringe samples.  In most
cases, the roads (along Duck Creek, Yellow Creek,, and road 18 adjacent to
gas-bag sampling station 18) approximate the sides of a triangle formed by
roads.

     Concentrations shown are plotted at road-sampling locations projected
onto the abscissa.  Road locations include cross-creek valley sampling
traverses for both main roads and two-wheel side trails.  Main roads cross
                                     11-7

-------
valleys at projected locations of about 2.6  and  10 km along Duck Creek and
2 5 and 6.5 km along Yellow Creek.  Two-wheel  side roads cross valleys at
projected locations of about 2.5, 3.7, 5.5 and 7.2 km along Duck Creek and
at 7.4 km along Yellow Creek.  ....     ...     	

                               Results, July 15

     Meteorological data are shown as a function of time in Figure 3 for the
first evening tracer experiment.  Since the  frequency of temperature inver-
sions is relatively high, the occurrence of  temperature inversions during
each tracer experiment is noted;  i.e., T60 m - Tin m is Positive at
MET  1.  A temperature inversion began about 1900 MDT.   Other important
times noted are tracer release and tracer gas sampling  periods.  Tracer
release was from 1500 to 2100 MDT.

     At present, these data are  included for completeness since airborne
tracer concentration data reduction is incomplete.  Analysis of the data col-
lection record will need to be completed before  the airborne tracer concen-
trations can be reported as a function of sampling time.  Record analysis  is
required since this is the experiment  in which an electrical short caused
transmission of spurious control  signals to  the  radio-controlled gas-bag
sampling stations.
                    i
 6O


 40


 20


360


340


320


300


280


260


240


220


200

 14
 12
 10
 8
 6
 4
 2
 0
 +2
 +1
 0
 -1
 •2
                           MET-1.Ttom-TiOm
                         16  16  17 18  19  20 21  22  23 24
                                     TIME. MOT
          FIGURE 3.   Wind Direction, Speed and Temperature  Gradients
                     on July 15, 1981
                                       11-8

-------
                               Results, July 17

     Meteorological data are shown as a function of time in Figure 4a for
the second evening tracer experiment.  Temperature inversions below 60 m
were intermittent and were only established at about 0945 and 2045 MDT.
Other important times are tracer release and the three tracer gas sampling
periods.  Tracer release was from 1800 to 2100 MDT.  Tracer release was
stopped due to a local storm.  There were three tracer sampling time periods:
two gas-bag sampling periods from 1945 to 2000 and 2043 to 2053 MDT and one
syringe sampling time period from 1946 to 2235 MDT.

     It is significant to note the first gas-bag sampling period was con-
ducted between 1945 and 2000 MDT.  As shown in Figure 4a, this was a time
period of rapid wind direction change with a variable wind direction at the
tracer release site.  Surface and upper wind directions were decoupled.
Upslope winds at the tracer release site were decoupled from upper wind
directions, even at 10 m.  The 10- and 60-m wind directions at MET. 1,
MET. 2, and MET. 3 were significantly different (about 330 to 0° at
2000 MDT) from the 2.1-m surface wind direction (30 to 165° at 2000 MDT) at
the tracer release site.

     Decoupled air flow directions are reflected in concentrations shown in
Figures 4b and 4c for both gas-bag sampling periods.  As shown in Figure 4b,
tracer plume transport was to the northwest.  Significant airborne tracer

in
Ul
HI
K
0
HI
a
2
g
u
Ul
K
5
Q
z
I


1
E
Q
HI
CL
CO
Q
Z
i


y_
(-"
80


360

280


200


120

40
0
-
_

, .. -
ff
- !

1 •
"" A r i ~

».X/ //'
- 	 3£] /t -

'1.1,1.1'














—


-

_

I
$

i/
K- \ ~
K * .
V
1- —
ll'l
S I-
w» ^* «'-
^> M K '
' \y p ~

V.
~ ' VARIABLE 2

/>_
r/v-J ,]
18 19
16
14
12
10
8
6
4
2

™ ^
"" »

- V //fVvy -
- ^h3r 	 v •
_ 	 ("\ _
- » » ~— ! -
r T T .sv', i














20 21
TIME. MDT



MET-1
— —60m
	 10m


+1
-1
- MET 1, Teom- Tt0m -
" I . I . i . I ~




SITES
MET-2 MET-3
——10m 	 10m

TRACER
RELEASE
	 1.5m

                          18  19   20  21
                            TIME, MDT
         FIGURE 4a.  Wind Direction, Speed and Temperature Gradients
                     on July 17, 1981
                                     11-9

-------
                                               l -v   1  -•  .
                                                   1.6 km (1 MILE) GRID SPACING
                                                   (20 ft CONTOUR SPACING)
                                                     RADIO-CONTROLLED BAG
                                                   CONCENTRATIONS SHOWN
                                                   NEXT TO EACH SITE ARE
                                                   pptv
 FIGURE 4b.   Airborne  Tracer Concentrations from 1945  to 2000  MDT,
               Collected at the  Radio-Controlled  Sampling Sites
               During the Evening  of July  17, 1981
                    J--Z PROPOSED. 4 •  - -  .•.
                  £f SURFACE- SPENT ,^S '   "  \
                                                         MILE) GRID SPACING
                                                   '2° *CONTOUR SPACING)
                                                   SF, SAMPLING

                                                   4 RADIO-CONTROLLED BAG
                                                     SAMPLING SITES       I ,

                                                   CONCENTRATIONS SHOWN  / i
                                                   NEXT TO EACH SITE ARE
                                                   pptv
FIGURE  4c.  Airborne Tracer Concentrations from 2043  to 2053  MDT,
             Collected  at the Radio-Control led  Sampling  Sites
             During the Evening  of July  17, 1981

                                     11-10

-------
concentrations were measured at gas-bag sampling sites 1, 11, and 10 (refer
back to Figure 2 for site identification).  The maximum concentration was
2400 pptv at site 1.

     Indices of both horizontal and vertical diffusion were measured during
this experiment.  Since tracer concentrations at site 11 were 58 pptv at
0.3 m and 42 pptv at 5 m, concentrations decreased over one order of magni-
tude in the 1.3 km horizontal distance between site 1 (2400 pptv) and site 11
(58/2400 = 0.02).  The tracer plume tended to remain adjacent to the ground
during upslope flow.  The concentration decrease with height at site 11 was
0.7; i.e., the ratio of 42 pptv at 5 m compared to 58 pptv at 0.3 m.

     The direction of the maximum tracer concentration shifted and concen-
trations decreased in the subsequent sampling time period, from 2043 to
2053 MDT.  The tracer plume flowed counter-clockwise from the tracer release
site.  Tracer concentrations were essentially constant at 2 pptv west of the
MDP area.

     For this time period, upslope plume flow from the tracer release site
spilled into Corral Gulch.  Vertical concentration profiles at sites 25 and
27 reflect plume spillage into Corral Gulch.  A maximum concentration of
680 pptv was measured at site 25.  In contrast to the previously discussed
vertical concentration profile decrease for upslope flow at site 11, the
relative concentration increase with height at site 25 was a factor of 1.3;
i.e., the ratio of 680 pptv at 5 m compared to 510 pptv at 0.3 m.  At this
site, the tracer plume was overriding drainage flow in Corral Gulch.  How-
ever at down-gulch site 27, vertical mixing was complete for the two heights
investigated.

     Plume flow inferences are made since the tracer-plume path was almost
certainly between sampling sites 26 and 31.  Concentrations were 0 pptv at
site 26 and 1 pptv at site 31.  Although the tracer plume rotated counter-
clockwise, the main plume width was either relatively narrow (the 1.5 km
between sites 26 and 31), or the main plume was elevated at sampling sites
with 0 or 1 pptv concentrations.

     Concentrations from the more spacially frequent syringe samples shown
in Figure 4d confirm, in more detail, plume-location and plume-width results
from gas-bag samples shown in/igure 4b.  Syringe samples were rollected
along Ridge Road, which is adjacent to sampling sites 1 and 11.  The main
plume width was narrow; concentrations decreased over two orders of magni-
tude from the maximum along a plume width of about 0.7 km.

     Concentrations from syringe samples reflect tracer plume segments
extending across Ridge Road, especially between 4 and 7 km.  These rela-
tively high concentration plume segments were of short time duration since
corresponding 15 min gas-bag sample concentrations were less, e.g., 1 and
2 pptv at site 10 in Figure 4b.

     Tracer plume flow extended into the bottom of Big Duck Creek Valley, a
valley immediately north of Ridge Road.  Concentrations are shown in Fig-
ure 4d;  data for a time period between 2145 and 2235 MDT.  Samples were col-
lected during heavy rain after tracer release was stopped at 2100 MDT.


                                      11-11

-------
      FIGURE 4/1.   Airborne SFs  Concentrations Along Roads for Sampling
              "~   Times  from 1946  to  2235  MOT on July 17, 1981


                               Results.  July 19

     Meteorological  data are shown as a  function of time in Figure 5a for the
first morning tracer experiment.   Establishment of a temperature inversion is
unknown since the 60-m height wind and temperature instruments at MET. 1
were not operated.  Important times are  tracer release and the tracer
gas-sampling periods.   Tracer release was  from 0330 to 0930 MDT.  There were
seven tracer sampling  time periods:   five  gas-bag sampling periods were 0533
to 0548, 0700 to  0715, 0745 to  0800,  0930  to 0845 and 0937 to 0952 MDT; and
two syringe sampling time periods  were 0537 to 0652 and 0704 to 0946 MDT.

     Concentration data reflect little tracer-plume spread during drainage
flows before sunrise,  plume spread and dilution after sunrise, and subsequent
plume lifting.  A coherent plume is shown  in Figure 5b with relatively high
centerline concentrations, 2200 to 2300  pptv, along sampling sites 18, 19 and
20.  As shown in  Figure 5c between 0700  to 0715 MDT (an hour after sunrise at
0615 MDT, sunrise was  delayed by a cloud), plume spread is reflected  in con-
centrations of 14 and  19 pptv at sampling  site 10.  Plume spread may reflect
increased wind direction variability at  the tracer release site.  Even with
increased plume spread,  the concentration  remained 2200 pptv at site 18.
This high concentration was caused by interactions of a main drainage flow
plume with drainage flow containing tracer.  (More detailed evidence for
plume interaction was  obtained  in  the next experiment on July 21).  Further
plume spread and  also  plume lifting are  shown 45 min later in Figure 5d.
Concentrations at site 18 decreased an order of magnitude to 171 pptv.

     Concentrations continued to reflect increased surface heating effects
with increasing time.   As shown in Figure  5e for 0830 to 0845 MDT, continued
                                        11-12

-------
                                   6    7    8     9    10


                                    TIME. MOT



        FIGURE 5a.   Wind Direction  and Speed  on  Ouly  19,  1981
                      a PROPOSED:  .57'  -- •*.
                     SURFACE • SPENT . «Z300 " \
                     RETORT . SHALE ".  ,«2


                             '*""' "
                                                    \        .M»W." ». Wi.S -^


                                                    .6 km (1 MILE) GRID SPACING U

                                                    tn *• nnhiTnuR GDAPIURI   i A
                                                   (20 ft CONTOUR SPACING)


                                                   SF. SAMPLING


                                                   • RADIO-CONTROLLED BAG /
                                       cASCtf



FIGURE  5b.  Airborne Tracer Concentrations from  0533 to  0548 MDT,

             Collected  at the Radio-Control led Sampling Sites

             During the Morning  of July  19, 1981
                                        11-13

-------
                                                      ^i'ki'.w?^
                                 9  ,:
                     S"5.FA«- •«"! 2so   "  *
                                                    1.6 km (1 MILE) GRID SPACING
                                                    120 ft CONTOUR SPACING)
                                                    SF. SAMPLING

                                                      RADIO-CONTROLLED BAG
                                                      SAMPLING SITES

                                                    CONCENTRATIONS SHOWN
                                                    NEXT TO EACH SITE ARE
                                                    pptv
                                                                i  -J-*  '
FIGURE  5c.
Airborne Tracer Concentrations from 0700  to 0715  MDT,
Collected at  the Radio-Controlled  Sampling Sites
During the Morning of  July  19, 1981
       '.  "-  "  .4j^i^^}jU^=«gSi$g,
                      .k^^rP^ff-':  >'

           :       '   T"  ^^^—
                     -'L v.j^s5L*r* ;..'  ,<-,
                      ^^_J,9,,.;

              "'y^x'^ff/V'i^'  -' \\ •••''
      -0
          ©»!;••••;'    .,
         ^t:LV-;r^
              ^»aK*
              ; -LEASE TRACT C-.i
              l! '»."•'>-^!.. •••.•!•>
              \  .•-_•!• .»»>1«>'.-' v ^-.rs^
             1.6 km (1 MILE) GRID SPACING Vj
             (20 ft CONTOUR SPACING)    ^

~-j~  Y~'~*  i  SF»SAMPL|NG          *-
 >• ' I   '  f  •  RADIO-CONTROLLED BAG  •/
 "&'  ' '*•' V -"-   SAMPLING SITES       ;;
 •'"''l1 ;  
-------
                            <3*^
                          j-|-3S**-,)>-Tr,
                   '    {-   ^K^"~'
                •    .^f^^
              '.<*?*&*..  ff/X-' - ::\ -,
               -•]*?•    ."X^Uu..:_.[-i™e..
           W^- . I-1 -"yJ-^SR.'*. 3oi"°' -A
                      SURFACE- SPENT ,,in'  "  V
' 910 DUCK CHEEK
               ^^••'&M.['t''^ -x
                _> •./  > JF'i* -.- A .' : • -»•*. i
  <^:^fffe	
  -r^?Wl^!f--
,  T-7'-Y*$Hf'-VV;"f
v.     •  • •-.IjfcT*^*«™"v.  .-.;•/.  •
)-K -  5-^>i'A--^/'.••!.- ••-'
                                   1.6km(1 MILE) GRID SPACING
                                   (20 ft CONTOUR SPACING)

                                   SF. SAMPLING

                                    RADIO-CONTROLLED BAG
                                    SAMPLING SITES

                                   CONCENTRATIONS SHOWN
                                   NEXT TO EACH SITE ARE
                                   pptv
                                                              l'\
     FIGURE 5e.  Airborne Tracer Concentrations from 0830 to 0845 MDT,
                Collected at the Radio-Control!ed Sampling Sites
                During  the Morning of  July 19, 1981
plume spread  included  site 1.  However by 0937 to 0952 MDT, plume lifting
resulted  in low-surface  concentrations  shown in Figure  5f.

     Concentrations determined from syringe samples confirm that tracer-plume
spread was limited during drainage flows before sunrise.  As shown in Fig-
ure 5g near site 18,  syringe-sample concentrations were similar to 15-min
gas-bag concentrations shown in Figure  5b.  The main-plume width was narrow;
concentrations decreased over two orders of magnitude from the maximum along
a plume width of about 1.2 km.

     Except for concentrations along Ridge Road, syringe data reflect that
the tracer plume flowed  beyond the terrain investigated with gas-bag sam-
plers.  Syringe-sample concentration data in Figure 5g  and 5h show the
following:

  •  a time-history of pollutant concentration changes  in gulches
     before, during and  after sunrise

  •  identification of major drainage transport flow paths at greater
     distances than investigated with gas-bag sampling  sites

  •  pollutant mixing at the confluence of two valleys.
                                   11-15

-------

         BIG BUCK CBEEK^
       -j^/-,. -•.:; •  > 1* PROPOSED ^ 2,
        / '/ "    x-S'SURFACE- SPENT ,
                 -  RETORT SHALE
                             CORRAL GULCH
                                  ' 'SV1--' Vv '"•* :' i W':; A^V*-;
                                  	It'-l' j -vi i . i	„! liCi' .Jj 'vJ w J>-
                                  ^^CTsasraS^SftN ^\
                                  •.^vIr;1/i--?'-/f^/..tV'»-
 2  39 ,-^S  '   •    ^T/tveaowcBEEK?-,-  - '•.' ,
rr;  '/s,'-.. ^^.Si!?'.'"":,''
                                  'J7x' 'f '•'   1.6 km (1 MILE) GRID SPAC
                                  //•;• I" '^    (20 ft CONTOUR SPACING)
                                 FV1 '
                                  '^'
             /
      -r I 2
      !-J * '
        .f -H.
 ."•' •   ,«;Jl I
_:	,.  'i ''So 1
;'-v*il
                                             SF. SAMPLING
RADIO-CONTROLLED BAG ,'
SAMPLING SITES
                                          1 ^  CONCENTRATIONS SHOWN
                                             NEXT TO EACH SITE ARE
                                             pptv
FIGURE 5f.   Airborne Tracer Concentrations from 0937  to  0952 MDT,
            Collected at the Radio-Control!ed Sampling Sites
            During the Morning of July 19, 1981
 FIGURE  5g.  Airborne SF6 Concentrations  Along  Roads for Sampling
            Times from 0537 to 0652  MDT  on  July  19, 1981
                                11-16

-------
      FIGURE  5h.   Airborne SF6  Concentrations  Along  Roads  for  Sampling
      -   Times  from 0704  to  0946  MDT  on  July  19,  1981


     In addition   Figure 5g shows  that the effects of  surface  heating imme-
diately- alter sunrise9 (06ll MDT) directed  a surface ff%£%^*£«
Horse Road.  Upslope flow was present between  0629 and 0639 MDT,  the tracer
o?ume divided along Ridge Road into two identifiable plumes.  The high  con-
?en?rat on near site 17  at 0629 MDT continued  to reflect the tracer following
drainage flSwT  In contrast, concentrations near 3.5 km reflect  surface
%  o I flow Iron, surface heating.  Surface wind velocities for upj slope
flows are significant since tracer was measurable (0639 MDT) even at the
r ad junction9 between sites 1 and  2.   Upslope flows al so transports   the
tracer near to site 15;  concentration is indicated by the dashed  line.
     Concentrations along Duck Creek also reflect plume splitting j^o major
and secondary plumes.  The major plume passed across the creek near 9 km
while the secondary plume was near 3 km.  The secondary plume followed a
ma or tributary Into Duck Creek.  In contrast, the major plume passed over
ridges along Duck Creek.  The major plume was uniformly mixed across Duck
Creek Valley at the confluence of Duck and Yellow Creeks.

                               Results, July  21

     Meteorological data are shown as a function of  time in  Figure 6a for
the second morning tracer experiment.  A temperature Aversion dissipated  at
about 0645 MDT.   Important times are tracer release  and the  tracer gas-
  am  in   eHods.  Tracer release was from 0300  to  1217 MDT   There  were   en
tracer sampling periods:  seven gas-bag sampling periods from 0500 to 051b,
0700 to  OTIS]  0745 to  0800, 0830 to 0845, 0937 to 0952, 1100 to  1115; and
                                       11-17

-------
                         360


                         340


                         320


                         300


                         280


                         260


                         240


                         220


                         210


                         140


                         120
- — — 60m	
    10m
      TRACER
      , RELEASE
10m
    flu
                            34567
                                            9  10  11 12
                                      TIME. MOT
         .FIGURE 6a.  Wind Direction, Speed and Temperature Gradients
                     on July 21, 1981
 1200  to 1215 MDT and three syringe sampling time periods from 0500 to 0647,
 0548  to 0748, and 0800 to 0938 MDT.

      Tracer-plume flow trends were similar to those during the first morning
 experiment.  Concentration data reflect little tracer-plume spread during
 drainage flows before sunrise, plume spread and dilution after sunrise  and
 subsequent plume lifting.  A coherent plume during 0500 to 0515 MDT is
 ononnCted ^ n Fi9ure 6b by 1 ar§6 centerline concentrations, 1800 and
 20800 pptv, along sampling sites 22 and 20, respectively.  Since tracer
 concentration was an order of magnitude greater at downwind site 20,  the
 tracer-plume axis passed either to the side or over site 20.

 10  oS SCe]aol° is Su99ested to explain relative concentrations at sites 19,
 18, 20, and 22,  concentrations of 130, 210, 20800,  and 1800 pptv,  respec-
 tively in Figure 6b.  In the region of sites 19 and 18, there are  two
 significant drainage plumes with concentrations reflecting partial mixing of
      tw° ^n^396 Plumes.   The smallest of the four 1 istid concentrations!
       -TJ3?,PptV-at,site 19»  reflects little mixing of the tracer drainage
       J   9 J6 mT drainage plume'  The main drainage plume  is directed 9
to f   H   \     re^10ns  near site 19 b* the deePest local  gully,  a dis-
tance for developing drainage  flow depths which is  almost twice as long as
the distance from the tracer release site.   Drainage flow from the tracer
                                      11-18

-------
                                     w^-'^ym-i-:-
                                     !;:--•  ^t''?i
                                     H. '- /,• V^^TTF^V^ - - • -
	 v^>^- ••.,,'  • .1^;PROPOSED. 7. — '- -v
f-^' / '.- "   ! jfftZ SURfACE SPBNT , . • '    V
.'   / /     .,/*   RSTORT SHALE '.  ,.1800 •'
   y•':  -x:  •   li 7 r'r/^^••••-'•  -,«.«
     FIGURE 6b.  Airborne Tracer Concentrations from 0500 to 0515 MDT,
                Collected at  the Radio-Control led Sampling Sites
                During the Morning of July 21, 1981


release site may override the  main drainage flow until the two plume mix
near site 18, a concentration  of 210 pptv.  The main plume is deflected
northward near site 22, a concentration of 1800 pptv, by drainage flow  from
the tracer release site.   The  "undiluted" tracer-plume axis passes near
site 20, a concentration  of 20800 pptv.

     There is subsequent  supporting evidence for this two-plume scenario
from the succeeding gas-bag sampling period between 0700 and 0715 MDT (an
hour after sunrise at 0605 MDT); see Figure 6c.  The tracer plume axis  was
deflected and lifted by interaction with a main drainage flow of less rela-
tive strength than earlier because of surface heating.  All concentrations
were significantly reduced. Nevertheless, concentrations at downwind site  20
remained on an order of magnitude greater than at site 22; i.e., 2070 pptv
compared to 235 pptv.  Tracer  plume spread to the north included sites  10,
9, and  17.  The conclusions are that the "undiluted" tracer plume axis  was
near site 20 and plume interactions resulted in significant tracer
concentrations at sites 10, 9, and 17.

     Concentrations continued  to reflect increased surface heating effects
with increasing time.  As shown in Figure 6d for 0745 to 0800 MDT and in Fig-
ure 6e  for 0830 to 0845 MDT, plume spreading and interaction continued.
Differential plume heating effects became important. The "major" drainage
flow became relatively less significant because of valley side-wall
heating.  In contrast, the "secondary" drainage plume persisted from the
tracer  release site.  This persistence directed tracer plume flow into  the
gully between sites 17 and 18.  After drainage flows dissipated, plume
                                     11-19

-------
                             300'   "*2070
                     V. PROPOSED.  .570
                  -'* SURFACE • SPENT .  * '
                                                    . HAOIO-CONTROLLED B*O   -"'=- S
                                                     SAMPUNG SITES      t l"'
                                                                   J? • j >
                                                    CONCENTRATIONS SHOWN
                                                    HEX! TO EACH SITE ARE
                                                F^jOWvmi^t
FIGURE  6c.   Airborne Tracer Concentrations  from 0700 to 0715 MDT,
             Collected at  the Radio-Control!ed Sampling Sites
             During the Morning of  July 21,  1981
                                        /•V.'"''  V        HB>1«.J ^ •fV»'
                                       -/Y ' ' !.' ,'   V« Urn (1 mll.| GRID SPACING    >9
                                       t-  .  f  S   120 ft CONTOUR SPACINGI       ^
FIGURE  6d.  Airborne Tracer Concentrations from  0745 to  0800 MDT,
             Collected  at  the Radio-Control led Sampling Sites
             During the  Morning of  July 21, 1981
                                      11-20

-------
                   • LITTLE DUCK CREEK
                        .'•.PROPOSED  .80
                      •*" SURFACE SPENT , *260 "  '
                        RETORT SHALE '.  «7
                                 . '   MO
           ".-,-fpK": pS;
         >-""'•. KV 'm
                  k^::sM*Mi
           . .. „  W.^...Sp^,Jffi\,^Jg£-
                 _,^y. - |) ^/^flrOwy -
     FIGURE 6e.   Airborne  Tracer Concentrations from 0830 to 0845 MDT,
                 Collected at the Radio-Control!ed Sampling Sites
                 During  the Morning of July 21, 1981
lifting occurred.   Plume  lifting  is shown by low concentrations in Figure 6f
at times from 0937  to  0952 MDT.


     Gas-bag samplers  were subsequently collected in zone 1; see zone loca-
tion in Figure 2 to further  investigate pollutant transport during upslope
flows.   As shown in Figure 6b, downslope winds ceased at about 1000 MDT at
the tracer release  site,  with subsequent variable wind directions.  As a
result  of this variation, upslope surface winds transported the tracer west-
ward towards sampling  zone 1.  Concentrations for two time periods are shown
from 1100 to 1115 MDT  in  Figure 6g and from 1200 to 1215 MDT in Figure 6h.
Concentrations above 1 pptv were not detected in zone 1.

     Syringe samples were collected to investigate in detail tracer plume
segmentation,  effects  of  tracer plume flow through side tributaries along
Yellow  Creek,  and cross-valley tracer plume flows into Duck Creek.  The
syringe-sample concentration data in Figures 6i through 6k show the
following:

  •  a  time-history of pollutant concentration changes in gulches
     before,  during and after sunrise

  •  further identification of major drainage transport flow paths at
     distances greater than those investigated with gas-bag sampling
     sites

  •  cross-valley concentration profiles along Duck and Yellow Creeks
                                    11-21

-------
                                  •vi .1  x\ ''••"   ; ^;C }:^v*:<»   •
                                 o^^^iS£Si*§ v^ •

                                 ^Tl^J^^{T;:;
                                  0:lV';^Li(:W: -:
FIGURE 6f.   Airborne Tracer Concentrations from 0937  to  0952  MDT,
            Collected at the Radio-Controlled Sampling Sites
            During the Morning of July 21, 1981
             1  
-------
                                        'W1--  VI '"•*  •" My- 'i'S#? 'v<>
                                       •~^'^jD^EEij^^j^fr;-1^& V ^;:-
      - ->5*- •  . ,.•'  •  yfepnoposeD:  .
         '  "      Z SUKFACE-SMNT . .»    "
               .     RETOIIT  SHALE -
        -.-'!
  •  - ,•>• '/r-   ;r-
"*      --'    -   --•'  •-
                                      f'jf. '(  {.'    1.6 km (1 MILE) GRID SPAC
                                      //. I" ''1     (20 ft CONTOUR SPACING)

                                       y"'I. ;.Jr'..'<:..'  i  S

                                      •*"•  ' \'  ' -.  i  •
            SF« SAMPLING

              RADIO-CONTROLLED BAG
              SAMPLING SITES

            CONCENTRATIONS SHOWN
            NEXT TO EACH SITE ARE
            pptv

                    '...ft .!»._»£
FIGURE 6h.   Airborne Trace Concentrations from  1200 to  1215 MDT,
              Collected at  the Radio-Control!ed Sampling  Sites
              During  the Morning of  July 21,  1981
FIGURE  6i.   Airborne SFe Concentrations Along Roads for  Sampling
              Times from 0500  to 0647  MDT on  July  21, 1981
                                      11-23

-------
                           O——O ALONG MAIN ROADS

                           O—O ALONG CftOSS-V ALLEY T« AILS
FIGURE 6j.  Airborne SF§ Concentrations  Along  Roads from Sampling
            Times from 0548 to 0748 MDT  on  July 21, 1981
FIGURE  6k.   Airborne SF5  Concentrations Along Roads  for  Sampling
             Times  from 0800 to 0938 MDT on July 21,  1981
                                  11-24

-------
  •  pollutant flow from side tributaries leading  into Yellow Creek and

  •  pollutant mixing and channeling in valleys at valley confluences, and
     downwind of valley confluences.

     Concentrations determined from syringe samples confirm that tracer-plume
spread was limited during drainage flows before sunrise.  As shown in Fig-
ure 6i near site 18, the maximum syringe-sample concentration is comparable
to the maximum 15-min gas-bag concentration of 20800 pptv shown in Figure 6b.
The main-plume width was narrow, and concentrations decreased rapidly as a
function of cross-wind distance.  Relative concentration decrease rates fur-
ther support the scenario concept:  a major drainage flow interacting with a
smaller drainage flow containing tracer;  Concentrations decreased over two
orders of magnitude in a distance of about 0.1 km  at a projected road dis-
tance of 2.7 km.  Hence, the "edge" of the main drainage flow is near 2.7 km.
In contrast, the gradual concentration decrease from 3 to 6 km reflects the
diverted prinicipal plume flow from the tracer release site.

     Tracer plume flow segmented into multiple tracer plume flows as plumes
followed along terrain depressions toward Duck and Yellow Creeks.  Terrain
effects were investigated by collecting some samples either near side-
tributaries along main roads or along cross-valley traverses.  Cross-valley
concentrations were investigated by using side trails from the main roads
along Duck and Yellow Creeks.

     A more detailed description of main road and  side-trail locations is
needed to emphasize concentration measurement locations, more detailed than
can be read from the figures.  Main roads were usually located along either
side of each valley.  In Duck Creek Valley, the main road was along the val-
ley's northern.edge.  Side trails extended approximately half-way across the
valley towards the tracer release site and terminated at the creek.  Beyond
the confluence of Duck and Yellow Creeks, the main road continued along the
northern edge of the valley to 7 km along Yellow Creek.  Beyond 7 km, the
main road crossed to the southern side of the valley.  Several syringe sam-
ples were usually collected along this valley crossing.  For the remainder
of Yellow Creek Valley used in these experiments, the main road was located
along the southern edge of Yellow Creek Valley between 0 and 2.4 km and along
the northern edge between 2.4 and 7 km.  At 7.5 km, a side trail extended
completely across the valley.

     Concentrations along Duck Creek shown in Figure 6i reflect tracer plume
impingement on a valley side wall at 2.5 km; a major tracer plume with a cen-
ter! ine near 5.5 km, and a secondary plume near 8.8 km.  Impingement was
along the northern side of a major gulch leading in to Duck Creek, near the
juncture with Duck Creek.   Impingement occurred prior to the cross-valley
traverse:   concentrations are indicated by the dashed line.  At the traverse,
the plume axis was near the center of this major gulch.  The traverse was
near the confluence with Duck Creek Valley.  Beyond the confluence, the
tracer plume impacted drainage flow from upper Duck Creek.  Consequently, the
tracer plume flowed along the southern edge of Duck Creek.  At 4 km, concen-
trations along the center of Duck Creek were an order of magnitude greater
than they were along the northern edge of Duck Creek.  Nevertheless, plume
mixing continued down valley.  However, mixing was incomplete since concen-
trations were always greater near the center of Duck Creek.  Also, these
                                    11-25

-------
greater center!ine concentrations probably reflect contributions from  seg-
mented tracer plumes flowing over the southern edge of Duck Creek.

     The tracer plume segmented in the region between Duck and Yellow  Creeks.
Segmentation is confirmed by concentrations along Yellow Creek.  Along Yellow
Creek, the maximum concentration was at 4.5 km, a location adjacent to a
side tributary which directed a tracer plume segment into Yellow Creek.

     Beyond the confluence of Duck and Yellow Creeks, the tracer plume flowed
along the northern valley edge.  At 7.5 km along Yellow Creek, concentrations
were nearly an order of magnitude less than along the southern edge.   It is
noted for subsequent referral in Figure 6j that although sunrise was 0605 MDT
at the tracer release site, this southern edge was still shaded at 0618 MDT.
Concentrations were nearly uniform across the valley further down creek,
about 10.2 km along Duck Creek.

     Tracer plume segmentation and interaction with ambient air plumes are
further confirmed with concentrations shown in Figure 6j.  First discussed
are concentrations near site 18.  Between about 2 and 2.5 km, concentrations
cycled about an order of magnitude between succeeding sampling locations.
This cycling may have occurred because samples were collected near the undu-
lating interface between the tracer plume and the major drainage plume from
near site 2, and the interface width was greater than reflected in Figure 6i.
Although plume-interfacing characteristics were changing, a portion of the
tracer plume continued to flow as a secondary plume near 1.2 km.

