United States    Office of
        Environmental Protection Research and Development
        Agency      Washington, DC 20460
EPA-600/R-93-019

January 1993
*EPA Characterization of PM-10
        Emissions From Antiskid
        Materials Applied To
        Ice- And Snow-Covered
        Roadways
   Prepared for Office of Air Quality Planning and Standards
               Engineering Research Laboratory

-------
                        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 Informa-
tion Service, Springfield, Virginia 22161.

-------
                                                      EPA-600/R-93-019
                                                      January 1993
                       CHARACTERIZATION OF PM-10 EMISSIONS
                       FROM ANTISKID MATERIALS APPLIED TO
                        ICE- AND SNOW-COVERED ROADWAYS
                                   FINAL REPORT
                                    Prepared by:

                                   John S. Kinsey
                              Midwest Research Institute
                                425 Volker Boulevard
                             Kansas City, MO 64110-2299
                             EPA Contract No. 68-DO-Q137
                               Work Assignment No. 12
                        EPA Project Officer: Charles C. Masser
                         U.S. Environmental Protection Agency
                    Air and Energy Engineering Research Laboratory
                          Research Triangle Park, NC  27711
                                   Prepared for:
U.S. Environmental Protection Agency            U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards       Office of Research and Development
Research Triangle Park, NC  27711               Washington, DC  20460

-------
                                  ABSTRACT
      Several areas of the country in violation of the National Ambient Air Quality
Standard for PM-10 have conducted studies which identify  the resuspension of
antiskid material applied to paved roadways as an importan. source. The application
of antiskid materials creates a temporary but substantial increase in the amount of fine
particulate on the road surface over and above that which is normally present.
Measured emission data are lacking for all types of antiskid materials; therefore, an
appropriate field program was undertaken whose objective was to establish a
predictive model for PM-10 emissions. Source-oriented emissions sampling was
conducted on a section of US 53 near Duluth,  Minnesota, during March and April of
1992. The measured emission factors varied from 1 to 11  g/VKT for the three tests
conducted. The data were not sufficient, however, to develop any specific correlation
between the measured  emission factors and source parameters.  The only general
observation made was that PM-10 emissions appear to increase with the amount of
antiskid material applied.  A comparison of measured emission factors with  those
predicted by AP-42 indicated that most of the measured factors are higher than those
which would be predicted from silt-loading values alone. Due to the marginal test
conditions during the storm events, no definitive assessment of this increase can be
made until additional data are obtained.

-------
                                  CONTENTS
Abstract	    ii
Figures	    iv
Tables	    v

      1.     Introduction	  1-1
      2.     Site Selection	  2-1
                   2.1    Screening methods  	  2-1
                   2.2    Site survey and selection	  2-2
      3.     Overall Study Design	  3-1
                   3.1    General air sampling equipment and techniques  ....  3-1
                   3.2    Testing procedures  	  3-5
                   3.3    Chemical analyses	  3-9
                   3.4    Ancillary sample collection and analysis 	3-10
                   3.5    Emission factor calculation procedure	3-12
      4.     Field Sampling Program  	  4-1
                   4.1    Modifications to study design	  4-1
                   4.2    Source description and activity	  4-3
                   4.3    Exposure profiling results	  4-3
                   4.4    Results of ancillary sampling and analysis	4-11
                   4.5    Discussion of results 	4-17
      5.     Quality Assurance	  5-1
                   5.1    Performance audit	  5-1
                   5.2    Data audit	  5-1
                   5.3    Data assessment	  5-2
                   5.4    Report review  	  5-3
      6.     Study Conclusions	  6-1
      7.     References	  7-1

Appendices

      A.     Material sampling and analysis	  A-1
      B.     Modifications to ASTM methods for LA  abrasion loss and Vickers
             hardness  	  B-1
      C.     Example data forms used in SHRP Project  H-208A for monitoring
             site conditions	  0-1
      D.     Sample calculations	  D-1
                                       in

-------
                                   FIGURES
Number                                                                   Page

  1-1       Diagram of street surface/atmospheric exchange of paniculate
           matter	  1-3
  1-2       Decision "tree" for antiskid material selection	  1-4
  2-1       Historical average snowfall data for Duluth, Minnesota	  2-3
  2-2       Historical maximum and minimum temperature data for Duluth,
           Minnesota	  2-4
  2-3       Approximate location of original and final test sites on US 53
           near Duluth, Minnesota 	  2-5
  3-1       Sampler deployment scheme	  3-2
  3-2       Diagram of high-volume cyclone sampler	  3-4
  4-1       Silt-loading history for northbound lanes (sand/salt mix)	4-14
  4-2       Silt-loading history for southbound lanes (rock salt only)	4-15
  A-1       Example data form for storage piles  	  A-4
  A-2       Sample riffle dividers	  A-6
  A-3       Example moisture analysis form	  A-8
  A-4       Example data form for paved roads   	A-11
  A-5       Example silt analysis form  	A-13
                                       IV

-------
                                   TABLES

Number

  1-1       Material selection criteria for antiskid abrasives 	   1-5
  3-1       Sampler deployment  	   3-3
  3-2      Quality control procedures for sampling media	   3-6
  3-3      Quality control procedures for sampling flow rates	   3-6
  3-4      Quality control procedures for sampling equipment	   3-7
  3-5      Criteria for suspending or terminating a test	   3-8
  3-6      Test methods for antiskid materials	3-10
  4-1       Source test summary	   4-4
  4-2      Results of gravimetric analyses	   4-5
  4-3      Summary of experimental results  	   4-7
  4-4      Results of emission factor calculations  	   4-8
  4-5      Results of chemical analyses  	4-10
  4-6      Properties of antiskid material samples	4-11
  4-7      Results of road surface sampling  	4-13
  4-8      Predicted PM-10 emission factors for measured surface silt
           loadings  	4-16
  B-1       Modifications to ASTM Method C 131-89 for aggregate material
           < 2.36 mm  	   B-3
  B-2      Modifications to ASTM E 384-89 to incorporate fine aggregate
           materials	   B-5

-------
                                  SECTION 1

                                INTRODUCTION
       Several areas of the country that are in violation of the National Ambient Air
Quality Standard for PM-10 (airborne particles less than or equal to 10 p.mA in
diameter) have conducted studies to determine the sources of these emissions. One
source of PM-10 emissions identified in a number of these studies is the resuspension
of antiskid material that is applied to paved roadways.  Antiskid materials may consist
of abrasives, such as sand, stone, cinders, or other materials, applied to the road
surface to improve traction or "deicers," which serve to restore pavement traction by
preventing the formation of ice films, weakening the ice/pavement bond, and/or by
melting ice and snow.

       The application of antiskid materials creates a temporary, but substantial,
increase in the amount of fine particles on the road surface over and above that which
is normally present.  Prior research has established a direct relationship between the
loading of silt-size fines (particles < 75 urn in physical diameter) and the PM-10
emission generated by vehicular traffic. The empirical relationship between silt loading
and PM-10 emissions is reflected in the EPA-recommended PM-10 emission factors
for paved urban roads. This relationship was developed from a data base encom-
passing the  results of tests conducted at eight sites, ranging from a freeway to a rural
town road.

       According to  EPA publication AP-42, the quantity of dust  emissions from vehicle
traffic on a paved roadway (per vehicle kilometer of travel) may  be estimated  using the
following empirical expression (EPA, 1990):
                                ..£.£•                          (")
where:      e =  PM-10 emission factor (gA/KT)
            s =  surface silt content (fraction of particles < 75 ^m in physical
                 diameter)
            L =  total road surface dust loading (g/m2)

      The total loading (excluding litter) is measured by sweeping and vacuuming
lateral strips of a known area from each active travel lane. Using a modified version


                                       1-1

-------
 of ASTM C 136, the silt fraction is determined by measuring the proportion of loose
 dry road dust that passes a 200 mesh screen.  Silt loading is the product of total
 loading and silt content.

       Recently, in the absence of specific emission test results for antiskid materials,
 PM-10 emission factors from sanding and salting were estimated as a "gap filling"
 exercise (Grelinger et a/., 1988).  The emission factor for sanding assumed that all of
 the PM-10 content of typical road sand would be suspended.  In the case  of salting,
 the emission factor was based on the assumption that  (a) 5% of the salt takes the
 form of a dried film on the pavement, and (b) 10% of the film is liberated as PM-10
 particles.  Because of the uncertainties in these underlying assumptions, a low quality
 rating  (E) was assigned to the estimated emission factors.

       It is clear from Equation (1-1) that techniques for controlling PM-10  emissions
 resulting from antiskid  materials should be aimed at minimizing silt loading on the
 traveled portion of the  roadway (Figure 1-1). Specifically, reduced silt loading may be
 expected to result from snow/ice control programs that encompass improvements in
 three areas: the properties of antiskid materials applied, the application protocols and
 procedures, and the procedures for removal of the antiskid material from roadways.

       In a recent EPA study (EPA-450/3-90-007), a literature search, engineering
 analysis, and laboratory-testing program were performed to provide air pollution
 controJ agencies with information on how  to identify appropriate antiskid materials that
 are both durable and effective and produce lower PM-10 emissions (Kinsey,  1991).
 The primary objectives of this study were  to provide guidance on methods to deter-
 mine:  (a) the physical properties and durability of antiskid material selected for use on
 ice- and snow-covered roadways, and (b) criteria for defining the elements of an
 effective PM-10 emission control strategy associated with use of antiskid materials.
 The material selection criteria and decision  process developed in this study are
 provided in Table 1-1 and Figure 1-2, respectively.

       Although the above program provided guidance for the selection of  antiskid
 materials, no direct information was developed regarding the actual  PM-10 emissions
 relat- d to their use, the changes in surface silt loading resulting from such application,
 or th   degree of control actually achieved by compliance with the material selection
 criteria developed in the study.  Although  measured emissions data are lacking for all
types of antiskid mat^'ial, deicing chemicals have received the least amount of
 attention.  Therefore, an appropriate field  program was needed to establish a suitable
 method for predicting the PM-10 emissions from the use of antiskid materials. This is
the primary objective of the work reported here.

       To adequately determine the fate and transport of antiskid materials, ideally, a
 mass balance sampling and analysis approach would be most appropriate. However,
 in the case of fugitive sources (e.g., snow- and ice-control materials), this  is
impossible since many of the "outputs" can be neither sampled nor quantified.


                                       1-2

-------
     PARTICIPATE  ENTRAINMENT  FROM URBAN  STREETS

                         Background
         Local
         Vehicles
         (Exhaust)
 Ground- Level
11 Suspended
   Parti culates"
    Urban
    Sources—
             DEPOSITION
     Sanding,
     Salting,
     Spills
   Conventional
   &  Fugitive
            ENTRAINMENT
            (By Wind & Vehicle  Motion)
Accumulated
Street  Deposits
.Vehicular Deposits
 (Carryout from Unpaved-
  Areas, Tire* Wear,Oil,etc.)
                   Runoff    ,   Mechanical Removal
                   (Sewers)     (Street Cleaners)
Figure 1-1.  Diagram of street surface/atmospheric exchange of paniculate matter.
                                  1-3

-------
  IDENTIFY
  SERVICE    <
CONDITIONS
                       Level of Antiskid
                      Control Required
Traffic Volume,
 Lanes, Speed
                      Temperature and
                       Environmental
                        Conditions
  UseAASHTO,
Federal, and State
   Guidelines
Define Candidate
  Materials for
 Consideration
                                      FOR EACH CANDIDATE MATERIAL
   Define Material
  Specifications to
 Reduce Silt Loadings
        I
                                                 I
                  Define Application Levels
           and Procedures to Reduce Silt Loadings
                                                                        1
                                    Define Clean-up and
                                 Other Mitigation Procedures
                                  for Reducing Silt Loading
    • Particle Size
 • Particle Morphology
     • Hardness
     • Durability
 > % Insoluable Matter
      • Others
                          1
               • Application Rate of Abrasives
                 • Application Rate of Sails
             1 Equipment Calibration/Maintenance
               • Development of Protocol and
                    Documentation for
                   Equipment Operators
                • Monitoring of Compliance
                     w/Specificalions
                           • Clean-up Frequency ..  . lime After Storm
                                  • Type of Equipment Used
                             • Monitoring of Clean-up Effectiveness
                                           Select Material for
                                       Field Study and Evaluation
                                                                      SEVkinsgr 11/3*89
                    Figure 1-2.  Decision "tree" for antiskid material selection.

-------
Instead, source-oriented emissions sampling was employed in the program using
MRI's "exposure profiling" approach. The data obtained by this technique were
coupled with the results of road surface sampling and materials application data in an
attempt to develop a method for predicting PM-10 emissions.
TABLE 1-1. MATERIAL SELECTION CRITERIA FOR ANTISKID ABRASIVES8
Measurement
parameter
Modified Los Angeles
abrasion loss
Initial silt content6
Vickers hardness
Particle shape index
Units
Weight %
Weight %
kg/mm2
Dimensionless
Mean value for
acceptable materials
3
0.1
1,000
10
Mean value for
unacceptable materials
11
6
800
9
a If 4l-irt rivsirisir4isif* r\f o r\*ir+i/M flor* motAriol f^tl K^fiaiaan fho no^*rȣsr\f oKlo11 r*r\f4
    "unacceptable" ranges, the material is considered "questionable" and good engineering
    judgment should be used.

  b  This parameter is coupled to LA abrasion loss and thus included in the material
    selection criteria.
      Originally, the goal of the study was to develop a predictive emission model for
the use of "straight" deicing salts.  However, due to site-specific conditions, this was
not possible. Instead an abrasive/salt mixture was tested in the field program.  This
was deemed to be representative since relatively few transportation  agencies use
straight salt because of runoff, vegetative damage, and other environmental concerns
(Kinsey, 1991).  However, appropriate chemical analyses of selected filter samples
were performed in an attempt to isolate the contribution of deicing salt to the total
PM-10 emissions from the test  road.

      The remainder of this report is structured as follows: Section 2 describes the
site selection process; Section 3 describes the overall study design; Section 4
describes the results of the field sampling program; and Section 5 discusses quality
assurance.  Conclusions reached from the experimental data are included  in
Section 6, and the references cited in the  report are listed in Section 7.
                                       1-5

-------
                                  SECTION 2

                               SITE SELECTION
      This section describes the site selection process used in the study. Screening
methods are described first, followed by details related to the specific location
selected.

2.1   SCREENING METHODS

      Based on almost 20 years of testing fugitive emission sources, MRI has
developed a number of site selection criteria for most generic source categories.
These criteria are useful as screening tools for evaluating candidate test locations
during the site survey.  The following selection criteria apply to  roadway source
testing:

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

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

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

      4.    The height of the nearest upwind obstruction should be less than one-
            third the distance from the road to the obstruction.

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

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

      7.    The test road should have an adequate number of vehicle passes per
            hour to enable completion of a test  in less than 3  h in order that testing
            can be safely completed during daylight hours.

