United States
                Environmental Protection
                Agency
Atmospheric Research and Exposure
Assessment Laboratory
Research Triangle Park NC 27711
                Research and Development
EPA/600/S3-91/065  Dec 1991
EPA       Project Summary
                A  Regional  Fine  Particle  Field
                Study:  Data Base  and  Initial
                Results
                Roy L. Bennett and Leonard Stockburger
                  The Fine Particle Network (FPN), a
                group of rural sites located In the north-
                ern two-thirds  of the eastern United
                States, has been used to determine fine
                particle concentrations during selected
                periods between August 1988 and Feb-
                ruary 1989 and then continuously from
                July 1989 through May 1990.  Samples
                from the network, collected 24 h each
                day, were analyzed for the most abun-
                dant elements by wave length disper-
                sive x-ray fluorescence (WDXRF). About
                30 sites were In operation during each
                sampling period.   These included 17
                sites that were operating during the en-
                tire program schedule and others that
                were available only during parts of the
                schedule.
                  The FPN  data base can be used to
                enhance the understanding of the sea-
                sonal variation  of fine particle concen-
                trations  in the  eastern United States.
                The data base, combined with meteoro-
                logical data, will be useful for evalua-
                tion of regional aerosol models and the
                development of methodologies to esti-
                mate regional contribution to urban par-
                tlculate concentrations.  The 14 sam-
                pling sessions  produced  12,859
                samples for the 44  stations that were
                analyzed by WDXRF.  Seasonal aver-
                ages of fine partlculate mass and el-
                emental sulfur concentrations were de-
                termined for each of the seven seasons
                in which samples were collected. Con-
                centrations of both were highest in the
                summer.  The average concentrations
                for all sites and all collection dates were
                15.5 and 1.9 u.g/m3 for the fine particu-
                late mass and elemental sulfur, respec-
                tively.
  This Project Summary was developed
by EPA's Atmospheric Research and
Exposure Assessment Laboratory, Re-
search Triangle Park, NC, to announce
key findings of the  research project
that Is fully documented In a separate
report of the same title. (See Project
Report ordering Information at back).

Introduction
  A sampling program has  been  spon-
sored  by the Environmental Protection
Agency to sample ambient air fine par-
ticles daily at 44 rural sites in the northern
two-thirds of the  eastern United States
during periods of  1988, 1989, and  1990.
This Fine Particle Network (FPN) consisted,
of a subset of sampling locations selected
from the Acid Model Operational and Di-
agnostic Evaluation Study (Acid MODES)
surface sampling program.   The Acid
MODES monitoring program was designed
to evaluate regional acid deposition mod-
els. Since the FPN operation was subor-
dinate to the larger Acid MODES pro-
gram, changes in the latter during Decem-
ber 1989 necessitated shifts in the FPN
sites. Samples were delivered to EPA by
the contractor responsible for  collection
and mass measurements,  ENS5R, Inc.
Analyses for elemental composition by x-
ray fluorescence were made by ManTech
Environmental Sciences.
  The objective of the fine particle field
study was to obtain a  data base for fine
particle (less than 2.5  u.m aerodynamic
diameter) concentrations that would rep-
resent extensive synoptic measurements
of fine particle mass.  This  fine particle
                                                                  Printed on Recycled Paper

-------
 data base, along with other chemical spe-
 cies and  meteorological measurements
 made in the Acid MODES  program, is
 useful to
   (1) develop methods for estimating and
       understanding regional contributions
       to urban fine particulate concentra-
       tions,
   (2) evaluate regional particulate mod-
       els, and
   (3) enhance the understanding of the
       spatial, and temporal variations of
       fine participate concentrations within
       each season in the eastern United
       States.
   Estimates from previous experimental
 field studies indicate that regional back-
 ground particulate concentrations are ma-
 jor contributors to urban particulate  con-
 centrations. Because this contribution is
 large, it is desirable in urban particulate
 field studies to have upwind sampling sites
 to estimate the regional background input.
 This often is difficult because the back-
 ground sites may not  be free of local
 source contributions. One alternative ap-
 proach is to use concentrations from mea-
 surements made  at other times  during
 comparable  meteorological  conditions.
 However, backgrounds may vary, even
 for the same wind directions if the  par-
 ticles that  arrive at the  urban  boundary
 have traveled along different trajectories.
  This report describes the FPN sampling
 network, the sampling  procedures,  and
 the x-ray fluorescence analytical method
 for elemental content determination.