     The main concentration trends (cross valley concentration differences
and segmented tracer plume flows along side tributaries) continued along
Duck and Yellow Creeks.  However,  significant cross-valley differences also
occurred along Yellow Creek.  At 2.5 km, concentration near the northern
edge was about four times greater than near the southern edge of the valley.
At 7.5 km, where the entire valley was not being heated by the sun, there
was no longer a cross-valley concentration difference.

     Concentrations in Figure 6k are more influenced by surface heating
effects since samples were collected two or three hours after sunrise.  Sur-
face heating directed a segmented tracer-plume flow between 1 to 3 km along
Ridge Road.  Near site 18, the tracer plume was also directed between 3 km
(the road along site 18) and 6 km along Ridge Road.  This last plume direc-
tion confirms in more detail the main tracer-plume direction concluded from
gas-bag concentrations shown in Figure 6e for times between 0830 and
0845 MDT.
                                 CONCLUSIONS


     Fugitive emissions transport and dispersion were simulated in tracer
experiments during conditions of relatively low wind speeds; i.e., wind
speeds less than the threshold wind speed causing soil suspension.  Although
spent^shale will probably suspend only for greater wind speeds, fugitive
emission rates will  be nearly independent of wind speed if suspension were
mainly caused by mechanical  stresses.  For instance, constant vehicular
traffic on spent shale can produce  a relatively constant fugitive emission
                                     11-26

-------
flux which is independent of wind speed.  If this proposed emission scenario
occurs, the maximum downwind airborne concentrations will similarly occur
during low wind speed conditions.  Airborne concentrations will remain rela-
tively large since airborne pollutant concentrations will not be diluted by
high wind speeds,  u, i.e., downwind airborne concentration decreases are
proportional to 1/u.

     A multiplicity of plume flows and plume interactions were identified
for fugitive emissions transport from the proposed spent shale disposal
site.  The tracer plume segmented.  The greatest concentrations occurred
during drainage flows.  Plume interactions were between traced plumes and
ambient air plumes.  Since multiple plume flows and plume interactions were
identified, model  predictions of downwind transport of fugitive emission
plumes on a local  scale must include predictions of plume segmentation and
interaction.

     Predictions on a local scale are essential if maximum "fence-line" con-
centrations are considered.  Two parameters in model predictions are fugitive
emission rates, Q, and airborne concentrations, x.  The only controllable
parameters in these models are pollutant release heights, release times, and
release rate, Q.  Predictions of X/Q ratios both at the fence-line and
downwind are used by industry and in regulations.  Although further analysis
is planned for evaluations of predictive models, the x/Q ratios can
nevertheless be estimated from the reported tracer concentrations; x with
units of pptv, and-the constant tracer release rate, Q = 26 kg/hr.


                                  REFERENCES
Clements, W. E., S. Goff, J. A. Archuleta and S. Barr.  1981.  Experimental
Design and Data of the August 1980 Corral Gulch Nocturnal Wind Experiment,
Piceance Basin, Northwestern Colorado.LA-8895, Los Alamos National Labora-
tory, Los Alamos, New Mexico.

Sehmel, G. A.  1981.  "A Dual-Tracer Experiment to Investigate Pollutant
Transport, Dispersion, and Particle Dry Deposition at the Rio Blanco Oil
Shale Site in Colorado."  PNL-SA-9327.  In Proceedings of the Second
Conference on Mountain Meteorology, November 10-31, 1981, Steamboat Springs,
Colorado.Published by the American Meteorology Society, Beacon Hill,
Massachusetts,  pp. 137-146,

Whiteman, D. C., N. S. Laulainen, G. A. Sehmel and J. M. Thorp.  1981.  Mete-
orological Data Report for Field Studies Conducted in August 1980 in the Oil
Shale Region of Northwestern Colorado.PNL-3734, Pacific Northwest Labora-
tory, Richland, Washington.
                                       11-27

-------
        Laboratory Testing To Improve Rail Car Sealant Spray and Loading
                Techniques for the Abatement of Fugitive Coal Dust

            Colin J. Williams  and William F.  Waechter, MHTR Ltd. *

                               ABSTRACT

The technique of spraying coal rail cars with latex or similar sealants to con-
trol fugitive emissions has been used for several years.   There are serious
deficiencies in the systems used to load coal  cars and also in the application
techniques of the sealants.  This paper reviews these problems and gives prac-
tical advice,  resulting from laboratory tests  and field visits  to four coal load-
ing facilities,  on how to improve the loading and spraying methods  currently
used.

A series of laboratory tests has been conducted which subjected over 100  samp-
les of coal to simulated train  journeys.  These tests, unlike  previous tests on
sealants,  simulated the important parameters of the motion induced vibration
of the coal surface  and also wind effects on the erosion of the sealed surface.
Measurements of the actual vibration levels in  a full scale coal rail car were
used to design  the vibration table and MHTR's boundary layer wind tunnel was
used for the wind erosion tests.
The following variables affecting sealant performance were investigated:
sloped surfaces; centre humps; settling during  transit; concentration of sealant
solution; alternative sealants; various application techniques; effect of rainfall;
preferential spraying and ambient air temperatures of 15°C and 30°C.
(This is an abstract of a presentation for which a paper is not available.)
(*) Current address:  Rowan,  Williams, Davies, and Irwin, Ltd., 650 Wood-
   lawn Road, West, Guelph, Ontario, Canada NIK IB 8.
                                    12-1

-------
         Studies of Nontraditional Fugitive Particulate Control Techniques

                     Brock M.  Nicholson,  EPA/OAQPS-RTP
                 Maxine Borcherding, City of Portland (Oregon)
                        Gary Ekhardt, State of Minnesota
                         Ray Mohr. State of Colorado

                                 ABSTRACT

Demonstration studies have been conducted with funding assistance from EPA in
three U. S. cities to determine the effects of controlling nontraditional sources of
fugitive particulate on ambient air quality.  These studies involve application of
control measures which have not been traditionally required as part of the State
Implementation Plans (SIPs) for attaining ambient air standards. It is now gener-
ally recognized that many urban areas of the country will be unable to attain the
ambient standards solely with application of traditional or stack  emissions con-
trol.  Because of the magnitide of  this fugitive  problem and the difficulty in defi-
ning its character,  EPA has allowed states to  perform demonstration studies of
nontraditional fugitive control techniques as a  logical first step in a control stra-
tegy for attainment of ambient standards.
The study design,  execution,  and  results to date of these studies are described.
Attention is given to the technical and institutional aspects that were  followed in
an effort to ensure  results more reliable than  previously obtained in  such  studies.
The SIP development process, as envisioned by EPA in allowing studies, has not
resulted in the initiation of a  large number of studies. Therefore, it is hoped
that the knowledge gained by a few major studies, on the least understood of the
nontraditional control measures,  will be applicable in many areas throughout the
country.
In Denver, the study involves modified winter  snow and ice  control  practices,
such as sand cleanup and use of salt as an alternative. In Portland,  vacuum  sweep-
ing of heavily soiled paved road surfaces in industrial and commercial areas is
be.ing studied. In Minneapolis, the study focuses on the control of unpaved road-
ways  at construction sites and  the cleanup of mud and dirt carried onto  paved
roads around construction sites.

(This is an abstract of verbal  presentations for which papers are not available.)
                                  13-1

-------
  This paper has been reviewed in accordance with the U.S. Environmental
  Protection Agency's peer and administrative review policies and approved for
  presentation and publication.
  A  New Charged  Fog Generator
  for  Inhalable Particle Control  *
                        by
                   C. V. Mathai
          Arizona Public Service Company
           P. O.  Box 53999, Station 5680
          Phoenix, Arizona 85072-3999
(*) Although this is not the actual presentation made at the May 1982
    meeting, it closely resembles that presentation, which was en-
    titled, "Evaluation of the Efficiency of a Charged Fog Generator
    in Controlling Inhalable Particles at a Steel Plant, " by C. V.
    Mathai and Bradley M.  Muller (then of AeroVironment,  Inc.,
    Pasadena, CA 91107) and William B.  Kuykendal (then of  EPA/
    IERL-RTP,  but now of EPA/OAQPS/AQMD). The work was
    supported by EPA contract 68-02-3145.
                        14-1

-------
  A review of the literature shows that
control efficiency of inhalable particles
using water droplets can be improved
significantly if the droplets are electrically
charged. A spinning cup fog thrower
was developed initially to generate
electrically charged water droplets. The
poor performance of this device in wind
tunnel tests was attributed to the short
lifetime of the fine droplets generated.
the ineffective  ionizer  ring method of
charging the droplets, and the nonuni-
form charge  distribution  observed
along different regions of the fog spray
pattern.
  A new charged fog generator  (CFG)
was then  developed  by modifying  a
commercial rotary atomizer. In this
device, the droplets  generated  are
contact-charged to provide a high
charge-to-mass ratio of 1.2 fiC/g. The
droplets have a number concentration
median diameter of about 100 fjm and a
mass median  diameter of  about 200
/an. The water flow rate is variable (4-
70 l/h), and the fog spray pattern can
be easily changed from long and narrow
to broad and short, with a typical spray
coverage of 16-24 m3. The device uses
about  1  kW power (110 VAC) and is
portable.
  The CFG (at a bentonite ore loading
operation) was extensively field-tested
to determine  the dependence  of its
inhalable particle control efficiency
(PCE)  on various instrument settings
and field conditions. These tests show
that the overall mean PCE is 78% higher
than the corresponding value for
uncharged fog. Individual PCEs as high
as 88% were achieved. The lifetime of
the droplets seems to be the dominant
factor determining the PCE; and PCE
values were higher for higher applied
voltages and higher water flow rates.
The data suggest that,  under optimum
instrument settings.  PCE of water
droplets could be doubled by charging
the droplets.

Introduction
  Although the  total  particle  mass
loading of anthropogenic aerosols is only
about 10% of that from  natural sources,
their effects are significant and largely
detrimental. Recent reports indicate
that inhalable  particles (15 i/m  and
smaller in aerodynamic diameter), in
general, and fine particles (2-3 //m and
smaller), in particular, may be a human
health hazard, and degrade atmospheric
visibility. Devices such as electrostatic
precipitators and wet and dry scrubbers
have performed exceedingly well  in
controlling pollutants from conventional
industrial stack sources. However, effec-
tive and economically feasible methods of
controlling inhalable particles from non-
stack sources in open areas are lacking.
Recognizing this situation, the U.S.  EPA
has been encouraging the development
of new methods, such as charged fog
technology, to control fugitive emissions.
  Removing fine particles from  a gas
stream in an open area is  difficult
because of the particles' low mobility and
unfavorable inertia! properties, as well as
uncontrollable external factors (mete-
orological  parameters). The  most  com-
monly  used dust control (ordinary water
sprays) in mining  and other material
handling areas is only 30%-40% efficient
in controlling  inhalable particles.  Only
during the last few years has electrostatics
been used to augment particle collection
efficiency of water droplets. Numerous
studies have shown that most industrial
pollutants and naturally occurring  dust
particles acquire electric charges as they
are dispersed  into the air. Studies  have
also shown that the polarity  and magni-
tude of the charges on these particles
depend on their size and origin (coal, soil,
mineral, etc.). Therefore, the particle
collection efficiency  (PCE) of water
droplets can be significantly enhanced
via electrostatic forces of attraction if the
droplets are charged  to  the  opposite
polarity.
   Commercially available charged fog
devices for fugitive emission control have
certain disadvantages: the need for high
pressure air or water to properly atomize
the water; the possibility of spray nozzle
clogging if the water supply contains high
concentrations  of 'dissolved salts and
suspended solids; and poor charge-to-
mass  ratio of the droplets to provide a
high degree of fine particle control. Under
the sponsorship of  the U.S.  EPA, Aero-
Vironment,  Inc.,  has  developed  a  new
charged fog  generator  (CFG)  which
overcomes these difficulties. This research
and development project consisted of two
phases: Phase I involved examining the
theoretical  aspects  of  charged fog
technology, developing a bench-scale
prototype instrument and  evaluating its
inhalable  PCE in a  controlled laboratory
experiment, and an economic analysis of
the feasibility of the technique to control
fugitive emissions; and Phase II involved
construction  and field-testing of a
prototype to evaluate its PCE. This report
summarizes both phases.
                                      14-2

-------
Theoretical Background
  When an aerosol particle approaches a
water droplet with a relative velocity, it
may collide  directly with the  droplet
(impaction),  barely  touch the  droplet
(interception), or miss the droplet entirely.
The collection of an aerosol particle by a
charged droplet is the result of a number
of simultaneous mechanisms of interac-
tion between them; e.g., inertial impaction,
direct interception,  Brownian diffusion,
and electrostatic, diffusiophoretic, and
thermophoretic forces. The relative effect
of the mechanisms of interaction between
the droplet and the particle depends on
the size of the particle. For large particles
(aerodynamic diameter greater than 2-3
      the dominant mechanisms  of
                   10°
 particle collection by droplets are impac-
 tion  and  interception.  For particles
 smaller than 0.1 fjm, Brownian diffusion
 becomes  very important.  For particles
 between these two ranges, electrostatic
 forces are the dominant interaction
 mechanisms.
   The PCE of uncharged  water sprays
 (where inertial  impaction is  the major
 collection mechanism) is given by,
          1-exp  -.- -2k-.!=.•
                   2  QG  D
(D
 where QL and QG are the volumetric flow
 rates of the water (liquid) and air (gas)
 components of the dust  cloud, respec-
 tively; L is a characteristic length for the
 total capture  process; 0 is  the  mean
 droplet  diameter; and rj is  the  single
 droplet collection efficiency. Figure 1 is a
 plot of calculated E versus particle radius
 for a droplet radius, R, of 106 fjm and RH
 of 75% (curve  C). This curve  has  a
 minimum (whose magnitude depends on
 the droplet size) for particles with radius,
 r, near 1 fjm. This minimum is caused by
 the ineffectiveness of inertial and diffusive
 interaction mechanisms for particles in
 that size range.  However, if the droplet
 and the particle are oppositely charged,
 the minimum in E is eliminated (Curves A
 and B) as electrostatic forces become the
 dominant mechanism of particle collection
 in this  size  range.  This effect  is the
 fundamental principle on which charged
 fog technology is based. When  the
 droplets and the particles are charged, rj
 is given by,

     77 = - 4CQC Qp/24 77* c0r R2 pU0, (2)
                  10'
                  Iff
                     0.01
              0.1       1.0
           Particle Radius, tun
                                                10.0
               Figure 1.
           Calculated single droplet collec-
           tion efficiency in air of 10aCand
           900 mb as a function of panicle
           radius for 106fjm radius droplet
           at 75% relative humidity for
           droplet and particle charges of
           (A) ±20 esu cm'2. (8> ±2 esu
           cm'*, and (Cl zero charge.
 where C is the Cunningham slip correction
 factor; Qe and Qp are the charges on the
 droplet and particle, respectively; E0isthe
 dielectric constant; p is the viscosity; and
 U0 is the free-stream velocity. Thus, for a
 given  particle size distribution, the
 electrostatic  forces  are  proportional  to
 the  magnitude  of the  charges and
 inversely proportional to the droplet's size
 and  its free-stream velocity.
  Although  Figure  1  shows  that the
 addition of electric charges on particles
 and  droplets  eliminates the  minimum in
 PCE (near particle radius, r  = 1 um) and
 yields values of E which are 5 to 10 times
 higher in certain  size ranges  than for
 uncharged sprays, the overall PCE of an
 operating  system may not be that high.
 Increases in PCEs of about 15% for 1 -//m
 particles to over 45% for 0.3-//m particles
 (compared to uncharged droplets) have
 been reported. Values of 50%-80%, with
 charged droplets under controlled experi-
 mental conditions,  have  also  been
 reported.
  When sprayed into the air, the charged
droplets will  evaporate unless the air is
saturated with water vapor. The droplet's
lifetime determines the effective contact
time between the droplet and  particles
and thus strongly influences the overall
                                         14-3

-------
PCE  of  the system. The  lifetime of a
water droplet depends on the temperature
and relative humidity of the medium into
which it is introduced. To obtain the best
PCE. the droplets must  be  small enough
to provide both an adequate spray rate per
volume of gas treated and  sufficient
contact time, yet large enough so as not to
evaporate too quickly.
  The  maximum  charge  a droplet  can
carry before it disintegrates is reached
when the outward pressure produced by
the electric field at the surface of the drop
is equal to the inward pressure produced
by the  surface  tension.  This limiting
charge is called the Rayleigh limit, given
by,

         0,.y = 87r [CoffR3]1'2        O)
where Qp.y is the limiting  charge on the
droplet (coulombs); c0 is the permittivity of
the  medium  in  which the  droplet is
 located; a is the surface tension  of the
 liquid;  and R is the droplet  radius, in
 micrometers.
Charged Fog Generator
 Development
  Water droplets may be generated by a
spray nozzle or rotating cup. Droplets are
 generally charged by electrostatic induc-
 tion, ionized field, or contact charging.
 Commercially available charged fog
 devices use  water spray nozzles  and
 induction charging.
   In this study, a prototype unit, called the
 spinning  cup fog thrower (SCFT),  was
 initially developed in association with the
 University of Arizona in Tucson. The SCFT
 consisted of a rotary atomizer,  an ionizer,
 and a  vane-axial blower. The  rotary
 atomizer consists of a small hollow-shaft
 motor and a spinning cup. Water from a
 low-pressure source is introduced into
 the hollow shaft  and  flows toward the
 other end where the spinning  cup is
 mounted. Entering the rear of the cup, the
 water stream strikes a rotating "spider"
 which  deflects the water  to the sides. A
 sheet of water then flows toward the lip of
 the cup  where droplets  are  formed by
 centrifugal force and the air stream's
 striking the thin water layer. The droplets
 were then charged by a stream of positive
  ions produced by the  ionizer  containing
  numerous small  discharge  needles. It
  was expected that the ions produced in
  the region of the ionizer ring would follow
  the airflow from the vane-axial fan, mix
  with the droplets, and charge them. The
  charged droplets were then deflected and
  projected forward by  a  stream of air
  supplied by the vane-axial blower. These
  droplets had  a median diameter of about
  20 fjm and a charge-to-mass ratio of 1 x
    "
   To evaluate the PCE of the SCFT, tests
 were conducted in  a wind tunnel at the
 University of Arizona. These tests showed
 only about 50% inhalable PCE. This poor
 performance was attributed to the very
 short lifetime of the fine droplets gene-
 rated, the ionizer ring method of droplet
 charging (which proved ineffective), and
 the  nonuniform  charge distribution
 observed along different regions of -the
 fog  pattern. The SCFT was  therefore
 abandoned,  and a new  charged fog
 generator  (CFG)  was  developed by
 modifying a commercial rotary atomizer.
   Figure 2 schematically represents the
 new CFG. Water is introduced through
 the water tube into the 3600-rpm rotating
 cup, whose inside is  fabricated to  a
 gradual smooth taper. A small deflecting
 baffle is attached to the open end of the
 water tube  so that the water will be
 deflected 90° and strike the rear end of the
 rotating cup. Because of the centrifugal
 forces, the water becomes a thin film and
 moves  forward into a high velocity
 airstream from the vane-axial fan. The
 impact of the high velocity air on the thin
water film breaks the water film into fine
water droplets.


   The water tube, air cone, and rotating
 cup are made of nonconductive materials.
 The  water tube is attached to the rotating
 cup, thus rotating with the cup. The other
 end  of the water tube is attached to the
 water supply through a rotating seal. The
 water for  atomization  is  stored in  an
 electrically isolated reservoir (120-1
 capacity), and a small pump is used to
 pump it to the inlet of the rotating seal.
 The  water flow rate can be varied from
 about 4 to 70 l/h. The airstream from the
 fan,  controlled by an air butterfly setting,
 projects the fog forward. By adjusting this
 setting,  the airflow speed can be controlled,
 thereby controlling  the shape  of the fog
 spray pattern to  better conform to the
 shape and size of the source of dust upon
 which the charged fog  is to be applied.
 The  spray pattern covers a volume of 16-
 24m3.
   To achieve a high charge-to-mass ratio
 for  the droplets,  contact charging by
 directly connecting  a high voltage source
 to the inflowing water was found to be the
 most efficient method. However, this
 method requires that the entire water
 supply  (reservoir) and associated tubing
 be electrically isolated so that there is no
 current leakage.
   The size distribution of the droplets was
 determined using  a cloud optical array
 probe and a  precipitation optical array
 probe.  These measurements gave  a
                                       14-4

-------
        number concentration median diameter
        of about 100 fjm and a mass  median
        droplet diameter of about  200 fjm.
        Collecting the droplets on greased glass
        slides  and observing   them under a
        microscope yielded values consistent
        with the above results.
          The charge-to-mass ratio of the droplets
        was determined using a special sample
        train  (developed by AeroVironment),
        consisting of an insulated stainless steel
        probe tip mounted on a standard glass
        midget impinger. The probe was connect-
        ed  electrically to copper wool,  packing
        inside  the midget impinger, which was
        also connected by a shielded cable to an
        electrometer. The impinger was immersed
        in a Dewar flask containing dry ice.  An
        isokinetic sample of droplets  was then
        extracted  from  the fog  spray.  As the
        droplets moved through the impinger
        they were condensed, transferring their
        charge to the copper wool packing. The
        charge  transferred  to the copper wool
        was then measured by the electrometer.
        Knowing the  mass of droplets collected
        and the current produced by  them, the
        charge-to-mass ratio was calculated.
        This method gave a typical value of 1.2 x
        10~* C/g with an applied voltage of 15 kV,
        about 25% of the maximum allowed value
        (Rayleigh limit)  for 200-//m droplets.
           The charge-to-mass ratio was  also
        estimated using another experimental
        setup, in which droplets were allowed'to
          transfer their charges to a 125 um mesh
          size standard sieve placed in the path of
          the fog spray and the current generated in
          the mesh was  measured. Knowing the
          size of the droplets and the number of
          droplets generated by the CFG  per
          second, the  charge-to-mass  ratio was
          estimated. This  method  gave values
          consistent  with those  reported for the
          gravimetric method.
            In summary,  the CFG is fairly small,
          portable, and mounted  on  a moveable
          platform. The total power requirement is
          about 1 kW (110  V AC power supply). It
          has no nozzle-clogging problem and does
          not require compressed air. Using a small
          inverter, the unit  can be operated by an
          automobile battery. This feature may be
          important because the system has high
          potential application in locations  where
          commercial electric power is not available.
          The physical characteristics of the
          charged droplets from the CFG are such
          that high innalable PCE is expected.


          Evaluation of the Inhalable
          PCE of the CFG
            The CFG was field tested at the Kaycee
          Bentonite  Corporation's bentonite  pro-
          cessing plant in Worland, WY, during the
          first half of 1981. Bentonite  is a water-
          absorbing material used mainly to seal
          water leaks in oil  wells.  Bentonite ore is
          unloaded from front-end loaders onto the
                                     Air Fan
                                                                                          DC Power Supply
             Nonconductive Air Cone
 Water-Deflecting
     Baffle
    Nonconductive
    Spinning Cup
Nonconductive
 Water Tube  \
                                                                          DC Water Pump
Figure 2.    Schematic of the charged fog generator.
                                                  14-5

-------
grill of a hopper which is attached to the
west wall of the plant building. From the
bottom of the hopper (approximately 4 m
below the hopper grill level,  inside the
plant), the ore is carried by conveyor belts
to processing areas. The hopper  is
completely  enclosed except on the side
through  which  the front-end  loaders
unload. The hopper is 6.7 m wide, about 2
m deep and, from the grill level, about 3 m
high. The grill is inclined about 30° to the
horizontal.
  Bentonite ore from two large piles,
approximately 100  m to the  northwest
and southwest of the hopper,  are carried
to the hopper by front-end loaders and
dropped  on the grill.  The bucket of the
front-end loader is about 2.4 m wide and
successive loads are unloaded uniformly
over the 6.7 m wide hopper. It takes about
10 front-end  loader  dumps  to  fill the
hopper to the grill level. Those 10 dumps
are accomplished in about 25 minutes.
The bucket is removed from inside the
hopperjirea  in 20-25 seconds. One full
 hopper of ore will be carried away by the
 conveyor belt in about an hour.
   The land around the plant is fairly flat. A
 railroad  track is east of the plant and a
 paved road (very little traffic) beyond that.
 Bottom-dump trucks filled with bentonite
 ore arrive near the storage piles from the
 south and,  therefore, the dust in  the
 hopper area due to transport of ore to the
 plant area is negligible compared to the
 dust generated in the hopper  itself. In
 other words, the hopper can be considered
 as a fairly isolated source.
    The CFG and  the particle  sampling
 instruments were mounted on  the
 outside of the south wall of  the hopper,
 about 4 m above ground level. A platform
 was built to mount  these instruments.
  Ideally,  the particle sampler  inlet should
  have been mounted on the east (rear) wall
  of the hopper, but for practical reasons
  could  not be.  The CFG sprayed water
  droplets across the hopper above the grill.
  The total volume to be treated by the
  charged fog was about (6.7 m x 2 m x 3 m)
  40  m3, somewhat larger than  the max-
  imum coverage of the CFG fog. Unfortu-
  nately, only one prototype  CFG  was
  available for these tests. To have had a
  second unit, mounted on the north wall of
  the hopper  and operated concurrently,
  would have been  ideal.
    Particle samples were collected using a
  Sierra Model 230 CP cyclone preseparator
  followed  by a Sierra two-stage cascade
  impactor. The cyclone's air inlet protruded
  0.3  m into  the  hopper. The particle
  sampling system was operated at a flow
  rate of 0.85 mVmin (30 cfm), which was
calibrated at regular intervals during the
entire test program. At this flow rate, the
cyclone has a particle cut-point of about
7.3 //m. The  impaction plates of the
cascade  impactor were chosen so that
the upper  filter would collect  particles
larger  than 1.8 fan. The  lower backup
filter collects all particles smaller than 1.8
fjm. For this report, particles collected on
the backup filter are characterized as the
fine  particle fraction,  and those on the
upper  filter as  the coarse fraction. The
mass of coarse fraction collected was
often an order of magnitude smaller than
the fine fraction (possibly indicating a
particle bounce problem with  particles
larger  than the cut-point  of 1.8 /urn
reaching the backup filter).-Subdividing
the samples into more size ranges would
have required  either much longer sam-
pling time or much higher flow rates to
obtain  acceptable filter loadings. Neither
alternative was desirable.
 Field Test Design
  Three test scenarios were designed to
 determine the PCE of theCFG. In the first,
 no attempt was made to control the dust
 inside the hopper, except that the CFG's
 fan ran continuously (this  ensured that
 the mixing of the dust cloud was nearly
 identical among various runs in order to
 make direct comparison of tests with and
 without charged fog). In  the second,
 uncharged fog was applied on the dust
 cloud.  In the  third,  charged  fog was
 applied. Other governing  parameters
 were varied under each scenario to yield
 a statistically acceptable set of test data.
  The parameters varied were water flow
 rate, fog pattern  (short or long), applied
 high  voltage, wind conditions, and
 relative humidity. In addition, tests were
 performed with the polarity of charges on
 the droplets reversed. This test protocol
 called for 32 test runs. However, in actual
 practice, data had to be collected under
 prevailing  field conditions; hence,  96
 runs were performed.
  Particle  samples were collected  on
 preconditioned and preweighed glass
 fiber filters.  Each particle  sample was
 collected during  12-15 front-end loader
 dumps (roughly  30-40 minutes). Wind
 speed and direction and relative humidity
 were also recorded simultaneously. After
 each  sample was collected, the filters
 were  transferred to  special envelopes
 and brought to AeroVironment's labora-
 tory for analysis.
                                         14-6

-------
Data Processing
  The  mass  of the fine and coarse
fraction  and  sum of the  two were
obtained from the final and initial weights
of the filters. Using known values of the
sample time, number of dumps, and flow
rate, particle mass concentration for each
dump was calculated.
  The amount of dust generated in the
hopper fluctuated from dump to dump. To
overcome this variation,  each sample
was collected during 12-15  dumps and
the  mean  value for each  dump  was
calculated. The amount of dust generated
also varied from day to day, caused by
factors such  as changes  in ambient
conditions  and the moisture content of
the bentonite ore. The effect  of the latter
is not too significant since the ore stored
in each pile comes from the same  mine
and a pile is completely processed before
a new pile is started. To  overcome the
day-to-day dust level fluctuations, sample
concentrations were normalized for each
day with respect to the background value
(the fan-only value), and a percentage
particle collection (control) efficiency was
calculated. Percentage PCE is obtained
from:
 where C0 and C are the particle concen-
 trations corresponding to "fan only" and
 "fog test" scenarios, respectively.
   Values of E are calculated for the fine
 fraction, the coarse fraction, and the sum
 of fine and coarse. The report gives these
 percentage PCEs and the corresponding
 field condition data, water flow rate, and
 applied high voltage for the whole field
 test program.

 Comparison  Between Charged
 and Uncharged Fog
   Figure 3 shows the mean values of the
 measured percentage fine PCE and total
 PCE for  charged (striped  bars)  and
 uncharged (solid bars) fog. For this
 comparison, all  test  runs under all
 instrument settings and field conditions
 are included. The mean and standard
 deviations of the fine PCE are 48.1% and
 23.0%, respectively, for the charged fog
 and 27.8% and 25.3%, respectively, for
 the uncharged  fog. The  corresponding
 values for all the particles  (fine and
 coarse together) are 44.5%, 21.8%,
 25.0%, and  24.4%, respectively.  These
 numbers show that, even under average
 field conditions and instrument settings,
 inhalable PCE can be almost doubled by
 electrically charging the water droplets.
 However, under  optimum instrument
 settings and  favorable field  conditions,
50
Control Efficiency. %
8 S
Particle
K>
O

10
0
, • Uncharged Fog
0 Charged (+ & -) Fog
-
-



I
'/
I








I
\
'/>





-
-


  Fine Panicles Only Fine & Coarse Particles

Figure 3.   Mean inflatable PCE of all test
           runs (for various CFG settings
           and field conditions/ for charged
           and uncharged fog.


the improvement in inhalable PCE can be
expected to be  higher. Note  that the
volume  of  the dust cloud treated  was
somewhat  larger  than the maximum
coverage of the CFG.
  The size range of particles collected in
the coarse mode was fairly narrow, and the
mass collected was often an  order  of
magnitude smaller than the fine fraction.
Another problem which may have inhibit-
ed coarse particle collection is the particle
bounce  effect, by which some  of the
larger particles pass the upper impaction
plate and settle on the backup filter with
the fine particles. Consequently, most of
the ensuing discussion concentrates on
the sum of fine and coarse fractions of the
particles collected. Therefore, this partic-
ular experiment could not demonstrate
the full effect of charged fog on the coarse
particles.
  Figure 4 shows total PCE with E plotted
as a function of ambient RH for two sets of
instrument  settings for an applied high
voltage of 4 kV (positive charges) and a
water flow rate  of 60 l/h. The circles
represent a broad spray pattern,  and
squares represent a narrow spray pattern.
The method of least squares was used to
fit a  straight line to the data sets, shown
in the figure,  yielding  a correlation
coefficient of 0.93 for the broad spray and
-0.19 for the narrow  spray. The corre-
sponding slopes  are  0.99 and -0.09,
respectively. Although the wind conditions
were not identical for all data points, the
                                            14-7

-------
I as
u
s
•575
 [65
o
•45
!•
      CFG Settings
      Water flow: 60 Iph
      Voltage: 4 kV

      Spray Type
      Broad o
      Narrow  n
'Slope: 0.395
 Corr.
 Coeff.: 0.927
                      ° Slope: -0.087
                        Corr.
                        Coeff.: -0.189
    35   45   55   65    75   85   95
        Ambient Relative Humidity. %
Figure  4.    PCS of the CFG platted as a
            function of ambient relative
            humidity for broad to) and
            narrow (at spray patterns.
fine PCE increases with increases in
ambient relative  humidity for  a broad
spray, but is fairly independent of RH for a
narrow spray. The generally lower values
of E for the narrow spray can be easily
explained: a board spray covers most of
the dust cloud  in the hopper, while the
narrow spray covers a  smaller portion,
resulting in lower E values.
  The difference in  the dependence of E
on RH for the broad and narrow sprays is
also explainable. For  a narrow  spray,
most droplets occupy a volume away from
the open side of the hopper. When fog is
applied continuously, this area becomes
more humid than outside the hopper or
near the hopper opening, assuming the
wind is not too strong. Thus,  there is
minimal droplet evaporation and. con-
sequently, fairly steady PCE. However, for
a broad spray, the droplets are distributed
from the rear wall of the hopper to outside
the open side of the hopper. Thus, when
the ambient RH is  high, fewer droplets
will  evaporate,  leaving more droplets to
collect dust particles; on the other hand,
when  the ambient RH is low, more
droplets are lost due to evaporation near
the open side of the hopper and outside the
hopper.
  The effect of the longer droplet lifetime
in a higher RH atmosphere increases the
PCE. These test observations  are thus
consistent with the conclusion that the
droplets should  be small  enough to
provide high PCE, yet large enough not to
evaporate too quickly.
  Figure 5 shows PCEs for negatively and
positively charged fog for the fine fraction
separately, and  for both fine and coarse
particles combined,  with the same water


40
SS
i;
1 30
,£*
l£
1
1 20
U
a
•«• 10
0
S Charged H fog
0 Charged (+1 Fog
•
•
.


'

K
^
•^
\
\
V////////77/77777S






\


^
V
x
O
Y/////////////777/,

m


'

!
-
                                                            Fine Particles Only Fine & Coarse Particles

                                                        Figure 5.    Comparison of PCE of the CFG
                                                                    for positively  and negatively
                                                                    charged  fog.   with all  other
                                                                    parameters nearly identical.