      8.    The traffic mix during a test should  be representative of the type of
            vehicles that regularly use the road.
                                      2-1

-------
       In the case of the current pr-aram, a number of factors other than those listed
above were given special consideration during site selection.  First, most previous MRI
tests were performed during warm, dry weather to characterize worst-case emissions.
However, the present study required testing in cold, wet conditions, which complicated
both equipment placement and sample collection.

       Second, previous antiskid material emission studies paid relatively little attention
to the  material(s) being applied to the road, the amount being applied, or the fre-
quency of application (PEDCo, 1981; RTP  Environmental Associates, 1990).
Therefore, some means had to be provided for the collection of detailed data on
source conditions during testing.

2.2    SITE SURVEY AND SELECTION

       Based on the best available information, and the short time frame in which the
study was to be  conducted, it  was decided that one of the 14 sites currently active in
Strategic Highway Research Program (SHRP) Project H-208A should be used  for
emission testing. (SHRP is an independent federal agency that performs research in
conjunction with  various state  transportation departments.) This decision was based
on the fact that detailed information on winter maintenance practices and roadway
conditions was already being collected at these locations as part of that study.   Each
site is  also equipped with a permanently installed Roadway Weather Information
System (RWIS) that monitors  both ambient and pavement conditions on a continuous
basis.

       From the  14 candidate  sites, the SHRP "control" section located on US  53,
northwest of Duluth, Minnesota, was finally selected for testing.  (Note that the SHRP
"control" section  refers to that portion of the roadway that receives standard winter
maintenance practices and the "test" .section is that portion of the road which receives
treatment with only deicing chemicals.)  This site was suitable for a number of
reasons, including iis orientation with  respect to ambient winds, good cooperation by
the state transportation agency (Minnesota Department of Transportation—MnDOT),
and the typical snowfall and temperature expected in the month of March (see
Figures 2-1 and 2-2).

      The original test  site was located on northbound US 53 at  Mile Post (MP) 13.35
(or approximately Station No. 452+50).  This particular location has good orientation
with respect to the wind direction anticipated for periods after storm events, good
exposure to ambient winds (i.e., lack of trees in the upwind direction), and a relatively
flat median for installation of the air-sampling equipment. The approximate test area
is shown on the topographic map included  as Figure 2-3.
                                      2-2

-------
  18 -

  17 -

  16 -

  15 -

  14 -

  13 -

  12 -

  1 1 -

  10 -

   9 -

   8-

   7 -

   6 -

   5 -

  4 -

  3-

  2 -

   1 -
  0 -
 INCHES
        AVtKAGE SNOWFALL/MONTH
          QOQQ  OCT-riAY   OQGQ

                 DULUTH, NN
 1.2

I
               17.1
           13.7


//A


\

                        14.0
                       VA
                   11 7





                            6.7


QCT  NOV  DEC  JAN  FEB  MAR   APR. MAY
Figure 2-1. Historical average snowfall data for Duluth, Minnesota.
                     2-3

-------
          AVERAGE  MAXIMUM\M!NIMUM TEMPERATURE
                        OCTOBER - APRIL
                          OULUTH . MN
60-
55 -

SO -

45-


40-

35 -
30-
25-
20-
15 -
10-
05-
00 -
05-

53










r
x
X
X
X
/
X
f.
x
\\\\\\X\\\\\\\\\\X\\\NX\\\XvJ








35 33
V//////////////////////////A


^ \//////////////////\

k\\\\\\\\\\\\\\\\\Vi »
06
1











13
I
-3
22
f




47






02
1

13

71
x
y
y
y
y
y
y
y
y
y
y
y
\
/
/,
/.
/,






23
1

           OCT
DEC
JAN
FEB
MAR
APR
Figure 2-2.  Historical maximum and minimum temperature dai« for Duluth, Minnesota.
                             2-4

-------
                                                     Northbound Lanes:
                                                     Abrasive/Salt Mix
                                  ; .' -
                                     Original Site
                  7*  v     . M • ,'   •«••••.._.• *^^F* " "  *•("

                  -~.v—   x  Southbound Lanes: •  „    •)
                  3l^      -   CfrainK* Dnr-lr Colt    '. "* ~
Figure 2-3.   Approximate location of original and final test sites on US 53 near
             Duluth, Minnesota.
                                        2-5

-------
      Several shortcomings were noted with respect to air sampling at the original
location on US 53.  First, the expected prevailing winds after storm events allowed
sampling of only the northbound laru.s which use an abrasive/rock salt mix. Thus, the
PM-10 emissions from the application of straight deicing salts could not be directly
measured. In addition, certain upwind influences from  a nearby county road ("old"
US 53 or Miller Trunk Highway) could not be avoided.  This road is a low-volume route
located some distance from the site which mitigates its contribution. Nevertheless,
MRI was convinced that this location was the best site  of those readily available.

      However,  due to the mild winter in nuluth, the original test site could not be
used. The site was selected on the assL ..ption that the ground would remain frozen
throughout the sampling period, allowing equipment to  be safely placed in the median.
When the ground remained thawed, the sampling site was moved approximately
VA mile (0.4 km) northbound  (i.e., to the west) to an area near the intersection of
US 53 with a county road (Figure 2-1).  This area provided a firm surface for location
of the sampling equipment but also had substantial stand of trees bordering the
northern portion of the right-of-way which did not conform to selection  criterion 2 and 4
listed above.  For this and other related reasons, only one of the three test series
conducted met all of the MRI QC criteria for wind speed and direction.  This non-
compliance with established criteria adversely affected  data quality as  discussed
below.
                                      2-6

-------
                                  SECTION 3

                           OVERALL STUDY DESIGN
      The source-directed field sampling conducted in this study employed the
"exposure profiling" approach to quantify source emission contributions. After two
minor storm events, a series of exposure profiling tests were performed, beginning
approximately at the point when the pavement became dry.  This section describes
the overall study design used during source testing, testing procedures, chemical
analysis, ancillary sample collection, and emission factor calculations.

3.1   GENERAL AIR SAMPLING EQUIPMENT AND TECHNIQUES

      The "exposure profiling" technique for particulate  source testing  is based on the
isokinetic profiling concept used in conventional (stack) testing.  The passage of
airborne pollutant immediately downwind of the source is measured directly by means
of simultaneous multipoint sampling over the effective cross section of  the open dust
source plume. This technique uses a mass flux measurement scheme similar to EPA
Method 5 for stack testing rather than requiring indirect emission rate calculation
through the application of a generalized atmospheric dispersion model. Further details
of the exposure profiling method can be found in earlier technical reports such as the
recent EPA collaborative study (Pyle and McCain, 1986).

      For measurement of particulate emissions from roads, a vertical network of
samplers (Table 3-1) was positioned just downwind and upwind from the edge of the
road. The  downwind distance of 5 m was far enough that sampling interferences due
to traffic-generated turbulence was minimal, but close enough to the source that the
vertical plume extent could be adequately characterized with a maximum sampling
height of 5 to 7 m. In  a similar manner, the 10-m distance upwind from the road's
edge was far enough from the source that (a) source turbulence did not affect sam-
pling; and (b) a brief reversal would not substantially impact the upwind samplers.
The  10-m distance was, however, close enough to the road to provide  the representa-
tive background concentration values needed to determine the net (i.e., due to the
source)  mass flux.

      As shown  in Figure 3-1, the planned equipment deployment scheme made use
of four downwind vertical sampling arrays, D1 through D4. Downwind  arrays D1, D3,
and D4  (as well as upwind array U2) made use of high-volume (hi-vol) air samplers
equipped with constant-flow electronic controllers and cyclone preseparators.


                                      3-1

-------
                     High volume sampler w/cyclone
A                     High volume sampler w/Wedding inlet and
                       critical orifice flow controller
                     Warm wire anenometer

                     Wind vane
Array D3
     7m
Array D4
     7m
CO
                                                          North Bound Lanes-US 53
   Array U1
   3mA     Array U2
                                                                                                  Direction of Travel
                                                        N
                                                                                                         92-37 SEVkhl ton 0623K
                                         Figure 3-1. Sampler deployment scheme.

-------
Arrays D2 and U1, used hi-vols equipped with Wedding PM-10 inlets and critical
orifice flow controllers.  Also installed at the site were concrete barricades, sand
barrels, and related safety equipment required by MnDOT regulations.
                        TABLE 3-1. SAMPLER DEPLOYMENr
Sampler No. of
array ID instruments
U1

U2
D1, D3, D4


D4

D4
D2

1

2
4


2

1
1

Measurement
height(s) (m)
3

1.5,3
1,3,5,7


1,5

3
2.5

Type of sampler
or instrument
Hi-vol + Wedding
inlet
Hi-vol + Cyclone
Hi-vol + Cyclone


Warm wire
anemometer
Wind vane
Hi-vol + Wedding
inlet
Parameter(s)
measured
PM-10, Pb

PM-10
PM-10, Na+, Cr
(Selected arrays
only)
Wind velocity

Wind direction
PM-10, Pb

        "  Certain downwind sampling equipment could not be used during testing due to
          persistent generator problems (see text).

      For each profiling trailer (i.e., Array D1, D3, and  D4), PM-10 samples were
collected at four downwind measurement heights. Also, PM-10 (as well as Pb)
concentrations were determined at a single height both upwind and downwind of the
road using the EPA ambient reference method. The latter measurements were used
for comparison against data collected by the cyclone samplers.

      The primary air sampling device  in this program  was a standard high-volume air
sampler fitted with a Sierra Model 230CP cyclone preseparator (Figure 3-2).  The
cyclone exhibits  an effective 50% cutoff diameter (D50)  of approximately 10 microns
(u.m) in  aerodynamic diameter when operated at a constant flow rate of 40 dm
(68 m3/h).

      Throughout each test, wind speed was monitored by warm-wire anemometers
(Kurz Model 465) at two heights, and the vertical wind  speed profile was determined
assuming a logarithmic distribution.  Horizontal wind direction was also monitored  by
a wind vane at a single height, with 10-min averages determined electronically prior to
and during the test. The sampling intakes were adjusted for proper directional
orientation based on the approximate average wind direction.
                                      3-3

-------
                                                         Back-up Filter
                                                         Holder
0           5
I   I  I   I  I   I
 Scale - Inches
         Figure 3-2. Diagram of high-volume cyclone sampler.
                                 3-4

-------
3.2   TESTING PROCEDURES

3.2.1  Preparation of Sample Collection Media

      Paniculate samples were collected on Type AH grade glass fiber filters.  Prior
to the initial weighing, the filters were equilibrated for 24 h at constant temperature
and humidity in a special weighing room.  During weighing, the balance was checked
at frequent intervals with standard (Class  S) weights to ensure accuracy.  The filters
remained in the same controlled environment for a second 24-h period, after which
another second analyst reweighed them as a precision check.  If a filter could not
pass  audit limits, the entire lot was reweighed.  Ten percent of the filters taken to the
field were used as blanks.  The quality control guidelines pertaining to preparation of
sample collection media are presented in  Table 3-2.

      As indicated in Table 3-2, a minimum of 5% laboratory blanks and 5% field
blanks were collected for QC purposes (EPA, 1977). This procedure involved
handling at least 1 filter in every 10 in an  identical manner as the  others to determine
systematic weight changes. These changes were then used to mathematically correct
the net weight gain determined from gravimetric analysis of the filter samples. In the
case  of laboratory blanks, this involved only those procedures followed in MRI's  gravi-
metric analysis laboratory.  For field blank collection, filters were actually loaded into
samplers and then recovered without air actually being passed through the media.

3.2.2 Pretest Procedures/Evaluation of Sampling Conditions

      Prior to actual sample collection, a number of decisions were made as to the
potential for acceptable source-testing conditions. These decisions were based  on
forecast information obtained either from the local U.S. Weather Service office and/or
from the Roadway Weather Information System  (RWIS) located at the site.  If
conditions were considered acceptable, the sampling equipment was prepared for
testing.  Pretest preparations included calibration of the various air sampling
instruments, insertion of filters, and so forth.  The quality control guidelines governing
this activity are found in Table 3-3.

      Once the source testing equipment was set up and the filters inserted, air
sampling was conducted.  Information recorded on specially designed reporting forms
included:

      a.     Air samples—Start/stop times, wind speed profiles, flow rates, and wind
             direction relative to the roadway perpendicular (10-min average).  See
             Table 3-4 for QC procedures.

      b.     Traffic count by vehicle type and speed.

      c.     General meteorology—Wind speed, wind direction, and temperature.
                                      3-5

-------
  TABLE 3-2.  QUALITY CONTROL PROCEDURES FOR SAMPLING MEDIA
           Activity                       QC check/requirement
 Preparation                   Inspect and imprint glass fiber media with
                              identification numbers.
 Conditioning                   Equilibrate media for 24 h in clean controlled
                              room with a relative humidity of 45% (varia-
                              tion of  less than ±5%} and with a tempera-
                              ture of 23°C (variation of less than ±1%).
 Weighing                     Weigh  hi-vol filters to nearest 0.1 mg.
 Auditing of weights            Independently verify final weights of 10% of
                              filters (at least four from each batch).
                              Reweigh batch if weights of any hi-vol filters
                              deviate by more than ±2.0 mg.  For tare
                              weights, conduct a 100% audit. Reweigh
                              tare weight of any filters that deviate by more
                              than ±1.0 mg.
 Correction for handling effects3  Weigh  and handle at least one blank for each
                              10 filters of each type for each test.
 Calibration of balance          Balance to be calibrated once per year by
                              certified manufacturer's representative.
                              Check  prior to each use with laboratory Class
	S weights.	
 a  Includes both laboratory blanks and field blanks  (see text).
          TABLE 3-3. QUALITY CONTROL PROCEDURES FOR
                        SAMPLING FLOW RATES	

              Activity                       QC check/requirement
   High volume air samplers           Calibrate flows in operating ranges
                                     using calibration orifice upon arrival
                                     and every 2 weeks thereafter.
   Orifice and electronic calibrator      Calibrate against displaced volume
                                     test meter annually.
   Warm wire anemometers           Calibrate annually in standard wind
                                     tunnel.
                                  3-6

-------
          TABLE 3-4. QUALITY CONTROL PROCEDURES FOR
                        SAMPLING EQUIPMENT
         Activity
            QC check/requirement3
Maintenance
•  All samplers
Check motors, gaskets, timers, and flow
measuring devices prior to testing.
Operations
•  Timing

•  Isokinetic sampling
   (cyclones)
   Prevention of static
   mode deposition
Start and stop all downwind samplers during time
span not exceeding 1 min.
Adjust sampling intake orientation whenever mean
wind direction dictates.