                                                        flow rate, the same applied voltage, and
                                                        nearly identical  field conditions. The
                                                        higher value of E for negatively charged
                                                        fog suggests that inhalable bentonite
                                                        particles carry a net excess positive charge.
                                                          Figure 6 is a good example of the effect of
                                                        water flow rate in the CFG on its ability to
                                                        control particles.  The PCE in this case
                                                        decreased by 42% when the water flow
                                                        rate was reduced from  60 l/h to 30 l/h,
                                                        with all other  instrument settings and
                                                        field conditions nearly identical. This
                                                        effect is expected: increased water flow
                                                        rate  increases the number of charged
                                                        droplets available for particle  collection.
                                                        Water flow rate was  particularly signifi-
                                                        cant  in this experiment because the CFG
                                                        was being applied to control dust in a vol-
                                                        ume  larger than its maximum coverage;
                                                        therefore, any  decrease  in the water
                                                        flow rate further decreased the coverage.
                                                          Under a controlled experimental setup
                                                        one would expect the  PCE of charged
                                                        droplets  to decrease if the wind speed in
                                                        the volume being treated were increased.
                                                        The  increased  wind  would reduce the
                                                        time available for the droplet and particle
                                                        to interact. However, in this experimental
                                                        situation,  because of the location and
                                                        method of particle sampling, this effect is
                                                        not evident.


                                                       Conclusions
                                                         After  the initial  setbacks with a
                                                       prototype spinning cup fog thrower, a
                                                       new charged fog generator was devel-
                                                       oped. Water droplets are produced in this
                                         14-8

-------
as 40
1
•$
u
§30
•5
o
^ 20
1
2

-------
   industrial hygienist. Control of
   germs and possible decontamina-
   tion of an area using proper chemi-
   cal additives in the water  spray is
   also  of  interest to our defense
   forces. Preliminary investigation of
   this application  could be  coupled
   with (1), above.
(4) The potential of the CFG to control
   dust from a mobile source is worth
   examining. No alternative  method
   of control is available, other than
   ordinary water sprays,  which  are
   not efficient in controlling inhalable
   particles.
(5) Possible application of charged fog
   in  recovering airborne precious
   metals should be of great interest to
   the gold and other precious metal
   and mineral mining industries.
              14-10

-------
                ESTIMATION OF AMBIENT TSP IMPACTS OF COAL
                      STORAGE AND HANDLING FACILITIES
                              Robert C. Wells
                              Enviroplan, Inc.

                               Dennis C. Doll
                              Enviroplan, Inc.

                                John Hattrup
                        Baltimore Gas & Electric  Co.
           The work described in this paper was not funded by the U.S. Environmental
           Protection Agency. The contents do not necessarily reflect the views of the
            Agency and no official endorsement should be inferred.
                                   ABSTRACT
          This  paper is  directed to those concerned with  evaluating the  air
quality   impact  of   fugitive  particulate  emissions   from  coal  storage and
handling. Site-specific  analyses were conducted for the coal handling system
planned for Baltimore Gas  & Electric's Brandon Shores Power   Plant  in  Anne
Arundel   County,  Maryland.   Fugitive particulate emissions were considered
significant due to the presence of a designated nonattainment area  for  TSP
near  the  plant.  The analyses employed dispersion modeling with a modified
version of the  Industrial  Source Complex (ISC) Dispersion Model.   Emission
factors   and  control efficiencies  were  taken  from  available  published
information.  Available  emission  factors  were  critically  evaluated  and
specific  factors were required to be consistent with available measurements
and site-specific  conditions.    Significant  problems  developed  with  the
meteorological  data  base  using the standard U.S. EPA preprocessor.  Since
these problems  produce  unrealistic  prediction  under  some  meteorological
conditions,  approximations  were  employed  to  eliminate   the unreasonable
predictions  from  the  analysis.   Site-specific  analyses   have  aided  in
defining  the   appropriate  dust  control  equipment  for the coal handling
system.
                                     15-1

-------
     1.0  INTRODUCTION

          National  Ambient  Air  Quality  Standards   (NAAQS)   have   been
established   for   Total   Suspended   Particulates   (TSP)   which  define
concentrations of suspended particulates associated with health problems  and
with secondary impact such as property damage.   TSF  measurements  indicate
that  many areas of the United States do not attain NAAQS for TSF.  Further,
emissions from "traditional" sources of pollutants (e.g. smoke stacks) often
do not account for the observed TSF  concentrations.   Fugitive  particulate
emissions, including both natural dust and anthropogenic sources, contribute
to measured concentrations.  As such, quantification and control of fugitive
dust is an important part of achieving and maintaining the NAAQS for TSP.

          This  paper describes a case study of the application of available
estimates  of  uncontrolled  fugitive  emissions,  control  efficiency,   and
atmospheric  dispersion  for  environmental  analysis  of a coal-fired power
plant.  Regulatory criteria are  stated  in  terms  of  maximum  permissible
ambient   concentrations.   Thus,  dispersion  modeling  based  on  emission
estimates is the only suitable technique to adapt available  information  on
fugitive emissions to the specific situation of a proposed facility.

          A  wide  range  of  control  measures  is available to suppress or
contain fugitive emissions resulting from  material  storage  and  handling.
Costs associated with different control measures for fugitive emissions vary
greatly.  In  addition,  some  control  measures  can  affect efficiency  and
reliability  of  operations.   Therefore,  one  cannot  assume  that  it  is
automatically  desirable to control fugitive emissions to the maximum degree
possible.  Ambient impacts of fugitive emissions  under  various  levels  of
control  are  important  in determining cost-effective measures of attaining
ambient air quality goals as defined by the NAAQS.

          Additionally,  the  analyses  described  here  for  evaluation  of
fugitive emissions of a new facility are also useful in identifying emission
reduction   credits   (ERC)   for  use  in  emissions  trades.   Preliminary
investigations based on dispersion modeling are useful in defining the  most
cost-effective  controls associated with emission trade-offs, as well as  the
spatial effects  of  trades.«involving  both  fugitive  and  stack  sources.
Current  U.S.  EPA  policy     requires  that proposed emission trades which
control fugitive emissions rather than stacks demonstrate the equivalence of
emission reductions using dispersion modeling.  These trades will  typically
require  post-approval  monitoring  to  confirm  modeling  results; however,
pre-approval analyses are an  important  first  step  in  demonstrating   the
appropriateness of ERC.

          The  facility  evaluated in this study is the Brandon Shores Power
Plant being constructed in Anne  Arundel  County  by  the  Baltimore  Gas   &
Electric  Company.   Figure  1  details  the  location  of the plant and  air
quality in the area.  The plant was originally designed to burn  either   oil
or  coal.   Plans  initially  called  for  operation  with oil and the plant
received environmental permits on that basis.  During construction, however,
it was determined that it would be economically advantageous to  burn  coal.
                                  15-2

-------
Therefore, the plant has been treated for regulatory purposes as an existing
facility undergoing coal conversion  even  though  the  plant   is  currently
under construction.

          The Brandon Shores Power Plant consists of tvo units  each with 620
MW  generating  capacity.   Units  I and II are projected to come on-line  in
1984 and 1988, respectively.  The lay-out of the  pover  plant  and  on-site
coal handling system is detailed in Figure 2.
     2.0  STUDY DESIGN CONSIDERATIONS

     2.1  REGULATORY APPROACH

          There are tvo basic approaches to  regulating  fugitive  emissions
associated  with coal handling and storage.  First, specific controls can be
required as a condition of the appropriate permit.  With some  restrictions,
this  approach  can apply to federal Prevention of Significant Deterioration
(PSD) permits as veil as to construction and operating permits issued  by  a
state  government.  This is common practice in some states.  Controls can be
either specified by the regulator or proposed by the operator  and  approved
by the regulatory agency.

          Second,  a  facility  may  be  limited to specific maximum ambient
impacts.  These impacts are defined either as maximum  total  concentrations
including  background (e.g., NAAQS) or as maximum incremental concentrations
due to the facility (e.g., PSD increments).

          The Brandon Shores Pover Plant is subject to the  latter  form  of
regulation.  While  a  minimum  level  of  control  is  not  dictated by the
regulatory agency, the facility must account for ambient concentrations from
fugitive  emissions.   This  approach  potentially  provides  for  the  most
cost-effective  controls,  in  that  the  specific  control required vill be
directly related to the ambient impact of a new facility and to the existing
air quality.  In practice, the analysis is  complicated  by  limitations  of
available emission factors and dispersion models.

          Three ambient criteria potentially constrain fugitive emissions at
Brandon  Shores.  First, the NAAQS apply in any area of public access.  This
is generally taken to apply at any  point  beyond  the  facility's  property
line.

          Second,  in areas that are nov attaining the NAAQS (vhich includes
the Brandon Shores Plant site), incremental increases in TSF  concentrations
due  to   new  sources are limited.  As with the NAAQS, these PSD increments
apply in all areas of public access.  They include impacts from all  sources
constructed  after  a  specific  date.   In  this case, no other sources are
projected to consume PSD increments of TSP in the vicinity  of  the  Brandon
Shores  facility; therefore, PSD increments apply to concentrations from the
Brandon Shores facility alone.
                                    15-3

-------
          The third case,- involving areas that do not attain  the   NAAQS,   is
more  complicated.   The  Brandon  Shores  Plant  site   is   located   near  a
(secondary) nonattainment area for TSP.  If the facility does not  contribute
significantly to ambient concentrations in the nonattainment  area, then  the
facility  is  permissible.  If the facility does contribute  significantly to
the nonattainment area, it is then subject to two types of control.    First,
the facility must employ extensive controls to achieve the lowest  achievable
emission  rate  (LAER)  for particulates.  Second, the facility must  provide
for emission reductions in the nonattainment area to  offset   the  increased
emissions  at  the  facility.   While  it  is  possible to satisfy these  two
requirements in some cases, they represent a severe limitation on  the  siting
of new sources.   It is often preferable to voluntarily  control  sources   to
the point where they exert no significant impact in a nonattainment area.

          Criteria for ambient concentrations and  incremental  impacts   are
based   on  both  long-term  (annual)  and  short-term  (24-hour)  averages.
Specific concentrations applicable here are presented in Table 1.   Criteria
for  short-term  NAAQS  and  PSD compliance limit the largest  second-highest
concentration predicted based on approved  dispersion  modeling  procedures.
The   significance  criteria  for  nonattainment  area  impact  (de  minimis
increments) are subject to  interpretation  depending  upon   the  regulatory
context and the discretion of the regulatory agency.

          The  increments  were  proposed  by  U.S.   EPA  to   relate   to  the
incremental contribution to a violation of the  NAAQS.   Hence  they   should
apply  only  at the time and place of an actual violation of  the NAAQS.   For
long-term  averages  this  includes  the  entire   nonattainment   area    by
definition.   For  short-term averages, prediction impacts from the proposed
facility may or may.not coincide with short-term concentrations in excess  of
the NAAQS.

          In this instance, the State of Maryland is free  to  set  its   own
criteria  for  significant  impact  in  a nonattainment area.  The State  has
chosen to employ the numerical criteria suggested in federal  regulations   as
applicable  to the overall highest concentration predicted from the proposed
facility in the designated nonattainment area.  This  determination  ignores
the   coincidence   of   the   facility's   impacts    with  high  background
concentrations.

          Projection of TSP concentrations in the Baltimore   area  developed
by  the  State  of  Maryland  for the years 1982 and 1986 indicate that this
boundary of the nonattainment area is projected to change drastically  in  the
1980's (Figure 2).     The location of this boundary is an important part of
the  definition of significant impact.  Concentrations at  the  boundary   of
the  designated  nonattainment  area  ultimately  prove  to   be the limiting
concentrations,  which determine the level of  control  required.   As   such,
projecting  air  quality  to  the time of plant start-up would significantly
alter the conclusions of this analysis.
                                    15-4

-------
     3.0  METHODOLOGY

          Under the regulatory  criteria  outlined  in  Section  2,  ambient
impacts  from  the  operation  of coal storage and handling equipment at the
Brandon Shores Power Plant and resulting total ambient  concentrations  must
be  determined.   Projection  of  ambient  impacts  for  proposed facilities
requires the independent estimate of emission rates from  the  facility  and
dispersion   from   the   source.    In   addition,   projections  of  total
concentrations resulting from proposed facilities require some consideration
of other sources of a  pollutant.   These  three  issues  are  addressed  in
separate subsections below.

     3.1  EMISSION RATE ESTIMATES

          Available procedures for estimating emission rates from sources of
fugitive  particulate  emissions require (1) selection of an emission factor
which describes uncontrolled emissions from the source,  (2)  identification
of  a  control  efficiency  associated  with the specified controls, and (3)
specification of source extent (area or operating rate).

                                       Factors
          Published emission factors quantify emissions from  all  potential
sources  of fugitive particulates which are of concern here.  The quality of
these emission factors  varies  greatly,  however.   Quantitative  estimates
available  for some sources are merely assumptions.  Other sources have been
subjected to empirical  field  studies  and  emission  factors  exist  which
account  for  a wide range of potential conditions. In some cases, different
field studies employing  different  methodologies  have  been  conducted  to
quantify  similar  sources.   It  is  disturbing  to note that field studies
conducted with reasonable methodologies arrive at different results.

          Appropriate emission factors for a  particular  facility  must  be
selected from the wide range of available estimates.  These emission factors
should  be  based  on  two  criteria.  First, the emission factors should be
determined from a specific field  study  or  estimation  procedure  of  high
quality.   Second,  the  field  study  or  estimation procedure should be as
closely related as possible to the specific sources  of  concern.   A  third
criterion   often  raised  in  a regulatory context is conservatism within a
range of uncertainty.'

          An overall quality index has been suggested in past work conducted
by U.S. EPA contractors.     The proposed rating scale is presented in
Table 2.  Essentially, the quality index  asserts  that  (1)  any  data  are
better  than  no  data; (2) representative data are better than extrapolated
data; (3) accurate data are better  than  inaccurate  data;  and  (4)  large
quantities  of  accurate  data  are better than small quantities of accurate
data.  The qualitative scale is intuitively reasonable, and will  provide  a
useful  measure  of  the relative quality of different emission factors from
different sources.  Unfortunately, selection of alternate  emission  factors
is  often a choice between extrapolation of high quality data from different
                                    15-5

-------
processes  or  the  use  of  lower  quality  data  for  more  representative
processes.

          In practice, the authors know of no emission factors for   fugitive
particulate  emissions which deserve an A rating.  Some emission factors  are
available which would be rated B.  Most available emission factors would  be
rated  C  or D, and a few potentially significant sources must be quantified
using E-rated emission factors (assumptions).  Generally, D-rated or E-rated
data should be disregarded if better information  is  available.   B  and C
ratings  are  actually  a  continuum.   Use  of any emission factor  requires
"extrapolation  from  similar   processes."    The   relative   degrees   of
extrapolation  and  inherent  data accuracy must be balanced in selecting an
emission factor.

          The regulatory concern for conservatism  is  often  not  practical
with  fugitive  emissions estimates.  The range of emission factors  for some
sources can span more than three orders of  magnitude.   It  seems   entirely
unreasonable  to rely on a questionable assumption simply because it is more
conservative than the results are a measurement program.

          Three  sources  of  published  information   potentially   provide
emission  factors  useful  in this study.  First, work has been conducted by
Midwest Research Institute (MRI) under U.S. EPA contract.  Their work  began
in  1924-x   with  additional  useful  information  published  in  1978   and
1979.  '    Emission  factors  potentially  applicable  to  coal  have  been
developed  based  on  limited  sampling  of  fugitive  emissions  from  coal
handling, and  based  on  extrapolation  from  more  extensive  sampling  of
different materials.

          Second,  FEDCo  Environmental Inc. conducted a limited field study
of fugitive particulate emissions from coal mining  in  the  western  United
States.      While  coal  mines are not identical to power plants, the study
represents a moderately large data base specific to coal.  An  extrapolation
of  these  emission  factors  may  be  preferable to use of emission factors
developed for other materials and for a specific process.

          Climatic conditions must also be accounted for in  using   emission
factors  developed for coal mines in the arid west for operations at a power
plant in the eastern United States.  The effect of  climatic  conditions  on
fugitive   particulate  emissions  are  poorly  understood.   However,  both
empirical studies and theoretical considerations  '   suggest  that  climate
is important.

          Third,  a wide range of emission factors from a variety of sources
was summarized in 1977 under U.S. EPA contract.      This  summary   included
earlier  MRI  work,  as well as numerous limited field studies, and  emission
factors which had been  used  in  regulatory  determinations  in  the  past.
Generally, emission factors cited in this document which are not included in
work  discussed  above  would  be  rated D or E.  The compilation provides a
useful summary  of  emission  factors  which  may  be  considered  based  on
conservatism  of regulatory precedent.  However, field studies subsequent to
                                   15-6

-------
its publication  generally  have  found  lover  emission  rates  than  those
previously employed in regulatory determinations.

          The uncertainty noted in estimating fugitive particulate  emission
factors,  combined  with  regulatory concerns of precedent and conservatism,
precluded the selection of a  single  "correct"  set  of  emission  factors.
First, a set of emission factors was formulated based on adherence to recent
regulatory  precedent.   It was intended that these emission factors provide
an upper bound on ambient impacts associated with coal handling equipment at
the Brandon Shores Power Plant.  A determination that  the  plant  would  be
environmentally  acceptable  based  on  these emission factors would provide
strong support for issuance of permits.

          A second set of emission factors  was  formulated  based  on  what
seemed  to  be the most representative factors available.  An analysis based
on these emission factors provides the best assessment  of  ambient  impacts
from fugitive particulate emissions.  The substantial margin of conservatism
typical  of  the  former  set  of  factors  is  not  the case here, however.
Emission factors judged to be best are frequently at  the  low  end  of  the
range of factors quoted.

          At  the  Brandon  Shores  Power  Plant  there  are three important
processes with the potential  to  produce  fugitive  particulate  emissions.
First,  continuous  transfer  equipment  removes coal from barges, transfers
coal between conveyors, places coal in and  out  of  storage,  and  supplies
bunkers.   Second,  wind  erosion  from  storage  piles on-site is likely to
produce emissions.  Third, coal must be crushed  prior  to  combustion.   In
addition, some emissions would be expected from conveyors.

          Conservative  emission  factors  for continuous handling equipment
were based on a  number  of  emission  factors  that  had  been  used  in  a
regulatory  context.   The emission factors for barge unloading and transfer
houses are based  on  commonly  used  factors  with  unclear  origins.  '
Emission  factors  for load-in and loadr-out of storage piles were taken from
MRI's earlier work with stone quarries.     The  available  emission  factor
judged  the  best  to quantify emissions from continuous handling operations
was developed for fugitive emissions from the iron  and  steel  industry.
The emission factor is based on similar activities with different materials.
Site-specific  correction  factors are suggested to minimize the uncertainty
associated with extrapolation.

          A conservative estimate of wind erosion from  coal  storage  piles
was  taken from emission factors developed for stone quarries.     While the
factor is not particularly applicable to coal, it has frequently  been  used
in  a  regulatory  context.   The wind erosion emission factor developed for
coal mines    was judged to be more appropriate  for  an  accurate  emission
estimate.  This factor is based on data directly relevant to coal.  The data
were  collected  in  an arid region and corrections for climatic differences
and storage pile activity were assumed based on past work.     The  emission
factor  is  stated  in  terms  of a functional relationship with wind speed,
which is.intuitively reasonable though different from theoretical functional
forms.   '     This functional relationship  is  particularly  important  in
                                     15-7

-------
avoiding  severe  overpredictions  of  ambient impacts from coal storage and
handling under low wind speed conditions.

          Conservative emission factors for.coal crushing were based on U.S.
EPA emission factors for stone crushing.      Alternative  emission^xfactors
have  been  proposed  based  on  limited  data  specific to coal.      While
neither  factor  could  be  considered  ideal,  the  choice  of   the   more
source-specific  factor  seems  preferable  to  the assumed applicability of
stone crushing emission factors based on regulatory precedent.

          An available estimate for emissions from conveyors  was  based  on
the  assumption  that  251 of the emissions from "transfer and conyeving" or
332 of the emissions from transfers would be  due  to  conveyors.       This
estimate  is  based  on  visual  observations  and is employed here both for
conservative and best estimates.

          Emission factors employed in this study are presented in Table  3.
Comparison of emission factors is difficult because of different assumptions
for  the  two sets of factors.  The conservative emission factors often lump
sources together (e.g., all  transfers  and  conveyors),  whereas  the  best
emission  factors  are  specific  to  single  sources (e.g., single transfer
points).  Conservative  wind  erosion  estimates  are  based   on   material
throughout; best estimates are based on surface area of storage piles and on
the  wind  speed.   Table  4  summarizes the alternative emission factors in
terms of annual mass emissions.  This summary  illustrates  the  assumptions
inherent in selection of one factor over another.

                             Control Estimates

          The  uncertainties  noted  above  for  estimates  of  uncontrolled
fugitive particulate emission rates are  compounded  in  estimating  control
efficiencies.   With  the exception of baghouses, where control efficiencies
can be readily  demonstrated,  available  control  techniques  for  fugitive
particulate  emissions  are  not  quantified at all.  Available.estimates of
control efficiency are based entirely on subjective estimates.     While use
of these subjective estimates is not desirable, there is little choice.  The
added confusion associated with developing  alternative  "conservative"  and
"accurate"  estimates  of  control  efficiencies  is not warranted.  Neither
conservatism nor accuracy can  be  demonstrated,  and  extensive  regulatory
precedent does not exist with respect to assumed control efficiencies.

          It   is   important   to   specify  control  which  will  mitigate
unacceptable environmental impacts without excessive  costs  or  operational
limitation  on the facility.  An initial set of controls was suggested based
on engineering  judgement.   Selective  controls  were  added  as  potential
problems  were  identified  in  dispersion  modeling.   Table  5  summarizes
controls considered for  the  Brandon  Shores  Power  Plant,  together  with
estimates of control efficiencies employed in dispersion modeling.
                                   15-8

-------
     3.2  DISPERSION MODELING

          In  order  to  accurately  represent  dispersion  of   particulate
emissions from coal handling, a dispersion model should be able to represent
(1)  isolated  area  sources, (2) local turbulence associated with buildings
and obstructions, (3) dispersed emissions  associated  with  enclosures  and
conveyors,  and  (4)  settling  and  deposition  associated  with  the large
particles typically emitted as  fugitive  dust.   Regulatory  considerations
dictate  that  the  dispersion model also correspond to established D.S. EPA
guidelines on dispersion modeling for environmental permitting.

          Based on these criteria, the U.S. EPA's ISC Model appears to be  a
clear  choice.   The  model  is  flexible enough to account for a variety of
source geometries (points, lines, volumes, areas) and specifically  accounts
for  settling  and  deposition  associated with particulates.  The ISC Model
includes features to account for dispersion due to the mechanical turbulence
(downwash) in the wakes of buildings  and  similar  large  structures.   The
downwash  algorithm  has been extended here to include dispersion associated
with area sources which are, in themselves, obstructions to air flow.

          The flexibility in specifying source geometry available in ISC was
fully utilized in this application.  Storage piles were represented as  area
sources.   Sources controlled by baghouses were represented as point sources
(without buoyant  plume  rise)  which  are  subject  to  building  downwash.
Sources  with  controls which do not duct emissions to a specific point were
modeled as volume sources, where emissions are assumed to  be  emitted  over
the effective cross section of the source.

          The particle data needed to determine settling and deposition were
based  OB,sphysical diameters determined from microscopic analysis of hi-vol
filters.     This corresponds to the implicit definition of TSP as particles
subject to  hi-vol  capture.   The  continuous  particle  size  distribution
observed  was  divided  into  three  categories:  (1) aerosols, (2) slightly
settling  particles,  and  (3)  rapidly  settling  particles  based  on  the
assumption of stokes settling.  Table 6 details the particle size ranges and
particle characteristics employed to represent each group.

                             Receptor Selection

          Dispersion  modeling*  requires  identification  of locations where
concentrations are expected to be maximized.  Ground level sources will show
maximum ground level concentrations near the source.  Elevated releases  may
show  maximum  ground  level  concentrations  further downwind.  Preliminary
modeling of the significant sources of fugitive particulates at the  Brandon
Shores  Power  Plant  yielded  the  intuitively  reasonable  result that the
release heights of fugitive emissions at Brandon Shores were low  enough  so
that  maximum ambient concentrations would be expected at the property line.
This corresponds to the nearest source-receptor  distance  where  the  NAAQS
would  apply.   Thirty-four  receptors  were  therefore  located  along  the
property line.  Eight additional  receptors  were  located  near  the  barge
unloader  at  the  minimum source-receptor distance recommended for accurate
                                    15-9

-------
prediction modeling with ISC (100 meters).      A total of 42 receptors were
located to quantify overall maximum impacts.

          Maximum ambient impacts in the designated  nonattaimnent  area   to
the  vest  and  north of the plant are anticipated at the nonattainment area
boundary. Thirteen  receptors  were  thus  located  on  the  boundary.   Six
receptors  were  also  located  at nearby monitoring locations identified  in
Figure 1.  A total of 63 receptors were employed in this analysis.

                            Meteorological Data

          Dispersion modeling employed five  years  of  meteorological  data
(1964-1968).     Surface    observations   from   the   Baltimore-Washington
International Airport were combined with upper air data from Dulles  Airport
using the standard U.S. EPA meteorological preprocessor.

          Two  important  anomalies  were  noted which required corrections.
First, extremely low interpolated mixing depths were eliminated because they
are not reasonable and because  these  unreasonable  mixing  heights  affect
predictions  for  low-level  releases.  Mixing  depths  from  the  U.S.  EPA
meteorological preprocessor were uniformly truncated  to  a  minimum  of   75
meters.  Second,  periods  of  calm winds recorded in airport data presented
both  a  theoretical  problem  and  a  practical  problem.   In  theory,   a
steady-state  Gaussian  Dispersion Model does not apply where an established
mean flow (wind) does not exist.  This problem is typically  dealt  with   by
assuming a minimum speed of 1.0 meter per second.

          The  practical problem is created by the fact that typical airport
meteorological instruments cannot measure wind directions accurately at  low
wind  speeds associated with calm conditions.  Wind directions are typically
assumed to persist over the period of  observed  calms.   This  creates  the
unrealistic  situation  of  multiple  hours  of  persistent  wind  direction
accompanied by low wind speeds.

          In any case where prediction of  maximum  concentrations  resulted
from  this  unrealistic  representation of calm wind conditions, predictions
were reevaluated without the periods of calm winds.  The  deletion  of  calm
winds  provides  an. important perspective on the limitations of the current
state of the art of dispersion modeling.

     3.3  BACKGROUND CONCENTRATIONS

          Estimating total ambient concentrations associated  with  a  given
source  requires  the  inclusion of background concentrations from known and
unknown sources which also  emit  the  same  pollutants.   In  the  case   of
particulates,  it  is  impossible  to include all other potential sources  in
dispersion  modeling.    Ambient  concentrations  may  be  dominated  by   an
industrial facility, a highway, a farm, or a wide range of sources.  In this
situation  it  is  often  expedient to make the conservative assumption that
high measured background concentrations— which do not include the  proposed
facility —added to high predicted concentrations from the proposed facility
represent likely maximum total impacts.
                                   15-10

-------
           Background  concentrations  representative   of   the   Brandon  Shores
Plant   site were  taken  from  the  nearest  continuous monitoring station,  which
is  located in  the attainment area  identified  in Figure  1.  Measurement   data
from monitoring stations  located in  nonattainment areas would obviously show
violations of the   NAAQS   which  are   assumed not to  ezist  based on the
designated attainment status of  the  Brandon  Shores   Plant  site.    In   this
study,  monitoring data  for the calendar  year  1979 was employed.

           Differences  in background  particulate   concentrations  should be
expected  as a  function  of meteorology.   In the   case of  a   pollutant   with
continuous monitoring  data  available  (e.g.,  SO.), background concentrations
can be  related to specific meteorological  conditions.  In  the case   of   TSP,
which is  measured as  a  daily 24-hour average, this relationship  is  difficult
to  define.

           The   most    obvious   relationship   expected  is high ambient
concentration  associated  with a  limited  range of wind  directions.   Days  of
high  measured concentrations   were  examined  for persistent meteorological
conditions which  identify background concentrations  as  a  function   of   wind
direction.

           The  highest measured concentration  (114 ug/m  ) was  associated with
persistent winds from  the south,  therefore,  this concentration  was taken as
representative background for receptors  to the  north of the   Brandon Shores
Plant site.  The  second-highest  measured concentration  (106  ug/m )  could not
be  characterized with  respect  to wind direction.   The   second-highest
measured  concentration  was thus  used to  represent background   concentrations
for all   other   receptors,   as  none could be specifically excluded based on
meteorological considerations.
     4.0  RESULTS

          Results of dispersion modeling are evaluated in  terms  of  three
regulatory   criteria  identified  above.   Prediction  modeling  generally
proceeded from conservative assumptions,  which  would  have  unambiguously
supported  permitting, to better assumptions, which indicate best estimates
of projected impacts (but with less conservatism).

          Conservative  emission  factors  combined  with  initial  control
specifications   indicated  the  potential  for  exceedence  of  all  three
regulatory  criteria  noted  above.   While  annual  average  impacts, were
generally  acceptable,  appropriate  24-hour  impacts  were  230 ug/m , 124
ug/m , and 46 ug/m   for  the  NAAQS,  PSD  increments,  and  nonattainment
impact, respectively.  Examination of source contributions to the predicted
concentrations  indicated  that two sources (inactive storage pile and main
yard  conveyor)  were  primarily  responsible  for   the   high   predicted
concentrations.   At  this point, it was elected to identify more effective
controls for these sources based on prediction results.

          Additional controls (Phase Two in Table 5) were  applied.   These
controls suggested compliance with the secondary NAAQS; however, applicable
PSD  increments  at  the  plant  boundary and de minimis thresholds for the
                                   15-11

-------
nonattainment area were both  exceeded.   Relevant  maximum  concentrations
over  five  years  were  predicted  to  be  43  ug/m   and 16 ug/m   for  PSD
increments and nonattaimnent impact, respectively.  It was noted  that  calm
wind  conditions  contributed  significantly to the predicted exceedence of
the PSD increment.  The predicted PSD increment  consumption  without  calm
winds  was  40  ug/m .   Analysis  of  source  contributions  to  predicted
concentrations indicated that the active storage pile area  and  main  yard
belt were primarily responsible for predicted high concentrations.

          Additional  controls  were  specified  for the main yard conveyor
(Phase  Three  of  Table- 5).   Predicted  concentrations  with   additional
controls  were  40  ug/m   and  13  ug/m  for PSD and nonattainment  impact,
respectively.