Change the cyclone intake nozzle whenever the
mean wind speed approaching the sampler falls
outside of the suggested bounds for that nozzle.
This technique allocates no nozzle for wind
speeds ranging from 0 to  10 mph, and unique
nozzles for four wind speed ranges above
10 mph.
Cap sampler inlets prior to and immediately after
sampling.
a All means refer to 5- to 15-min averages.
                                 3-7

-------
      MRI has developed criteria for suspending or terminating a source test.  These
 criteria are presented in Table 3-5.  In the case of the current program, however, a
 number of these criteria could not be followed in most of the tests conducted for a
 variety of reasons unique to this project.  These are discussed briefly below.

      TABLE 3-5. CRITERIA FOR SUSPENDING OR TERMINATING A TEST

    A test may be suspended or terminated if:

    1. Precipitation ensues during equipment setup or when sampling is in
      progress.

    2. Mean" wind speed  during sampling moves outside the 1.3- to 8.9-m/s (2- to
      20-mph) acceptable range for more than 20% of the sampling time.

    3. The angle between mean wind direction and the perpendicular to the path
      of the moving point source during sampling exceeds 45 degrees  for two
      consecutive averaging periods.

    4. Daylight is insufficient for safe equipment operation.

    5. Source condition deviates from predetermined criteria (e.g., occurrence of
      wet pavement conditions.)

    a  "Mean" denotes a 5- to 15-min average.

      First, due to the extremely mild winter in Duluth, testing opportunities were, at
best, limited.  For this reason, testing was often conducted under less than optimum
wind conditions for the sake of practicality and economy.  Each decision to conduct a
sampling trip was discussed in advance with the EPA project officer to obtain his
concurrence.

      Second, to meet the objectives of the program, sampling had to take place as
soon as possible after the road became dry to characterize the  maximum emissions
potential  resulting from application of the  antiskid material.  Thus, if the wind speed
and direction could be in any way accommodated by the sampling array, testing was
performed. Also, testing under highly variable wind conditions, which would normally
preclude  sampling, was likewise conducted for the same reason.

      Finally, the criteria  listed in Table 3-5 were originally intended for source testing
under hot, dry conditions, which would  represent worst case.  In the current study,
however, frozen precipitation and wet road conditions were required preparatory to
sample collection. Therefore, although the - ad surface was slightly damp outside of
the wheel tracks, testing may have been  r   -ted.
                                      3-8

-------
 3.2.3 Sample Handling and Analysis

      To prevent particulate losses, the exposed media were carefully transferred at
 the end of each run to protective containers for transportation.  In the field laboratory,
 exposed filters were placed  in  individual glassine envelopes and then into numbered
 file folders. When exposed  filters and the associated blanks were returned to the MRI
 laboratory, they were equilibrated under the same conditions as the initial weighing.
 After reweighing, 10% of the filters were audited to check weighing accuracy.

 3.3   CHEMICAL ANALYSES

      Selected filters were extracted and chemically analyzed by an outside
 laboratory to determine the concentration of chloride (Cl~), sodium (Na+), or lead (Pb)
 in the particulate sample collected (see Table 3-1). The analytical procedures, and
 associated QA/QC, used for this purpose are described below.

 3.3.1 Chlorine Analysis

      Selected filter samples were extracted following the 40 CFR 50, Appendix G,
 procedure and analyzed for  Cl" using EPA Method 300.0 as published in EPA-600/
 4-84-017, dated March 1984.  These procedures involved the extraction of the sample
 using dilute nitric acid, followed by analysis using ion chromatography.  Replicates,
 spikes, spiked duplicates, split samples, blanks, calibration  checks, reagent checks,
 and detection limit checks were used to assure quality control during the analyses.

 3.3.2 Sodium Analysis

      Sodium content (Na*) of the filter samples was extracted following the 40 CFR
 50, Appendix G, procedure and determined using EPA Method 273.1 as published in
 EPA-600/4-84-017, dated March 1984. These techniques consisted of sample extrac-
 tion using dilute nitric acid, followed by flame atomic absorption spectroscopic
 analysis.  Replicates, spikes, spiked duplicates, split samples, blanks, calibration
 checks, reagent checks, and detection limit checks were used to assure quality control
 of the analyses.

 3.3.3 Lead Analysis

      Analysis of selected filters for lead (Pb) content were performed using the
40 CFR 50, Appendix G, procedure for extraction and  EPA Method 239.1 as published
in EPA-600/4-84-017, dated  March 1984, for analysis.  These techniques consisted of
sample extraction using dilute  nitric acid, followed by graphite furnace atomic absorp-
tion spectroscopic analysis.  Replicates, spikes, spiked duplicates, split samples,
blanks, calibration checks, reagent checks, and detection limit checks were used as
part of the QA/QC for the analyses conducted.
                                       3-9

-------
3.4    ANCILLARY SAMPLE COLLECTION AND ANALYSIS

      The types of ancillary samples and information to be collected fell into two
broad categories:

            Antiskid materials and roadway surface samples.

      •     Source activity levels.

Each category is described in greater detail below.

3.4.1  Material Sample Collection and Analysis

      In conjunction with the emissions tests, samples we   taken of the antiskid
material applied to the road and the dust remaining on the surface after it became dry.
These samples were needed not only to evaluate the performance of existing
emission models but also to develop improved models for antiskid materials.

3.4.1.1  Sampling and Analysis of Antiskid Materials-
      To characterize the antiskid materials applied during the study, appropriate
material samples were collected and analyzed for moisture and silt content.  Grab
samples were taken of both the stockpiled materials as well as the abrasive/rock salt
(NaCI) mixture that was actually distributed by the spreader trucks. The standard MRI
procedures used for the collection and analysis of antiskid material samples are
provided in Appendix A.

      To further characterize the antiskid material samples collected, a number of key
physical and chemical properties were also determined. These properties are listed in
Table ?  % along with the appropriate ASTM measurement method. The properties
shown  n Table 3-6 were determined to be good general indicators of the overall silt
production potential of antiskid materials as determined in the laboratory study
described earlier (Kinsey, 1991).

            TABLE 3-6. TEST METHODS FOR ANTISKID MATERIALS
Type of material
Abrasives




Deicers
Material property
LA abrasion loss
Initial silt content

Vickers hardness
Particle shape/texture
Insoluble matter-NaCI
ASTM method
C131
C136
C117
E384
D3398
E534
Method modified
Yes
Yes
Yes
Yes
No
No
                                     3-10

-------
      As indicated in Table 3-6, most of the ASTM methods have been specially
modified by MRI for the analysis of antiskid abrasives (Kinsey, 1991). Except for
Vickers hardness, these modifications are generally minor changes to the basic
methods to facilitate application to typical antiskid abrasives. Appendix B, Tables B-1
and B-2, provide the modifications made to the ASTM methods for LA abrasion loss
and Vickers hardness,  respectively.  The modified method for the determination of silt
content is also provided in Appendix A.

3.4.1.2  Road Surface Sampling and Analysis-
      Although exposure profiling was only performed on the emissions from the
northbound  lanes on US 53, surface sampling was conducted on both the northbound
and southbound sections of the highway. (Recall that an abrasive/salt mixture is used
on the northbound lanes and straight salt is used on the southbound lanes—see
Figure 2-3.) The specific procedures used to collect and analyze paved road surface
samples to determine silt loading are described in Appendix A.

3.4.2 Source Activity Monitoring

      Source extent and activity data were collected with a variety of tools.  For
example, in addition to visual observation and note taking, pneumatic traffic counters
were used to determine source activity on US 53.  A  radar gun was also  used to
determine the average  speed of vehicles passing the sampler array.

      Vehicle-related parameters were  obtained using a combination of manual and
automated counting devices. Pneumatic tube axle counters were used to obtain traffic
volume  data. Because these counters only record the number of passing axles, it was
also necessary to obtain manual traffic mix information (e.g., number of axles per
vehicle) to convert axle counts  to the number of vehicle  passes. Vehicle mixes were
observed visually.  Comparison of the observed vehicle  mix to the pneumatic counter
totals also allowed the accuracy of the axle counter to be assessed.

      Finally, as mentioned in  Section 2.2, the site selected for testing in the current
program was also used as part of SHRP Project H-208A. In conjunction with the
SHRP program, detailed information was collected by MnDOT personnel on the condi-
tion of the weather and pavement during the course of the storm, the types and
amounts of antiskid materials applied to the test  road, and a general indication of the
residual deicing chemical on the road surface  at various points in time. Although MRI
was not directly responsible for collecting these data, they were used to augment the
information obtained on source activity.  Also,  additional surface sampling was
conducted between storms to develop a silt-loading "history" during data  analysis.
Sample forms completed by MnDOT personnel in the SHRP program are included in
Appendix C.
                                      3-11

-------
3.5    EMISSION r  "TOR CALCULATION PROCEDURE

       To calculate ^mission rates, a conserv  ->n of mass -oproach was used. The
passage of airborne particulate (i.e., the quar.  / of emissions per unit of source
activity) was obtained by spatial integration of distributed measurements of exposure
(mass/area) over the effective cross section of the plume.  Exposure is the point value
of the flux (mass/area-time) of airborne particulate integrated over the time of
measurement, or equivalently, the net particulate mass passing through a unit area
normal to the mean wind direction during the test. The steps in the calculation
procedure are described below.

3.5.1  Particulate and Compound-Specific Concentration/Exposures

       The concentration of PM-10 measured by a sampler is given by:

                                  C =  103 -ZL                            (3-1)
                                           Qt

where:  C  =  particulate concentration  (ng/m3)
         m  =  particulate sample weight (mg)
         Q  =  sampler flow rate (m3/min)
         t   =  duration of sampling (min)

       The concentration (Cj) of Na+, CI", or Pb measured by the sampler is given by:

                                    C, = —'                              (3-2)
                                     1   Qt

where:  Cj  =  concentration of component i determined by filter analysis (^g/m3)
         m,  =  mass of component i collected on the filter (|ig)
         Q  =  sampler flow rate (m3/min)
         t   =  duration of sampling (min)

      To be consistent with the National Ambient Air Quality Standards, all
concentrations and flow rates are expressed in standard conditions (25°C and
101 kPa or 77°F and 29.92 in Hg).

      The isokinetic flow ratio (IFR) is the ratio of a directional (i.e., cyclone)
sampler's intake air speed  to the mean wind speed approaching the sampler. It is
given by:
                                      3-12

-------
                                   IFR = -Sf                              (3-3)
                                          aU

where:   Q   =  sampler flow rate (m3/min)
         a   =  intake area of sampler (m2)
         U   =  mean wind speed at height of sampler (m/min)

      The above ratio is of interest only in the sampling of total paniculate, since
isokinetic sampling ensures that particles of all sizes are sampled without bias.  Note
that because the primary interest in this program is directed to PM-10 emissions,
sampling under moderately nonisokinetic conditions poses no difficulty.  It is readily
agreed that  10 |im (aerodynamic diameter) and smaller particles have weak inertial
characteristics at normal wind speeds and, thus, are relatively unaffected by
anisokinesis (Davies, 1968). Therefore,  IFR was not calculated for the downwind
samplers in  the current program.

      Exposure represents the net passage of mass through a unit area normal to the
direction of plume transport (wind direction) and was calculated by:

                                E10  = ID'7 x CUt                          (3-4)

where:   E10 =  PM-10  exposure (mg/cm2)
         C   =  net concentration (pig/m3)
         U   =  approaching wind speed (m/s)
         t    =  duration of sampling (s)

Compound-specific exposures (i.e., for Na+, Cl~, and Pb) can be found analogously.

      Exposure values vary over the spatial extent  of the plume.  If exposure is
integrated over the plume effective cross section, then the quantity obtained
represents the total passage of airborne paniculate  matter (i.e., mass flux) due to the
source.

      For the test roadway, a one-dimensional integration scheme was used:

                                        £,0 *
where:   I   =   integrated PM-10 (or compound-specific) exposure (m-mg/cm2)
         E10 =   PM-10 exposure (mg/cm2)
         h  =   vertical distance coordinate (m)
         H  =   effective extent of plume above ground (m)
                                      3-13

-------
The effective height of the plume (H) in Eq. 3-5 is found by linear extrapolation of the
uppermost net concentrations to a value of zero.

      Because exposures are measured at discrete heights of the plume, a numerical
integration is necessary to determine I.  The exposure must equal zero at the vertical
extremes of the profile (i.e., at the ground where the wind velocity equals zero and at
the effective height of the plume where the net concentration equals zero).  However,
the maximum exposure usually occurs below a height of 1  m so that there is a sharp
decay in exposure near the ground. To account for this sharp decay, the value of
exposure at ground  level is set equal to the value at a height of 1 m. The integration
is then performed using Simpson's rule.

3.5.2  PM-10 Emission Methodology

      The emission model for PM-10 generated by vehicular traffic on roadways
treated with  antiskid materials, expressed in grams of emissions per vehicle-kilometer
traveled (VKT), is given by:
                                  e = 104                                (3-6)
                                           N

where:   e  =   PM-1 0 emissions (g/VKT)
         I   =   integrated PM-10 exposure (m-mg/cm2)
         N  =   number of vehicle passes (dimensionless)

Similar results can also be generated for NaCI and Pb by substituting the appropriate
integrated exposure into the above calculation  scheme.
                                     3-14

-------
                                  SECTION 4

                          FIELD SAMPLING PROGRAM
      The preceding section described the study design as it was originally
envisioned. However, a number of unanticipated events required that portions of the
original plan be altered to accommodate project-specific conditions.  This section
discusses the program as it evolved in response to these events, as well as the
results of the field sampling conducted in Duluth, Minnesota.

4.1   MODIFICATIONS TO STUDY DESIGN

      Discussions with EPA regarding the current study began in late November of
1991 with field testing expected to begin in late January or early February of 1992.
However, due to administrative delays, assignment, all required planning and
equipment preparation was conducted during February with field testing starting in
early March.  A decision to test during this period, based on historical snowfall and
temperature data obtained for the Duluth area  (see Figures 2-1 and 2-2), was
discussed and approved by the EPA Work Assignment Manager.

      In keeping with the extremely mild winter nationwide, Duluth experienced less
than 1 in of total precipitation during the month of March, which substantially elimi-
nated source testing.  (Note that only  one aborted attempt was made during the  entire
month.) Therefore, in an attempt to "salvage" the experimental program, special
measures were taken (with the full concurrence of the EPA Work Assignment
Manager) to extend field testing into April.  Each of these measures is described
below.

      First, rather than maintain a full-time sampling crew on-site (with its associated
cost impact),  personnel were dispatched from Kansas City on  an as-needed basis.
This was usually accomplished within 18 h  of receiving a prediction of measurable
precipitation either from the U.S. Weather Service, the Roadway Information System,
or a contract forecasting service (i.e.,  Accu-Weather)  provided to MnDOT.  Such a
procedure allowed the maximum possible opportunity for testing within the time
available; however, it did not allow for recalibration of the sampling equipment prior to
one test (Test AY-4).  Based on MRPs previous experience, this noncompliance with
normal QC procedures would not be expected, in itself, to substantially affect data
quality.
                                      4-1

-------
      The second special measure taken during the program involved the protocol
 used for application of antiskid materials to the road surface.  Normally, antiskid
 material is only applied to US 53 when significant frozen precipitation occurs and road
 conditions become hazardous.  In these  conditions, multiple passes are usually made
 while applying a sand/salt mixture at typical rates of 400 to 500 Ib/lane-mi (110 to
 140 kg/lane-km). (Note that the application rates used by MnDOT are substantially
 lower than those set by many other transportation agencies, which can be as high as
 2,000 Ib/lane-mi or 560 kg/lane-km.)

      However, due to the extraordinarily mild winter in Duluth and the low number of
 storm events, arrangements were made for antiskid material to be applied to the road
 surface anytime that precipitation occurred, regardless of type or quantity, even
 though  conditions might not actually warrant the application of such material.
 Although this procedure resulted in only a very limited amount of material being
 applied to the road, it was at least marginally adequate for test purposes.  The
 frequency and amount of material applied during testing cannot, however, be
 considered typical of common winter maintenance practices. Therefore, the results
 of this study might better be regarded as a conservative "best case" (rather than
 "representative") estimate of the PM-10 emissions resulting from the application of
 antiskid materials to the test road due to the relatively small amount of antiskid
 material applied (and thus available for resuspension) during testing.

      Special measures were also taken during the field sampling program with
 respect to the acceptability of test conditions.  Because of the relatively low amounts
 of antiskid material applied to the road (and its low traffic volume of < 5,000 vehicles/
 day), testing had to be started as soon as possible after the road surface became dry
 in order to obtain measurable downwind  concentrations. This necessitated the
 collection of air samples during periods when wind conditions (i.e., speed and
 direction) were not normally considered appropriate for exposure profiling (see
 Table 3-5). In this study, testing was performed whenever the ambient winds
 were from any northerly direction, as long as the other required test conditions
 could be met.  This procedure adversely affected data quality as  will be discussed
 later.

      Finally, ongoing generator problems plagued the field sampling program, which
 resulted in either samplers not being operated and/or samples being lost.  Repeated
 attempts by the equipment leasing company to rectify these problems proved to be
 unsuccessful.  Special operational procedures were implemented, therefore, to collect
the maximum amount of useful data within the limitations of available generator power.

      During each test conducted, efforts were d' led towards the operation of at
 least one downwind profiler array and the Weddi,    VI-10 sampler (Figure 3-1). In
 certain instances, two profilers may have been Oj-   ,ied for some period of time but,
 generally, only one was able to operate for the entire test.  The lack of generator
                                      4-2

-------
power adversely affected both sampling efficiency and completeness, which severely
limited the amount of experimental data collected during the program.

4.2    SOURCE DESCRIPTION AND ACTIVITY

       As stated in Section 2.2, the test site used in the experimental program was
located on northbound US 53 on the outskirts of metropolitan Duluth, Minnesota.
US 53 is an unlimited-access, four-lane, high-speed roadway carrying commuter traffic
to and from  Duluth at  an approximate volume of 5,000 vehicles/day.  Data collected
during field sampling showed that the majority of the traffic were two-axle,  light-duty
vehicles traveling between 88 and 97 km/h (55 and 60 mph). Surface loadings deter-
mined both visually and  by sampling were generally very low, with normal silt loadings
in the range of 0.2 g/m2.

       Exposure profiling was performed after two minor storm events that occurred on
April  10 and on April 21  and 22, 1992. One test series was conducted on April 11,
with additional testing  performed on April 23 and 26.  A summary of the test conditions
for each sampling period is provided in Table  4-1.

       During the April 10 storm, approximately 4 to 6 in of wet snow fell on US 53
during a 24-h period.  Four applications of an abrasive/salt mixture were made to the
road  over a  period of approximately 10 h, totaling 395 kg/lane-km (1,400 Ib/lane-mi).
For the second storm, a combination of wet snow and freezing  rain fell during
approximately a 36-h period. In this case, only  one application of 197 kg/lane-km
(700  Ib/lane-mi) was made to each lane especially  for test purposes (see Section 4.1).

       As noted above, the application rates used by MnDOT are far below those of
most other transportation agencies as determined in previous research conducted by
MRI (Kinsey, 1991). Also, much of the material applied was either cast off by the
snow plow during clearing operations or eliminated by melted precipitation. What
material that did remain  on the surface was quickly lost to the atmosphere by the
action of passing vehicles. Thus, surface loadings were generally low with higher
loadings observed in the passing (left) lane as compared to the driving (right) lane.

4.3    EXPOSURE PROFILING RESULTS

      A summary of the three exposure profiling tests conducted on US 53 was
provided previously in  Table  4-1. The test results are discussed in detail below with
the particulate sampling  data described first, followed by the results of the chemical
analyses and ancillary sampling/analysis.

4.3.1  PM-10 Sampling Results

      The results of the gravimetric analyses performed on the filter samples
collected in the field are  summarized in Table 4-2.  Note that the data provided in this
                                      4-3

-------
                                                                     TABLE 4-1.  SOURCE TEST SUMMARY
Test Date of Sampler
ID test array*
AY-3 4/11/92 U1
U2
D1

D2
D4
AY-4 4/23/92 U1
U2
D1
D2

D4
AY-5 4/26/92 U1
U2
D2
D4
Operating period
Start time (h)
1435
1435
1519*
1612
1527
1503
0841
0841
0914
0857*
1035
0924
0829
0829
0936
0936
Stop time (h)
1909
1909
1524
1904
1612
1804
1438
1438
lutr
1017
1103
1410
1417
1417
1316
me
Mean wind speed (m/s)*
1.0-m height
—
—
3.8

1
3.9
_
—
0.58
0.56

0.54
__
—
1.4
1.4
5.0-m height
—
—
6.1

1
6.6
_
—
1.2
1.2

1.2
_
—
1.9
1.9
Mean
wind
direction
(degrees)
—
—
294

1
293
	
—
38
50

46
—
—
336
336
Total vehicle
passes during
test period
1.419
1.419
1,175

—
983
1,200*
1,200s
220"
380*

962
1,030"
1,030"
650
650
Total antiskid
material applied
to roadway
(kg/lane-km)
395d
395"
39511

395-1
3951
197"
197*
197*
197s

197*
197*
197*
197*
197*
Time since
last antiskid
application
(h)
20.6
20.6
21.3

21.5
21.1
22.9
22.9
23.5
23.2

23.7
94.7
94.7
95.9
95.9
Approx. number ol
vehicle passes
since last
application'
5.200
5,200
5.400

• .. •
5,r.;r:
4.800
4.800
4.900
4.900

5,000
16,800
16,800
17,000
1- .'
•  Array D-3 uio ' lor ((old blanks due to generator (allure (see text).
k  Average ol 10-mhi Integration periods over duration ol test.
•  Based on vehicle counts obtained during testing.
'  Two passes ol 113 kg/lane-km (400 b/tane-ml) ol a 90%: 10% abrasive/rock salt mixture plus two passes ol 85 kg/lane-km (300 Ib/tane-ml) of a 80%:20% percent abrasive/rock salt mixture.
•  Sampler stopped and restarted due to generator failure.
1   Sampling aborted due to generator failure.
•  Vehicle counts estimated from available data.
"  One pass of 197 kg/tane-km (700 Wane-mi) ol a 90%:10% abrasive/rock salt mixture.

-------
               TABLE 4-2. RESULTS OF GRAVIMETRIC ANALYSES'
Test ID
No.
AY-3
(4/11/92)











AY-4
(4/23/92)












AY-5
(4/26/92)










Array ID
No.
D1


D3
(Blanks)

D4



U1
U2

D1



D2
D3
(Blanks)
04



U1
U2

D2

D3
(Blanks)

D4



U1
U2

Sampling
height (m)
1
3
7
1
3
5
1
3
5
7
3
1.5
3
1
3
5
7
2.5
1
3
1
3
5
7
3
1.5
3
2.5

1
3
5
1
3
5
7
3
1.5
3
Filter ID
No.
9221024
9221025
9221027
9221009
9221010
9221011
9221020
9221021
9221022
9221023
9221013
9221028
9221029
9221031
9221032
9221033
9221034
9221046
9221042
9221043
9221038
9221039
9221040
9221041
9221035
9221037
9221036
9221051

9221059
9221060
9221061
9221052
9221053
9221054
9221055
9221057
9221058
9221056
Filter tare
weight (mg)
4,303.75
4,308.05
4,313.75
4,253.15
4,227.05
4,237.20
4,214.60
4,238.45
4,237.80
4,353.20
4,197.75
4,275.75
4,318.80
4,296.35
4,300.45
4.338.00
4,293.00
4,281.55 '
4,298.40
4,290.90
4,314.00
4.311.85
4,333.30
4,308.00
4,287.15
4.353.00
4.339.70
4,330.20

4,265.30
4.284.10
4,339.85
4,291.35
4,311.80
4,332.80
4,336.20
4,334.65
4,314.75
4,270.10
Filter final
weight (mg)
4,309.50
4,313.30
4,319.05
4,252.30
4,226.30
4,237.80
4,231.90
4,246.90
4,240.40
4,354.50
4,199.80
4,276.90
4,322.45
4,303.30
4,305.05
4,341.15
4,295.85
4,293.00
4,298.75
4,295.30
4,357.25
4,313.80
4,336.45
4,313.05
4,298.40
4,363.45
4,346.15
4,333.40

4,266.10
4,285.70
4,331.15
4,296.20
4,315.65
4.335.95
4,339.70
4,337.80
4,318.00
4,273.90
Weight
difference
{mg)
5.75
5.25
5.30
-0.85
-0.75
0.60
17.30
8.45
2.60
1.30
2.05
1.15
3.65
6.95
4.60
3.15
2.85
11.45
. 0.35
4.40
43.25
1.95
3.15
5.05
11.25
10.45
6.45
3.20

0.80
1.60
-8.70
4.85
3.85
3.15
3.50
3.15
3.25
3.80
Includes only valid samples.
                                   4-5

-------
table include only those samples that were considered valid and subsequently used to
develop PM-10 emission estimates.

       Using the raw data provided in Table 4-2, the measured (i.e., blank-corrected)
PM-10 concentration was determined for the various sampling locations using the
calculation scheme outlined in Section 3.5.  In these calculations, the net sample
weight for each filter was first determined by subtracting the average filter blank value
(from Array D3) from the gross weight difference (Table 4-2). The resulting values
were then entered into Eq. 3-1, along with the applicable sampler flow rates and
operating times to obtain the measured PM-10 concentration at each location.  The
results of these calculations are provided in Table  4-3, along with any comments
relevant to the experimental data.

       Using the data shown in Table 4-3, net (i.e., upwind-corrected) PM-10
concentrations were calculated at each height by subtracting estimates from a  straignt
line fit of the measured upwind concentrations. (Note that net concentrations were
calculated only for arrays with  adequate information for integration purposes.)  Using
these net concentrations, the  net PM-10 exposure was calculated for each measure-
ment location using Eq. 3-4.  Exposure integration was then performed by the two-
step process described in Section 3.4.1 with the effective plume height (H) defined as
that height (possibly extrapolated) at which the net PM-10 concentration was zero.
Finally, PM-10 emission factors were calculated from the data using Eq. 3-6. The
results of this analysis are shown in Table 4-4 with a sample calculation for Test AY-5
provided in Appendix  D.

      Several factors should be  noted with  regard to the experimental results shown
above. First, the field blank values used to  correct for handling were generally higher
than normally would be expected for this type of sampling and were neither consis-
tently positive or negative. The influence of emissions from the diesel-powered
generator, as well as site logistics, might explain some of the high blank values, and
steps should be taken to mitigate these influences in any future testing.

      Second, the net filter catches measured during the program were generally very
low, thus enhancing systematic sampling errors. Net sample weights of approximately
1  to 10 mg for most filters (Table 4-2) are only slightly above blank correction values,
which complicates data analysis and interpretation.

      Another factor of interest involves the exposure profiles themselves.  As shown
by Table 4-4, some of the profiles (e.g., Test AY-4) are essentially flat (i.e., little
difference in exposure with height) over the first 7  m of plume height.  Flat exposures
are generally indicative of a highly undefined plume (and associated mass flux) and
poor test conditions during sample collection. The QC criteria developed for exposure
profiling (Table 3-5) are based on those conditions that MRI has found to be
necessary for a well-defined plume and associated exposure profile.
                                       4-6

-------
                                                        TABLE 4-3.  SUMMARY OF EXPERIMENTAL RESULTS*
Test Array
Test ID date ID No."
AY-3 4/11/92 U1
U2

D1



D4



AY-4 4/23/92 U1
U2
D1



D2
D4



AY-5 4/26/92 U1
U2

D2
D4



Sampling ~
height (m)
3
1.5
3
1
3
5
7
1
3
5
7
3
1.5
3
1
3
5
7
2.5
1
3
5
7
3
1.5
3
2.5
1
3
5
7
Net filter weights (mg)
Gross
catch
2.05
1.15
3.65
5.75
5.25
Void
5.30
17.30
8.45
2.60
1.30
11.25
10.45
6.45
6.95
4.60
3.15
2.85
11.45
43.25
1.95
3.15
5.05
3.15
3.25
3.80
3.20
4.85
3.85
3.15
3.50
Blank-
corrected'
2.38
1.48
3.90
6.08
5.58
Void
5.63
17.63
8.78
2.93
1.63
8.87
8.07
4.07
4.57
2.22
0.77
0.47
9.07
40.87
Nil
0.77
2.67
5.25
5.35
5.90
5.30
6.95
5.95
5.25
5.60
Sampler
flow rate
(std mVmin)"
1.31
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.28
1.03
1.35
.26
.26
.28
.52
.33
.53
.10
.38
1.16
1.27
1.16
1.16
1.30
0.948
1.21
1.14
1.19
Sampling time
(min)
274
274
274
177
177
177
177
181
181
181
181
357
357
357
63
63
63
63
108
286
103*
103*
286
348
348
348
220
220
220
220
220
Measured PM-10
concentration
(ug/ms)
6.63
4.54
12.0
28.9
26.5
—
26.7
81.8
40.8
13.6
7.57
19.4
21.9
8.44
57.6
28.0
9.55
4.91
63.1
93.4
Nil
5.42
8.05
11.9
13.3
14.6
18.5
33.3
22.4
20.9
21.4
Comments/problems
Array D1 did not run the entire test period
due to generator trip. Array D2 could not
be operated due to inadequate generator
capacity.







Arrays 01 and D2 did not operate entire
period due to generator trip. 3-m and 5-m
samplers on Array D4 failed to operate
due to generator problems. Run time for
these units estimated from available
information.







Continuing generator problems prevented
the use of Arrays D1 and D3.