          At this point, it was clear  that  the  conservative  assumptions
implicit  in  emission  factors  employed  thus  far  would not support  the
permitting  of  a  coal  handling  facility  with   conventional   fugitive
particulate  controls.  The identical analysis was then conducted using  the
best estimates of fugitive emission rates identified above.

          The control efficiencies for some of the sources were revised  for
this analysis.  These revisions were necessary because, in a few  cases,  the
emission  factor  itself  implicitly  accounts  for  some  control.   Also,
additional  information  regarding  the controls considered for the  Brandon
Shores Plant became available  which  necessitated  some  revision.   These
revisions are identified in Phase Four of Table 5.

           The    predicted    largest   second-highest   24-hour    average
concentrations at the property line were now 2 ug/m .   The  maximum first
highest  24-hour  average  concentration  predicted  at the boundary of  the
                                 2
nonattainment  area  was  2  ug/m .    Both  of  these  concentrations   are
significantly below  the  appropriate  regulatory  criteria.    Addition  of
appropriate  background  concentrations  suggests  that ambient air quality
standards will also be maintained.

          Prediction modeling based  on conservative emission factors  would
have  strongly  supported  environmental  permitting.    Prediction modeling
based on the best estimates does  not present as convincing a  demonstration
of the acceptability of environmental impacts.   The analysis  does, however,
suggest  that requirements of extremely stringent controls (e.g. ,  silos for
active storage)  are not consistent  with  the  best  available  information
concerning fugitive dust emissions.
                                    15-12

-------
     5.0  DISCUSSION

     5.1  RESOLUTION OF PERMITTING FOR THE BRANDON SHORES PLANT

          The analysis based on best  emission  estimates  described  above
demonstrates  that  the  environmental   impact  from  fugitive  particulate
emissions from coal handling at the Brandon Shores Plant is  likely  to  be
veil  below  established  regulatory  criteria  for ambient impacts and  low
enough to maintain the NAAQS.   Analysis  based  on  conservative  emission
estimates,  which  are  more consistent with regulatory precedent than with
scientific evidence, suggests the possibility  of  significant  short  term
impacts  under  some  conditions.  Unfortunately, the conditions associated
with maximum impacts are also conditions not treated  well  by  either   the
conservative emission factors or Gaussian dispersion modeling.

          While  the  possibility  of  ezceedence  of applicable regulatory
criteria exists, it would  seem  unreasonable  to  require  extensive  (and
expensive)  controls  on  the  basis  of  this  analysis.  The more prudent
approach would be to issue environmental permits based on best estimates of
environmental impacts.   A  monitoring  program  could  be  established  to
confirm  that  environmental  impacts are, indeed, acceptable.  If problems
were identified due to fugitive emissions from  coal  handling,  mitigative
measures  (e.g.  modification  of  management  practices,  retrofitting  of
control equipment, obtaining emission offsets) could be implemented.  While
there is a risk that more expensive controls would be  ultimately  required
if  problems  were  identified,  this  risk is more than compensated by  the
likelihood that environmental impacts will be acceptable.

          This course of action was, in fact, pursued at the Brandon Shores
Power  Plant.   The  facility  received  environmental  permits  with    the
stipulation  that  TSP  monitors  be  sited  to  identify potential maximum
impacts from the coal handling facilities.
     5.2  LIMITATIONS OF THE METHODOLOGY EMPLOYED

          The methodology discussed above did not conclusively predict  the
environmental  impact  of  fugitive  emissions  from  coal  handling.   All
predictive analyses will be similarly limited by the quality  of  available
emission data and dispersion models.

          The  requirement  of  specific  controls is a viable alternative.
The problem with this approach is that the possibility  of  minimal  impact
with  little  control  is  not  explored  and  the  cost  effectiveness  of
alternative controls for specific problems is not considered.

          The regulation of coal handling based on ambient criteria for TSP
is difficult.  Analytical tools for  the  analysis  must  be  improved  for
reasonable and timely permitting based on predicted impacts.  Several areas
can   be   readily   identified   where  available  analytical  tools  need
improvement.  These can be generally classified as problems  of  estimating
emission factors, of control effectiveness, and of dispersion modeling.
                                    15-13

-------
                             Emission Factors

          Emission  factors specifically  for coal handling  and  storage need
 to be  improved.  Demonstrated measurement techniques  are  probably   adequate
 but  the  amount of data is not.  Material handling emission factors  should
 be based on the full range of  typical  operating  parameters  (e.g.,  drop
 height).   Parameters  related to surface moisture and  silt content need to
 be extensively measured in  varying  climates  and/or  related   to readily
 available  parameters.   Some  potentially significant  sources  (e.g., train
 unloading, crushing) need much more study.

          Emission factors for vind erosion need to be  related  to   climate,
 pile    configurations   and  cumulative   erosion.   Furthermore,   currently
 available emission factors, which are linearly dependent  on wind speed,   do
 not agree with either theoretical relationships or empirical data  from wind
 tunnel testing.  At a minimum, more extensive data to quantify  the range of
 variation expected are needed.

                           Control Effectiveness

          Quantification  of  control  effectiveness  is extremely  important
 and urgently needed.  With the possible exception of  baghouses,  no controls
 are quantified.  Even baghouses may not be accurately  characterized   by  a
 simple percentage  reduction.  Control effectiveness needs to  be  specified
 over the range of options available for   generic  types   of  control.   For
 example, spray systems should be distinguished on the basis of  droplet size
 (fog vs. spray) and wetting agents.

                            Dispersion Modeling

          Dispersion  models  employed  in  regulation  development have  not
 been extensively validated for impacts from low-level,  non-buoyant releases
 of  particulates  near  flow  obstructions.   The  area   source    algorithm
 (critical to this analysis) has not been validated at all.

          Several  problems  with  conventional  dispersion modeling   are
 magnified  by  the  situation  here.   Plume  dimensions  based on 3-minute
 average  data  produce  extremes  (high  and  low)  of  impact   at    short
 source-receptor distances when assumed representative of  one-hour  averages.
 Near field dispersion from mechanical turbulence near ground  level  (typical
 of  industrial facilities) is completely neglected except in  the context  of
 building downwash.   Treatment of extremely low  wind  speed   conditions   is
 neither physically reasonable nor empirically accurate.   These problems  are
 of   lesser   importance   for   stack-type  emissions  than  for   fugitive
 particulates.   It is  not surprising that  new  applications   of  dispersion
models lead to new problems in their application.

          In  summary,   the  current  state  of  the  art   of fugitive dust
 analysis has provided an  adequate  quantitative  description  of   fugitive
 emmisions  and  resulting  ambient  impacts  from  a proposed coal handling
 tacility.   An informed  regultory decision which  accounted  for  inevitable
uncertainties   resulted.    Much productive work remains,  however, which  can
narrow the  range  of uncertainty and expedite the consideration of  fugitive
dust in environmental permitting.                                      s^^vc

                                    15-14

-------
CJ1
I—1
01
               Designated TSP Nonattainment Area
               Projected 1982 Nonattainment Area
      ———— Projected 1986 Nonattainment Area
            A TSP Monitor Site
            }|C Brandon Shores Plant
    Figure 1: Location of the Brandon Shores Plant and TSP Nonattainment Area

-------
01
 i
               ^

             PROPERTY LINE
                     DUST SILO.
           •
           •
CRUSHER
BUILDING
                         INACTIVE
                         STOCKPILE
                                          COAL BUNKER UNIT +2

                                          COAL BUNKER UNIT + 1
CONVEYOR
  INACTIVE STOCKPILE
         ACTIVE STOCKPILE
          STACKER/RECLAIMER
          TRANSFER HOUSE +2
                                                          TRANSFER
                                                          HOUSE +1
                                                               """«««„
    Figure 2:  Coal Handling System at the Brandon Shores Plant

-------
               TABLE 1:    REGULATORY CRITERIA FOR EVALUATING
                                AMBIENT IKS ACT
          Criterion                Critical  Concentration (ug/m )
                                   24-Hour Basis       Annual  Basis

          Clean Areas
            NAAQS (Secondary)            150*                60
            PSD Increment                37*                19

          Nonattainment Areas
          Impact Threshold                5**                1
            Second-Highest
**  Maximum First-Highest
        TABLE 2;  CRITERIA FOR FUGITIVE PARTICULATE EMISSION FACTORS


Quality Index          Criteria	
      A                  Based on a statistically representative number of
                         accurate field measurements of a  specific  process
                         vhich span expected ranges of parameters

      B                  Based   on  a  limited  number  of  accurate  field
                         measurements of a specific process

      C                  Based on a limited number of field measurements  of
                         a specific process with undetermined accuracy — or
                         on  — extrapolation of B-rated data from a similar
                         process

      D                  Estimate based on professional judgment

      E                  Assumption
                                    15-17

-------
          TABLE 3;  SUMMARY OF EMISSION FACTORS FOR COAL HANKLING
Source
AND ASH HANDLING EMISSIONS
Conservative
Factor
Ub/ton)
iader 0.4
0.15(2)
0.05(2)
0.2
0.15
.rage .028(4)
Estimated
Quality
Index

D
D
E
E
E
D
Best
Factor
Ub/ton;
2.56 x 10-3(1)
1.31. x 10-3(1)
4.32 x 10"4(3)
0.06
3.60 x 10-3(1)
2.18 x 10"4(1)
Estimated
Qual ity
Index

C
B
E
D
B
B
Transfers

Conveyors

Crusher

Bunkers

Active Sti
  Pile Load-In

Active Storage         .035
  Pile Reclaim

Active Storage         .320
  Pile Wind
  Erosion

Inactive Storage       .320
  Pile Wind
  Erosion

Fly Ash Unloading      0.30

Fly Ash Storage        0.30
(5)
(6)
(6)
(8)

(8)
D

D
         2.18 x 10~4(1)
             .091u(7)
            (Ib/acre-hr)
             .091u(7)
            (Ib/acre-hr)
5.08 x 10

    .288(1)
                  -4(2)
                                        B
,(9)
Notes to Table 3:

(1)  Emission factor determined from;

     (0.0018(5/5)(D/5)(H/10))/(M/2)2
                                     15-18

-------
          TABLE 3;   SUMMARY OF EMISSION FACTORS FOR COAL HANDLING
                       AND ASH HANDLING EMISSIONS

                                (Continued)
     vbere:

     S • Silt content of material in percent (varies)
     U » Average wind speed (9.5 mph)
     h - Drop height in feet (varies)
     M - Surface moisture content in percent (varies)

(2)  Based on 0.2 Ib/ton for all transfers and conveyors with the assumption
     that 25Z of the total emissions are due to conveyors.

(3)  Assumed to be 33Z of transfer emissions  detemined  vith  the  emission
     factor  equation in Note 1.  This corresponds to 25Z of total, emissions
     from transfer and  conveying,  as  assumed  for  conservative  emission
     factors.

(4)  Emission factor determined from:

     0.04/(PE/100)2 (Ib/ton)

     where

     PE   *  Thornthwaite's  precipitation  -  evaporation  index  (118  for
     Baltimore area)

(5)  Emission factor determined from:

     0.05/(PE/100)2 (Ib/ton)

(6)  Emission factor determined from:

     (0.11/(PE/100)2) (D/90) (Ib/ton)

     vhere D - duration of material in storage (365 days)

(7)  Emission factor determined from:

     A (1.6u) (repEDCo/PEBALTIMORE)2 db/acre-hr)

     vhere:

     A « Activity factor •  .33
     U - Hourly mean wind speed  (m/s)
     PEPEDCo * Mean PE index for PEDCo field studies - 49
     PEBALTIMORE " Mean PE  index for Baltimore area - 118

(8)  Personal communication with Mr. William Wagner of U.S. EPA Region IV.

(9)  Substantial extrapolation from another material involved.

                                    15-19

-------
     TABLE 4:  ESTIMATES OF ANNUAL EMISSIONS WITH ALTERNATIVE  EMISSION
                  FACTORS BASED ON FINAL CONTROL
Source            Conservative Estimate               Best  Estimate
                       tons/year                       tons/year

Barge Unloader            6.0                            0.2

Transfers                 1.1                            0.3

Conveyors                 6.4                            0.2

Crusher                   1.5                            0.4

Bunkers                   0.9                            <.l

Active Storage Pile       5.4                            <.l
  Load-in

Active Storage File       5.4                            <.l
  Reclaim

Active Storage Pile       1.0                            0.7
  Wind Erosion

Inactive Storage Pile     4.3                            1.2
  Wind Erosion

Fly Ash Unloading         1.5                            <.l

Fly Ash Storage           0.2                              .2

Total                    33.7                            3.6

Notes to Table 4:

(1)  Based on an annual average wind speed of 9.5 miles per hour.
                                   15-20

-------
                TABLE 5;  SPECIFICATION OF CONTROLS FOR THE
                             BRANDON SHORES PLANT
Source

Phase One

Barge Unloader
Conveyor (Barge
 Unloader to
 Transfer House #1)

Conveyor (Transfer
 House £l to Transfer
 House #2)

Transfer Houses fl
 and #2 and Crusher
 Building

Main Yard Conveyor

Active Storage File
 Load-in

Active Storage File
 Reclaim

Conveyor (Crusher to
 Unit No.l Bunker
 House)

Coal Bunkers
Fly Ash Storage


Fly Ash Handling
Control Method
Enclosed with baghouse
dust collection system
and vater sprays

Enclosed conveyor and
gallery
Hooded conveyor with
wind break return
Totally enclosed with
baghouse dust collection
Uncontrolled

Stacker spray system


Reclaimer spray system


Enclosed
Enclosed with baghouse
dust collection

Enclosed in silo with
baghouse dust collection

Wet suppression from
enclosed rotary unloader
Efficiency (Z)
     99
     99
     90
     99
     75


     80


     70



     99


     99


     95
                                    15-21

-------
 Inactive  and  Active
  Storage  Pile Wind
  Erosion
TABLE 5:  SPECIFICATION OF CONTROLS FOR THE
             BRANDON SHORES PLANT

                (Continued)

         Uncontrolled
 Phase Tvo

 Inactive Storage Pile
  Wind Erosion

 Main Yard  Conveyor
         Crusting Agent
         Water spray
 95


 50
 Phase Three

 Main Yard Conveyor
         Foam suppressant
75
 Phase Four

 Barge Unloader
Conveyors  (except
 main yard conveyor
 and the conveyor
 from Transfer House
 #1 to Transfer House
 #2)

Main Yard Conveyor
Transfer Houses,
 Coal Bunkers and
 Crusher

Active Storage Pile
 Load-in
         Enclosure with vetting
         agents

         Enclosed  conveyors  and
         galleries
        Foaming  agent, vind
        shields  and hooded
        conveyor/wind break  return

        Enclosed with baghouse
        dust  collection
        Stacker spray  system
                                              (1)
95


99.5
90/87.5
99.5
                                                   (2)
67
                                              (3)
                                   15-22

-------
                TABLE 5:  SPECIFICATION OF CONTROLS FOR THE
                             BRANDON SHORES PLANT

                                (Continued)
Active Storage Pile      Reclaimer spray system             67
 Reclaim

Fly Ash Storage          Enclosed in silo with              99.5
                         vith baghouse dust
                         collection
                                                              (4)
Fly Ash Unloading        Vet suppression from               90
                         enclosed rotary unloader

Active Storage Pile      Wind activated spray system        90
 Wind Erosion            vith vetting agents

Notes to Table 5;

(1)  Based on the average control for the two transfers,  conveyor  and  the
     scoop which are part of the barge unloader.

(2)  First  figure  refers  to  load-in.   Second  figure refers to reclaim.
     Differences arise because different fractions of the  conveying  system
     are enclosed during each operation.

(3)  These  figures  are  less than those used for the conservative emission
     estimates (75Z load-in,  802  reclaim)  because  part  of  the  control
     mechanism  is  explicitly accounted for in the "uncontrolled emissions"
     determined from the best emission estimates.

(4)  Control is for enclosure only.  Control from uniform moisture  addition
     is accounted for in estimating "uncontrolled emissions".
              TABLE 6:  REPRESENTATIVE PARTICLE PROPERTIES USED
                            IN DISPERSION MODELING
                     Representative
    Mass             Particle            Reflection          Settling
 Fraction (Z)         Diameter	      Coefficient         Velocity (m/sec)

     16                  7.07 urn             .97              2.0075 x 10~3

     54                 19.31 urn             .7375            1.498  x 10"2

     30                 40.33 urn             .5812            6.533  x 10~2
                                    15-23

-------
                               REFERENCES
(1)  U.S.  EPA,  Emission Trading Policy Statement, 47 FR 15076,  April  7,
     1982.

(2)  Maryland  State Department of Health and Mental Hygiene, Air Quality
     Program.  Plan  for  Implementation  of  the  National  Ambient  Air
     Quality   Standards   for   Total   Suspended   Particulate  Matter.
     Photochemical Oxidants and  Carbon  Monoxide  for  the  Metropolitan
     Baltimore  Intrastate Air Quality Control Region. December 15, 1978.

(3)  Bohn,  R.,  Cusci.no,  T.  and  Cowherd,  C. "Fugitive Emissions from
     Integrated Iron and Steel Plants,"  U.S.  EPA  Contract  #68-02-2120
     (Midwest  Research  Institute),   Research Triangle Park, N.C., March
     1978, EPA-600/2-78-505.

(4)  Cowherd  C., Jr. et al.  Development of Mission Factors for Fugitive
     Dust  Sources. U.S. EPA,  OAQPS, EPA-450/3-74-037, June  1974.

(5)  Cowherd, C., Jr. et al.  Iron and Steel Plant  Open  Source  Fugitive
     Emission Evaluation. U.S. EPA, IERL, EPA-600/2-79-103, May 1979.

(6)  PEDCo-Environmental  Inc.  Survey  of Fugitive Dust from Coal Mines.
     U.S.    EPA,    Region   VIII,   Office    of    Energy    Activities,
     EPA-908/1-78-003,  February 1978.

(7)  Chepil, W.S. "Influence  of Moisture on Erodability of Soil by Wind,"
     Soil  Science Society Proceedings, pp. 289-292, 1956.

(8)  PEDCo-Environmental,   Inc.   Technical   Guidance  for  Control  of
     Industrial Process Fugitive Particulate Emissions. U.S. EPA Contract
     #68-02-1375,   Research    Triangle   Park,    N.C.    March    1977,
     EPA-450/3-77-010.

(9)  United  Nations.  Air  Pollution  bv  Coking  Plants. United Nations
     Report SECE/COAL/26, 1968.

(10) Johns Hopkins University, Applied Physics Laboratory. Fugitive  Dust
     Emissions   from  the  Proposed  Vienna  Unit No. 9. Prepared for the
     Maryland Power Plant Siting Program, JHU/PPSE 8-14, February 1981.

(11) Gillette,  D. "Tests with a Portable Wind Tunnel for Determining Wind
     Erosion Threshold Velocities," Atmospheric Environment. Vol. 12, No.
     12, pp. 2309-2313, December 1978.

(12) Chepil, W.S. "Dynamics of Wind Erosion:   II.   Initiation  of  Soil
     Movement." Soil Science. Vol. 60, No. 5, P.397-411, 1945.
                                     15-24

-------
(13)  U.S.   EPA.  Compilation  of  Air  Pollutant  Emission Factors. Third
     Edition (including Supplements  1-10),  U.S.  EFA,  OAQPS,  February
     1980.

(14)  Jutze, G.A. et  al.  Evaluation  of  Fugitive  Dust  Emissions  from
     Mining.  U.S. EPA, IERL, EPA-600/9-76-001 (Contract Nos. 68-02-1321,
     Task  No.  36), June 1976.

(15)  W.P.   Reefe.  Emission  Factors  For  Mining  Operations.   Colorado
     Department  of  Health,  Air  Pollution  Control  Division,  Denver,
     Colorado  , Unpublished, March 1978, Cited by Currier, E.L., and B.D.
     Neal.   "Fugitive  Emission  from  Coal-Fired  Power   Plants,"   Air
     Pollution  Control  Association,  Proceedings  of  the  72nd  Annual
     Meeting (Cincinnati, Ohio, 1979), Publication 79-11.4.

(16)  U.S.  EPA. Regional Workshops on Air  Quality  Modeling;   A  Summary
     Report. U.S. EPA, OAQPS, April 1981.

(17)  Bowers,  J.F.,  et  al.  Industrial  Source Complex (ISC) Dispersion
     Model User"s Guide. U.S.  EPA,  Source  Receptors  Analysis  Branch,
     EPA-450/4-79-030, December 1979.
                                      15-25

-------
The work described in this paper was not funded by the U.S. Environmental
Protection Agency.  The contents do not necessarily reflect the views of the
 Agency and no official endorsement should be inferred.
              INHALATION PATHWAY  RISK ASSESSMENT
         OF  HAZARDOUS WASTE  INCINERATION FACILITIES*
                       GREGORY  A.  HOLTON**
                    THE MAXIMA  CORPORATION
                     OAK RIDGE,  TENNESSEE
                       CURTIS  C.  TRAVIS
                      ELIZABETH  L.  ETNIER
                     FRANCIS R.  O'DONNELL
                 OAK RIDGE  NATIONAL LABORATORY
                     OAK RIDGE,  TENNESSEE
 (*) Although this is not the actual presentation made at the May 1982
     meeting, it closely resembles that presentation, which was
     entitled. "A Determination of the Impact of Fugitive VOC Emis-
      sions from a Municipal Hazardous Waste Incinerator on the Sur-
     rounding Community, " G. Holton  (Oak Ridge National Laboratory)
     and L.Staley (EPA/IERL-Cin).
 (**) Current address: First Environment. 314 W. Broadway. Lenoir
     City.  TN  37771.
                              16-1

-------
                                 I.  INTRODUCTION

One  objective of  quantitative  risk  assessment  is  to  arrive  at  a  value or, more
realistically, a range of values that describe the possible adverse effects on human health
associated with low, chronic exposure to a known or suspected toxic substance.  There is
increasing impetus to use risk assessment  to  prioritize  research needs in health  and
environmental areas.  Quantitative risk assessment typically consists of two components,
the measurement or estimation of the concentrations of pollutants to which a population
will be exposed  and an estimation of resultant health effects based on available dose-
response relationships. These two components are called exposure assessment and health
effects assessment, respectively.

Exposure assessment  is  defined as  the  determination of  the  concentration  of toxic
materials in space and time  at  the  interface  of  target  populations.   This description
should include identification of all major  pathways (air, water, and  soil) for movement
and transformation of a toxic material  in a selected environmental setting.  The present
assessment  focuses only on atmospheric concentrations and exposure via the inhalation
pathway.

Health effects assessment consists  of relating population  exposure  to such potential
human   health   impacts   as   carcinpgenicity,   mutogenicity,  teratogenicity,   or
neurotoxicity.  Unfortunately, adequate data for assessing excess human  risk  resulting
from exposure to most lexicological substances do not exist.

The U.  S. Environmental Protection Agency  is conducting, research  to determine  air
concentrations,  public inhalation exposure, and  health risk  resulting from chemicals
emitted during incineration of hazardous  wastes.  Incineration facilities produce stock
and fugitive (non-stock) atmospheric emissions.  The  magnitude of public exposure to
such  emissions   is  dependent on  both  plant  design  and operation  as  well  as  the
physicochemical  properties of the waste.  Engineering variables  include facility design,
size, destruction and removal  efficiency (ORE), and emission source strength (stack vs
fugitive).   Physiocochemical  waste  variables include  incinerability,  volatility,  and
toxicity. The  present assessment was performed to determine the relative importance of
these variables to human inhalation exposure and health risk.

It should be emphasized that both the exposure assessment and health effects assessment
methodologies presented in this paper  are very generalized, and caution should be
exercised in interpreting the results.   Site-specific application of  our results would
require careful evaluation of  the extent to which the models and parameter values used
in this paper are representative of conditions prevailing at  the specific site.


                        2.  INCINERATION FACILITY DESIGN

In the United States, there  are at  least  219 facilities operating 284 hazardous waste
incinerators of varying design and capacity.   The  liquid injection (LI) design is the most
common type of hazardous  waste incinerator  (51%) employed  in the United  States.
Liquid injection  incinerators can be applied to virtually  all liquid  wastes (liquid,  slurries,
sludges). Liquid injection systems employ a waste nozzle (burner)  which atomizes the
                                      16-2

-------
waste and  mixes it with air to form a suspension that promotes complete  combustion.
Atomization is achieved mechanically or by pressure atomization systems using high-
pressure air or steam.

The rotary kiln (KK) design is employed at 7% of the incineration facilities in the United
States.'  A rotary kiln is a cylindrical refractory-lined shell  mounted at a slight incline
from the horizontal plane.  Rotation  of  the  shell provides for movement of  the waste
through the kiln as wejl as for enhanced mixing.  Rotary kilns can be applied to  both
liquid and solid wastes.

Three incinerator  sizes (I,  10,  ISO  x  10^ Btu/h)  were chosen for  analysis to provide a
profile  of  existing incinerator capacities.   Two  facility designs (liquid injection  and
rotary kiln) for each of the three sizes were selected.

Facility design may be important in  determining fugitive emmisions, since such emissions
are dependent  on  the number of pumps, tanks,  valves and flanges in operation. Each
incineration  facility was designed following a review  of existing-practice incinerators.
For example, the hypothetical  ISO x 10° Btu/h incinerator used in this study was assumed
to possess  one  receiving tank for each of four categories of waste: clean or dirty high-
Bru waste and clean  or dirty  low-Btu  waste (dirty waste is  any waste that  requires
pretreatment to enhance viscosity).  Two additional tanks were  included to provide extra
storage capacity for irregular  shipments.  The storage area of  the facility was designed
with sufficient capacity for 14 days continuous operation.  The feed area was assumed to
have sufficient storage for  three days of operation.  Other design factors such as piping
or numbers of pumps were also considered.

A compilation of important  facility design parameters is presented in Table I. Equipment
quantities  and operating assumptions for the  receiving area are identical  for  both
incinerator types,  but  vary slightly with size.  The number of storage tanks in the storage
and feed areas  remains the  same for all sizes of incinerators, but the holding capacity of
the tanks  increases.   The  number  of pumps, valves, and flanges are identical in the
storage and feed areas of the 10 and ISO  x  10° Btu/h incinerators.  Since these areas are
the major  contributors to fugitive  emissions, one should not expect  a large  change in
fugitive emissions  as a function of incinerator size.


               3.  WASTE CHARACTERIZATION  AND EMISSION RATES

Operation  of a commercial incinerator is characterized by receipt of waste of widely
varying composition.  Except in settings where an incinerator  is  dedicated to a particular
chemical  process  waste stream, a  detailed qualitative or quantitative makeup of the
waste being burned is usually unknown.  The EPA has  funded several surveys to  determine
the composition of hazardous waste streams currently being incinerated.  A  total of 237
different  constituents  have been identified as  present  in  one  or  more  of  the  413
hazardous  waste streams reviewed.  Table II lists  the ten most prevalent constituents of
hazardous  waste streams  currently being  incinerated, according  to  the most recent
survey.3 Unfortunately, data from  this survey were not available when this assessment
was performed.

To deal with the  uncertainties inherent  in  the characterization of incineration waste
streams, it was decided to  examine  the behavior of representative  chemicals from three
generic waste classes.  To  choose these  waste classes, data from  an  earlier survey of
industrial   wastes4  were   analyzed  and  grouped  by  incinerability  characteristics.
Incinerability in this report  is defined as the inherent heating value (Btu/lb) of the waste
                                      16-3

-------
before fuel oil is added. The three waste groups selected for study, in order of increasing
incinerability  were:   (I) pesticide-related  chemicals (3,021  Btu/lb); (2)  API separator
sludge chemicals (4,049 Btu/lb); and  (3) phenol/acetone  distillation  chemicals (15,850
Btu/lb). Four  chemicals from each waste group were selected for analysis.  Criteria for
choosing specific chemicals  were  high  and low volatility, high and low toxicity,  and
prevalence of  the chemical in U. S. industrial waste streams.  The chemicals chosen for
the three  waste groups were:  (I) chloroform,  ethylene bichloride, hexachlorabutadiene,
and  1,1,2,2-tetrachloroethane; (2) chromium, lead, arsenic, and  phenol; and (3) toluene,
pyridine, phthalic anhydride, and methyl styrene.

The  rate  of release (mass per  unit time)  of  specific chemicals  in stack emissions  is
controlled by three facility variables:  waste throughput, chemical concentration in the
waste stream, and ORE.  Waste  throughput in an incineration facility is  determined by
the percent contribution of the  waste to the total  waste stream  after  supplementary
addition of No.  2 fuel oil to insure combustibility.  If the waste has a  high enough  Btu
content (10,000 Btu/lb) to bum without the addition of supplementary fuel (as is the case
with phenol/acetone distillation waste), no additional fuel oil was assumed.

To determine  the effect of ORE on human  exposure  and risk estimates, three scenarios
were chosen.  These scenarios were 99.99, 99.9, and 99.0%.  An  additional  analysis of
95.0%  ORE  for arsenic, chromium,  and  lead  was performed.   Predicted  annual
incineration rates (MT/y), and annual stock emission rates (g/y) for the 99.99% ORE
scenario are given in Table III.

Fugitive emissions  consist of  tank,  pump,  in-line  valve, open-ended  valve, flange,
instrument connection, and sampling connection emissions from the receiving, storage,
and feed areas of the incinerator.  Fugitive .emissions were calculated using Monte Carlo
procedures reported  elsewhere.    Expected  total  fugitive  emission  rates for each
chemical are given in Table  III.   It should  be emphasized that  these chemical-specific
fugitive emission  rates are  only first  approximations.   An  exact methodology  for
estimating volatilization rates of specific chemicals in complex mixtures does not exist.


                            4.  EXPOSURE ASSESSMENT

4.1  SITE DESCRIPTION

The  location  chosen for this  assessment was a hypothetical northern Midwest site (S-l)
located in Marathon County, Wisconsin at 44° 55' latitude and 115° 30* longitude.  This
site  is located in a rural  area adjacent  to  several large cities.  Thus,  even  though  the
total population within 100 km of  the  site is  0.45 x  10° people,  there  ore no people
residing within 875 m of  the  incinerator site (see Table IV).  This  location is neither a
current nor a planned incineration facility site.

A  circular area of  100-km radius around the  incinerator facility was assumed for  the
assessment. Exposure and  risk due to long-ronge transport (greater than 100 km) was not
performed in this assessment.  Although  long-range transport can significantly contribute
to total population  exposure by incorporating large population  centers, the average
individual  exposure to such transported  materials is  very low (orders of magnitude  less
than background in some cases).
                                    16-4

-------
4.2 ATMOSPHERIC TRANSPORT MODELS

Annual-overage  ground-level air concentrations of representative chemical pollutants
were  estimated  using  IEM, an automated inhalation  exposure methodology.    This
methodology employs a slightly modified version of the  Industrial .Source Complex Long
Term Model (ISCLTM), a Gaussian plume model developed for EPA.7

A  circular area of 100-km radius around  the  incinerator  facility was assumed for the
assessment.  Gaussian  plume models  are  generally applied for  distances of 20-50 km
around a site.  However, this type of dispersion model has been validated out to ISO km
over flat  terrain at one location," predicting annual-average air concentrations within a
factor of three of those measured.

The assessment area was divided into sector segments consisting of 20 concentric circles
about the origin and sixteen radial direction vectors. The circles had radii of 0.15, 0.2S,
0.50, 0.75, 1.0, 2.0, 3.0, 4.0, 5.0, 7.5, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100 km.  Each
radial vector was separated by  22.5° intervals with  the first sector being centered on due
north (0°), and  proceeding clockwise.    ISCLTM  was then  employed  to  calculate
concentration estimates at segment centroids.

The total exposure in a sector  segment is  calculated by multiplying the annual-average
ground level air  concentration at the segment centroid by the number of persons located
in the segment,  as obtained from I960 Census  data tapes using an  adaptation of the
APORT computer code.'