•  Results of all calculations to three significant figures.
b  Array D3 could not be operated due to reoccurring generator problems which were not solved despite repeated attempts to do so.
0  Corrected by average of field blank values from Array D3.
4  All airflows based on posttest flow rate audits except for Test AY-3 which used the calibration values. Test AY-4 did not meet 'every two week* QC check for flow rate recalibration.
'  Run time estimated.

-------
                       TABLE 4-4. RESULTS OF EMISSION FACTOR CALCULATIONS'
  Run
  No.
Array
 No.
Sampler
 height
  (m)
 Net PM-10
concentration
 (ng/std m3)"
Wind
speed
(m/s)
Net PM-10
 exposure
 (ug/cm2)
 Integrated
 exposure
(m-ug/cm2)c
No. of
vehicle
passes
PM-10 emission
 factor (g/VKT)
 AY-3
 D-1
   1
   3
   5
   7
    26.9
    14.5
 3.8
(5.4)
 6.1
(6.7)
   109
    83.2
   (41.6)
     0
                                                   459
              1,175
                3.91
          D-4
           1
           3
           5
           7
               79.8
               28.8
                0
                0
                3.9
                (5.7)
                6.6
                (7.3)
            338
            178
              0
              0
               1.038
                983
                10.6
 AY-4
 D-1
   1
   3
   5
   7
    31.2
    19.6
      1.11
      0
 0.58
(1.0)
 1.2
(1.4)
     6.84
     7.41
     0.50
     0
    31.8
  220
      1.44
 AY-5     D-4
           1
           3
           5
           7
               20.4
                7.8
                4.5
                3.3
                1.4
               (1.7)
                1.9
               (2.1)
             37.7
             17.5
             11.3
              9.14
                149
                650
                                                                                                2.29
*  Parentheses denote inter/extrapolated values.

b  Net concentration calculated as difference between measured downwind and upwind concentrations.  Upwind data
   inter/extrapolated for different heights.

c  Integration scheme assumes constant value of exposure from 0 to 1 m height, with Simpson's rule used for
   integration between 1 m and effective plume height (H). Maximum effective plume height (H) assumed £ 9 m.
                                                   4-8

-------
       Finally, certain inconsistencies were noted with the individual emission factors
shown in Table 4-4.  During the first storm tested (Test AY-3), the emissions appear to
decrease with time after application (i.e., the emission factor for Array D4 is higher
than that for Array D1).  The opposite trend was observed, however,  during the
second storm (Tests AY-4 and AY-5), which showed emissions  increasing with time
after application (i.e., the emission factor for Test AY-5 is higher than for Test AY-4).
Therefore,  no constant relationship was found between the emission factors developed
in the study and the time after application of the antiskid material.

       Taking all of the above factors into consideration, it might be expected that the
emission factors shown in Table 4-4 are probably within the same order of magnitude
as the "true" PM-10 emissions from the test road.  The interpretation  of these results
is discussed further in Section 4.5.

4.3.2  Results of Chemical Analyses

       As mentioned in Section  3.3, selected filters (including blank filters) were
submitted for chemical analysis for either lead (Pb) or sodium (Ma*) and chloride (Cl~)
content. The concentration of particulate lead was determined both upwind and
downwind of the road during Tests AY-4 and AY-5 using samples collected by the two
Wedding PM-10 instruments. Similar analyses could not be  performed for Test AY-3
due to an downwind sample that was invalidated by generator problems.

       In the case of sodium and chloride, filter sets from one downwind profiler array
for each of the three tests were submitted for chemical analysis. The purpose of
these analyses was to determine the relative contribution of rock salt to the total
PM-10 emissions from the roadway.  The results of the chemical analyses performed
are summarized in Table 4-5.

      As shown by the data in  Table 4-5, the analyte mass found  on most of the
filters did not exceed the blank values for either of the three  species.  Of the few filters
that did show a net increase in analyte mass over the blank  value, the quantity
determined was generally so slight as to be negligible. There is no particular reason
for these high blank values except for the pervasive existence of all three elements in
ambient air near roadways as well as the relatively high elemental  background
associated with the filter media  used  for sample collection. Therefore, for the purpose
of this  study, it was assumed that the contribution of both particulate  Pb and NaCI to
the total PM-10 emissions from  the road was not significant and thus could be
ignored. For this reason, no specific emission factors were developed for either Pb or
NaCI in the current program.
                                      4-9

-------
                 TABLE 4-5.  RESULTS OF CHEMICAL ANALYSES
Test
Analyte No.
Lead o) AY-4

AY-5

Blank
Sodium AY-3
(Na*)


AY-4



AY-5



Blank
Chloride AY-3
(CO


AY-4



AY-5



Blank
Array
No.
U1
D2
U1
D2
D3
D4



D1



D4



D3
D4



D1



D4



D3
Sampling
height
3
2.5
3
2.5
5
1
3
5
7
1
3
5
7
1
3
5
7
1
1
3
5
7
1
3
5
7
1
3
5
7
1
Filter No.
9221035
9221046
9221057
9221051
9221011
9221020
9221021
9221022
9221023
9221031
9221032
9221033
9221034
9221052
9221053
9221054
9221055
9221059
9221020
9221021
9221022
9221023
9221031
9221032
9221033
9221034
9221052
9221053
9221054
9221055
9221059
Total
paniculate
mass on
filter (mg)a
11.25
11.45
3.15
3.20
0.60
1730
8.45
2.60
1.30
6.95
4.60
3.15
2.85
4.85
3.85
3.15
3.50
0.80
17.30
8.45
2.60
1.30
6.95
4.60
3.15
2.85
4.85
3.85
3.15
3.50
0.80
Mass of
analyte per
filter (mg)b
0.0027
0.0011
0.0011
0.0008
0.0012
14.37
12.70
13.63
12.95
12.20
9.999
11.73
11.82
12.40
11.85
13.10
12.10
10.53
1.067
0.887
0.456
1.166
1.044
1.049
1.138
1.021
1.156
1.005
1.131
1.059
1.140
"  Not blank-corrected.

b  Includes background frc
   boldface type.
-lass fiber substrate.  Values above blank inr   red in
                                     4-10

-------
4.4    RESULTS OF ANCILLARY SAMPLING AND ANALYSIS

4.4.1  Antiskid Material Samples

       Samples of both the abrasive/rock salt mixture and straight rock salt were
collected from the MnDOT storage piles and/or spreader trucks according to the
procedures outlined in Appendix A.  These samples were then analyzed for moisture
and silt content as well as for the various key material properties listed previously in
Table 3-6. The analytical results obtained for the antiskid material samples collected
are summarized in Table 4-6.
                  TABLE 4-6. PROPERTIES OF ANTISKID MATERIAL SAMPLES"




Sample
origin
Storage
pile






Sample
composition
90:10
abrasive/
rock salt
mix6


Average
moisture
content
(wt %)
3.6




Modified
LA
abrasion
loss
(wt %)
1.9





Average
silt
content
(wt %)
2.1





Average
Vickers
hardness
(kg/mm2)
461



Weighted
average
particle
shape index
(dimension-
less)
7.23





Average
percent
insoluble
matter
N/A



Storage
pile
Spreader
truck

Spreader
truck

Rock salt

90:10
abrasive/
rock salt mix
8020
abrasive/
rock salt mix
0.31 — 0.56 —

3.9 — 1.4 —


4.5 — 1.3 —


— 1.28

— N/A


— N/A


    '  N/A «= Not applicable.
      — indicates no data available.

    b  This material is considered as "questionable" quality according to the criteria specified in
      EPA-450/3-90-007.
       Using the data in Table 4-6 for the abrasive/salt mixture, a comparison was
made of its properties with the material selection criteria listed previously in Table 1-1.
This comparison showed that the abrasive/salt mixture (storage pile sample) used by
MnDOT did not meet three out of the four key parameters thought to be important for
high material durability.  For this reason, the abrasive/salt mixture analyzed in the
current program was classified as "questionable" quality,  based on the results of
previous MRI testing (Kinsey, 1991).

      With regard to the rock salt samples obtained from the MnDOT storage pile, a
preliminary selection criterion of < 2% insoluble matter was also established by Kinsey

                                      4-11

-------
 (1991). As noted from Table 4-6, the average insoluble content of the two salt
 samples analyzed is well below the 2% value and thus easily meets the above
 criterion for chloride deicers.

 4.4.2  Road Surface Sampling

       Road surface sampling was performed throuahout the period that field testing
 was attempted in Duluth, Minnesota.  Samples we-  collected and analyzed to deter-
 mine silt loading using the procedures outlined in Appendix A. Surface samples were
 collected from both the driving (right) and passing (left) lane on  each side of the
 highway near the air sampling site with the majority of the samples collected from the
 northbound lanes. (Recall that a 90:10 [weight percent] sand/salt mix was used on
 the northbound lanes and straight rock salt was used on the southbound lanes as part
 of the SHRP project—see Section 2.2.) The silt-loading values  obtained from the
 various samples collected are shown in Table 4-7.

       Using the data provided in Table 4-7, a "silt-loading history" was developed for
 both the northbound and southbound lanes during the entire period that samples were
 collected.  These histories are shown in Figures 4-1  and 4-2.  Also shown on these
 figures are the storm event periods and the days on which source testing was per-
 formed.  (Note that a fairly substantial storm hit Duluth on February 24, which was
 also included on Figures 4-1 and 4-2 as a starting point.)

       As shown by the above figures, the silt loadings on US 53 tended to be
 relatively low, as is typical for high-speed roadways. Also, the northbound lanes
 (which received the abrasive/salt mixture) tended to  have generally higher silt loadings
 as compared to the southbound lanes (which  received straight rock salt). In both
 cases, the  loadings generally dropped off fairly rapidly after a storm event, which was
 also verified by on-site observation.

       Using the silt-loading data shown in Table 4-7, PM-10 emission factor estimates
 were calculated for the various samples using the AP-42 predictive model provided
 previously as Eq. 1-1.  Emission factors were predicted  for both northbound and
 southbound lanes in terms of g/VKT.  The results of these calculations  are shown in
 Table 4-8.

      As indicated by Table 4-8, the predicted emission factors were generally low,
 ranging from less than 1 to a maximum of 4 g/VKT.  The highest  emission factors
 (resulting from the highest silt loadings) were for samples collected from the
 northbound lanes after a fairly major storm event occurring on February 24,1992.
 After that time, the predicted emissions generally dropp    o < 1 g/VKT until the  next
 storm in early April. This decrease in loading would be    pected due to the effects of
traffic, which tends to reestablish a silt-loading "equilibrium" on the road surface after
 external influences (e.g., the application of antiskid material) have been eliminated.
                                      4-12

-------
                              TABLE 4-7. RESULTS OF ROAD SURFACE SAMPLING'
Date
2/26



2/28



3/2



3/11



3/19
4/1

4/22


4/24


Lane
sampled
NB-Driving
NB-Passing
SB-Driving
SB- Passing
NB-Driving
NB-Passing
SB-Driving
SB-Passing
NB-Driving
NB-Passing
SB-Driving
SB-Passing
NB-Driving
NB-Passing
SB-Driving
SB-Passing
NB-Driving
NB-Driving
NB-Passing
NB-Driving
NB-Passing
NB-Tuming
NB-Driving
NB-Passing
SB-Driving
and passing
Sample
bag No.
9017
9019
9711-23
9711-22
9012
9010
9013
9011
9016
9022
9020
9021
9008
9014
9015
9018
63
62
119
74
75
61
98
96
93
Road surface area
sampled
ft*
576
576
576
576
576
576
576
576
576
576
576
576
576
576
576
576
575
1,150
1,150
646
646
53.3
648
648
1.300
m2
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.5
53.4
107
107
60.0
60.0
4.95
60.2
60.2
121
Sample
mass
(0)
336.8
114.4
306.5
104.4
147.1
1.4
130.0
204.1
92.3
113.3
23.1
162.0
45.5
63.3
114.0
112.0
2.3
76.0
1575
44.0
314.0
1107.1
181.1
161.2
65.7
Total surface loading
oz/yd2
0.186
0.0628
0.169
0.0575
0.0811
0.0008
0.0717
0.113
0.0510
0.0625
0.0127
0.0894
0.0251
0.0348
0.0628
0.0616
0.00127
0.0209
0.0437
0.0216
0.154
6.61
0.0885
0.0790
0.0160
gin?
6.30
2.13
5.73
1.95
2.75
0.0262
2.43
3.82
1.73
2.12
0.432
3.03
0.850
1.18
2.13
2.09
0.0431
0.710
1.48
0.733
5.23
224
3.00
2.68
0.543
Silt
content
(wt %)
16.5
23.5
9.23
13.9
12.4
N/A
18.3
7.74
4.05
3.12
8.85
4.03
23.5
13.9
9.77
16.9
N/A
11.1
10.6
8.43
10.4
4.96
4.97
4.96
9.68
Road surface silt
loading
oz/yo*
0.0307
0.0148
0.0156
0.00799
0.0101
0.0008
0.0131
0.00870
0.00206
0.00195
0.00112
0.00360
0.00590
0.00484
0.00613
0.0104
0.00127
0.00233
0.00460
0.00183
0.0160
0.327
0.00442
0.00392
0.00156
g/m2
1.04
0.501
0.529
0.271
0.341
0.0262"
0.445
0.295
0.0701
0.0661
0.0382
0.122
0.200
0.164
0.208
0.354
0.0431"
0.0788
0.156
0.0618
0.544
11.1
0.150
0.133
0.0526
* Calculations to three significant figures.



" Assumed to be 100% silt.
                                                 4-13

-------
      SILT LOADING HISTORY - NORTHBOUND LANES
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
T
-
-
-
-
-
-
-
-
j 	





•



i









li
24 26 28
February









HI

KEY
Q = Driving Lane y = Storm ! vent
[] = Passing Lane • = Emissions Testing









I I 1 1
24 6 8 10











I I I 1$ I ^
T
•
(abode








i i i I I
12 14 16 18 20 22 24 26 28 30
March

T TT
• • •
i)








i i i i i i i i i i i




Ml i i i
24 6 8 10 12 14 16 18 20 22 24 26 28 30
April
Figure 4-1. Silt-loading history for northbound lanes (sand/salt mix).

-------
                                   SILT LOADING HISTORY - SOUTHBOUND LANES
en
    1.1
    1.0
oT  0-9
I  °'8
.e  °7
|  0.6
£  0.5
    0.4
    0.3
    0.2
    0.1
             li
       24 26 28
        February
I
                                     KEY
                          0= Driving Lane y =
                          fj = Passing Lane • =
                          | = Both Lanes
                                      Storm Event
                                      Emissions Testing
                                             T
                                             •
                                           (aborted)
                                     '   i   i   t   i   i   i	1	1	L
                     2  4  6  8
             10 12 14 16 18 20 22 24 26  28  30
                     March
                                                                  j	L
          		l__i	L
                                                                                                           i   i   i
                                                                                                                   I
24  6  8  10 12 14 16 18 20 22  24 26 28 30
                     April                     I
                                 Figure 4-2. Silt-loading history for southbound lanes (rock salt only).