4.3 MODEL INPUT PARAMETERS

IEM  input parameters include model plant  descriptors, pollutant  behavior variables, and
region-specific meteorological  data.  Each of the six hypothetical incineration facilities
(two designs in three sizes) was assumed to contain four separate emission sources:  a
stack and three  area sources (the receiving, storage, and  feed  areas). The stack was
located at  the  center  of the  circular grid  (site  origin),  and the  area sources  were
represented as square  areas centered on  the origin.   This  potentially artificial area
source configuration of squares was necessitated by IEM modeling constraints and will
not affect pollutant concentration estimates unduly.  Each  area was centered  on the
stack to  simplify computation  and  to prevent  possible  bias of  atmospheric dispersion
estimates.  Stack and area source quantities are summarized in Table I.  Inspection of
these values shows that stack  and receiving parameters are  identical  for both designs.
However, differences in area, width, and generic gas emission rates  for the storage (10
and ISO x 10° Btu/h sizes) and feed (ISO x 10° Btu/h size) areas do occur between designs
because of  size and number of  storage tanks.  The specific parameters employed are
summarized in Table V.

Region-specific  meteorological data were  obtained from Stability  Array (STAR) data
tapes"' and from a compendium of weather statistics''. The STAR data were organized
into  six  Pasquill stability categories (A through  F) and six wind  speed classes with
average  wind speeds  of  0.75, 2.5,  4.3,  6.8,  9.5,  and  12.5  m/s.   The  remaining
meteorological  parameters were obtained  or  derived from  Ruffner  (1978).  For the
principal  site they include an  average air temperature of 280.2° K and mixing  layer
heights of  1819.5, 1213.0,  1213.0, 1024.3, 10000.0, and 10000.0 m for stability categories
A  through F, respectively. Mixing layer heights were assumed equal  for all wind speed
classes within each stability category.
                                  16-5

-------
Site-specific population data were obtained from  1980 Census data tapes that hod been
reformatted into coarse and  fine rectangular latitude/longitude grids. Approximately
0.45 x 10° people reside within c 100-km radius around the generic site.


                        5. HEALTH EFFECTS ASSESSMENT

A  key component  of  health effects assessment is the determination of human health
response to a given exposure.  Quantification of the exposure-response relationship can
be made in  an number of  ways, depending on the availability of data and the health
impact  under  consideration.  There are many potential health endpoints in  humans,
including carcinogenicity,  mutagenicity, and  teratogenicity.  Unfortunately,  adequate
human data  for  assessing risk to most toxicological endpoints do not exist, although a
wealth of animal data are available for selected endpoints.

Toxicity data  employed  in  this assessment are taken from a health effects assessment
summary  for  300 hazardous  organic jaxistituents  prepared  by  EPA's Environmental
Criteria Assessment  Office  (ECAO).     Chemicals  studied were  classified as either
carcinogens  or non-carcinogens, and separate measures of  toxicity were developed for
these two classes.  Carcinogenic toxicities are etimated  in  terms of  lifetime excess
cancer  risk  factors,  while noncarcinogenic  toxicity  is estimated by acceptable  daily
intake (ADD, and threshold  limit  values (TLV).

Health  effects data  for compounds considered in this  assessment are summarized  in
Table VI.   It is apparent  that  not all  health  risk estimators are available  for  each
chemical studied.  The estimate  of excess human lifetime risk  to cancer was used for all
compounds in  the pesticide-related waste.  In the absence of cancer risk estimates, the
available ADI  value was  used for all compounds in  the API separator sludge waste except
for arsenic;  the TLV1J was converted to an ADI  for arsenic and for  compounds in the
phenol/acetone distillation  waste for which  there is no reported AOI.


                          6. RESULTS ANO CONCLUSIONS

The purpose of this assessment was to determine the relative importance of plant design
and  waste physicochemical  variables on  human inhalation exposure and  health risk
resulting from hazardous waste  incineration.  A hypothetical waste incineration site  in
the northern Midwest was  chosen for analysis. This site  has a population of 0.45 x 10°
persons, with the closest individuals residing 1500 m from the incineration site.

To account for engineering variables, two facility designs (liquid injector and rotary kiln)
of three sizes (I, 10, and  ISO x 10° Btu/h), each burning  three generic wastes,  were
selected.  These are designated LI-1,  LI-IO, LI-150, and  RK-I,  RK-IO, and  RK-150,
respectively.  Three  levels  of destruction and removal efficiency (ORE) were considered
(99.0, 99.9 and  99.99% ORE). The three waste groups selected for study, in  order of
increasing incinerability were:   (I) pesticide-related  chemicals (chloroform,  ethylene
dichloride, hexachlorobutodiene, and 1,1,2,2-tetrachloroethane); (2) API separator sludge
chemicals (arsenic,  chromium,  lead,  and  phenol) and  (3) phenol/acetone distillation
chemicals (toluene, pyridine, phthalic anhydride, and methyl  styrene).

Annual-average  ground  level  air concentrations of representative chemical pollutants
were  estimated  using  IEM,  an automated  inhalation  exposure  methodology.5   Air
concentrations were estimated for both stack and fugitive emissions using region-specific
meteorological and climatological data.
                                     16-6

-------
Estimates  of  individual and  total  population exposure resulting from incineration of
hazardous  materials were calculated  by multiplying ground-level air concentrations by
site-specific population  data obtained from  the  I960  Census.  Estimates of  risk
associated with incineration of hazardous materials were obtained by multiplying the
population  exposure  estimates   by  health  risk  estimators  supplied  by  the  U. S.
Environmental  Protection  Agency  (EPA) Environmental  Criteria  Assessment Office
(ECAO). Chemicals studied were classified as either carcinogens or noncarcinogens, and
separate measures of  toxicity  were used for  those two classes.  Carcinogenic toxicities-
are measured  in terms of  lifetime excess cancer risk, while noncarcinogenic toxicity is
measured  in terms  of acceptable  daily intake (ADI).  Estimates of  life-time excess
cancer risk were used for  chloroform, ethylene dichloride, hexachlorobutadiene, 1,1,2,2-
tetrachloroethane, and arsenic.   The AOI value was  used far chromium, lead, phenol,
toluene, pyridine, phthalic anhydride, and methyl styrene.

Major conclusions to be drawn from  this report are:

I.   Fugitive emissions may be  an important  contributor  to  total po!Mutant emission
     rates. The largest contribution occurs when stock emission rates are lowest, that is
     for small  incinerators (I x I06 Btu/h) with OKE's ranging from 99.9 to 99.99%.  For
     large incinerators, fugitive  emissions are a relatively  unimportant contributor to
     total emissions  at  all  ORE'S  studied  (Table  VII).   Caution should  be used in
     interpreting these results as it appears that the fugitive emission predictions in this
     report may be higher than those actually experienced at incinerator sites.

2.   Total  population exposure  is  relatively insensitive to changes in ORE  for small
     incinerators (I  x 10° Btu/h), increasing  by only a factor of  7 from 99.99 to 99.0%.
     Total population exposure to volatile compounds is sensitive to incinerator size for
     low ORE'S (99.0%) and is relatively insensitive to incinerator size for high ORE'S
     (99.99%) (see Table VIII).

3.   For  all chemicals  studied,  human health risk  from  incineration  was small.   For
     incineration of pesticide-related waste  at  a site with a population of  0.45 x  10°,
     expected  number of  cancers over' 70 years are less than 1.6 x IO"3  for a ORE of
     99.99% (see  Table IX).   The excess cancer risk over  70 years to the maximally
     exposed individual at site S-l  from incineration of pesticide-related waste is  less
     than  1.2 x IO'7 for a ORE of 99.99% (see Table X).

4.   None of the four selected constituents of the phenol/acetone distillation wastes are
     known carcinogens.  Therefore, using the AOI as a measure of noncarcinogenic risk,
     estimates of average daily  intake  as a fraction of the AOI were mode.  These  risk
     estimates, which account for both stack and  fugitive emissions,  are  highest for
     pyridine releases from ISO x 10° BTU/lb plants, but even these risk values are  less
     than  lO'7 of the AOI  (see Table XI).

5.   Risk to exposed individuals (as a fraction of AOI) from incineration of heavy metals
     is  small (see Table XII).  Even for the maximally exposed individual, all inhalation
     exposures to heavy metals  result  in average  daily intakes which are  at least 6
     orders of magnitude less than the acceptable daily intake.
                                    16-7

-------
                               7. REFERENCES

'Keitz,  E. L., L. J. Boberschmidt, and  C.  C. Lee, "A Profile of Existing Hazardous
     Waste Incineration Facilities," MITRE Corporation, McLean, Virginia (1982 — in
     preparation).

2Bonner, T., 8. Oesai,  J. Fullenkamp, T. hughes, E. Kennedy, R. McCormick, J. Peters,
     and O. Zanders,  Engineering  Handbook  for Hazardous Waste Incineration, EPA
     Contract No. 68-03-2550, Monsanto Research Corporation, Dayton, Ohio (1980).

3MITRE  Corporation,  "Composition   of   Hazardous  Waste  Streams  Currently
     Incinerated." Working paper (1983).

*U.  S.  Environmental  Protection  Agency,  Background  Document, Resource  and
     Recovery Act. Subtitle C. Identification and Listing of Hazardous Waste, Office
     of  Solid Waste, Washington, D.C. (1980).

Vtolton, G. A. and C.  C. Travis, "A Methodology for Predicting Fugitive Emissions for
     Incinerator  Facilities," Proceedings,  1983  Notional  Meeting of the American
     Institute of Chemical Engineers, Washington, U.C. (November, 1983).

6ODonnell, F. R., and G. A. Hoi ton, "An Automated Methodology OEM) for Assessing
     Inhalation Exposure to Hazardous Waste  Incineration Emissions," Incineration and
     Treatment  of  Hazardous  Waste;   Proceedings of  the 9th  Annual Research
     Symposium  at Fort  Mitchell, Kentucky,  May  2-4,  1983, U.  S. Environmental
     Protection Agency, Cincinnati, Ohio (in press).

7Bowers, J. F.,  J. R. Bjorklund, and C. S. Cheney, Industrial Source Complex  (ISC)
     Dispersion  Model  User's  Guide.  (Volume   I),  • EPA-450/4-79-030,   U§7
     Environmental Protection Agency, Research Triangle Park, North Carolina (1979).

^uckner, M.  R.  (ed.),  Proceedings of the 1st SRL Model Validation Workshop. Nov. 19-
     20, 1980, Hilton Head, S. C., Savannah River Laboratory, DP-1597 (1981).

'Fields,  O.  C.  and  C.  A. Little,  APORT  —  A  Program  for  the  Area-Bosed
     Apportionment of County Variables to Cells of a Polar Grid. ORNL/TM-6418. Oak
     Ridge National Laboratory, Oak Ridge, Tennessee (I97T5E

'"National Oceanic and Atmospheric Administration, Seasonal ond Annual Distribution
     by Posquill Stability Classes STAR Program. National Climatic Center, U. S.
     Department of Commerce, Asheville, North Carolina,  1974.

"Ruffner, J. A., Climates of the States. Vols. I and II, Gayle Research Company, Book
     Tower, Detroit, Michigan (1978).

I2'VI. S. Environmental Protection Agency, Environmental Criteria Assessment Office,
     Health Effects Assessment Summary for 300 Hazardous Organic Constituents in
     Support  of  Regulatory Impact Analysis  of the  Land  Disposal Branch (LOB) and
     Interim Final Incinerator Regulations of the Technical Branch (TB) of the Office
     of Solid  Wastes,"  OSWER/EPA (1982).

13American Conference of Governmental Industrial Hygienists, TLVs - Threshold Limit
     Values  for Chemical  Substances and  Physical  Agents   in  the  Workroom
     Environment with Intended Changes for 1974, Cincinnati. Ohio (1974).
                                   16-8

-------
Table I. Expected equipment quantities and operating assumptions
              for each incineration facility design
Rotary
kila
liqvid iaj*etor
Six*. 10* Bttt/n

•BCEIVDW AREA:
Capacity, track*
Capacity, taak car*
Pwpa (pain)
la-lia* fair**
Op*a-*ad*d ralr**
Flaagea
laatr. eoaacctioaa
1

1
0
1
13
2
24
4
Saatpliag ooaneotioas 0
Storage taaka
S1DRA6E AREA:
Poapa (pain)
Pvapa (aiagl*)
Ia-lia* talrea
Op*a-*ad*d >*lr**
Flaagea
laatr. coaaectioaa
0

1
0
13
.2
24
4
Saatpliag coaaectioaa 4
Storage taaka
FEED AREA:
Pnpa (pairs)
la— liae fair**
Opea-eaded fair* a
Flaagea
laatr. eoaaectioa*
4'

1
13
2
24
4
Saatpliag aoaaectioaa 1
Storage taak*
•1.0 x 104 gal
b2.0 x 104 gal
e1.5 x 10S gal
"1.25 x 104 gal
•2.5 x 104 gal
ll
capacity.
capacity.
capacity.
capacity.
capacity.
10

2
0
3
39
6
72
12
2
2*

1
2
39
<
72
12
4
4b

1
13
2
24
4
1
1*





150

2
1
4
52
1
96
16
6
'*

1
2
39
6
72
12
4
4«

1
13
2
24
4
1
1
*2.0 103
15.0 103
•1 .2 103
i(.Q JO3
Jl.5 103
1

1
0
1
13
2
24
4
0
0

1
0
13
2
24
4
4
4d

1
13
2
24
4
2
2
«•!
gal
gal
gal
gal
10

2
0
3
39
6
72
12
2
2«

1
2
39
<
72
12
4
4«

1
13
2
24
4
2
2d
capacity.
capacity.
capacity.
capacity.
capacity.
150

2
1
4
52
8
96
16
6
<*

1
2
39
6
72
12
4.
4*

1
13
2
24
4
2
2-i





                         16-9

-------
   Table II. Ten most prevalent constituents of hazardous
                    waste streams
Cam«titB»t            iBoamt  Incinerated (JfT/y)
lUtkuol                         133.168
Ao«tomitril»                      St.646
TO!M««                           it.620
Btianol                           55.090
4«rl «a«t*t*                      54.926
Ae«tom«                           51.535
XTl»n«                            49.453
lUtiyl «thyl t«too«               42.520
Adlple told                       36.135
Etiyl toctatt                     32.576
                       16-10

-------
                                                     Table III. Consumption, concentration, and emission rates of
                                                                 incinerated chemicals (99.99% ORE)
                                                                                                    lM Ml«,  0/7
                             •allBIIBt
     I FiMlloB
 |BB!B-   •••la
•r«t«4*  • tfB»B
 mitt     it)
                                                                                    UU
                                                                                                                            UjMlor
                                                                                 10
                                                                                               ISO
                                                                                                                            10
                                                              BUck P«|lll» ll««l FBflllt* lock P«(IU>B link PB|lllf* BlBBk FB|ltl» luck  F<|lllt<
OJ
 i
PESnCIDE-BELATBb BASTE
Cklorof or»
ElkfltBi «Ukloil4B
B«»eklorokBU4l»*
I.I1E*4
7.1JE.4
4. If 1.4
l.l.l.l-l«tr>cklo(0*tkBB«l.lOE*4
API IEPABATOI
SLUDGE BASTE
Cklo.lue
Ui4c
AI»B!CC
Pk.B.I
PBENOL/ ACETONE
DISTILLATION BASIC
TalB»*
PycHlBB
PklktIU (BkytfrKt
lUlkjrl |.) 0 1.441.4 0 1.101.) 0 l.(*l»l 0 1.4*1.4 0 1.10B*! O
1.101*1 0 1.101*4 0 I.lOBtl 0 1.101*1 0 1.101*4 0 1.101*1 0
4.001*1 0 4.001*1 0 (.001*4 0 4.001*1 0 4.001*1 0 4.001*4 0
(.111-1 1.4)1-1 (.171*0 *.(BB-1 1.0)1*1 I.IOB-I 4.I7B-1 1.I1B-1 (.111*0 ».IIB-1 l.OJf.l 1.I1B-1


1.»1B«1 4.4)B») I.*1B«4 1.1(1*4 1.171* ) 1.111*4 l.UB.l (.111*1 l.»ll*4 1.101*4 1.171*1 1.40E*4
2.41E.1 1.411*1 l.(BB*l ». 411*1 4.011*4 1.111*1 l.(IB*l 1.441*1 1.411.1 ».(»•*! 4.011*4 1.111*1
l.Olt-1 1.141-1 1.011*0 1.4(1-1 1.0*1*1 1.44B-1 1.01B-1 1.401-1 1.01B*0 1.4*E-1 1.0*1*1 1.101-1
1.121*1 *.»11*1 1.711*4 1.111*1 1.111*1 4.1*1.) 1.111*1 1.011*1 1.111*4 1.1*1*1 1.1JB.J 3.1JE.J
                        •fiom oo.pll>llo> •( ml* l>l«n>llo>  (DIETA.  1*M«).

                        ^Calcvlklcd ky •*«r«|l«| f«ill«l4*-v*lft*«4 •••!• Bka«l

                        •i ME af »1.M ll B1IBM4 101 Ik* k«fj MUU.

-------
Table IV. Cumulative population by distance from the incinerator site
        Distance to Ring Center (m)

                    200
                    375
                    £25
                    875
                  1,500
                  2,500
                  3,500
                  4,500
                  6,250
                  8,750
                  15,000
                 35,000
                 55,000
                 75,000
                 95,000
Population

      0
      0
      0
      0
    203
    949
   2,378
   4,053
   6,039
   7,437
  76,576
 126,614
 263,097
 349,969
 448,187
          Table V. Stack and area source parameters employed in atmospheric
                               dispersion modeling
                                       Rotary kiln
               Liquid  iajcetoz
                                     Sis*.  10* BtWh
               Six*, 10*  Btu/h
                                            10
                                                    150
                                                                      10
                                                                             150
STACI:

    H*i|At. •                         15.24
    Exit «a* t*»p«rator*.  I          355
    Exit fa* f.locity,  •/*           1.10
    Di*B*t*z. •                       0.41
    8»*rlc (a*  taiui OB rat*.  (/•   1.0

RECEIVING AREA:

    H.iffct, •        2
    Ar*a (iqvar*),  •
    Width of ar*a.  •
    Q*a*xic |t«  Minion rat*,  f/t

STOKAQE AREA:
    B*i(Jit. .        2
    Area (iqnar*),  •
    lidth of az*a.  •
    0«n«rie |*«  ••ii*ion rate,  f/i

FEED AREA:
Araa (tqnaz*). •
Widti of u*a. •
G«n«rie (*• aaiitiioa  rat*,  1/1
                                      4.1
                                      16.2
                                      8.73
                                      76.2
30.49
355
11.0
0.61
1.0
6.1
130
11.4
130
30.41
355
41.1
1.22
1.0
15.24
355
1.10
0.41
1.0
                    30.48
                    355
                    11.0
                    0.61
                    1.0
f.l
S»9
2«.3
689
S.I
76.2
8.73
76.2
                     f.l
                     130
                     11.4
                     130
                                                                            30. 4«
                                                                            355
                                                                            41.1
                                                                            1.22
                                                                            1.0
6.1
74.3
8.6
74.3
6.1
130
11.4
130
6.1
390
19.8
390
6.1
74.3
8.6
74.3
6.1
130
11.4
130
6.1
390
19.8
390
6.1
325.2
18.0
325.2
6.1
345
18.6
345
6.1
2165
46.5
2165
6.1
325.2
18.0
325.2
6.1
423
20.6
423
6.1
2737
52.3
2737
                                                                            S.I
                                                                            S66
                                                                            29.4
                                                                            866
                                    6-12

-------
       Table VI.  Health risk estimators for chemicals in three generic wastes
                                           Be«lth li«k E«ti««tor
            Pollutant
              Exce**  tiik*     ADIb
                          TLV«
     PE3TXCIDE-IBLATED WASTE

Cklorefoa
Btarlea* dicaloride
BexsoAlorobatadlene
1.1.2.2-tetr»ckloroet««ae

    API SHPAIAIOK SUOBB WASTE

      BB 71
         III
Uad
ArMaie
Pkcaol

FUMOL/ACETDNE DISTILLATION IASTE

TelMM
Ptridlm.
Pkti«He tniydrid*
       .tyr.n.
 1.8E-1
 3.7B-2
7.7SB-2
 2.0B-1
   NA
   NA
   NA
 1.4E+1
   NA
   NA
   NA
   NA
   NA
                              I. SEX)
                                NA*
                              1.4B-1
                                NA
                              1.4E-1
                              1.3E+J
                              l.OE-1
                              7.0B-2
                              6.IB+0
                              1.3E+2
                              5.4B+0
                              2.1B+0
                              1.7E+1
                                                                5.08+1
                                                                4.0B+1
                                                                  NA
                                                                3.JB+1
                                                                5. OB- 2
                                                                5.0B-1
                                                                1.3E-1
                                                                2.0B-1
                                                                1.9E>1
                                                                3.75E+2
                                                                l.JEfl
                                                                6.0E+0
                                                                4.JOE+2
         c»ae«r tf t«r » lifetia* «rpo»r(
i«t»k« (OSEPA.  1983).
     •Ese««* ri»k of
(USEPA. 1982).

     *Aee«pt«bl« dail
     °Thz*«hold li»it t*la« 1* U« ti««-»«i»ht»d conc*ntr«tion for »n
S-h verkd>7 »d 40- h workweek to be ued t* i |«ide in tie control of
       kascrda (ACOIH, 1980).

     'Not t»«iUbl».
                                16-13

-------
    Table VII.  Percent contribution of fugitive emissions to total
      emissions of chloroform at a liquid injection incinerator
« ME

99. 99
99.90
99.00

LI-1
93.5
51. •
12.5
CamtribatiOB (%>
U-10
71. <
37.1
3.6

U-150
24.1
3.2
0.3
     Table VIM.  The effect of ORE on total population exposure
            (person-ug/m-') to chloroform emissions from
                   a liquid injection incinerator
« MB
99.99
99.90
99.00
LI-1
2.5B+0
3.7B+0
1.7E+1*
U-10
7.1E+0
1.7E+1
1.1B+2
U-150
2.08+1
1.2S+2
1.2B+3
       •l»«d >i  1.7 z 101
    Table IX.  Expected number of excess cancers over 70 years
      from incineration of pesticide-related waste at a liquid
        injection incinerator (for a population of 0.45 x 10*)
* ME         U-l              U-10             U-150
99.99
99.90
99.00
1.3B-4
2.4B-4
1.6B-3
4.0B-4
1.4B-3
l.lfi-2
1.6B-3
1.2B-2
1.2E-1
                               16-14

-------
      Table X.  Excess cancer risk over 70 years to the maximally
       exposed individual from incineration of pesticide-related
          waste at liquid injection incinerators (99.99% ORE)
            U-l
                                LI-10
                                                   U-150
           2.SB-8
                               7.8E-8
                                                   1.2E-7
Table XI.  Average daily intake of selected phenol/acetone distillation
     waste constituents released by hazardous waste incinerators
                (presented as a  fraction of the ADI)



Tolnca*
Pyridin.
Piti. tafcyd.
Methyl > tyrant

AOI

134
5.4
2,1
171


U-l
2.7E-10
3.9E-10
3.2E-13
9.1E-11
Inoin*

U-10
1.2E-9
1.8E-9
1.4E-12
5.8B-10
r*tio« Facility (99.

U-150
8.4E-9
1.4E-8
2.SE-11
5.7E-9

LI-1
2.8E-10
4.1E-10
3 .2E-13
9.3E-11
99% ORE)

U-10
1.2E-9
1.9E-9
2.4E-12
5.8E-10


Ll-150
8.5E-9
1.4E-8
2.5E-11
S.7E-9
    Table XII. Average daily intake of heavy metals (as a fraction
    of AOI) for the maximally exposed individual from incineration
             at a liquid injection incinerator (95% ORE)

Chraaia'
Ari«nic
Utd
U-l
7.4E-8
3.8E-8
1.4E-7
U-10
3.0E-7
1.5E-7
5 .3E-7
U-l 50
7.4B-7
3.8E-7
1.4E-4
  •Tb« AOI f»la»  for ehroaini  IV *•* u«d.
                              16-15

-------
             Impact of Fugitive Emissions on PM-10 Concentrations
                    Thompson G.  Pace, EPA/OAQPS-RTP
                                   ABSTRACT
The National Ambient Air Quality Standard for Particulate Matter is cur-
rently being reviewed.  Indications are that a revised standard for inhal-
able particles (IP) will be proposed.  This will consider only  those par-
ticles smaller than 10 jion aerodynamic diameter.  Knowledge of the bi-
modal distribution of particles in ambient air suggests that, while fugi-
tive sources were a major component of TSP,  they will be much less sig-
nificant under an IP standard.

Ambient chemical and size distribution data and  assessment of source
contributions based on a number of ambient studies will be consolidated
into an appraisal of the impact of fugitive emissions on the inhalable par-
ticle fraction of Particulate Matter.  Empasis will be on size segregated
data from dichotomous sampler studies where the  coarse  mode contribu-
tion to ambient IP can be isolated.
                                                           »
(This is an abstract of a verbal presentation for  which a paper is not available.)
                                 17-1

-------
                     APPLICATION OF DISPERSION DICTATED
                        MASS BALANCE FOR CALCULATING
                          FUGITIVE  DUST EMISSIONS
         The work described in this paper was not funded by the U.S. Environmental
         Protection Agency. The contents do not necessarily reflect the views oi the
         Agency and no official endorsement should be inferred.
CLIFFORD F. COLE,  TRC  ENVIRONMENTAL CONSULTANTS, INC.
JOHN G. MOLDOVAN,  ANACONDA MINERALS COMPANY
PAUL B. KUNASZ,  CONSULTANT
                                   ABSTRACT
      This  paper  describes  the  development  and  application  of  a  new
receptor-source model  which  provides  estimates  of  emission  rates  from
fugitive  dust  sources  by  employing  a dispersion dictated  mass  balance
(DDMB)   technique.   Unlike  other   chemical  or  morphological   receptor
models,  the  DDMB  method  considers  only  one  constituent, which in  most
practical  applications will  be  total suspended  particulate matter.   One
equation  is written for  each  receptor, such  that the total  concentration
of  particulate collected  at  each receptor is  set equal  to  the  sum  of  a
contributing   source's  emission   rate  times   a  dispersion  factor.   The
dispersion  factor  must  be calculated  from a dispersion model, and  in this
sense,  the DDMB  method  relies  on  meteorological data.   Given  measured
particulate concentrations, values  of  emission rates can be computed  by
solving a set of simultaneous  linear equations.

      The  DDMB method  was  applied  to compute  the emission rates and the
contribution   of   TSP   levels   of  a  major  mining   operation  in  Butte,
Montana.  The DDMB results provided an estimate  of the mine's "effective
emissions"  which  differed  greatly  from  emissions  determined  by  the
traditional  emission   factor   technique.    The  Butte  study  is   used  to
demonstrate the DDMB methodology, and findings  of  the study are discussed.
                                   18-1

-------
1.0   BACKGROUND

      In airsheds  characterized by  high particulate  concentrations,  the
quantification  of  each  major  source's  contribution   to   ambient  air
particulate levels is critically  important.   Only when  it  is known where
particulates come  from,  and how  much an  individual  source  adds  to  the
particulate burden, can controls be  applied  in  a  cost-effective way.  The
large number  of  areas in  the  U.S.  in which concentrations  still exceed
particulate  ambient   air  standards,  despite expensive  control  efforts,
attests  to   the  difficulty   of   identifying   and    quantifying   dust
sources(l).

      Source apportionment—the determination  of  the  mass  of particulate
matter   at  a   receptor   that   emanates   from   various   sources—has
traditionally been accomplished with source modeling.   Emission rates of
sources  are  estimated,  and input  along  with meteorological  data  into  a
dispersion model  that  analytically  simulates  the  transport,  diffusion,
and  deposition  of  particulate matter.    In the  vicinity   of  discrete,
well-defined  point  sources, this  modeling method  appears  to  work well.
However, in  regions  where a large portion of  the particulate  matter is
fugitive dust  the source modeling  method  has proven  inadequate^).  The
failure  of   traditional  modeling  to   properly  account  for  observed
particulate   concentrations  is  due  in   part   to   the   difficulty   in
quantifying  fugitive dust  emission rates.   Emission  rates  from  dust
sources  are  hard  to  measure,  and  change dramatically  with  meteorological
conditions, type and  extent of  source activity,  and other parameters.

      In the last few years receptor modeling has  been touted as a remedy
to  the  fugitive  dust  source  apportionment problem(3).   Proponents  of
receptor modeling  correctly point out that  chemical mass  balance (CMB),
enrichment   factor,   time  series,   and  multivariate   receptor   models
eliminate  some  of  the  uncertainty  associated   with  source  modeling.
However;  there   are   two   distinct  drawbacks  that  limit  most  receptor
models' use in regions where fugitive dust is suspected  of  inducing large
hi-vol   concentrations.    First,   receptor   models   are   unable   to
differentiate physically  or chemically  similar constituents  that  may be
entrained from different  sources;  and second, receptor models assume that
the  relative  mass fractions of all  chemical  species are  conserved,  and
that  no   selective   deposition  occurs(^).   To   illustrate   that  these
drawbacks  can  be  severely  limiting, consider  a  TSP nonattainment  area
where   road   dust   is   found   to  be  a   significant   contributor   to
concentrations.     If   the  road   dust  is   chemically   and  physically
homogeneous, and  if  deposition and  resuspension  of the  dust occur, then
the receptor model cannot pinpoint which  are the offending roadways(2).

      This paper  discusses  a new  receptor model method  termed Dispersion
Dictated Mass  Balance  (DDMB).   The DDMB method  utilizes features from
both source  and  receptor  modeling technology, thereby overcoming  some of
the  deficiencies  in   either modeling approach  used alone.   Because DDMB
considers only one constituent, usually  total suspended particulate,  and
because  the  effects   of  deposition  and  reentrainment can  be  taken into
account, the  method   is  especially  well suited to  fugitive  dust source
apportionment.  The  DDMB  method is  simple  and  inexpensive to  apply,  and
yields statistically  sound results.

                                   18-2

-------
2.0   ANALYTICAL METHOD

2.1   GOVERNING EQUATIONS

      Like all receptor  models,  the basis for the DDMB method is a simple
model  in  which  the  pollutant  contribution  from  a  given  source   at  a
receptor  is  set  equal   to  the  emission  rate  of  that  source times  a
dispersion factor:

      Xj - (X/Q)j  Qj                                                  (1)

      where X j      is the concentration  at  the receptor due to source j

            Qj      is the emission rate  from  source j

           (X /Q)j   is  the dispersion  coefficient  from  source  j  to the
                    receptor

Equation (1) can be summed over a number  of  sources, P, so that

                P

        X -    , *r i   0(/Q) .  Q.                                       (2)

The only restrictions imposed  by equation (2) are that concentrations are
additive,  and  directly  proportional  to  emission rates.  If  there  are no
more  than  P  sources  in  the  vicinity  of  the  receptor  that  contribute
pollutant,  then  equation  (2)  will  account  for  all  of  the  pollutant
concentration  detected.   More likely,  it will be  impossible  to identify
all  of  the  sources  that  impact the  receptor,  but  the  contribution of
unknown   sources   can   be  accounted   for   by  an   assumed  background
concentration, B, which  is added to  the  right-hand  side  of equation  (2).
Obviously,  equation  (2) can  be  written for  any  number  of  independent
receptors  to  produce  a set  of linear  equations.   Generalizing  for  n
receptors,
      Xn
                      (x/Q)
nj
        +  B
(3)
If  the  x/Q  values  and  the  background concentration  are  known,  if  the
measured concentrations at  each  receptor  are available, and if the number
of  receptors  equals or exceeds  the number  of  sources,  then equation (3)
yields a fully determined  (or  overdetermined)  system  that  can be solved
to find  source emission rates.

      While   the   above  list   of  necessary   data   conditions  appears
restrictive, in fact for many regions  these data may already be available
or  can be  readily calculated.  Consider an area deemed nonattainment for
TSP.   Annual   average  hi-vol  concentrations  at  several   locations  will
undoubtedly   be   available,   and   an  annual   average   background   TSP
concentration  can  easily   be  assumed  or   can  be  determined  from  these
measured  concentrations  using  a variety  of methods^).   The  X/Q values
may already be known if  adequate modeling  studies  have been performed in

                                18-3

-------
the area.  If not,  model  simulations of TSP  concentration at each of the
hi-vol  receptors  must be  performed.   The  X/Q values  are  determined by
dividing modeled  concentrations  by  the  source emission rates  assumed in
the model.