-------
        TABLE 4-8. PREDICTED PM-10 EMISSION FACTORS FOR MEASURED
                           SURFACE SILT LOADINGS8
Date
2/26

2/28

3/2

3/11

3/19
4/1
4/22
4/24

Lane sampled
NB — Driving
NB — Passing
SB— Driving
SB — Passing
NB — Driving
NB — Passing
SB — Driving
SB — Passing
NB — Driving
NB— Passing
SB — Driving
SB — Passing
NB — Driving
NB— Passing
SB — Driving
SB — Passing
NB — Driving
NB— Driving1"
NB— Passing6
NB— Driving"
NB— Passingb
NB — Driving
NB — Passing
SB — Driving and
Passing
Road surface silt
loading (g/m2)
1.04
0.501
0.529
0.271
0.341
0.0262
0.445
0.295
0.0701
0.0661
0.0382
0.122
0.200
0.164
0.208
0.354
0.0431
0.0788
0.156
0.0618
0.544
0.150
0.133
0.0526
Predicted PM-10 emission
factor (g/VKT)
4.10
2.28
2.38
1.39
1.68
0.215
2.08
1.50
0.473
0.452
0.291
0.738
1.10
0.935
1.13
1.73
0.321
0.520
0.900
0.428
2.44
0.870
0.790
0.376
* PM-10 emission factors predicted using Eq. 1-1 and silt loadings shown in Table 4-7.

6 Loadings measured after application of antiskid material.
                                    4-16

-------
      As also shown, the estimated emission factors (Table 4-8) are of the same
approximate magnitude as the test results provided in Table 4-4. This lends  some
additional credibility to the data obtained by exposure profiling in the current study.

4.5   DISCUSSION OF  RESULTS

      As shown by the above test results, the measured emission factors varied from
1 to 11 g/VKT for the three  tests conducted.  The data were not sufficient,  however, to
develop any specific correlation between the  measured emission factors and  source
parameters such as quantity of antiskid material applied, time since application, and
silt loading.  The only general observation that can be made from the data is  that the
PM-10 emissions appear to increase with the amount of antiskid material applied
during the two storms tested (i.e., the emission factors for Test AY-3 are higher than
for Tests AY-4 and AY-5).  (Also note that Test AY-5 had the most acceptable wind
conditions of the three tests conducted and thus represents the  most reliable  emission
factor determined in the study.)  Otherwise, the results do not seem to follow a
consistent relationship. A major cause for the lack of association was the marginal
test conditions discussed above.

      A  comparison of the  measured emission factors (Table 4-4) with those
predicted by the current AP-42 equation (Table 4-8) also  yields some interesting
results.  With only a  relatively few exceptions, most of the measured factors are  higher
than those which would be  predicted from silt-loading values alone.  In the case  of the
test (i.e., northbound) lanes themselves, the measured factors either equal or exceed
the lane average emissions predicted by the AP-42 equation, except for the
February 26 samples that were collected after a fairly major storm. This suggests that
the application of antiskid material results in a short-term increase in PM-10 emissions
from the  roadway in an amount greater than that which would be predicted by silt
loading alone. Although  no definitive assessment of this  increase can be made  based
on available data, the PM-10 emissions could be almost a factor of three higher (i.e.,
11 vs. 4 g/VKT) than would be predicted from the silt loading. This proportional
increase  in emissions is considerably higher than the precision factor of approxi-
mately 2  for the current AP-42 predictive equation and thus worthy of further study
(Cowherd and Englehart, 1984).  Therefore, a more explicit evaluation is needed to
determine the  exact extent  of the emissions increase due to the application of antiskid
material to paved roadways.
                                      4-17

-------
                                  SECTION 5

                             QUALITY ASSURANCE
       An independent evaluation of the field and analytical activities on this work
 assignment was performed by the Midwest Research Institute Quality Assurance
 Manager (QAM). The evaluation procedure included the review of the field and
 analytical data.  The field work was performed by Midwest Research Institute and the
 elemental analyses were conducted by Galbraith Laboratories, Inc., Knoxville,
 Tennessee.

 5.1    PERFORMANCE AUDIT

       The analytical laboratory, through its internal quality control program, analyzed
 replicate performance samples using a NIST standard reference material (SRM 1643c)
 for lead. The results for the performance audit samples were 99.4% and 102.8%.
 The precision for the performance audit samples, calculated as a percent range, was
 3.36%. These results are within the data quality objectives as stated in trie Quality
 Assurance Project Plan (QAPjP).  In addition to the performance audit samples, the
 laboratory analyzed quality control samples (method spikes) prepared at a theoretical
 concentration of 2.0 ng/mL for each analyte. The average results for the method
 spikes were 101% for the sodium analyses, 103% for the chlorine analyses, and 108%
 for the lead analyses.

 5.2    DATA AUDIT

      Two data audits were performed for this work assignment, one on the field data
 and the second on the analytical data. A summary of the audit findings are given in
 the following subsections.

 5.2.1  Field  Data

      The sampling procedures followed in this field testing program were subject to
 quality assurance/quality control (QA/QC) guidelines.  As a part of this program,
 quality assurance audits were performed to demonstrate that the measurements were
 made within acceptable control conditions for paniculate source sampling and to
 assess the reliability of the field data with respect to the established criteria.  The use
 of specially-designed reporting forms for sampling and quality control data obtained in
the field aided in the auditing procedure.


                                      5-1

-------
      Two source activity factors pointed out by the Work Assignment Leader (and
 mentioned previously in this report) that had major effects on the quality of the field
 data collected were the atypical amount of antiskid material applied by MnDOT and
 the antiskid material itself. These are discussed in Sections 4.1 and 4.4.1.

      The quality control criteria established for this program are given in Table 3-3,
 "Quality Control Procedures for Sampling Flow Rates," and Table 3-4, "Quality Control
 Procedures for Sampling  Equipment." During this work assignment, the  calibration of
 the equipment was checked by the field personnel to ensure that the equipment was
 prope  / calibrated.

      The criteria used to define the unacceptable conditions for the collection of
 reliable test data are given in Table 3-5, "Criteria for Suspending or Terminating  a
 Test."  Because of the mild winter in  Duluth, adherence to the criteria for suspending
 the test was not always followed.  However, each decision to conduct sampling was
 discussed in advance with the EPA project officer. The reliability of the data collected
 during these periods needs to be carefully reviewed  with  respect to the conditions at
 the time of the tests.

 5.2.2 Laboratory Data

      A data audit was conducted to evaluate the analytical data generated. The
 quality of the analytical data was evaluated  against the QA indicators for the
 measured data presented in the QAPjP, the analytical methodology, and the project
 Standard Operating Procedures (SOPs).

      The samples (filters) were initially analyzed using MR! SOP EET-610 to
 determine the weight change between prefield weights and postfield weights. The
 samples were equilibrated for 24 h in a clean, controlled temperature and humidity
 room. The filters were analyzed as described in the SOP and were within the data
 quality indicators as given in the SOP and the QAPjP.

      The air samples analyzed for lead were digested using EPA SW-846
 Method 3050 and analyzed using EPA 600  Series Method 239.2.  The air samples
 analyzed for sodium and chloride were extracted using the leaching procedure
 described in 40 CFR 50, Appendix G. The  analytical procedure used for chlorine was
 EPA Series 500 Method 300.1; Method 273.1 was used for sodium. The procedures
were follower as described, and the  associated quality control data met the
 measurement requirements of the analytical procedures and the QAPjP

£     DATA ASSESSMENT

      Although the analytical data generated met the quality control criteria
esiaolished for this work assignment, the field data were  affected by environmental
factors, many of which were beyond  the control of the Work Assignment Leader.


                                      5-2

-------
Since the samples were collected under environmental conditions that did not meet all
of the applicable quality control criteria, the quality of the data was adversely affected.
Because there was a failure of field equipment that resulted in the loss of samples, the
completeness of the sample collection process also affected the reliability of the data.
Therefore, the data generated for this work assignment should be considered only as
an estimation of PM-10 emissions from the roadway.

5.4   REPORT REVIEW

      The document was reviewed for consistency in reporting the analytical data,
and the report was found to reflect the analytical data generated for this activity.
                                      5-3

-------
                            SECTION 6

                      STUDY CONCLUSIONS


The following conclusions were reached as a result of the current study:

1.    Because the measurements conducted were performed under difficult
      environmental conditions that did not meet all of the applicable QC
      criteria for exposure profiling, data quality was adversely affected.

2.    Although the emission factors determined in the program ranged from 1
      to 11  g/VKT for the three tests conducted, these values should be used
      with extreme caution due to the  lack of suitable test conditions during
      sample collection and resulting poor data quality.

3.    The contribution of both particulate lead (Pb) and sodium chloride (NaCI)
      to the total PM-10 emissions from the road did not appear to be
      significant based on the limited chemical analyses performed in the
      program.

4.    The dry silt loadings determined on US 53 tended to:  (a) be relatively
      low; and (b) drop off rapidly after a storm event which is typical of high-
      speed roadways. The amount of silt available for resuspension was also
      found to be very low due to the  minimal application of antiskid material
      during the experimental program.

5.     Emission factors determined in the program were generally higher than
      those predicted by the current AP-42 equation, based on road surface
      silt loading.  The magnitude of this increase could not be definitively
      determined from available data,  but it may be almost a factor of 3.
                                6-1

-------
                                 SECTION 7

                                REFERENCES
American Society for Testing and Materials. Standard Method of Preparing Coal
Samples for Analysis.  Method D 2013-72. Annual Book of ASTM Standards.  1977.

Cowherd, C., Jr., and P. J. Englehart.  Paved Road Participate Emissions:  Source
Category Report.  EPA-600/7-84-077 (NTIS PB84-223734), U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, July 1984.

Davies, C. N.  The Entry of Aerosols in Sampling Heads and Tubes.  British Journal of
Applied Physics. 2:921, 1968.

Grelinger, M. A., G. Muleski, J. S. Kinsey, C. Cowherd, Jr., and D. Hecht. Gap Filling
PM-10 Emission Factors for Selected Open Area Dust Sources. EPA-450/4-88-003
(NTIS PB88-196225), U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina, February 1988.

Kinsey, J. S., P. Englehart, M. A. Grelinger, K. Connery, C. Cowherd, Jr., and
J. Jones. Guidance Document for Selecting Antiskid Materials Applied to Ice- and
Snow-Covered Roadways. EPA-450/3-90-007 (NTIS  PB90-183658), U.S.
Environmental Protection Agency, Research Triangle  Park, North Carolina, January
1990.

PEDCo Environmental,  Inc. Denver Demonstration Study. Contract No. C297337,
Colorado Division of Air Pollution Control, Denver, Colorado, October 1981.

Pyle, B. E., and J. D. McCain. Critical Review of Open Source Particulate Emission
Measurements: Field Comparison.  EPA-600/2-86-072 (NTIS PB86-239789), U.S.
Environmental Protection Agency, Research Triangle  Park, North Carolina, August
1986.

RTP Environmental Associates.  Street Sanding Emissions and Control Study.
Boulder, Colorado, July 1990.

U.S. Environmental Protection Agency. Compilation of Air Pollutant Emission Factors,
Volume I: Stationary and Area Sources. AP-42 (4th Edition plus Supplements A-D),
Research Triangle Park, North Carolina, T991.


                                     7-1

-------
Von Lehmden, D. J., and C. Nelson.  Quality Assurance Handbook for Air Pollution
Measurement Systems, V-'-jme II, Ambient Air Specific Methods.  EPA-600/4-77-027a
(NTIS PB-273518), U.S. E   ironmental Protection Agency, Research Triangle Park,
North Carolina, May 1977.
                                    7-2

-------
          APPENDIX A






MATERIAL SAMPLING AND ANALYSIS
             A-1

-------
                                 APPENDIX A

                     MATERIAL SAMPLING AND ANALYSIS


A.1    SAMPLING PROCEDURES FOR ANTISKID MATERIALS

       For stockpiled antiskid materials, the following steps were used to collect each
sample:

       1.     Sketch plan and elevation views of the pile to be sampled. Indicate if
             any portion is inaccessible.  Use the sketch to plan where the
             N increments will be taken by dividing the perimeter into N-1 roughly
             equivalent segments.

             a.     For a large pile, collect a minimum of 10 increments as  near to
                   the mid-height of the pile as practical.

             b.     For a small pile, a sample should consist of a minimum  of
                   6 increments evenly distributed among the top, middle, and
                   bottom.

             "Small" or "large" piles,  for practical purposes, may be  defined as those
             piles which can or cannot, respectively, be scaled by a  person carrying a
             shovel and pail.

       2.     Collect material with a straight-point shovel or a small garden spade.
             Take increments from the portions of the pile which most recently had
             material added and  removed. Collect the material with  a shovel to a
             depth of 10 to 15 cm (4 to 6 in). Do not deliberately avoid larger  pieces
             of aggregate present on  the surface.  Store the increments in a clean,
             labeled container of suitable size (such as a metal or plastic 19-L [5-gal]
             bucket) with a scalable polyethylene liner.

       3.     Record the required information on the sample collection sheet
             (Figure A-1). Note the space for deviations from the summarized
             method.

      The sample mass collected should be at least 5 kg (10 Ib). When most
materials are sampled with 10 increments, a sample of at least 23 kg (50 Ib) is typical.
Note that storage pile samples usually require splitting to a size more amenable to
moisture and silt analysis.

       For the abrasive/salt mixture, grab samples were collected directly from the rear
of the distributor truck while the vehicle was stationary.  Samples were collected from
                                      A-3

-------
                 SAMPLING DATA FOR  STORAGE  PILES
Date Collected
Recorded by
Type of material sampled	

Sampling location*

METHOD:

  1. Sampling  device: pointed shovel  (hollov sampling  tube  if
     inactive pile is to be sampled)
  2. Sampling depth:
     For material handling of  active piles:  10-15 cm  (4-6  in)
     For material handling of  inactive piles:  1 m  (3  ft)
     For wind erosion samples:  2.5  cm  (1  in)  or depth  of  the
     largest particle (whichever is less)
  3. Sample container: bucket  with a sealable  liner
  4. Gross sample specifications:
     For material handling of active or inactive piles:  minimum of
     6  increments  with  total  sample  weight of 5  kg  (10  Ib)  [10
     increments totalling 23 kg  (50. Ib) are  recommended]
     For wind erosion samples: Minimum of  6  increments  with  total
     sample weight of 5 kg  (10 Ib)

Refer to procedure described  in Section 4 of "Open  Source  PM-10
Method Evaluation" for more detailed instructions.

Indicate any deviations from the above:    	
SAMPLING DATA COLLECTED:
Sample
No.