      It is  important  to  note that  the emission rates of each source or
source grouping ultimately determined  by the DDMB method  depend  upon the
X/Q dispersion  coefficients  that  are  introduced in  equation (3).   For
this  reason  every  attempt   should  be  made   to  accurately  simulate
dispersion in the initial modeling phase.   If deposition  is thought to be
an  important  phenomenon  in the  area being  studied, then a model  which
accurately refects  deposition  should be used.   Similarly, meteorological
data  input  to  the  dispersion  model should  be  screened   carefully.   If
sources  are  combined,  then  the resultant  x/Q  values depend upon  the
intial  choice  of  emission rates  assumed  in the  dispersion  model,  and
special  care  should be  taken to  estimate these initial  emission rates
accurately.

2.2   METHODS OF SOLUTION

      The set of  fully determined or overdetermined  DDMB  linear equations
can be  solved with  almost  any matrix solution  software package,  although
.software  packages  that  provide  statistical  analysis   of  the data  are
preferable.   For  overdetermined  systems the  standard  deviation   of  the
predicted emission  rates,  and  the matrix of correlations  between  emission
rates,  are  extremely   valuable  statistics.    The  standard   deviation
provides a measure  of how "good" a  computed fugitive  dust emission rate
is, and allows one  to compute  statistical  confidence intervals  associated
with each emission  rate.    Given  the  uncertainty in  fugitive dust  emission
factors, confidence intervals  should be  demanded by every investigator.
The  correlation  matrix   between  emission  rates  provides  guidance  in
combining sources, a technique which  greatly enhances the DDMB method.

2.3   SOURCE COMBINATION

      In order for  equation  (3)  to  be  solvable the  linear system  must be
fully determined  or overdetermined,  that  is,  there must  be as   many or
more  hi-vol   receptors  than  sources.    One  way  to achieve this is  to
combine  sources  by grouping  them  together for  the purpose of the  DDMB
computations.  For  example,  two  adjacent unpaved  parking  lots could be
considered a  single fugitive  dust source,  thereby  reducing  the order of
the  linear  system  of governing equations  by  one.   Of   course,  source
combination  like  this  reduces the resolution of  unknown  emission rates,
but it  may be required  to make the matrix fully determined.  Even more
important,   careful  source  combination  can  dramatically  improve  the
accuracy of the computed emission rates, as will  be  demonstrated  later in
this paper.
                                18-4

-------
      In  deciding  which  sources  to  combine,  common sense  dictates that
sources which  have similar impacts  on  the hi-vol  receptors  be combined.
The hi-vol data cannot  discriminate  one  of these sources from another, so
it  is  not  reasonable  to ask  the  DDMB method  to  make a  distinction.
Frequently  such sources  will  be located  near  one another,  as  in  the
adjacent  parking  lot example above.  More rigorously,   one  can determine
which  sources  are linked  by  examining   the correlation  between predicted
sources'  emission  rates.   When  the   correlation  between  two  sources
approaches -1.0 this  indicates  that the two emission rates  are linked in
such a way that if one  rate is  varied,  the other must react  oppositely in
order  to  minimize  the  sum  of  squares  of the  residuals.   Two  strongly
negatively correlated sources cannot  be  resolved  by the  data at hand,  and
attempts  to  do so  would yield  arbitrary results  which  would be extremely
sensitive to any errors in the data.

3.0   DDMB APPLICATION

      The DDMB method was used to determine source apportionment and  TSP
emission  rates  from  various  sources  in   Butte, Montana for the  year 1978.
Butte  provides an especially  attractive study  area for the DDMB  method
because there  are a  number  of  fugitive dust sources  in and  near Butte
whose   emission   rates   have   proven   very   difficult  to   quantify.
Conventional fugitive dust inventories have yielded drastically different
emission  rates   for  the  same  sources^).    This is  illustrated   in
Table 1.,  which shows a summary of emission rate  estimates  determined by
three independent studies.

                                  TABLE  1.

               1978 ANNUAL ESTIMATED EMISSION RATE (tons/year)
                  FROM ANACONDA SOURCES  AND BUTTE SOURCES
EMISSION SOURCE
ANACONDA
BUTTE
STUDY 1
15,000
11,000
STUDY 2
5,612
3,402
STUDY 3
26,142
11,564
      In  the  following  sections  the  application of  the  DDMB  method  at
Butte is  illustrated.   The major  sources  in Butte  are  described,  hi-vol
and  x/Q  input data  are  tabulated, and  the  sequential steps  in the  DDMB
process are discussed.

3.1   SOURCE AND RECEPTOR DESCRIPTION

      Figure  1.  shows the  location of  major particulate  matter sources
and hi-vol samplers in Butte.   The  hi-vol  samplers,  indicated by stars in
Figure   1.,    provide   every-sixth-day   measures   of  total  suspended
particulate.   Of  the  nine  hi-vols displayed  in  Figure  1.,  three  are
operated  by  the  Montana  Air Quality  Bureau,  and  the  remainder  are
operated by Anaconda Minerals Company.
                             18-5

-------
   1.  ALPINE           6. KAW
   2.  HILLCREST        7  GREELEY
   3.  MONTANA POWER    8. KEBGEN
   4.  MT. CON          9  DR.  CANTY
   5.  YATES
9  J   234   5   6 Km.

          SCALE
                            FIGURE 1

                      SOURCE  AND HI-VOL
                            LOCATIONS
                                18-6

-------
      Anaconda operations  are located-from  the  northeast to  the  east of
the city  of Butte,  and  involve  the  mining  and  concentrating  of  copper
ores  from   the   Berkeley  Pit.   Drilling,   blasting,   and   removal  of
overburden  and copper ore by  shovels  are dust producing  activities that
are confined to  the  550  meter deep Berkeley Pit.   The dumping and storage
of ore,  as  well as  crushing and  concentrating  activities,  occur  at the
southern end  of  the  pit  perimeter.   Overburden is  temporarily  stored in
primary  dumps to  the  northwest,  northeast,  and  southeast  of the  pit
before being  returned in a backfilling operation.   Haul  roads and access
roads link  the Berkeley Pit with  the crusher, concentrator, and dumps.

      The city   of  Butte, adjacent  to  the pit,  is  itself  a  source  of
fugitive dust resulting  from  many different activities.   Previous  studies
have  identified  paved and unpaved roads, fuel combustion, motor  vehicle
exhaust, construction,  burning,  and  wind erosion  from  cleared areas  as
major particulate sources.

3.2   INPUT DATA

      Two separate  types of  input data  are  required by  the  DDMB  model:
measured TSP  concentrations  at hi-vol  receptors,  and the x /Q values for
each source-receptor  pair.

      The 1978 measured  TSP  concentrations  at nine hi-vols in  the Butte
area  are   shown  in   Table   2.    A  uniform  annual  average  background
concentration of 20..0 y  g/m^  is assumed  to  characterize   the  airshed,  so
that  the  contribution  of local   sources  to  measured concentrations  is
determined  by .subtracting 20.0 from the measured concentrations.

                                  TABLE 2.

                   1978 BUTTE AMBIENT TSP CONCENTRATIONS
                   MEASURED                             MEASURED MINUS
                CONCENTRATION        BACKGROUND           BACKGROUND
HI-VOL
GREELEY
HILLCREST
YATES
ALPINE
RAW
MT. CON
HEBGEN
DR. CANTY
MONTANA POWER
(Ug/m3)
79
37
73
89
54
34
70
56
47
(y g/m )
20
20
20
20
20
20
20
20
20
(yg/m )
59
17
53
69
34
14
50
36
27
      The  x/Q  values  used in  this  study were determined  from previously
run  ISCLT  dispersion model  simulations  of  the  Butte  area.   Using  1978
                                  18-7

-------
00
00
                                                   TABLE 3
                                  X/Q VALUES FOR SOURCE-RECEPTOR COMBINATIONS

HI -VOL
GREELEY
HILLCREST
YATES
ALPINE
KAW
MT. CON
HEBGEN
DR. CANTY
MONTANA POWER

INPIT
0.4426
0.1190
0.4077
0.5391
0.1097
0.2098
0.2636
0.0792
0.0738
_ T<
CRUSHER
0.7797
0.1161
0.8189
0.4287
0.2464
0.1972
0.7001
0.1415
0.1298
?P FMT^TON 9
STORAGE
0.8525
0.1195
0.8092
0.4919
0.2083
0.2264
0.5294
0.1338
0.1145
OIIRrFS — 	
HAUL ROADS
0.4896
0.1390
0.4544
0.6426
0.0925
0.1412
0.2132
0.0668
0.0585

BACKFILL
0.1566
0.0523
0.1603
0.1204
0.1583
0.9977
0.5135
0.3044
0.1262

DUMPS
0.1351
0.1574
0.1283
0.1823
0.0352
0.0702
0.0603
0.0296
0.0263

CITY
0.6002
0.1850
0.5490
0.6781
0.6817
0.2560
0.6457
0.5517
0.4408
          NOTE:  x/Q values  in sec/m3 x 106

-------
meteorological  data,  the  ISCLT  model  computed  the  impact  of  seven
distinct sources:

      1.  Berkeley Pit
      2.  Crusher dump
      3.  Active storage
      4.  Anaconda haul roads and access roads
      5.  Backfill
      6.  Primary dumps
      7.  Butte (unpaved roads, fuel combustion, wind erosion, etc.)

The  first  six sources  are  controlled by  Anaconda,  while  the  seventh
source   represents   the   city  of  Butte.    The  source  grouping  option
available in the  ISCLT model was used  to  compute the  impact  of combined
sources  where  appropriate.  The  city  of Butte, for  example,  was modeled
as 134 discrete area sources.

      Values   of  x /Q,   determined   by  dividing   the  model-predicted
concentrations  by  the model-assumed  emission rates,  appear in  Table  3.
Each x/Q value  in  Table  3.  represents a measure  of  the  TSP dispersion
coefficient  from a source  to a hi-vol receptor.

3.3   DDMB SOLUTIONS

      Initially,  attempts were  made  to compute  TSP emission  rates  for
each of  the individual  seven major  sources in  Butte   (pit,  haul  roads,
crusher,  storage,   backfill,  dump,  and city of  Butte).    The  effective
emission  rates and  confidence  intervals resulting  from this effort  are
shown in Table 4.

                                  TABLE 4.

                     TSP EMISSION RATES AND CONFIDENCE
                      INTERVALS  FOR  UNCOMBINED SOURCES
                          EMISSION RATE       95% CONFIDENCE INTERVAL
  SOURCE	(gm/sec)	(gm/sec)
INPIT
HAUL ROADS
CRUSHER
STORAGE
BACKFILL
DUMP
TOWN
-176
47
-35
196
11
-6
52
+ 1,540
+ 237
+ 224
+ 1,230
-1- 139
+ 230
+ 41
Clearly, the  DDMB  method is unable to  resolve  each of the seven sources,
as indicated  by the physically  unrealistic  emission  rates,  and  the very
wide  confidence intervals  which suggest  that  there  is  a great  deal  of
uncertainty associated with each predicted emission rate.   An explanation
for such large  error bars is found in  the matrix of correlations between
emission rates, displayed in Table 5.

                                18-9

-------
                                  TABLE 5.

                             COFRELATION MATRIX
CRUSHER
INPIT -.68
CRUSHER
STORAGE
HAUL ROADS
BACKFILL
DUMPS
STORAGE
.40
-.92




HAUL ROADS
-.995
.71
-.45



BACKFILL
-.978
.66
-.40
.98


DUMPS
.11
-.06
-.08
-.17
-.15

CITY
-.30
-.10
.16
.27
.23
-.13
The  correlation  between  the  inpit  and haul road  sources  is very close  to
-1.0,  meaning  that  these  two  sources cannot  be  distinguished  by the
available  data.   In  essence, the  hi-vols  "see"  these  two  sources as the
same  source.   This  is not  surprising  since  the haul roads  are  located
within and  surround  the  Berkeley Pit.  Similarly,'the crusher and  storage
sources  are correlated,  as  are the  inpit  and  backfill  sources.   These
strong negative  correlations suggest  that  emission  rate accuracy could  be
greatly improved by combining sources.

      The   seven  sources  were  systematically   combined  and  a  matrix
solution achieved for  each combination, partly in  an attempt to find the
most  accurate  emission rates,  and partly  to  experiment with the source
combination  method.   Table   6.   illustrates   the   improvement  in  DDMB
performance as  the  souces are combined.   In  Case  "B"  the pit  and  haul
roads  are   combined  as if  they constituted  a single,  indistinguishable
source, and the  confidence interval  associated  with  the  combined source
is  smaller than for  either  individual  source.    Case  "C"  in  Table   6.
illustrates  the  effect  of  combining  the  crusher  and  storage  sources,
while  Case  "D"   further  combines the backfill  and  primary  dump sources.
In Case  "D" all  but  one  of  the physically unrealistic negative emission
rates  have  been  eliminated,  but confidence intervals  still remain large.
This  condition  is not remedied until  all  of the  Anaconda  sources are
combined   (Case   "E").   When  they  are,   confidence  intervals  shrink
dramatically,  demonstrating  the resolution sought  between  Anaconda and
the  city  of Butte,   but  at  the  expense of  lack  of  resolution  among the
Anaconda sources.

      Using the  emission rates  shown for  Case "E"  of Table  6.,  the TSP
contribution from the  Anaconda  sources  and from the  town  can be computed
at each  hi-vol   receptor.   The  results  of these  computations  appear   in
Table  7.   which   lists the  concentration  contribution  in  percent  for
Anaconda  sources,  for  the  city  of  Butte,  and   for  background.   The
Anaconda  contribution  is   generally  lower   than   that   predicted   by
conventional source  modeling  techniques,  and  is  in  good  agreement  with
recent Anaconda  estimates of  source contribution^5).
                                  18-10

-------
                                                 TABLE 6


                           TSP EMISSION RATES AND CONFIDENCE INTERVALS FOR VARIOUS

                                          CASES OF COMBINED SOURCES
oo
 i




SOURCE
IN. PIT
HAUL ROADS
CRUSHER
STORAGE
BACKFILL

DUMP
CITY
CASE "A"
NO SOURCES
COMBINED


-176 ±1540
47 ± 237
- 35 ± 224
CASE "B"
CASE "C"
CASE "D" CASE "E"
PIT & HAUL ROADS PIT & HAUL ROADS PIT & HAUL ROADS, ALL ANACONDA
COMBINED


61.8 t 64

25.2 - 82
196 t 1230 23.1 ± 97
11 t 139 -7.4 ± 14


CRUSHER & STORAGE CRUSHER & STORAGE, SOURCES COMBINED
COMBINED

53.0 ± 63

5.5 ± 31

-7.0 ± 15

- 6 ± 230 -2.1 ± 110 -3.4 - 121
52 ± 41 50.0 * 19 52.4 * 19
BACKFILL & DUMPS
COMBINED
57.4 ± 37

4.7 t 25


-21.1 ± 38


59.1 - 2?.. 9




52.4 t 16 48.3 ± 13.7
                      Note: All emission rates expressed In gm/sec

                            All confidence Intervals at 95% level

-------
                                  TABLE  7-

                 1978 TSP CONTRIBUTIONS  AT  SELECTED HI-VOLS
           ANACONDA CONTRIBUTION  BUTTE CONTRIBUTION  BACKGROUND CONTRIBUTION
   HI-VOL           (%)	(%)	(%)
GREELEY
HILLCREST
YATES
ALPINE
RAW
MT. CON
HEBGEN
DR. CANTY
MONTANA POWER
37.7
20.8
37.4
38.7
11.2
25.3
24.0
9.5
9.6
36.8
24.6
35.8
38.1
55.3
28.6
46.4
51.7
44.9
25.4
54.6
26.9
23.2
33.5
46.1
29.6
38.8
45.6
4.0   DISCUSSION

      The DDMB method  provides  a means  of  estimating emission  rates and
source  apportionment  from a  number of  simultaneously emitting  fugitive
dust  sources.   Analytically  the  DDMB  method finds  that combination  of
source emission rates  that  yields  the  best fit  (in  the  "least  squares"
sense)  between  measured and  modeled concentrations.   The method  enjoys
several  advantages  over  current  fugitive  dust  source  apportionment
techniques:

      o  DDMB  does  not  require   the   detailed   laboratory   work
         associated with  many receptor  models  such  as   CMS,  SEM,
         X-ray diffraction,  or optical microscopy.

      o  DDMB can discriminate among sources that emit  chemically
         and physically identical  particulate matter.

      o  Unlike  many  spatial  receptor  models such  as  trajectory
         analysis,   cluster  analysis,  and  pollution  wind  roses,
         DDMB   provides    quantitative    measures    of    source
         contribution.

      o  Depending  upon  the  software  package  used   to  solve  the
         linear  system  of  equation, the DDMB method can  "police
         itself"   by   providing   confidence  intervals   for   each
         computed emission rate.

      o  The  DDMB  method   is    an  inexpensive,  cost-effective
         analysis.
                                 18-12

-------
Naturally  there  are disadvantages,  too.   The number  of  distinct sources
that can  be  resolved is constrained  by  the number  of hi-vols available.
Furthermore,   the   improvement   in   confidence  intervals   achieved   by
combining  sources  is made at  the expense of  the  number of  sources that
can be resolved.   It may  be  that  the DDMB method will be most useful when
investigators are willing to partition all  sources  into just a few source
categories, as in the example application presented above.

      Will  the  DDMB method  always work?   Application of  the procedures
outlined in  this paper  will always yield emission  rate estimates as long
as  the  governing  equations are fully determined or  overdetermined.  But
there  is  no  guarantee  that the  answers  will  be  "correct,"  or  that  the
confidence  intervals  will  not  be  so  wide  as  to  make   the  answers
meaningless.   Based on this  study,  and  other ' applications of  the DDMB
method, it  seems  that there are  some criteria  that  make success  of  the
DDMB method more likely:   first,  there must be no large systematic errors
in  the dispersion model used to compute X/Q values.  Random errors in the
dispersion  model   are  tolerable,   and their  effect can  be  minimized  by
combining  sources.  Second,  if  the  measured  hi-vol  concentrations  are
appreciably   greater  than   the   background   concentration,   then  the
unavoidable  errors  inherent  in  the hi-vol  measurements  will  have  the
least impact  on computed  emission rates.   In other  words, the DDMB method
probably  works  best in  areas  where  large  fugitive dust  sources  are
present.   Finally,  if particulate sources are  spatially separated, and if
hi-vol  data  are   available  from  many  different  locations,  then  the
method's ability to discriminate  sources will be enhanced.
                                 REFERENCES
1.  Cooper, J.  A.  and J.  G.  Watson,  Jr.,  "Receptor Oriented  Methods  of
    Air Particulate Source Apportionment," JAPCA 30; 1,116 (1980).

2.  Watson, J.  G.  and  J.  C. Chow,  "An  Overview of Source-Receptor Source
    Apportionment,"    presented    at    APCA    meeting,    Philadelphia,
    Pennsylvania, June 21-26, 1981.

3.  Core,  J.  E.  and  T.  G.  Pace,  "Receptor  Models—How  Great  Thou
    ArtJ(?),"  presented  at  APCA  meeting,  Philadelphia,  Pennsylvania,
    June 21-26, 1981.

A.  Yocom, J.  E., E.  T.  Brookman,  R.  C.  Westman, and  0. P.  Ambardar,
    "Determining  the  Contributions  of  Traditional  and  Nontraditional
    Sources of Particulate Matter," JAPCA 31: 17 (1981).

5.  Moldovan, G.  J.  and  P-  A. Doughty,  "The   Butte, Montana  Particulate
    Apportionment  Analysis,"  presented  at APCA   meeting,  New  Orleans,
    Louisiana, June 20-25, 1982.
                               18-13

-------
                                                                       *
                                                                       **
        Modeling the Emission of Aerosols in and Around a Metallurgical
                                 Plant

        B.  Vanderborght, I.  Mertens, and J. G. Kretzschmar, SCK/CEN,
        Belgium; F.  Adams,  UIA, Belgium;  and R. Dams,  INW, Belgium

                              ABSTRACT

Dust emissions from a metallurgical plant can cause considerable nuisance and
health hazard to the environment.  The cost to benefit analysis of pollution
abatement investments can only be performed by proper modeling of the  emis-
sion-immission situation. The problem was  thoroughly examined in a case
study by daily antimony immission measurements around a metallurgical plant
over a period of one year. It soon appeared that low  level fugitive emissions
and not the high stacks were the major sources of the high immission levels.
These sources cannot be  monitored continuously so that direct emission  data for
dispersion model calculations are not available. For every process in the plant
an emission factor was obtained by means of  a combination of tracer releases,
immission measurements, reversed modeling for the fugitive emissions and in
stack measurements for chimney emissions.  Emission factors, inventory of
the plants'  production and meteorological observations formed the input data
for bi-Gaussian dispersion modeling. In the  calculations, the contributions
from the fugitive and point source emission can be discerned in the immission.
Calculated values fitted the measurements of suspended Sb-aerosols very well.
Validation is done by means of scatter diagrams,  the  comparison of measured
and calculated cumulative frequency distributions,  time series and pollution
roses. Not only the global statistics but also  individual measuring periods show
very good agreement. Although the global Sb-dust deposition can be well des-
cribed by wet and dry deposition  parameters, and source depletion, some indi-
vidual measuring sites show systematic deviations from the calculations.  It
seems that local turbulence conditions were responsible.

(This is an abstract of a presentation for which a paper is not available.)
(*) Nuclear Energy Research Center, B-2400 Mol, Belgium.
(**) Institute for Nuclear Sciences, B-9000 Gent,  Belgium.
                                19-1

-------
                The work described in this paper was not funded by the U.S. Environmental
                Protection Agency. The contents do not necessarily reflect the views of the
                Agency and no official endorsement should be inferred.
    NEW YORK STATE INDUSTRIAL COAL  PILE  DRAINAGE REGULATIONS AND GUIDELINES
             By Carol  Hornibrook, Senior Project Manager,  NYSERDA
                    2 Rockefeller Plaza,  Albany, NY  12223

REGULATIONS
Existing  waste water  discharge  standards  applicable  for New  York  Coal  Pile
Drainage  are  found in  the  State Pollution  Discharge  Elimination  System  Permit
(SPDES).  At  this  time regulations applicable to  the industrial sector  control
only the  total dissolved solids  at  .5  mg/liter/day.  This  unlike  utility  coal
pile  drainage  which,  in  addition  to   total  dissolved  solids,   requires  the
regulation of  total suspended solids and pH.

GUIDELINES
In  1980  the New  York  State  Department  of  Environmental  Conservation  (NYSDEC)
published  a document  entitled  "Coal Pile Guidance  for SPDES  (State  Pollution
Discharge Elimination  System)"  which lists  a  number  of requirements for  new and
existing  coal  piles.   These  requirements  call  for  the monitoring of  possible
toxic  pollutants  emitted from  coal piles and  the  treatment facilities  required
to  prevent  contamination of ground water by  leachate and the  contamination  of
surface waters by runoff.  Specifically, the  New  York State guidelines  require
natural and/or manmade  liners  under new and  some existing  coal  piles,  ground
water  monitoring  systems,  storm  collection  ponds   to  contain  runoff  for  pH
neutralization and heavy metal  removal.

At  this  point I  would  like   to   clarify  that  a  10  year-24  hour storm  is  a
statistical  term  referring to  a  certain  size storm  event (expressed  in  inches
of  rain)  occurring over a 24  hour  period, which has  a recurrence interval of 10
years.

The specific  limitations for each  of the requirements previously  mentioned are
listed  in Table  1.    The  reader  should  note  that  these  guidelines   are  being
implemented   in   the  same  fashion  as  regulations  and  specific   discharge
limitations are being  set in  new and renewed SPDES permits.
                                  20-1

-------
                                    TABLE 1

                         COAL PILE GUIDANCE FOR SPDES



To  comply  with  surface  and  ground   water   classifications,  standards  and

limitations,  (Title  6,  Official Compilation  of Codes,  Rules  and  Regulations,
Part  701  and  703)  for  the  protection  of  the   State's  water  resources  and

compliance   with   promulgated   federal   best  management  practices    (BMPs)
regulations  (40 CFR,  Part  125)  designed  for  the  protection  of said  resources,

the following requirements  will apply.


I.  New Coal Piles


    A.     Site Considerations


           The  control of  wastewater  containing  runoff  and  leachate from new

           coal piles shall be as follows:


           1)   The site  shall be designed  to prevent  runoff from entering  the
               pile.

           2)   The  coal  pile shall  be  placed  on  an  impervious  surface.   This
               surface  may be  a  liner  or  any in-place  material  which  meets
               Department  criteria  for  impermeability.

           3)   Liner  material  shall be  selected so  as not to  deteriorate after
               contact  with  coal pile  wastewater and shall  be  protected  from
               puncture   from  operating  equipment.    Liner   protection   and
               repairability  shall  be evaluated in the liner  design.

           4)   Wastewater  shall be  collected  and treated.

           5)   Collection  ponds,  if  used, shall  be  lined  with an  imperivous
               material, and  designed  to contain  the runoff from  the ten-year,
               24-hour storm.

           6)   A  ground water  monitoring  system  may  be  required to  evaluate
               and/or monitor  the integrity of the facility's impervious  lining.
                                     20-2

-------
          7)    Discharge  from treatment facilities to surface waters are  not  to
               contravene  Part   701,   (see  Table   2)   'Classifications   and
               Standards  of Quality and Purity" and discharge to  the ground  are
               to  meet Part  703,  (see  Table  3)  "Ground Water  Classifications
               Quality Standards and Effluent  Standards  and/or  Limitation".

          8)    Coal  pile  wastewater may be:

               a)   Treated separately;

               b)   Recycled for other uses at  the  facility;

               c)   Combined with other  waste streams  for  treatment except  for
                   the purposed of dilution.
    B.     Permit Conditions (SPDES)

          1)    The  permittee  shall  prepare  an  engineering  report  and  final
               plans  for  construction  of  facilitites  designed  to  prevent,
               control, and treat  runoff and leachate  from  coal piles  for  the
               purposes of  maintaining and protecting  water  quality  standards
               and compliance  with any applicable federal requirements.

          2)    Organic and  inorganic  substances will  be limited  and  require
               treatment,  depending   on  the  coal  pile  characteristics.   Most
               heavy  metals  may  be   found  in coal  pile  wastewater.   Organic
               substances  which  have  been  detected  in  coal  pile  wastewater
               include chlorobenzene,  dichlorethylene,  chloroform,  pthalates,
               and  methylene  chloride.  Monitoring  and treatment  requirements
               for  these  compounds   will   be  evaluated  on  a  case  and  site
               specific basis because wastewater characteristics are  dependent
               on the characteristics of the coal used  and  discharge  objectives
               relating to the receiving water.
II.  Existing Coal Piles

    A.     Site Considerations

          Existing  coal  piles  will  be  evaluated  on  a  case-by-case  basis
          utilizing  new  coal  pile  considerations.   The   permittee   may  be
          required  to  undertake  a  ground  and/or  surface  water  monitoring
          program  to  determine  the extent  of  possible   contamination.   As  a
          minimum, ground water monitoring and treatment  of  contaminated runoff
          will be  required.  Ground water  monitoring results will  be  evaluated
          to determine applicable new coal  pile considerations.

                                     20-3

-------
   B.    Permit Conditions (SPDES)

         Specific permit conditions may  include  the  following:

         1)   Permittee to  submit a proposed plan  of study designed  to assess
              coal  pile operations  at  the  applicant's  facility,  including  a
              study  program  to evaluate the operation's effect on  surface  and
              ground water quality.

              The  study program  would be subject  to  Department  approval  and
              shall  consist  of a  ground and surface water monitoring  program
              to determine  compliance  with applicable water quality standards,
              and monitoring  of  coal pile runoff and leachate  to determine and
              assess wastewater  characteristics.

          2)   Based  upon NYSDEC  review of the study  program,  the permittee may
              be required  to  prepare an engineering  report  and final plans foe
              construction  of  facilities  designed  to  prevent,  control,  and
              treat  runoff and  leachate from  coal  piles  for  the  purposes of
              maintaining   and   protecting   water   quality   standards   and
              compliance with any applicable federal requirements.


III.   Imposition of  limitations and monitoring  requirements and/or any partic-
      ular  policy guidelines  if deemed  appropriate for the protection of  the
      environment may precede  any of the above  considerations for  either  new oc
      existing coal  piles.


Taken from:   Bureau of Industrial Program
              Division of  Water

Dates:    April  14,  1980
It becomes immediately evident  that  compliance  with these regulations,  from the
engineering design  study  to placement of  the liner material and monitoring  of
its  effectiveness,  can be  fairly  costly.   NYSERDA  sponsored  the  R&D  project
presented  today  with  specific  intent  of better understanding the  hydrology  of
coal  piles through extensive  data collection  and verifying  and calibrating  a
coal  pile  drainage  model,  in order  to potentially decrease  industries  cost  of
compliance with these guidelines.

I hope that the information presented  today will prove useful in future efforts
to comply with these guidelines and to obtain and  renew SPDES permits.


                                  20-4

-------
              Calibration and Verification of a Coal Pile Drainage Model

                        John G. Holsapple, New York Power Pool
                        Gordon T. Brookman, John A. Rlpp and
                        Pamela B. Katz, TRC-Environmental
                        Consultants, Inc.

                                      Abstract


     In 1980  the New York State Department of Environmental Conservation  (NYSDEC)
published a document entitled "Coal Pile Guidance for SPDES" which lists  a number
of requirements for  new and existing coal piles.  These requirements emphasize the
monitoring of possible toxic pollutants emitted from coal piles and the treatment
facilities required  to control leachate contamination of ground water and  runoff
contamination of surface waters.  Specifically, the New York State regulations
require natural and/or manmade liners under new and some existing coal piles, ground
water monitoring,  storm collection ponds to contain runoff from the ten-year 24-
hour storm, and treatment facilities for pH and most heavy metals.

     Because  of these new requirements, New York State utilities must plan for the
most efficient coal  pile drainage treatment and avoid the unneeded expense of
constructing  new coal pile runoff storage capacity to collect  drainage that does
not require treatment.

     This paper will present a summary of a research study being performed by
TRC-Environmental  Consultants for ESEERCO (Empire State Electric Energy Research
Corporation)  which has the following objectives:

     1)  The  calibration and verification of a mathematical
         model to  simulate coal pile drainage flows and pollutant
         loadings  under historical and design storm conditions.
         The  model will be used to size and design treatment
         systems to  meet NYSDEC specifications.  The model will
         also simulate the changes in coal pile runoff due to
         changes in  coal source and coal characteristics.

     2)  The  development of a data bank on coal pile drainage
         from two  plants which will be useful in characterizing
         which pollutant parameters appear in significant
         quantities.


     The presentation will include a detailed description of the sampling programs
at the two plants  and an overview of the mathematical model Including its uses and
its limitations.
              The work described in this paper was not funded by the U.S. Environmental
              Protection Agency. The contents do not necessarily reflect the views of the
               Agency and no official endorsement should be inferred.
                                      21-1

-------
INTRODUCTION



     The United States Environmental Protection Agency (EPA) presently  requires




that area runoff from coal storage piles at steam electric plants  be  controlled




to prevent the discharge of acidic water and total suspended solids  (TSS).




Accepted practice dictates that storage/treatment be designed to accommodate




all of the runoff from a 10-year, 24-hour storm.  Also, TSS must not  exceed




50 mg/1 and the discharge pH must fall within the range 6.0 to 9.0.




     Most state water  pollution regulatory agencies have adopted the  EPA limi-




tations for permitting purposes.  However, some states, including  New York, are




requiring more stringent controls for coal pile runoff.  The New York State




Department of Environmental Conservation's (NYSDEC) "Coal Pile Guidance for




SPDES" requires natural and/or man-made liners under new and some existing coal




piles, as well as groundwater monitoring, and 10-year, 24-hour storm  runoff collection




ponds and treatment facilities for pH, solids and several trace metals.  New York




utilities must submit an engineering report with their SPDES permit applications.