Time






Location* of
Sample Collection






Device
Used
S/T **






Depth






Mass
of Sample





••
  *  Use  code  given  on  plant  or  area  map  for  pile/sample
     identification.  Indicate each sampling location on map.
                 Figure A-1. Example data form for storage piles.
                                A-4

-------
as many loads as possible with the samples split (if required) prior to laboratory
analysis.  The following describes the sample splitting procedure used.

      The main principle in sizing the laboratory sample for subsequent silt analysis is
to have sufficient coarse and  fine portions to be representative of the material and to
allow sufficient mass on each sieve so that the weighing is accurate. A laboratory
sample of 400 to 1,600 g is recommended because of the scales normally available
(1.6 to 2.6 kg capacities).  A  larger sample than this amount may produce "screen
blinding" for the 20 cm (8 in)  diameter screens normally available for silt analysis.
Screen blinding can also occur for small samples of finer texture. Finally, the sample
mass should be such that it can be spread out in a reasonably sized drying pan to a
depth of < 2.5 cm (1 in).

      Two methods are recommended for sample splitting—riffles and coning and
quartering.  Since a riffle was used in the current study, only this procedure is
described.

       Figure A-2 shows two  riffles for sample division.  Riffle slot widths  should be at
least three times the size of the largest aggregate in the material being divided.  The
following quote from ASTM Standard Method  D2013-72 describes the use of the riffle
(ASTM, 1977).

             "Divide the gross sample by using a riffle.   Riffles properly used
      will reduce sample variability but cannot eliminate it.  Riffles are shown in
      Figure A-2.  Pass the material through the riffle from a feed scoop, feed
      bucket, or riffle pan having a lip or opening the full length  of the riffle.
      When using  any of the above containers to feed the riffle, spread the
      material evenly in the container, raise the container, and hold it with its
      front edge resting on top of the feed chute, then slowly tilt it so that the
      material flows in a uniform stream through the hopper straight down over
      the center of the riffle into all the slots, thence into the riffle pans, one-
      half of the sample being collected in a pan.  Under no circumstances
      shovel the sample into the riffle, or dribble into the riffle from a small-
      mouthed container. Do not allow the material to  build up  in or above the
      riffle slots. If it does not flow freely through the slots, shake or vibrate
      the riffle to facilitate even flow (ASTM, 1977)."

A.2   ANALYSIS OF ANTISKID MATERIAL SAMPLES

      All  antiskid materials were oven dried to determine moisture content prior to silt
analysis.  The following procedure was used for this purpose:
                                       A-5

-------
Feed Chute
                SAMPLE DIVIDERS (RIFFLES)
 Rolled
 Edces
      Riffle Sampler

           (b)
     Riffle Bucket and
Separate Feed Chute Stand
           (b)
                                                    ssv r*
                  Figure A-2.  Sample riffle dividers.
                                A-6

-------
       1.     Preheat the oven to approximately 110°C (230°F).  Record oven
             temperature.

       2.     Record the make, capacity, and smallest division of the scale.

       3.     Weigh the empty laboratory sample containers which will be placed in
             the oven to determine their tare weight. Weigh containers with the lids
             on if they have lids.  Record the tare weight(s). Check zero before each
             weighing.

       4.     Weigh the laboratory sample(s) in the container(s). For materials with
             high moisture content, ensure that any standing moisture is included in
             the laboratory sample container.  Record the combined weight(s). Check
             zero before each weighing.

       5.     Place sample in oven and dry overnight. Materials composed of
             hydrated minerals or organic material like coal and certain soils should
             be dried for only 11/2 h.

       6.     Remove sample container from oven and (a) weigh immediately if
             uncovered, being careful of the hot container; or (b) place the tight-fitting
             lid on the container and let cool before weighing.  Record the combined
             sample and container weight(s).  Check zero reading on the balance
             before weighing.

       7.     Calculate the moisture as the initial weight of the sample and container
             minus the oven-dried weight of the  sample and container divided by the
             initial weight of the sample alone. Record the value.

       8.     Calculate the sample weight to be used in the silt analysis as the oven-
             dried weight of the sample and container minus the weight of the
             container.  Record the value.  (See Figure A-3.)

A.3    PAVED ROAD  SURFACE SAMPLING

       In comparison to unpaved road sampling,  planning for a paved road sample
collection exercise necessarily Involves greater consideration as to types of equipment
to be used.  Specifically, provisions must be made to accommodate the characteristics
of the vacuum cleaner chosen.  For example, paved road samples are collected by
cleaning the surface using a vacuum  cleaner with "tared" (i.e., weighed before use)
filter bags.  "Stick broom" vacuums use relatively small, lightweight filter bags, while
bags for "industrial-type" vacuums are bulky and heavy.  Stick brooms are thus well
suited for collecting samples from lightly loaded road surfaces because the mass
collected is usually several times greater than the bag tare weight.  On the other hand,
the larger industrial-type vacuum bags are not only easier to use on heavily loaded


                                      A-7

-------
                           MOISTURE ANALYSIS
Date:
Sample No:
Material:
Split Sample Balance:
  Make	
  Capacity,
  Smallest Division
Total Sample Weight:
(Exd. Container)
Number of Splits:	
Split Sample Weight (before drying)
Pan + Sample:	
Pan:	
Wet Sample:
                                    By:
                                    Oven Temperature:
                                    Date In	
                                    Time In	
Date Out
Time Out
                                    Drying Time
                                    Sample Weight (after drying)
                                    Pan + Sample:	'
                                    Pan:	
                                    Dry Sample:
                                    MOISTURE CONTENT:
                                      (A) Wet Sample Wt_
                                      (B) Dry Sample Wt
                                      (C) Difference Wt  __
                                       0x100
                                          A
       % Moisture
                  Figure A-3. Example moisture analysis form.
                                   A-8

-------
roads but also can be more readily used to aggregate incremental samples from
several road surfaces.  In this study, an industrial-type vacuum cleaner was used.

      The following steps describe the collection method used for the individual
samples collected:

      1.     Ensure that the site offers an unobstructed view of traffic and that
            sampling personnel are visible to drivers. If the road is heavily traveled,
            use one crew member to "spot" and route traffic safely around another
            person collecting the surface sample (increment). (Note that a vehicle-
            mounted arrow board was also used in the  study as an extra safety
            precaution.)

      2.     Using string  or other suitable markers, mark the sampling width across
            the road.  (WARNING:  Do not mark the collection area with a chalk line
            or in any other method likely to introduce fine material into the sample.)
            The widths may be varied between 0.3 m (1 ft) for visibly  dirty roads and
            3 m (10 ft) for clean roads.  When using an industrial-type vacuum to
            sample lightly  loaded roads, a width greater than 3  m (10 ft) may be
            necessary to meet sample specifications unless increments are being
            combined.

      3.     If large, loose  material is present on the surface, it should be collected
            with a whisk broom and dustpan.  NOTE:  Collect material only from the
            portion of the road over which the  wheels and carriages routinely travel
            (i.e., not from berms or any  "mounds" along the road centerline).  On
            roads with painted side markings, collect material "from white line to
            white line" (but avoid any centerline mounds).  Store the swept material
            in a clean, labeled container of suitable  size (such as a metal or plastic
            19 L [5 gal] bucket) with a scalable polyethylene liner.  Increments of the
            same sample may be mixed within the container.

      4.     Vacuum sweep the collection area using a portable vacuum cleaner fitted
            with an empty  tared (i.e., preweighed) filter bag.  NOTE:  Collect material
            only from the portion of the road over which the wheels and carriages
            routinely travel (i.e., not from berms or any  "mounds" along the road
            centerline). On roads with painted side  markings, collect  material "from
            white line to  white line" (but avoid  centerline mounds).  The  same filter
            bag may be  used for different increments for one sample. For heavily
            loaded roads, more than one filter bag may be required for a sample
            (increment).

      5.     Carefully remove the bag from the vacuum  sweeper and check for tears
            or leaks. If necessary,  reduce samples  from broom sweeping to a size
            amendable for analysis (see Section A.1).  Seal broom-swept material in
                                      A-9

-------
            a clean, labeled plastic jar for transport (alternatively, the swept material
            may be placed in the vacuum filter bag).  Fold the unused portion of the
            filter bag, wrap a rubber band around the folded bag, and store the bag
            for transport.

      6.    Record the required information on the sample collection sheet
            (Figure A-4).

      Broom-swept samples (if collected) should be at least 400 g (1 Ib) for silt and
moisture analysis (see Section A.2). The vacuum-swept sample should be at least
200 g (0.5 Ib); in addition, the exposed filter bag weight should be at least 3 to 5 times
greater than the weight of the empty filter bag.  Additional increments should be taken
until these sample mass goals have been achieved.

A.4   ANALYSIS PROCEDURES FOR PAVED ROAD SAMPLES

      Paved road samples are not normally oven dried because vacuum filter bags
are used to collect the samples.  After the sample has been recovered by dissection
of the bag, it is  combined with any broom-swept material for silt analysis.  The
following procedure was used for sample analysis.

      For the paved road samples, the broom-swept particles and the vacuum-swept
dust are individually weighed on a beam balance. The broom-swept particles are
weighed in a container. The vacuum-swept dust is weighed in the vacuum bag, which
was tared prior to sample collection. After weighing the sample to calculate total
surface dust loading on the traveled lanes, broom-swept particles and the vacuum-
swept dust are combined. The composite sample is usually small and probably will
not require splitting in preparation for sieving.  The following steps were followed to
analyze the resulting surface sample:

      1.     Select the appropriate 20 cm (8-in) diameter, 5 cm (2-in) deep sieve
            sizes.  Recommended U.S. Standard Series sizes are:  3/8 in, No. 4,
            No. 40, No. 100, No. 140, No. 200, and a pan. Comparable Tyler Series
            sizes can also be utilized.  The No. 20 and the No. 200 are mandatory.
            The others can be varied if the recommended sieves are not available or
            if buildup on one particular sieve during sieving indicates that an
            intermediate sieve should be inserted.

      2.     Obtain a mechanical sieving device such as vibratory shaker or a Roto-
            Tap without the tapping function.

      3.     Clean the sieves with compressed air and/or a soft brush. Material
            lodged in the sieve openings or adhering to the sides of the sieve should
            be removed (if possible) without handling the screen roughly.
                                     A-10

-------
Date Collected
 SAMPLING DATA FOR PAVED ROADS

	           Recorded by
Sampling location*
                       No.  of Lanes
Surface type  (e.g., asphalt, concrete, etc.)

Surface condition  (e.g., good, rutted, etc.)
* Use code given on plant  or road map for  segment identification.
Indicate sampling location on map.

METHOD:

  1. Sampling  device:  portable  vacuum  cleaner  (whisk broom and
     dustpan if heavy loading present)
  2. Sampling  depth:  loose surface material  (do-not sample  curb
     areas or  other untravelled  portions of the  road)
  3. Sample  container:   tared  and  numbered vacuum  cleaner  bags
     (bucket with sealable liner if heavy  loading present)
  4. Gross sample specifications:  Vacuum  swept  samples  should be
     at least  200 g  (0.5 Ib),  with the exposed  filter bag weight
     should be at  least 3  to 5  times greater than the  empty bag
     tare weight.

Refer to procedure  described in Section  3 of "Open Source PM-10
Method Evaluation" for more detailed instructions.
Indicate any deviations  from the  above:
SAMPLING DATA COLLECTED:
Sample
No.





Vacuum Bag
ID





Tare Wgt (g)





Surface Area
Sampled





Time





Mass of
Broom-Swept
Sample +





  +  Enter "0" if no broom sweeping  is performed.
              Figure A-4.  Example data form for paved roads.
                               A-11

-------
4.    Obtain a scale (capacity of at least 1,600 g or 100 Ib) and record make,
      capacity, smallest division, date of last calibration, and accuracy.

5.    Weigh the sieves and pan to determine tare weights.  Check the zero
      before every weighing. Record weights.

6.    After nesting the sieves in decreasing order with pan at the bottom,
      transfer dried laboratory sample (preferably immediately after moisture
      analysis, as applicable) into the top sieve.  The sample should weigh
      between ~ 400 and 1,600 g (0.9 to 3.5 Ib). This amount will vary for
      finely textured materials; 100 to 300 g may be sufficient when 90% of the
      sample passes a No. 8 (2.36 mm) sieve.  Brush fine material adhering to
      the sides of the container into the top sieve and cover the top sieve with
      a special lid normally purchased with the pan.

7.    Place nested sieves into the mechanical sieving device and sieve for 10
      min. Remove pan containing  minus No. 200 and weigh.  Repeat the
      sieving in 10-min intervals until the difference between  two successive
      pan sample weighings (where the tare weight of the pan has been
      subtracted) is less than 3.0%.  Do not sieve longer than 40 min.

8.    Weigh each sieve and its contents and record the weight.  Check the
      zero reading on the balance before every weighing.

9.    Collect the  laboratory sample and place the sample in a separate
      container if further analysis is expected.

10.   Calculate the percent of mass less than the 200 mesh  screen (75 u.m
      physical diameter).  This is the silt content (see Figure  A-5).
                                A-12

-------
                              SILT ANALYSIS
Date	
Sample No:
Material:
Spilt Sample Balance:
Make
3y  	
Sample Weight (after drying)
Pan + Sample:	
Pan:	
Capacity	
Smallest Division
Dry Sample: .

Final Weight:
                                  snt = Ngt Weih  <200
                                          Total Net Weight
                              100 =
               SIEVING
nme: Start
Weiant(Pan Only)
Initial (Tare}: 1
20 min: I
30 min: I
4Q min:



Screen
3/8 in.
4 mesh
10 mesh
20 mesh
40 mesh
100 mesh
14Q mesh
200 mesh
Pan
Tare Weight
(Screen)









Final Weight
(Screen + Sample)









Net Weight (Sample)









%









                   Figure A-5. Example silt analysis form.
                                 A-13

-------
                   APPENDIX B
MODIFICATIONS TO ASTM METHODS FOR LA ABRASION LOSS
              AND VICKERS HARDNESS
                      B-1

-------
                                              APPENDIX B

            MODIFICATIONS TO ASTM METHODS FOR LA ABRASION  LOSS
                                     AND VICKERS  HARDNESS
        Modifications made to ASTM  Methods C 131 and E 384 for the analysis of
antiskid  abrasives are shown in Tables  B-1  and B-2, respectively.

                         TABLE B-1.  MODIFICATIONS TO ASTM METHOD C 131-89
                                  FOR AGGREGATE MATERIAL < 2.36 mm
  The standard ASTM methodology should be used with the following modifications:

  5. Apparatus


          5.4.1  The weight of the test sample placed in the Testing Machine shall be 5,000 ± 10 g along with a
  charge of six (6) standard spheres.

  8. Procedure

          8.1  Split the sample into two equal fractions according to Method C 702. Sieve the first fraction
  according to the procedure outlined in Table A and  record the results.  Weigh the charge of spheres to be
  introduced into the testing machine and record the results.

          8.2  Place the second sample fraction (as received) and the charge in the Los Angeles testing
  machine and rotate the  machine at a speed of 30 to 33 rpm for 500 revolutions. After the prescribed number
  of revolutions, discharge the material from the machine, separate the charge, and sieve the sample material
  using the procedure shown in Table A. Record the results.

                                      TABLE A. SAMPLE SIEVING PROCEDURE

          1. Select the appropriate 8-in diameter. 2-in deep sieve sizes. Recommended U.S. Standard Series sizes are 3/8 in.
             No. 4, No. 20, No. 40, No. 100, No. 140, No. 200, and a pan. Comparable Tyler Series sizes can also be utilized.
             The No. 20 and the No. 200 are mandatory. The others can be varied if the recommended sieves are not available
             or if buildup on one particular sieve during sieving indicates that an intermediate sieve should be inserted.

          2.  Obtain a mechanical sieving device such as vibratory shaker or a Roto-Tap.

          3.  Clean the sieves with compressed air and/or a soft brush. Material lodged in the sieve openings or adhering to the
             sides of the sieve should be removed (if possible) without handling the screen roughly.

          4.  Obtain a scale (capacity of at least 1,600 g) and record make, capacity, smallest division, date of last calibration,
             and accuracy (If available).

          5.  Tare sieves and pan. Check the zero before every weighing.  Record weights.

          6.  After nesting the sieves in decreasing order with pan at the bottom, dump dried laboratory sample (probably
             immediately after moisture analysis) into the top sieve.  Brush fine material adhering to the sides of the container
             into the top sieve and cover the top sieve with a special lid normally purchased with the pan.

          7.  Place nested sieves into the mechanical device and sieve for 20 min. Remove the pan containing the minus
             No. 200 fraction and weigh. Replace pan beneath the sieves and sieve for another 10 min.  Remove pan and
             weigh. When the differences between two successive pan sample weighings (where the tare of the pan has been
             subtracted) is less than 3.0%, the sieving is complete.

          8.  Weigh each sieve and its contents and record the weight. Check the zero before every weighing.

          9.  Collect the laboratory sample and place the sample in a separate container if further analysis is expected.
                                                    B-3

-------
                                       Table B-1 (Continued)
9. Calculation

        9.1  Determine the abrasion loss as a percentage using the following equation:
                                         aL= Si~ S'x100
where:

                aL      m       LA abrasion loss (weight percent)

                S,       -       Weight percent of material > 200 mesh (75 urn) in original sample

                S,       =       Weight percent of material > 200 mesh (75 |im) in sample after exposure in
                                Los Angeles Testing Machine
                                                B-4

-------
                Table B-2.  MODIFICATIONS TO ASTM E 384-89 TO
                  INCORPORATE FINE AGGREGATE MATERIALS
The standard ASTM methodology should be used with the following modifications:

7. Test Specimens

       7.1  Mount each aggregate sample in an appropriate epoxy resin mold. Grind and
       polish the surface of the mold according to standard techniques.

8. Procedure

       8.1  Select at least four (4) particles from each mold for analysis.  The particles
       should be representative of the bulk material present in the specimen. Variations
       in reflectance, polishing hardness, color, and anisotropism should be noted and
       particles excluded which  do not represent the bulk of the  material present.  For
       materials of greatly varied lithologic character, particles with different lithology
       should be selected which represent the entire  sample.

       8.2  Delete

       8.4.7  Measure a total of five (5) indentations on each of the four particles selected
       for analysis to obtain a total of 20 indentations for each polished section. Read the
       two diagonals of the indentation to within 0.25 \im or 0.4%, whichever is larger, and
       determine the average of the diagonal lengths. For materials with very low
       reflectance values (e.g., calcite, quartz, etc.), indentation measurements should be
       performed using an appropriate microscope equipped with dark field illumination or
       other special types of lighting arrangements.

       8.4.8  Indentations which exhibit obvious problems shall not be measured.  (For
       example, an indentation that shows cracks at its edge.)  Instead, repeat the
       indentation elsewhere on the same particle or select another representative particle
       from the same polished section.

       8.5.2  Compute the mean HV for the polished section by  calculating the Vickers
       hardness number for each of the 20 indentations from the equation given in Section
       4.3.1.1 or from the information given in Table 2. Average the 20 measurements to
       obtain the mean for the sample.  To obtain the hardness  number from the table,
       read the HV (1 gf) corresponding to the average of the two measured diagonal
       lengths (in micrometers) and multiply by the test load in grams-force.
                                       B-5

-------
                 APPENDIX C
EXAMPLE DATA FORMS USED IN SHRP PROJECT H-208A
        FOR MONITORING SITE CONDITIONS
                    C-1

-------
     NAME:
                           WEATHER AND PAVEMENT CONDITION LOG
                                                                AGENCY:   M'.
Dale
•»/.**







VNfta
Time
Q8-.I5
»*H
o8: R7
««
oB'. 3%
<«•.»
OS-.S,
*'.&7
«-*
' Type of
Precipitation
Drizzle









i









tt









I









LLSnow



*
X
X
y
x
/
1
x
»
»






Blowing Snow









o









Section
Observed
Teal









Control
*
x
*
X
*
y,
x
X
X
Location
(Route and
Mllepoat)
T»5^3
r*s-an*ai.^
VW 53 f-M* 2.S .^
T*ssKXf*a.*
•VV\6-i fAP 2.2, .S
Lane No.
and
Direction
lA\J>- OviviVi^
MB - Qv-»v»\^
Mfo- Dv,V.H^
HE* - bv'iViVt\
WV^r S>
-------
OPERATOR ACTIVITY LOQ (CONTROL/SECTION)
                             V .

     STATE OF  Mt'yA^cvka	
                   ~Y>

•J\y»    Swv,T3    /lUS.

   Spreader Make and Model

   Was spreader calibrated prior to us/JYeT^r No):
                          Typo of Spreader Control

                                II so. whop:	
                                        :fr
                                                  •> •

     X
               AppBod
                •*
                                                                                 WMIhw/Povomart Condhttoiw '
1*06
                                 too
                 2.
                                                                                        "571
f *w7cuM 01. Cl. Pi., Ot.'V Ot. or CL «• PL. DL « CdMnf t«n«; C/. «

2 hrfteate P, C. A. S. PtC. a'f + A P - P/«»*IJ»; 0 « Ctmmliial; A "
                                            ; PL - P«M/OB

                                          ; S •
rf$&H
                                         IXMnj and r«i*fn0 Un«»; or Ct + R - Ctnlor Un» utd
                                 m PlcyhgV C/iamfcftf ttoatint^ or P •» X « Plawliy t


                                                         Pago	of.

-------
SOBO/CORALBA DATA RECORDING FORM 2
                                                  Dale:
                                             vA-   Zl.\
               Agency:
Hc-J^rs  Test Location: TV\
                                                         H-IC -
                                    \vV.W
Name:
Time:
General Weather Condition:   SN^.-.^O _
General Pavement Condition: _
Type of Pavement: _ Open Graded Asphalt  _ Dense Graded Asphalt    \   Concrete
No. of Lanes  5.    Direction of Travel: Test Section  _   Control Section
Direction of Cross Slope:
Test Location Relative to Nearest RW1S Site:
TEST SECTION
Sampling
Point
1
2
3
4
5
Scale
Factor





Sobo
Reading





Coralba Reading: Lane 1 : 	 ;
Lane 2: : Lane 3:





CONTROL SECTION
Sampling
Point
1
2
3
4
5
Scale
Factor
0,!
<3>\
o . i


Sobo
Reading
£
IS
L^


Coralba Reading: Lane 1 : o.g-i :
Lane 2: : Lane 3:

COMMENTS
Vo>KLow Spwd Lara
                                                 (Lara2)        (Lan« 1)
                                                     FOUR LANE
                                                           _
                                                           "Trawsl Larw
                                                           Sampling
                                                           "Points
                                                         Wh««l
                                                        'Track
    DIRECTION
    OF TRAVEL
                  Inckto
                        Mwfian
%>
^
SN
^
xK
Jh
3pMd
^^
^
Lar
>
• >
%
N
(
                                    (Un«3)
                                   (Lan.2)

                                 SIX LANE

SAMPLING POINTS FOR DIFFERENT ROAD CONFIGUi

r
i
N
(1
>
K
")
\L


r*


SpMd

1
JU
>
^^
.•*
•>
Edgaof
»-*" Travel Lane
Sampling
-*** Points
Wheel
^-- Track
•^ tUff •¥*.•» UK
                                             (Lan.1)
                                       C-5

-------
     APPENDIX D






SAMPLE CALCULATIONS
        D-1

-------
                             MIDWEST RESEARCH INSTITUTE                       ***''"
   TITLE
,o           ..
 /     .     / I ,      I   P«OJICT PEV£LOPMENT^MTCH     ^         ,       >.    .

                                                     ,  /\«^l  >f
   PROJECT NO. J/fl-/2-        DRAWN            APPR.   4S.^    DATE
I.    -            ,   „  -   _        .
                 .  I   /^ /     -•»*p"~ .  I     J  "-^       »  ^  "^           »	    **9  If    *


              "'f     "^    "  ""    "    ""      "    ^'^
    -or   I -
                                 «\

                     I


                    ; x  1,000


                                ^1



   For *7- ^   ^/*-ip^r '.



                        ^ I ,OOO
^- •   «-— -  —.         .
                  ^^:
                                                     f  -
                                                         e,
                                       D-3

-------
      u
TITLE	L±
                  MIDWEST RESEARCH INSTITUTE
                _      PROJECT DEVELOPMENT SKETCH
    -7    c   Q      i      fl i   i  '
     d.   oV    V'-'-'p'c*    (_Cilc-~>(
PROJECT NO	DRAWN	APPR	DATE.
                        v i*K»»/^.                                   .
                                        >'
                                                . -  I'Z. f  ^.97(7^ =  /8.I
                                                                        '


     Kiel
                     .-,i«r-. C2Z-4--
 -7 --  %



/ e jci («1>ol«L-V<.   corner   \jetocr' i«" -^v,   ^-V  _ S- 'i  "7-
                                                                         «*n
                                       D-4

-------
TITLE.
                      9
                        MIDWEST RESEARCH INSTITUTE
                            PtOJECt, DEVELOPMENT SKETCH
PROJECT NO
                       . DRAWN
                                          APPR
                                                          DATE.
             OC
              -o
   V-  M  "7-^  u«.^4-, ^^^  _\J«~  ^ $?&£&-„
___
l-o
     or  -
                 .'. "7. S
For
                                                       see
   ve<*z>
                                                rfi

                                 D-5

-------
                                                  MIDWEST RESEARCH INSTITUTE
                                                     Fugitive Dust Ejmlsslon Test
                                                       -l£hl  T?U  Avl4
                                                          Velocity Profiled
                                                                  Dale
                                                                  Recorded
                                                                            ^^ __ .
                                                                             v\ Ci*^ -- .
o
              20   19
18
17    16    15    14   13
12    11    10
  (miles/hr)
                                                                                                              IMIMVUihilIMM

-------
                        MIDWEST RESEARCH INSTITUTE
TITLE.
           4-  r
            \  (-yf-
      _,. DEVELOPMENT SKETCH
     n <   i  t
PROJECT NO..
                          DRAWN.
                                         APPR._
                                DATE.

Klo.
                                           Sr   +U? -fT-
     ^-
             l
            -5-7.^ =
                               D-7

-------
TITLE.
                             MIDWEST RESEARCH INSTITUTE
                                 PIOJECU DEVELOPMENT SKETCH
                                     H  '
                                                                                  MtMi
PROJECT NO
DRAWN
                  APPR
                                                                    DATE.
                                                                    Vi
                                  6^0
                                        D-8

-------
                                                                      1
                                                                      »-= •< •
                                 TECHNICAL REPORT DATA
                          (Please read Instructions on the reverse before completing)
 . REPORT NO.
  EPA-600/R-93-019
                                                      3. RECIPIENT'S ACCESSION- NO.
4. TITLE AND SUBTITLE
 Characterization of PM-10 Emissions from Antiskid
 Materials Applied to Ice- and Snow-Covered
 Roadways	
                                                       5. REPORT DATE
                                                         January 1993
                                                       6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
  John S. Kinsey
                                                      8. PERFORMING ORGANIZATION REPORT NO.
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Midwest Research Institute
  425 Volker Boulevard
  Kansas City, Missouri  64110-2299
                                                       10. PROGRAM ELEMENT NO.
                                                       11. CONTRACT/GRANT NO.
                                                        68-DO-0137,  Tasks 12 and
                                                         71
 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
                                                        Task Final;  3-10/92
                                                       14. SPONSORING AGENCY CODE
                                                        EPA/600/13
 is. SUPPLEMENTARY NOTES AEERL project officer is Charles
                                                    C.  Masser,  Mail Drop 62, 9197
 16. ABSTRACT
          The report gives results of a field program to establish a predictive model
 for PM-10 (particulate matter with diameters = or < 10 micrometers) emissions.
 (NOTE: Several areas of the U. S. in violation of the National Ambient Air Quality
 Standard for PM-10 have conducted studies that have identified the resuspension of
 antiskid material applied  to paved roads as an important source of PM-10.  The ap-
 plication of antiskid materials creates a temporary but substantial increase in the
 amount of fine particulate on the road surface over and above that which is normally
 present. Measured emission data are lacking for all types of antiskid materials.)
 A source-oriented emissions sampling procedure was  conducted on a section of US
 53 just west of Duluth, MN, during March/April 1992.  The measured emission fac-
 tors varied from 1 to 11 g/VKT (vehicle kilometer traveled) for the three tests con-
 ducted. The data were not sufficient, however,  to develop any specific correlation
 between the measured emission factors and source parameters. The only general
 observation made was that PM-10 emissions appear to increase with the amount of
 antiskid material applied. A comparison of measured emission factors with those
 predicted by an EPA compilation of air pollutant emission factors indicated that
 most of the measured factors are higher than those predicted from silt-loading.
 7.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                           b.IDENTIFIERS/OPEN ENDED TERMS
                                                                   c. COSATI Field/Group
 Pollution          Deicers
 Particles
 Mathematical Models
 Skid Resistance
 Pavements
 Abrasives
                                           Pollution Control
                                           Stationary Sources
                                           Particulate
                                           Antiskid Material
13B
14G
12A
13L

UK
01C.08L
 8. DISTRIBUTION STATEMENT

 Release to Public
                                          19. SECURITY CLASS (This Report)
                                          Unclassified
                                                                    21. NO. OF PAGES
   84
                                          20. SECURITY CLASS (This page)
                                          Unclassified
                                                                   22. PRICE
EPA Form 2220-1 (9-73)
                                        D-9

-------