The report must contain sufficient field data to show what pollutants are being




emitted in existing coal pile drainage and the means to predict quantitative and




qualitative characteristics for new coal piles.  Surprisingly, there  is little




information on the characteristics of coal pile runoff.  There is  an  incentive




and a need, then, in New York State to develop a technique to accurately charac-



terize coal pile drainage.




     In June of 1981, TRC Environmental Consultants, Inc. initiated a research




project for the Empire State Electric Energy Research Corporation  (ESEERCO)




involving the calibration and verification of a coal pile drainage model.  The
                                21-2

-------
project is also partially funded by the New York State Energy Research and




Development Authority  (NYSERDA).  The objectives of this study include 1)




field testing and verification  of a coal pile drainage model and 2) the develop-




ment of a large data base on coal pile runoff and coal pile characteristics.




The scope of work, involves sampling at two sites.  Sampling at Site 1 was com-




pleted in the fall of  1981 at the Greenidge Station of New York State Electric




and Gas Corporation.   Site 2 sampling is presently being performed at the




Russell Station of Rochester Gas and Electric Corporation.




     Upon completion of this project in late summer/early fall of 1982, TB.C




will provide ESEERCO with a calibrated and verified model to simulate coal




pile drainage flows, pollutant  loadings under historical and design storm




conditions, and a model user's  manual.  The model will be a useful tool for




sizing and designing coal pile  drainage treatment systems.  It will also be




useful for simulating  changes in coal pile runoff due to changes in the coal




source and in coal characteristics.




     This paper presents an overall of the project to date, including the site




selection process, a detailed discussion on the sampling program, a description




of the model and the calibration and verification process.






SITE SELECTION




     In 1981, there were nine major utility plants burning coal in New York




State.  As part of the site selection process, TRC visited each plant and




developed a subjective ranking  system, which included the following factors:




     - coal pile size




     - percentage of sulfur in  coal




     - coal pile containment




     - ease of sampling on pile




     - variability of pile shape and size




     - climate




     - availability of data         21-3

-------
     The Russell Station (Site 2) was the highest ranked plant for  the  following




reasons:



     - The plant has an average size of 96,000 tons, which is near  the  average




       size of the piles evaluated;



     - The average sulfur content is 2.4 percent, which was one of  the  highest




       of all plants evaluated;




     - The plant has a partial concrete-lined gutter around the pile which




       collects most of the runoff;



     - Approximately 25 percent of the pile is reserve, which facilitates the




       installation of in-pile  piezometers;




     - The pile size and shape remains fairly constant throughout the year.




     The Greenidge Station (Site 1) ranked second for the following reasons:




     - The plant has an average size of 110,000 tons, which approximates the




       survey average of 112,000 tons.




     - Runoff and leachate are completely contained by an impervious plastic




       liner under the pile.   All runoff and leachate are directed  to a single



       discharge point.




     - Approximately 50 percent of the pile is reserve and, therefore,  facili-




       tates the placing of in-pile monitoring equipment.




     - Size and shape of the pile does not vary significantly during the year.




     •These two sites were  then recommended to ESEERCO as test sites and permission



was received from the member utilities to proceed.
                                     21-4

-------
FIELD MONITORING PROGRAM


     The field monitoring program was designed co provide input data to calibrate


and verify the drainage model and to collect an extensive data base on pile


hydrology and chemistry along with site meteorology and runoff quantity and


quality.  The scope of work for the field program was based on a TRC report to


EPA and EEI entitled                           "Planning Study to Model and


Monitor Coal Pile Runoff."1



DATA REQUIREMENTS


     Figure 1 illustrates the categories of data necessary to the modeling


effort and the data base.  Site description data includes general information


on pile size and shape, proximate analysis of delivered coal, methods of stacking


and reclaiming and pertinent existing data on site geohydrology.   Meteorological-


data being collected includes historic information from local U.S. Weather Service


stations and continuous monitoring of precipitation, air temperature, relative

                                                                      i
humidity, solar radiation and pan  evaporation during the survey period.  Runoff


data being measured includes continuous recording of all drainage flows from the


coal pile and when flow does exist continuous measurements of pH and conductivity.


For each storm event samples are collected on a 30-minute interval a't the start


of each storm and after 4 hours of runoff on a 60-minute interval for the fol-


lowing analysis:


     - Acidity/alkalinity


     - Total suspended solids


     - Total dissolved solids


     - Total organic carbon
 G. T. Brookman, at al., "Planning Study to Model and Monitor Coal Pile
 Runoff," EFA-600/7-81-016 (NTIS PB81-152530), February 1981.
                                      21-5

-------
                                                                SITE DESCRIPTION UAIA
                       GFNERAL  PI ANT
                        INFORMATION
 COAL PILE(S)
DATA (PHYSICAL)
    COAL(S)
DATA (CHEMICAL)
 GROUND  MATER
      DATA
                                                   II
           METEOIIOLOGICAL DATA
to
i—j

0>

1
PRECIPIIATION

HISTORICAL DATA
(3 MONTHS PRIOR
TO FIEIO PROGRAM)



1 II
AIR
flMI'ERAIURE

SOIAR
RADIATION
EVAPORATION




1
PRECIPITATION



FIELD DATA
(DURING FIELD
PROGRAM)


1
I
IEMPU'MUR.

III 1 RUNOFF DAI A









1
1 1
SOl Alt 1
j i.nn f WAD All AT 1 fUJ 1
HAD 1 AT 1 ON Vnr UHA 1 1 Uff 1
*

                                      DRY DAYS  (INCLUDE STORM
                                     EVENTS WITH 
-------
     - Sulfate




     - Hardness




     - Total and dissolved  forms of  20 metals




Not illustrated in Figure 1 is  the geohydrological data being collected at each




site to define pile moisture conditions.   At the  start of  each site monitoring,




a drilling program is conducted on the pile.  During  that  program, hollow stem




auger borings are made with split spoon  sampling  through the pile at 5-foot




intervals.   Coal samples are then analyzed for moisture content.  Where pos-




sible, in-pile piezometers  are  installed to observe water  level fluctuations at




the pile base.  Simultaneous with the drilling program a resistivity survey is




conducted on the pile to map the saturated areas  of the pile.  Throughout the




10 weeks of runoff monitoring,  piezometers are observed for water levels, pH,




conductivity and periodically sampled for the runoff  parameters.  Pile perme-




ability and infiltration rates  are also  measured  throughout each monitoring




season to correlate with runoff percentages.






LOGISTICAL CONSIDERATIONS




     As mentioned above,  2  sites are being monitored for 10 weeks each.  The




field program is built around a portable laboratory in which solids and acidity




analysis are performed on all runoff samples.  Also this laboratory serves as




of base of operations for calibrating instruments, preserving and storing samples




and cleaning equipment and  glassware.  An instrument  shelter is also located on




each site to house recorders for flow, pfl and conductivity, along with the auto-




matic sampling equipment.   The  samplers  are programmed to  start collecting




samples at the start of each storm through an actuator switch which senses an




increase in runoff flow.  This  insures the complete sampling of each storm event.




A technical support person  is kept on-site during the entire monitoring period




to maintain all equipment and perform on-site analysis.
                                      21-7

-------
RESULTS TO DATE

     The monitoring program at Site 1 has been completed and a  total  of  4

runoff events were measured.  Storms ranging in size from 0.19  inches to 1.8

inches of rain were measured.  A series of hyetograph/hydrographs  is   plotted

for the 4 storms and drainage coefficients were determined.  As  illustrated in

Figures 2 to 6, drainage coefficients never went higher than 39  percent.   This

indicates that at least 60 percent or more of the rain hitting  this pile was

absorbed by the coal and lost to either evaporation, pyrite oxidation or base

flow seepage.  Two observations made during the monitoring at this site  lend

to the fact that the pile is acting like a large sponge.  First, the  pile  was

continuously growing with new coal from 38,000 tons to 120,000  tons and  second,

the moisture of the pre-fired coal was significantly greater than  that of  the

delivered coal.  These measured drainage coefficients are substantially  lower than the

reported in literature, i.e., 70 to 80 percent.

     Most of the chemical analysis data has not been finalized  to  date,  but a

number of obvious observations can be made concerning the quality of  the runoff.

Briefly they include:

     1.  The pH of both storm runoff and base flow seepage did  not vary.   All

         pH's were in the 2.0 to 2.4 range.

     2.  The conductivity of the runoff was depressed during storm events, as

         low as 4,000 micromhos per centimeter (umho/cm) and during base flow

         seepage remained at a higher concentration, 8,000to 9,000 umho/cm.

     3.  Conductivities correlated well with dissolved solids.

     4.  Because of the low pH, most of the metals found in the runoff were in

         a dissolved form rather than suspended.

     5.  Analysis of metals data during storm events can indicate  the interflow

         hydrograph.

The completed results of the data collected at Site 1 will be included in  the

final report to ESEERCO.
                                    21-8

-------
to
I—1
CO
o
P   0.1

£L Z 0.2
          is
     10

     a

     6

     4

     2
                   TOTAL PRECIP: 0.19 IN.
                   DRAINAGE COEFFICIENT:  0.06
                   SITE: G STATION
                   DATE: SEPT. 8, 1981
                 i—i—r
                 10   12
                    NOON
                             HYETOGRAPH/HYDROGRAPH
                                   STORM EVENT 1
                                        FIGURE 2

-------
Z
O
     0.1
t ~i
Q. Z 0.2
O ~
g   0.3
Q.
     60
     40
     30
     20
     10 -
                                        I
TOTAL  PRECIP:  0.61 IN.
DRAINAGE COEFFICIENT: 0.06
SITE: Q STATION
DATE: SEPT.  14, 1082
                                 10    12
                                      NOON
                          HYETOQRAPH/HYDROGRAPH
                                STORM EVENT 2
                                     FIGURE 3
                                                        10

-------
DO
t—i
        z
        o
        t
        D.

        O
        UJ
        oc
        o.
         *•: o.i
       g
            70
            60
            60
40-
            30
            20 -
            10 -
      TOTAL PRECIP: O.08IN.

      DRAINAGE COEFFICIENT: 0.12

      SITE: Q STATION

      DATE: SEPT. 21, 1081
                     6
              0
i  i
  10
i  i   i  i   i
  4    6
10
                                MIDNIGHT

                               HYETOGRAPH/HYDROGRAPH

                                     STORM EVENT 3
i  i
  12
NOON
T
                                         FIGURIL

-------
N3
I—i
I
                   z
                   o
                       0.1
                   CL Z   0.2
                   O ~

                   UJ    o o
                   tE    O.3

                   O.
                       30O
                  g£
                  n!2
                       200
                       100
TOTAL PRECIP: 1.6 IN.

DRAINAGE COEFFICIENT: 0.29


SITE: G  STATION


DATE: OCT. 27, 1982
                                 1  4
                    2
    •   1  i  1  *  .1  i  i
       6    a    10    12

                   NOON


HYETOGRAPH/  HYDROGRAPH

STORM EVENT 4 - 1st PHASE
                                             FIGURE 5

-------
             o
             UJ
             DC
             Q.
to
(—i

t—i
CO
                6OO
                50O
            £$400
            ^& VM
            "• ^-
                3OO




                2OO




                10O
12
                         T  I  I
                         24
6
8
I   I  I   I   I  I
  1O   MID-   2

       NITE
                                  HYETOGRAPH/HYDROGRAPH
                                      STORM 4 2nd  PHASE
                                             FIGURE 6

-------
MODEL DEVELOPMENT



     A 1976 T5.C study of nonpoint source pollution at a utility  site  in  Penn-




sylvania attempted to address sheet washoff from coal storage piles using  the




SWMM/RECEIV II model.  This model was eventually thought deficient in the  sim-




ulation of coal pile runoff because it did not take into account:




     1.  Erosion of coal from the pile surface,




     2.  Percolation of storm-water through the pile,




     3.  Pyrite oxidation/acid production in the coal pile.




     In 1979, a follow-up study was conducted by TRC to determine what physical




and chemical phenomena associated with coal pile runoff needed to be  addressed




so that the model could be used as a prediction tool.  A number  of phenomena




were identified as characteristic of runoff and should, therefore, be included




in the model.  They are:




     1.  Precipitation in the form of rain and snow,




     2.  Surface runoff and.infiltration through the pile,




     3.  Moisture content of the pile,




     4.  Snowmelt from piles subject to heavy winter accumulations,




     5.  Groundwater infiltration through the base of the pile,




     6.  Pyrite oxidation and acid production,




     7-  Freeze/thaw cycles,




     8.  Gully erosion along the sides of the pile,




     9.  Washoff of coal from the pile surface.




A comprehensive search of existing models found that the Ohio State University




(OSU) version of the Stanford Watershed Model was a good basis for a  coal  pile




runoff model.  With extensive modification of the OSU model, the new  model




treats the coal storage pile as a small unvegetated watershed with many  features




similar to larger watersheds.  A TRC coal pile runoff model was  then  developed




and consists of two major components:  the hydraulic and the qualitative.
                                    21-14

-------
BASIC FUNCTION OF HYDRAULIC COMPONENT




     A diagram of the hydrologic c-ycle of a coal pile is shown, in Figure 7.




The model reads hourly precipitation data in the form of rain or snow input on




a meteorological data tape or cards.  A continuous water balance simulation in




the model uses precipitation, pile moisture content and pile water storage as




additions and surface runoff, interflow and base seepage as reductions.  The




hydraulic component of the model also predicts pile erosion caused by the gullying




of the pile slopes during intense storms.




     Hydraulic data are presently being measured in New York State at a utility




coal pile for the calibration of the hydraulic model.






HIGHLIGHTS OF A WORKING HYDRAULIC MODEL




     Once the hydraulic model has been calibrated and verified using data from




several coal pile monitoring programs, it will become a working design tool.




Input ^111 consists of two data sets:  1) site specific data such as volume and




acreage of the coal, average moisture content of the delivered coal and pile




compaction data and 2) meteorological data either in the form of historical




weather data vailable from the National Climatic Center (NCC) or local on-site




measurements taken by the owner of the coal pile.  Site-specific data are




entered into the model by card deck while the historical weather data are




entered on magnetic tape.




     The working model lists daily, monthly and annual runoff flows plus the




optional output of selected precipitation events with hourly or subhourly data.




     Optional output from the hydraulic model are also available.  The output




are described below.






           BASIC OUTPUT                             OPTIONAL OUTPUT




1.  Table of average daily rainfall.       1.  All input data.
                                     21-15

-------
                                                                             UPPER ZONE STORAGE -
                                                                              DEPRESSION  STORAGE
to
                  RUNOFF ^
                  STREAM
                                                           DEPRESSION
                                                             STORAGE
                                                           LOWER ZONE
                                                            STORAGE
                                         -S-^WATER TARIE
                                                    •	.

                                                     GROUND WATER
RUNOFF
STREAM
                                                                  TO DEEP  STORAGE
                                        HYDROLOGICAL CYCLE OF  COAL STORAGE  PILE
                                                       FIGURE 7

-------
           BASIC OUTPUT

2.  Summation of daily runoff rates
    for each month.
3.  Monthly and annual totals of
    runoff, seepage and total flow.

4.  Monthly and annual total of
    frozen precipitation.

5.  End-of-month values of pile,
    surface and ground-water moisture.

6.  End-of-month values for snowpack.

7.  Annual balance of unaccounted
    moisture.
         OPTIONAL OUTPUT

2.  Echo of recorded runoff (when
    measured)  as a comparison with
    simulated flows.

3.  Daily values of snowpack, snow-
    melt and snowfall.

4.  Average daily temperatures.
5.  Number of freeze/thaw cycles
6.  Simulated gully erosion

7.  Plots the hyetograph and hydro-
    graph for selected storms.
BASIC FUNCTION OF THE QUALITATIVE COMPONENT OF THE MODEL

     Stored coal exposed to the atmosphere undergoes the physical and chemical

process of pyrite oxidation.  Products of these reactions are acid, iron and

sulfate.  The acid further reacts with the coal by dissolving trace metals and

elements.  During wet weather, these materials are washed off the surface and

moved through the interior of the coal pile.  Runoff during storm events and

seepage from the pile base during wet and dry weather carry these products of the

pyrite oxidation to produce a Ughly acidic wastewater.  The qualitative component

of the model simulates these reactions; using the water balance produced in the

hydraulic model, it will predict the loading of acid, sulfates and trace elements

in the runoff.

     The model looks at both dry weather reactions and the transport of the

products of oxidation during wet weather.  For dry weather, the model simulates

the total pollutants in the coal and those in a dissolved state available for

transport.  For wet weather periods, the model simulates the distribution of a

component of the dissolved metals, sulfates and acid from the surface of the

pile to the interior and directly to surface runoff.  In addition, the  removal
                                      21-17

-------
of materials from the interior of the pile by dry weather seepage  is  simulated.




The results of the qualitative model will be a calculation of daily,  monthly  and




totals of acid, metals and sulfates.






HIGHLIGHTS OF A WORKING QUALITATIVE MODEL



     Once the qualitative component of the model has been calibrated  and  veri-




fied, it will be available for use.  Input to this model are from  tvo  sources:




     1.  A magnetic tape created by the hydraulic model containing data on




         rainfall, transport routes, flow rates from seepage and runoff and




         average daily air temperature.




     2.  Card input describing factors of pyrite oxidation including:  percent




         of pyrite in the coal, density of the coal pile, amount of trace metals




         in the coal and the solubility of sulfate and iron compounds.




     As data from the ongoing field program are evaluated, the oxidation  reaction




rates will be adjusted.  Output of the working qualitative model will  be:




     1.  A calendar for any selected year showing dry versus wet days,




     2.  A calendar listing daily runoff flow volumes,




     3.  Daily total loadings of acid, sulfate, iron and trace metals,




     4.  Optional  plots on a detailed storm basis showing the relationships




         of precipitation, flows and pollutant loadings.






CALIBRATION EFFORT TO DATE




     The calibration effort being completed has found that the model  can  easily




simulate the water balance of large storm but usually overpredicts the runoff




volume from short intense storms.  In addition, gully erosion, which  is very




sensitive to the intensity of precipitation, is difficult to simulate using




precipitation amounts averaged over an hourly period.
                                     21-18

-------
SUMMARY




     The ultimate purpose of simulating coal pile drainage is for use in the




design of systems to treat acid drainage before discharge.  Industries storing




coal may base treatment system design on short-term monitoring programs; however,




that data may not have any relationship to average or worst-case runoff condi-




tions.  The runoff model will allow an industry to simulate several runoff




conditions and thereby have a better basis for design.




     The current coal pile runoff program in New York State has two objects:




     1.  To provide detailed information on meteorological, coal storage, pile




         moisture, runoff quality and hydrologic conditions to the developing




         model so that key coefficients may be adjusted;




     2.  To provide approximately 10 weeks of continuous  data from each of two




          coal pile.monitoring sites as a data bank useful for comparative




         studies and treatment design.




     One important issue to be resolved during this program is whether a runoff




condition occurs producing a discharge "clean" enough to  bypass treatment.




Runoff data collected at two utility sites in Pennsylvania in 1976 indicates




that a first flush did indeed occur.  However, initial  data collected to date




showed that runoff acidity and sulfate were very high during all phases of the




storm and during dry weather seepage.  The analysis of  storms of greater volume




may help resolve this first flush issue.




     The field data collection programs are to be completed in May of 1982 with




the runoff model completed in September of 1982.
                                     21-19

-------
                                            *« U'5' Environmental
           Agency and n0 oici                      i-et "" "™ oi 
-------
                              INTRODUCTION


Various  approaches for characterizing  coal pile leachate  and  runoff have
been used  by  other researchers;  these  approaches have included sprinkling
pans of  coal  with a rainfall simulator  (7), recirculating distilled water
over an inclined box packed with  sample (4),  and collecting  actual
leachate and  runoff on site (2).   As the methodologies  and  coals used  in
different  tests varied, so did  the results.

One general conclusion of a study (5)  concerning the treatability of coal
pile leachate and runoff was that heavy metals and other trace elements  in
leachate were direct functions  of acidity.  As acidity  increased,  so did
the concentration of heavy metals and  trace elements in coal pile leachate
and runoff.   Verification of this conclusion was an important  objective  of
the study  reported herein.

Generally,  it would be  preferable  to  collect  samples  from  actual
facilities in order to assess the characteristics of runoff and leachate.
However, because  the coal pile  under consideration did  not  yet exist,  it
was simulated by  the  test methodology described in this  paper.   These
efforts were designed to  represent  actual  conditions  as  closely  as
possible,  including actual dimensions  and design density of the full-scale
coal pile.

This study was conducted  for a  facility converting to  coal  as a primary
fuel source.   An  on-site coal storage  pile was therefore required.  A  coal
pile simulation test was conducted to  determine  the quality  of leachate
and runoff resultant from rainfall contacting  the facility's coal pile.
Testing  was  conducted  on four  types of coal being considered as fuel
sources:  Coals 1, 2, 3, and 4.   The simulated leachate  and  runoff were
compared to applicable effluent  criteria.

                            TEST METHODOLOGY

A coal  pile  simulation  apparatus  was built to  the same scale  as  the
proposed coal pile:   10  ft (3  m) in height, 14-ft (4-m)  sides sloped
45 degrees  to the horizontal.   The  test apparatus,  shown  in Fig.  1,
consisted of vertical columns  and  open-top,  sloped  troughs.   Only
polyvinyl  chloride (PVC) pipe,  rubber  hoses, fiberglass  netting, plastic
duct tape, nylon  cord, and wood  were used to prevent metal contamination,
since heavy metals often are the  primary concern in coal  pile runoff and
leachate.   Ten-ft  (3-ra)  high vertical  columns,   constructed   of
4-in. (102-mm) diameter  PVC pipe  and packed  with coal  at  the  design
density, were intended to simulate rainfall  passing through the entire
10-ft (3-m)   depth of  the coal  pile.  Fourteen-ft  (4-m)  long troughs,
constructed of 10-in. (254-mm)  diameter PVC  pipe cut in  half and packed
with coal  at  the  design density of 55 pounds  per cubic  foot (Ib/cu ft)
(881 kg/m-3),  were designed to simulate  the side slopes of the coal
pile. Thus,  the  troughs  simulated runoff, the columns leachate.
                                 22-2

-------
                                 SOURCE: ENVIRONMENTAL SCIENCE AND ENGINEERING, INC., 1980
Rgure 1
COAL PILE SIMULATION APPARATUS
                                              Coal Pile Simulation Study
                             22-3

-------
Simulated  runoff was drained by two hoses  at  the  end  of  the trough.   The
troughs were  under-drained at a coal depth of 3 to 3.5  in. (76 to  89  nun)
to remove  the leachate, as depicted  in  Fig.  2.  If  the troughs had  not
been under-drained, water which infiltrated into  the  coal  pile would  have
continued  contacting the coal as it drained along the bottom of the trough
and contributed to  the volume of runoff.   The under-drains also allowed
separation of the volume of water which  would infiltrate  into a coal  pile
from the volume of  water defined as runoff.  These trough  leachate samples
were collected and  analyzed to provide additional data  describing  runoff
characteristics.

Prior to testing, the coal samples were  passed  through a  2.25-in. (57-mm)
sieve to ensure that the tested coal met  the size specifications  of  the
actual coal to be stored in the pile.  The coal was then  packed into  both
the PVC columns and the PVC troughs at the full-scale  design density  and
secured with  fiberglass netting, plastic  duct tape, and nylon cord.

Small amounts of certain inorganic salts  were added to raw ground water to
approximate the characteristics of  site-specific rainfall.  An average
site-specific rainfall  event of 0.66  in. (17  mm)  was selected  as  the
rainfall volume to  be applied for the simulation. The synthetic rainfall
was distributed over the PVC troughs via  a sprinkling system consisting of
plastic lawn  sprinkler  hoses.  The  majority of the  fiberglass netting
securing the  coal was removed after initial wetting had  occurred and  the
coal had settled into the trough.  Simulated runoff  and trough leachate
were collected in separate glass containers,  as shown in  Fig. 3.  Several
rainfall gauges placed under the apparatus were used  to measure the amount
of  applied synthetic rain  water falling on  the  coal  pile  apparatus.
Sprinkling was continued until the  average reading  of  these gauges  was
0.66 in. (17  mm)-

To generate simulated leachate from the  vertical  PVC  columns, a measured
volume of  synthetic rain water was poured  on the coal  at the top  of  the
columns until leachate appeared at the bottom.   All  resulting simulated
leachate  was collected  in  a glass  container.   Details of  leachate
generation are shown in Fig. 4.

Each coal  type produced  three streams:   runoff, trough leachate,  and
column  leachate.   Each  sample was  analyzed for  various parameters
addressed  in  applicable effluent criteria.  All  analyses were performed
according to United States  Environmental Protection Agency (EPA)
Method  60014-79-020 and  Standard Methods  for  Analysis of Water  and
Wastewater (1).

                                RESULTS

Table 1 presents the analytical results  of the  coal pile  simulation.   The
applicable effluent criteria are also presented  for comparison.

No consistent difference in concentration  was observed  between simulated
runoff  and simulated  leachate from the four  coal types in  the  test
apparatus.  The rainfall event applied to  the troughs  and the amount of
synthetic  rain water applied to the columns to  obtain  leachate resulted in
                                 22-4

-------
                            FRONT VIEW
       RUBBER
       DRAIN HOSES
                                                           COAL

                                                           PVC PIPE
         PERFORATED
         PVC DRAIN
                                SIDE VIEW
                        LEACHATE
                        COLLECTION'
RUNOFF
COLLECTION
                              SOURCE: ENVIRONMENTAL SCIENCE AND ENGINEERING, INC., 1980.
Figure 2

DETAILS OF TROUGH
                                          Coal Pile Simulation Study
                          22-5

-------
                                 SOURCE: ENVIRONMENTAL SCIENCE AND ENGINEERING, INC., 1980.
Figure 3
TROUGH UNDERDRAINS AND
RUNOFF COLLECTION
Coal Pile Simulation Study
                              22-6

-------
                                 SOURCE: ENVIRONMENTAL SCIENCE AND ENGINEERING, INC., 1980.
Figure 4
COLUMN LEACHATE GENERATION
AND COLLECTION
Coal Pile Simulation Study
                            22-7

-------
Table 1.  Results of Goal Pile Simitation Test and Applicable Effluent
            Standards (all values ng/L except pH)
Parameter
(1)
PH
fj Arsenic
to
' Total
CO
Chromium
lead
Mercury
Chlorides
Copper
Iron
Zinc
Effluent
Standard
(2)
6.0-8.5
0.05
1.0
0.05
H/D
500.0
0.5
0.3
1.0
Synthetic
Rain Hater
(3)
6.6
H/D
H/D
0.002
H/D
81.0
H/D
H/D
0.188
Coal Type 1
TO
(4)
6.3
H/D
0.003
H/D
H/D
116.0
H/D
H/D
0.016
CL
(5)
6.4
0.013
0.004
0.002
H/D
99.0
H/D
0.204
0.016
TL
(6)
6.5
0.01
0.004
0.002
H/D
124.0
H/D
0.204
0.024
SAMPLE nENTIFlGATICN
Coal Type 2 Coal Type 3
TO
(7)
6.1
H/D
0.004
0.012
H/D
109.0
H/D
H/D
0.013
a
(8)
6.2
H/D
0.002
H/D
H/D
92.0
H/D
H/D
0.01
TL
(9)
6.5
H/D
0.002
0.002
H/D
106.0
H/D
H/D
0.016
TO
(10)
6.1
H/D
0.002
H/D
H/D
103.0
H/D
H/D
0.013
a
(11)
6.2
H/D
H/D
H/D
H/D
88.0
H/D
H/D
0.013
TL
(12)
6.5
H/D
0.001
0.002
H/D
114.0
H/D
H/D
0.013
Coal Type 4
TO
(13)
2.7
0.221
0.068
0.132
H/D
109.0
0.687
361.0
4.52
CL
(14)
2.8
0.114
0.048
0.018
H/D
90.0
0.395
190.0
6.04
TL
(15)
2.8
0.131
0.040
0.406
H/D
96.0
0.469
214.0
4.93
TO  • Trough simulated runoff
CL  * Colum simulated leachate
TL  " Trough simjlated leachate
N/D - Hone detectable

NOTE:  1 mg/L - 1 ppra.

Source:  ESE. 1980.

-------
approximately the  same  solid-to-liquid ratio.  -This  ratio was 5.6:1  for
the troughs and  4.5:1 to 5.8:1 for the columns.

Samples of simulated  leachate and runoff from Coal 4  in the test apparatus
exceeded the criteria for  several  parameters.  Moreover, the simulated
leachate and runoff  resulting from this coal were highly  acidic, with  pH
values between 2.7 and  2.8.

                        SUMMARY AND CONCLUSIONS

Coal pile simulation tests were performed on four coal  types  in order  to
predict leachate and runoff quality.  Samples generated from the testing
were compared to applicable effluent criteria.  One  of  the  four coal types
(Coal Type  4) tested resulted  in  simulated runoff and leachate  which
consistently violated several criteria.  The  other  three coal leachates
met all effluent criteria.

The  testing agreed  with  previous  research, showing  that  runoff  and
leachate with low pH values could be expected to have high  concentrations
of heavy metals.  The testing  also demonstrated a  procedure  for  use  in
evaluating alternative coal supplies by  assessing  their  pollution
potential.
                                   22-9

-------
     CONTROL OF ACID PROBLEMS IN DRAINAGE

              FROM COAL STORAGE  PILES
                           by
Harvey Olem, Tracey L. Bell,  and Jeffrey J.  Longaker

               Office of Natural Resources
               Tennessee Valley Authority
                 Chattanooga,  TN 37401
 The work described in this paper was not funded by the U.S. Environmental
 Protection Agency. The contents do not necessarily reflect the views of the
 Agency and no official endorsement should be inferred.
                        23-1

-------
Abstract



     A method has been identified for controlling acid production and sub-



sequent dissolution of toxic pollutants in drainage from coal storage



piles.  Results of laboratory and field experiments indicate that it may



be possible to prevent, rather than treat, acid drainage by applying an



environmentally safe detergent formulation periodically to the coal.  A



mild solution of sodium lauryl sulfate (SLS) was found to effectively



block the activity of the bacteria that promote acid formation and chemi-



cal leaching.  Drainage from coal treated once with 50 mg/L of SLS remained



neutral for 60 days, about three times longer than the untreated control



sample.  Extrapolating results to an industrial-scale application revealed



that the cost of the SLS needed for a single application might be as high



as $500 per hectare of coal storage area  ($200 per acre).





Introduction



     As more and more utilities and industries throughout the United



States shift from the consumption of oil  and natural gas to the more



abundant supplies of coal, there should be an increase in onsite stor-



age of the coal.  It has been projected that by 1985 the amount of coal



in storage in the United States will increase from its current level of



128 million metric tons (1978 estimate) to 227 million metric tons.  By



the year 2000, the total coal stockpiled  could reach 680 million metric



tons  (1).  This projected increase in coal storage has generated renewed



interest in the environmental effects of  contaminated drainage from coal



storage facilities.




                            23-2

-------
     The most efficient method currently used for storing large quanti-



ties of coal is placement on the ground.  Because the coal is exposed to



the elements, rainwater falling on the pile can become contaminated by



the action of chemolithotrophic bacteria on pyritic materials (usually



iron disulfides).  This occurs by the same series of reactions that are



known to produce acid drainage from coal mines.  Consequently, pollution



of both surface and ground water is usually the major problem associated



with storage of this coal.



     Drainage from coal storage areas generally contains high concentra-



tions of metallic ions (aluminum, iron, manganese, mercury, nickel, zinc,



etc.), sulfates, and suspended solids.  It also usually has an extremely



low pH (pH 2 to 3).  Currently this drainage is typically collected in



a central pond and treated, usually by addition of hydrated lime.



     This paper presents results of laboratory experiments which indicate



that it may be possible to prevent, rather than treat, acid drainage by



applying an environmentally safe detergent formulation periodically to



the coal.  Laboratory experiments consisted of applying two anionic deter-



gents to columns filled with coal which had been rinsed with water until



the pH of drainage was neutral.  This was done to simulate freshly mined



coal or more nearly, coal that had been processed by coal washing.  A



third column was filled with coal and did not  receive any detergents as



a control.  Due to the limited number of tests performed to date, the




results are considered preliminary.






Background



     One of the most common and most  troublesome  impurities  in  coal  is  a



metallic sulfide, iron pyrite.  When  mining exposes  pyrite  (FeS2)  to water



and oxygen, a  chain of chemical  reactions begins.





                             23-3

-------
2FeS2 + 702 + 2H20 = 2Fe2+ + 4S042" + 4H+                         (D



4Fe2+ + 02 + 4H+ = 4Fe3+ + 2H20                                   (2)



Fe3+ + 3H20 = Fe(OH)3 + 3H+                                       (3)



           3+              2+       2"      +
     FeS2 + l4Fe+ + 8H20 = 15Fe   + 2S04" + 16H                     (4)


                                                                 2+
     In reaction 1 pyrite is oxidized to produce ferrous iron (Fe  ) , sul-



fate (SO,2"),  and some acid (H+) .   The ferrous iron can then be oxidized


                                                                   3+
with this acid and some oxygen (reaction 2) to form ferric ions (Fe  ).



Under certain conditions, the ferric irons will hydrolyze (reaction 3)



to form the yellow-orange ferric hydroxide precipitate (Fe(OH),j.) that



typically appears in acid mine drainage streams.  Ferric iron can also



react with more pyrite (reaction 4) to form more ferrous iron and acidity.



These reactions produce acid at a constant rate.



     Acid drainage would develop very slowly if this strictly chemical



chain of events were not greatly accelerated by the biochemical action



of certain bacteria.  First isolated in 1947, Thiobacillus ferrooxidans



were then believed to play some role in acid mine drainage formation, but



their exact role was not well understood (1).  More recently, Singer and



Stumm (2) showed that direct oxidation of pyrite by oxygen (reaction 1)



was too slow to generate the amount of acidity observed in nature and con-



cluded that the oxidation of ferrous iron to ferric (reaction 2) created



a cycle whereby the ferric ions produced in this reaction directly attack



pyrite to form considerable amounts of acid (reaction 4).  Reaction 2,



however, is extremely slow chemically under acid conditions.  The half



time for spontaneous ferrous iron oxidation in a sterile solution of



pH 3.5 has been estimated to be 2,000 days (3).  In the presence of



T. ferrooxidans, however, this  reaction was shown to accelerate by  a
                               23-4

-------
factor of 10 , thus producing significant quantities of ferric iron.

The subsequent oxidation of pyrite by these ferric ions produces more

ferrous ions to allow the cycle to continue, producing more and more

acidity until all the reactive pyrite is leached.


Bacterial Inhibition

     It is generally considered preferable to prevent the formation of

pollutants through at-source control rather than provide continuous treat-

ment of the resulting contaminated water.  It is with this objective in

mind that bacterial inhibition of acid formation was investigated to pre-

vent the contamination of rainfall-runoff draining through coal piles.

Anionic detergents, while normally considered as cleansers rather than

bactericides, do possess bactericidal properties.  Dugan (4) found that

anionic detergents can effectively stop  iron oxidation by T. ferrooxidans

at concentrations as low as 2 mg/L.  Kleinmann  (5,6) demonstrated the
           •*
effectiveness of one one type of detergent, sodium lauryl sulfate, in

reducing acid formation in  coal refuse piles and surface coal mines at

somewhat higher detergent concentrations.  The  reason for this inhibition

is not yet clear, but evidence indicates that the semipermeable properties

of the cytoplasmic membrane is altered so that  H  is allowed to seep  into

the .normally  neutral interior of the cell (7).  Thus, the bacteria are

appropriately attacked by the acid  they  helped  produce.

      In  this  study,  two  types of anionic detergents were evaluated, sodium

lauryl sulfate  (SLS) and neutralized benzene  sulfonate  (NBS).  The  common

use  for  each  detergent  is  shown  in  Table 1.



Materials  and Methods

      Three  acrylic plastic  columns  were  set up  in the  laboratory  to simu-

 late a  coal  pile  (Figure  1).  The  columns were  approximately 2 meters high


                               23-5

-------
               Table  1.   Detergents Tested to Prevent
                          Coal Pile Drainage Contamination
   Detergent
    Common Use
Percent Active  Ingredient
    in Product  Tested
Sodium Lauryl
  Sulfate

Neutralized
  Benzene
  Sulfonate
Active ingredient in
many hair shampoos

Active ingredient in
many laundry detergents
           30
           15
                                         PLASTIC
                                          CAPS-
                     s COLLECT!-.
                       BEAKERS
                              * COLUMN C
                                CUT AT
                              MIOSECTION
             Figure 1.  Laboratory columns used to simulate
                        coal  pile drainage.
                               23-6

-------
and 30 cm in diameter.  A 1.5-cm diameter drain valve was located in the



bottom of each column to collect leachate.  The underdrain system in the



columns consisted of a 20-cm depth of thoroughly washed pea gravel under



25 cm of washed sand.  To prevent the gravel from clogging the drain, a



capped 10-cm diameter, 13-cm long perforated plastic column was glued



over each drain hold.  Compressed air was passed through.water and fed



through an "L" shaped glass tube to simulate atmospheric conditions within



a coal pile.  The tube was located at mid-height in the sand layer.  The



sand also acted as a diffuser to distribute the air across the coal.



     Coal was obtained from the Tennessee Valley Authority (TVA) Cumberland



Power Plant located near Nashville, Tennessee.  Small samples of the coal



were collected from different locations and depths around the coal pile.



The samples were then composited.  The composite was rinsed with water



until the pH of drainage was neutral in order to simulate freshly mined



coal or coal that had been processed by coal washing.  A portion of the



washed coal was analyzed for total sulfur, moisture, volatile matter, ash,



fixed carbon, and Btu content.



     Approximately 0.057 cubic meters of  the coal was divided equally



between the three columns.  The procedure of the three coal samples was



identical except for the different solutions applied.



     Each coal sample was initially placed in a 5-gallon bucket and satu-



rated with the necessary solution for one hour.  The excess water was



poured off and the resultant slurry poured into the appropriate column.



In the first column, coal was treated with deionized water only.  Coal  in



another column was treated with a 50 rag/L solution of SLS and deionized



water.  A 50 mg/L solution of NBS and deionized water was added to  coal



in the remaining column.




                               23-7

-------
     The columns were dosed twice a week with 1,500 ml of deionized water
to simulate a 2.5-cm (1-inch)  rainfall contacting a coal pile.  Leachate
was then collected from the drain valves at the bottom of the columns for
analysis.
     Acidity, pH, conductivity, iron, manganese, and the most probable
number (MPN) of iron-oxidizing bacteria were routinely analyzed on the
collected samples.  Selected trace metals and detergent concentrations
were also analyzed periodically during the experiment.  The deionized
water used in the experiment was routinely analyzed for pH.  All chemical
analyses were performed according to procedures outlined in "Standard
Methods" (8).  Enumeration of iron-oxidizing bacteria was estimated by
employing the inorganic salts culture medium and the combination multiple
tube and microtitre plate procedure described by Olem and Unz (9).

Results and Discussion
     Coal Characteristics—Characteristics of the  coal used in these
experiments are  shown in Table 2.  The total sulfur content was typical
for  coal burned  at most TVA power plants.  Coal with these characteristics
has  been found  to produce  strongly acidic drainage exposed to the  elements.
     Simulated  Rainwater—The  deionized water used to simulate rainfall
had  a pH of  approximately  5.5.   The  acidity present in  the deionized  water
was  due  to  contact with atmospheric  carbon dioxide.
     Leachate Characteristics—The quality of  leachate  from each  coal col-
umn  three  days  after initial  application of detergents  is  presented  in
Table 3.   As  expected, the drainage  was  initially  pH  7.2  for  all  three
columns.   Most  other water chemistry characteristics  were  similar for
the  drainage from all three columns,  although  certain parameters  varied
widely.  After  remaining  nearly  neutral  for  about  20  days,  the  leachate

                              23-8

-------
         Table 2.   Analysis  of Coal Used in Laboratory
                   Simulation of Coal Pile Drainage
Parameter                    Units                           Value
Total Moisture
Volatile Matter
Ash
Fixed Carbon
Total Sulfur
Energy Content
%
% Dry Basis
% Dry Basis
% Dry Basis
% Dry Basis
Btu/lb Dry Basis
3.8
39.8
9 f\
.8
r* f\ 1
50.4
3.1
13,162
                           23-9

-------
                   Table 3.  Quality of Leachate from Laboratory Coal Columns
                             Three Days After Initial Application of Detergents
Parameter
Concentration In Leachate (MgA unless indicated)
Control          SLS Treated          NBS Treated
pH (S.U.)
Acidity (mg/1 CaCO )
Conductivity (jJmhos/cm)
Fe
Mn
Pb
Hg
As
Cd
Cu
Se
Cr
Ni
Zn
7.2
0.6
480
1,020
96
4
0.7
1
0.2
<1
1
2
<50
<5
7.2
0.3
210
299
22
<1
0.3
<1
<0.1
<1
<1
320
<50
<5
7.2
0.3
600
1,680
100
11
<0.2
1
<0. 1
4
2
1
<50
<5

-------
suddenly became ten-thousandfold more acid in the control column, as pH



decreased from 7 to about 3  (Figure!;.  The pH continued to decline for



another 20 days until it leveled off at about pH 2.




     Drainage from the test  column treated with NBS became slightly more



alkaline at first but remained approximately neutral only about five days



longer than the control.  Then it paralleled the same pattern of rapidly



increasing acidity, reaching pH 2 after about 50 days.



     Drainage from the coal  treated with SLS became even slightly more



alkaline for the first 20 days.  After a slight decrease in pH, it



remained neutral for two and one-half times longer than the control.



Even when the pH finally dropped after about 60 days, it hovered for



another 25 days between pH 4 and 5.  The acidity at this pH was approxi-



mately 100 mg/L, compared to 5,000 mg/L for the control column.  This



difference could be extremely important in terms of the treatment- that



would be needed to neutralize it.



     The coal treated with SLS eventually produced drainage with the same



low pH of about 2, but only  after 140 days, three and one-half times longer



than the 40 days for the untreated control.



     Other constituents generally followed the same pattern as the changes



in pH.  The quality of leachate from each column 56 days after initial



detergent" application is presented in Table 4.  At this point in the



experiment, only drainage from the SLS-treated coal was neutral.  Acidity



in drainage was over 100 times higher in the control and NBS-treated coal.



Significantly lower concentrations were also observed in drainage from



the SLS-treated coal for conductivity, Fe, Mn, As, Cd, Cu, Cr, Ni, and



Zn.  Only Pb and Se were higher in drainage from SLS-treated coal compared



to the control.  However, these concentrations were almost at the detec-




tion limits for these constituents.



                              23-11

-------
          8
                                     o CONTROL
                                     •SLS  TREATED
                                    A NBS  TREATED
   PH
to
CO
>—'
N5
                      20
40         60         80       100       120

  DAYS  AFTER  TREATMENT
140
                                   Figure 2.  Changes in pH in laboratory-generated
                                             coal pile drainage following application
                                             of 50 mg/L of sodium lauryl sulfate  (SLS)
                                             and neutralized benzene sulfonate (NBS).

-------
                            Table 4.   Quality of Leachate from Laboratory Coal Columns
                                      Fifty Six Days After Initial Application of Detergents
to
CO
Parameter
pH (S.U.)
Acidity (mg/1 CaCO )
Conductivity (jjmhos/cm)
Fe
Mn
Pb
Hg
As
Cd
Cu
Se
Cr
Ni
Zn
Concentration
Control
2.0
5,000
6,000
1,560,000
48,900
<1
0.3
200
310
1,050
<1
295
1,500
13,700
In Leachate (pg/1
SLS Treated
6.9
10
1,950
365
1,780
2
<0.2
1
2.3
7
4
<1
<50
<5
unless indicated)
NBS Treated
2.1
5,500
5,500
1,004,000
74,500
7
<0.2
220
340
1,160
1
2
3,300
16,700

-------
     Drainage from both treated columns had detergent concentrations



below detection limits (<0.1 mg/L measured as methylene blue active sub-



stance) .   This is important because these detergents are toxic to aquatic



life.  For example, Dailela et al. (10) found the LC5Q value for a cer-



tain species of fish exposed to SLS to be 11.2 rag/L.  Apparently, SLS was



not washed much out of the coal.   Most of it was probably adsorbed onto



the coal particles and remained there until its chemical structure was



broken down into simple degradation products such as carbon dioxide, water



and sodium and sulfate ions.



     Kleinmann and Erickson (6) performed a detergent adsorption capacity



test on refuse from a coal-cleaning plant and determined that one gram



of refuse adsorbed between 50 and 68 |jg of SLS.  These values were an



order of magnitude higher than results for similar tests performed on



surface-mined overburden material.  The same test was performed on the



coal used in the laboratory column experiments.  Adsorption capacity was



found to be 60 jjg SLS/g coal.  Because of the lower adsorption capacity



of overburden materials, Kleinmann (6) incorporated SLS into rubber



pellets which gradually release the detergent into infiltrating rainwater.



This method may not be needed for coal storage and coal refuse disposal



areas.




     The "NBS and SLS detergent applications apparently did not kill all



the iron-oxidizing bacteria, because small numbers of live, but appar-



ently inactive, bacteria were present even in samples collected after' the



coal was treated.  For example, bacterial numbers in the drainage 28  days



after treatment were 160,000 per  100 mL for the control column and  7,900



per 100 mL for the SLS-treated column.  Toward the end of the experiment,



when drainage from the treated and control columns were acid, bacterial





                             23-14

-------
numbers were more similar.  One hundred nineteen days after treatment,



drainage from the control and SLS-treated columns contained 1,100,000



and 920,000 iron-oxidizing bacteria per 100 ml.



     The reasons for the relative effectiveness of SLS treatment compared



to NBS are not known.  If detergents alter the semipermeable properties



of the cytoplasmic membrane and allow H+ to seep into the interior of



the cell, as has been reported, then it is possible that a detergent



that better attacks protein would be more effective.  SLS, being the



active ingredient in many hair shampoos, is designed to attack protein-



type buildup on hair.  On the other hand, NBS is the active ingredient  in



many laundry detergents and is designed to remove soil particles, not pro-



teinaccous matter.  This may be one. reason for the relative ineffectiveness



of NBS.



     Projected Costs—Roughly extrapolating results to an industrial-scale



application, the cost of SLS needed for a single application was estimated



at about $500 per hectare of coal storage area ($200 per acre).   This



figure assumes a delivered cost of $1.65 per Kg  ($0.75 per Ib) for the



SLS used in the experiments and allows for a 50-percent decrease in the



efficiency of detergent application in the field.  Because of the



limited number of experiments performed to date, these costs must be



considered preliminary.



     Methods of Application—Results of laboratory experiments showed



that SLS may be effective in preventing problems of acidity and dissolved



metals in coal pile drainage for about two months.  If an operating



facility stored less than a 2-month supply of  coal and always burned the



oldest coal first, a single application  to each  new coal  delivery might



prevent the development of acid drainage.  However, there are probably



few plants where these conditions would be met at  all times.  In the




                             23-15

-------
tests performed thus far,  it was assumed that it would be necessary to



treat all the coal from the top to the bottom of an entire pile each time



it is required.  There are some indications, however, that bacteria are



active mainly in the top layer of coal, where the temperature and avail-



ability of air best suit their needs.   If that is the case, it might be



necessary to treat only the surface of the pile, or it might be possible



to extend significantly the time between full-pile treatments by means



of supplementary surface layer treatment.



     It might even be economical to treat each new coal delivery before



it is added to the pile, either as a supplementary treatment or as the



only treatment required.  At many plants, the surface of the coal pile



is routinely sprinkled with water to control dust, and a small amount



of SLS might easily be added to the tanks of the sprinkler trucks.  A



spray system might be installed to treat every truckload or trainload



of coal with a small amount of SLS, either as it is dumped or as it moves



along conveyors to the pile so that the surface layer of new coal would



continually be protected against acid formation.  If significant time



is expected to elapse between mining and delivery, the coal might even



be protected with SLS as soon as it is mined and then treated again when



it is delivered.




     Additional Research Needs—Since drainage from existing coal piles is



often already  acid, there is a question as to whether it would be possible



to correct an  existing problem by application of detergents.  A modifica-



tion of  the methods described in these studies would likely be necessary



because  SLS breaks down rapidly in strong acid solutions.  It may be possi-



ble to correct an existing problem by  treating the coal with SLS  in combi-



nation with some buffering agent.  Further experiments using several buffers





                                23-16

-------
such as sodium bicarbinate and phosphate are currently in progress to



address this concern.




     It is not known whether SLS affects the burning characteristics of



coal.  This must be determined because laboratory results indicated that



SLS was readily adsorbed onto the coal particles.  Another potential con-



cern, the added sulfur due to the SLS, should not present a. problem.  The



sulfur content theoretically added to the coal due to the sulfate ions in



SLS is insignificant relative to the high-sulfur content of coal.





Conclusions



 1.  Controlled application of anionic detergents may be useful.in pre-



     venting acid drainage from coal storage piles.



 2.  Sodium lauryl sulfate was more effective than NBS in preventing acid



     production in laboratory-simulated coal piles.



 3.  Based on laboratory studies, it appears that a single application



     may last up to 60 days.



 4.  Extrapolating laboratory results to an industrial-scale application,



     the cost of SLS needed for a single treatment was calculated to be



     $500 per hectare of coal storage area ($200 per acre).






Acknowledgements



     We thank Robert L. P. Kleinmann and Patricia M. Erickson for per-



forming adsorption tests and Donald J. Rucker for critical review of the




manuscript.






Literature Cited



 (1)  Colmer, A. R.; Hinkle, M. E., Science, 1947,  106. 253.



 (2)  Singer, P. C.; Stumm, W., Science. 1970, 167, 1121.





                               23-17

-------
 (3)  Singer, P. C.; Stumm, W.,  "Oxygenation of Ferrous  Iron:   The
     Rate-determining Step in the Formation of Acid Mine  Drainage."
     Federal Water Pollution Control Administration, Report  14010,
     1969.

 (4)  Dugan, R. P., Ohio J. Sci., 1975, 75_, 266.

 (5)  Kleinmann, R.L.P., In "Proceedings, Symposium on Surface  Mining
     Hydrology, Sedimentology,  and Reclamation1!; Graves,  D.  H.,  Ed.;
     Report No. UKY BU123, University of Kentucky, Lexington,  1980;
     pp  333-337.

 (6)  Kleinmann, R.L.P., Erickson, P. M., In "Proceedings,  Symposium
     on  Surface Mining Hydrology, Sedimentology, and Reclamation";
     Graves, D. H., Ed.; Report No. UKY BU126, University of Kentucky,
     Lexington, 1981; pp 325-330.

 (7)  Hugo, W.  B., J. Bacteriol., 1967, 30, 17.

 (8)  "Standard Methods for the  Examination of Water and Wastewater."
     15th Ed.; American Public  Health Association, Washington, B.C.,
     1980.

 (9)  Olem, H.; Unz, R. F., Biotechnol. Bioeng.,  1977, l£,  1475.

(10)  Dailela,  R. C., et 4!., Water, Air and Soil Poll., 1981,  15_ 1.
63588/6650-B/6-82/214

                              23-18

-------
                        ATTENDEES
Antisavage, Joseph P.
Betz Laboratories, Inc.
4636 Somerton Rd
Trevose, PA 19047
215/355-3300 x377

Armstrong, James A.
Denver Research Institute
University of Denver
Denver,  CO  80208
303/753-2892
Bohn,  Russel
Environmental Services
 and Technology
11 E. 69th Ter
Kansas City, MO  64113

Boudreau,  Richard M.
Central Illinois Light Co.
300 Liberty St
Peoria, IL  61602
309/672-5471

Bowman, W. Alan
Applied Meteorology, Inc.
Suite 326
9000 Southwest Frwy
Houston, TX 77074
713/777-0106
Bramson, Mark
Syntech Products Corp.
520 E. Woodruff
Toledo, OH 43624
419/241-1215

Brass, Dwight
George A.  Rolfes  Co.
Box 458
Boone, IA  50036
515/432-3300
Brookman, Edward T.
TRC Environmental
 Consultants, Inc.
800 Connecticut Blvd
E. Hartford, CT  06108
203/289-8631
Campbell, Ivor E.
Smelter Control
 Research Associates
150 E. Broad St
Columbus, OH  43215
Chalmers, Jerry
S. C. Dept of Health and
 Environmental Control
2600 Bull St
Columbia, SC 29201
803/758-5406
Cole, Clifford F.
TRC Environmental
 Consultants, Inc.
8775 E.  Orchard Rd,
Suite 816
Englewood,  CO 80231
303/779-4940
Connolly, Michael
Minnesota Pollution
 Control Agency
1935  W.  County Rd B-Z
Roseville, MN  55118

Courtney,  Peter
Law  Engineering
2749 Delk Rd
Marietta, GA  30067
Cowherd, Chatten
Midwest Research Institute
425 Volker Blvd
Kansas  City, MO 64110
816/753-7600 x531

Cox, H. B.
U.S. Steel Research
(U. S. Mining, Inc.)
B St
Pittsburgh,  PA  15235
412/372-1212 x4291

Craig, Alfred B. (MD-54)
U.S. EPA/AEERL (lERL)
Research Triangle Park, NC 27711
919/541-2824
                                  A-l

-------
Attendees (cont.)
Crockett,  E. P.
American Petroleum Institute
2101 L St,  N.W.
Washington, DC  20036
202/457-7084

Dorsey, James A. (MD-62B)
U.S. EPA/AEERL (IERL)
Research  Triangle Park, NC 27711
919/541-2509
Dubrowski, John J.
Exxon Research and Engineering
Florham Park, NJ 07932
201/765-2101

Fletcher,  James K.
Kennecott Minerals Co.
PO Box 11248
Salt Lake  City, UT  84147
801/322-8262
Gatz, Donald F.
Illinois  State Water Survey
PO Box 5050,  Station A
Champaign, IL 61820
217/333-2512

Goebel, Gerald R.
DNR and EP/Div of Air
 Pollution Control
Ft. Boone Plaza,
18 Reilly Rd
Frankfort, KY 40601
502/564-3382
Hague,  William
Julius Koch USA,  Inc.
PO Box A-995
New Bedford,  MA 02741
617/995-9565

Harris, D. Bruce (MD-54)
U.S. EPA/AEERL  (IERL)
Research Triangle Park, NC 27711
919/541-7807
Harrison, Paul R.
Envirosol
1700 N.  Fiske
Pasadena, CA  91104
818/797-9581
 Hartshorn,  Wayne
 Sonic Development Corp.
 305 Island Rd
 Mahwah,  NJ 07430
 201/825-3030
 Holton, Greg (ORNL)
 First Environment
 314 W.  Broadway
 Lenoir City, TN 37771
 Hyatt, James
 Tennessee Valley Authority
 940 CSTZ-C
 Chattanooga, TN 37401
 Hyatt, John G.
 Exxon Co.,  USA
 Colony Shale Oil Project
 PO Box 440342
 Aurora, CO 80044
 303/695-2213
 Hyde,  Raymond G.
 Atlantic Research Corp.
 5390 Cherokee Ave
 Alexandria. VA  22314
"703/642-4193
 Ives, Jim
 Anaconda Minerals Co.
 PO Box 5300
 Denver, CO 80217
 303/575-7504
 Kaiser, Robert A.
 Ohio Edison Co.
 76 S. Main  St
 Akron, OH  44308
 216/384-5770
 Kawaters, Woody
 TRC Environmental
  Consultants,  Inc.
 800 Connecticut Blvd
 E. Hartford, CT 06108
 Kretch, Frank
 Sonic Development Corp.
 305 Island Rd
 Mahwah,  NJ 07430
 201/885-3030
                                  A-2

-------
Attendees (cont.)

Kuykendal, William B.  (MD-15)
U.S. EPA/OAQPS/AQMD (IERL)
Research Triangle Park, NC  27711
919/541-5372
Larson,  Alan G.  (TRC)
The Kent-wood Moore Co.
5690 Denver Tech Ctr Blvd
Englewood,  CO  80111
303/773-3399
Lawrence, Robert
KPN International,  Inc.
19 Pebble Rd
Newtown, CT 06470

Letizia,  Anthony P.
Envirosphere Co.
2 World Trade Center
New York, NY 10048
212/839-1088
Lewis, Ted P.
Iowa Electric Light and
 Power Co.
PO Box 351
Cedar Rapids,  LA 52406
319/398-4601

Lubas, Thomas
Port Authority of NY and NJ
1 World Trade  Center,
64 Floor E.
New York, NY 10048

Manning,  James
US EPA Region 4
345 Courtland St, N. E.
Atlanta,  GA  30365
404/347-3286
Martin, Dennis
TRC Environmental
 Consultants, Inc.
800 Connecticut Blvd
E.  Hartford, CT 06108

Mathai, C. V.
Arizona Public Service Co.
(AeroVironment,  Inc.)
PO Box 53999 (Sta. 5680)
Phoenix, AZ  85072-399S
602/371-6467
        Milhous, Madison N.
        Long Island Lighting Co.
        1775 E. Old Country Rd
        Hicksville,  NY  11801
        516/733-4385
        Mohr, Ray
        Colorado Health Dept
        1101 Bellaire St
        Denver,  CO 80220
        303/320-4180
        Moldovan,  John
        Anaconda Minerals Co.
        5555 17th St
        Denver,  CO 80433
        303/575-4312
        Mulloy,  Frederick W.
        Phillips  Petroleum Co.
        7 D 3 Phillips Bldg
        Bartlesville, OK  74004
        Nicholson,  Brock M.  (MD-15)
        U.S.  EPA/OAQPS/AQMD (CPDD)
        Research Triangle Park, NC 27711
        919/541-5517
        Nilsen, Glennyce C.
        United Illuminating Co.
        80  Temple St
        New Haven,  CT 06506

        Olem, Harvey
        Tennessee Valley Authority
        245 401 Bldg
        Chattanooga, TN  37401
        615/751-7338
        Pace,  Thompson G. (MD-15)
        U.S.  EPA/OAQPS/AQMD
        Research Triangle Park,  NC  27711
        919/541-5634
        Plaks, Norman (MD-61)
        U.S. EPA/AEERL (IERL)
        Research Triangle Park,  NC  27711
        919/541-3084
        Porter,  Richard  A.
        NAI
        1710 Firman Dr
        Richardson, TX 75081
        214/644-1616
A-3

-------
Attendees (cont.)
Rakes, Samuel L.  (MD-62)
U.S. EPA/AEERL (IERL)
Research Triangle Park, NC
919/541-2828
      27711
Ray, B.  Michael
Northrop Environmental Training
PO Box 12313
Research Triangle Park, NC  27709
919/549-0652
Reddy,  Tupili S.
Tennessee Air Pollution Control Div
150 - 9th Ave, N.
Nashville,  TN  37203
615/741-3651
Ripp, John A.
TRC Environmental
 Consultants, Inc.
800 Connecticut Blvd
E.  Hartford, CT 06108
203/289-8631
Romaine, Demarest (Dave)
Consolidated Edison Co. of
 New York,  Inc.
4 Irving PI
New York, NY 10003
212/460-4600
Rosbury,  Keith D.
PEDCo Environmental,
2420 Pershing Rd
Kansas City, MO  64108
Inc.
Rovell-Rixx, David C.
National Council for Air and
 Stream Improvement. Inc.
PO Box 11483
Gainesville,  FL  36204

Sando, Dick
Sonic Development Corp.
305 Island Rd
Mahwah,  NJ 07430
201/825-3030
Schumacher, Aileen
Environmental Science and
 Engineering
PO Box ESE
Gainesville, FL  32602
904/372-3318
Sehmel, George  A.
Pacific Northwest Laboratory
24000 Stevens 1100 Area
Richland,  WA 99352
509/375-6161
Shrock, John
Illinois Environmental
 Protection Agency
2200 Churchill Rd
Springfield,  IL 62706
217/782-1830
Simpson,  Charles
SynTech Products Corp.
520 E. Woodruff
Toledo, OH  43624
419/241-1215
Skoney, Davi d
Erie Co. Dept of  Environment
 and Planning
95 Franklin  St
Buffalo, NY  14225
716/846-8556
Smith, M. L.
Andersen  Samplers, Inc.
4215 Wendell Dr
Atlanta, GA  30315
404/691-1910
Staley, Laurel J.
U.S.  EPA/RREL (IERL)
Cincinnati,  OH 45268
513/569-7863
Stark, Phillip E.
Shell Oil Co. - Mining
PO  Box 2906
Houston, TX  77001
713/870-4263
                                    A-4

-------
Attendees (cont.)
Stone,  F.  Kenneth
S. C. Dept of Health and
 Environmental Control
2600 Bull St
Columbia, SC 29201
803/758-5406
Tucker, A. L.
Jones and Laughlin Steel Co.
3001 Dickey Rd
E. Chicago,  IN  46312
219/391-2571
Turetsky, William S.
Allied  Chemical Co.
PO Box 1139R
Morristown, NJ 07960
201/455-2690
Wells, Robert C.
Enviroplan, Inc.
59 Main St
West Orange, NJ  07052
201/325-1544
Williams, Colin J.
Rowan, Williams,  Davies,
 and Irwin, Inc.
(MHTR Ltd)
650 Woodlawn Rd,  West
Guelph, Ontario N1K1B8
519/823-1311
Wootten, John M.
Peabody Coal Co.
PO Box 14495
St. Louis, MO 63178
Yocom, John E.
TRC Environmental
 Consultants, Inc.
800 Connecticut Blvd
E. Hartford, CT  06108
203/289-8631
                                  A-5

-------
                                TECHNICAL REPORT DATA
                          (Please read Instructions on the reverse before completing)
1. REPORT NO.
  EPA-600/9-89-085
                           2.
                                                       3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
 Fifth Symposium on Fugitive Emissions:
  Measurement and Control (May 3-5, 1982,
  Charleston,  South Carolina)	
                                   5. REPORT DATE
                                    September 1989
                                   6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
 D.Bruce Harris and William B. Kuykendal,
    General Chairmen
                                                      8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 See Block 12
                                                       10. PROGRAM ELEMENT NO.
                                   11. CONTRACT/GRANT NO.
                                                       NA (Inhouse)
 12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Air and Energy Engineering Research Laboratory
 Research Triangle Park, NC 27711
                                   13. TYPE OF REPORT AND PERIOD COVERED
                                    Proceedings;  May 1982
                                   14. SPONSORING AGENCY CODE
                                    EPA/600/13
15. SUPPLEMENTARY NOTES  ^EERL project officer was D.  Bruce Harris,  Mail Drop 54, 919/
 541-7807.  Symposium was held May 3-5, 1982,  in Charleston,  SC.
is. ABSTRACT
              proceedj.ngs document presentations at the Fifth Symposium on Fugitive
 Emissions: Measurement and Control, May 3- 5, 1982,  in Charleston,  SC.  The Sym-
 posium was sponsored by the U. S. EPA's Air and Energy Engineering Research Lab-
 oratory (known then as the Industrial Environmental Research Laboratory)  in Re-
 search Triangle Park, NC,  as part of the Agency's continuing effort to develop me-
 thods for measuring and controlling airborne and waterborne fugitive emissions
 from energy and  industrial processes. The objective of the symposium was to
 bring together people  from industrial, academic,  research, and government orga-
 nizations with experience or interest in fugitive emissions problems to exchange
 information of mutual potential benefit. The program included presentations by
 individuals from a variety of organizations describing their experience and view-
 points regarding the impact, measurement, and control of  fugitive emissions. An
 international flavor was provided by presentations by authors from Belgium, Canada,
 and Sweden.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                          b.lOENTIFIERS/OPEN ENDED TERMS
                                                  COSATI Field/Group
 Pollution
 Measurement
 Emission
 Processing
 Leakage
 Hydrocarbons
Coal Storage
Coal Dust
Roads
Dust Control
Aerosols
Particles
Pollution Control
Stationary  Sources
Fugitive Emissions
Particulate
Dust Suppressants
Street Sweepers
13 B
14G

13H

07C
081
21D

05E
07D
 Release to Public
                                          19. SECURITY CLASS (This Report)
                                           Unclassified
                       20. SECURITY CLASS (This page)
                       Unclassified
                                                21. NO. OF PAGES
                                                    306
                                                                   22. PRICE
EPA Form 2220-1 (9-73)
                                        A-6

-------