 Network Description
  The EPA-sponsored  program, collec-
 tively named the Acid Model Operational/
 Diagnostic  Evaluation  Study  (Acid
 MODES), was  a part of a large coordi-
 nated field program to evaluate acid depo-
 sition models. The fine particle measure- •
 ment study sites were  a selected subset
 of the Acid MODES stations. Selection of
 sites for the Acid MODES stations was
 based on location in a high acid deposi-
 tion area, the northeastern United States,
 and attempts to place the sites at existing
 National Dry Deposition  Network (NDDN)
 locations when those sites met  certain
 siting criteria. These criteria included loca-
tions away from population areas, away
from large sources emitting sulfur and ni-
trogen oxides, and away from the Great
 Lakes and  the  Atlantic Ocean. Attempts
were made to select sites that were  ac-
cessible in all weather conditions, secure
from vandalism, and free from  potential
local contamination sources such  as
nearby plowed fields, roadways, and stor-
age facilities.
   The Acid MODES network  included
 three groups of stations: the model evalu-
 ation (ME-35) network, a gradient resolu-
 tion  (GRAD) group, and sub-grid variabil-
 ity (VAR) set of stations. The original 35
 sites of the ME-35 network were supple-
 mented with additional stations to provide
 uniformly distributed  site spacing for re-
 gional scale model  evaluation. The 12
 GRAD sites were located in Pennsylvania
 and  New York in  a relatively small area
 where steep concentration gradients were
 expected. The VAR network was five sta-
 tions in western Kentucky that were spaced
 in  a small  cluster to measure  variation
 within a model grid.
   The broad  range of measurements
 made at all Acid MODES stations included
 precipitation  chemistry (pH, conductivity,
 and the pollutant species sulfate, nitrate,
 chloride, ammonium,  sodium,  potassium,
 calcium and  magnesium) aerosol particle
 chemistry (sulfate, nitrate, and ammonium),
 and gases (sulfur  dioxide,  nitric acid, ni-
 trogen dioxide,  and ammonia).
   The Fine Particle Network did not in-
 clude all Acid MODES stations. Sites that
 were included at one time or another dur-
 ing the FPN program are listed in Table 1.
 Duplicate  samplers to determine collec-
 tion repeatability were placed at close prox-
 imity at three stations: State College, PA
 (306), Oxford, OH  (322) and Plainville, IL
 (359). The duplicate sites were designated
 by 100 units higher than their companion
 site (i.e., sites 408, 422, and 459).
   At  the start of the  FPN sampling pro-
 gram the original 32 sites were  selected
 from  the 68  Acid  MODES stations.  Be-
 cause of budget and program changes,
 29 Acid MODES stations were shut down
 in December 1989.  Of these, 16 were
 FPN  sites including one of the co-located
 stations (Plainville,  IL). However, fine par-
 ticulate samplers were transferred and in-
 stalled at other Acid MODES stations that
 were not previously included  in the FPN
 study. Therefore,  during  the last  four
 months of operation, the total complement
 of FPN sites was brought back to 28. One
 of the new  sites (Oxford,  OH,  site  no.
 322)  had  co-located  samplers.  Besides
 affecting the sampling continuity, the
 change of sites altered the  geographic
 distribution of the network.

 Sampling Procedures and
 Schedule
  The fine particle measurement samplers
were  assembled for  the FPN sites by
 ENSR Consulting and Engineering, Inc.
The sampling system consisted of a train
of the following sequence of components
positioned in a protective cabinet: a small
 cyclone for removing coarse particles, a
 47-mm filter holder for a fluorocarbon-type
 membrane filter, a critical flow orifice to
 maintain an 8.8-l/min flow, and the pump-
 ing system  with  flow-metering  devices.
 During the course of the sampling  pro-
 gram the type of fluorocarbon filters being
 used was changed. The high incidence of
 filter breakage with the fragile Teflo filters
 and its higher cost led to  its replacement
 with Zefluor in the last  11 sampling ses-
 sions.
   The Fine Particle Network sampling was
 started at the original 32 sites on August
 15, 1988  and concluded at a  different
 combination of sites on May 31,1990, but
 sampling was  not continuous during the
 entire period. The first three sessions were
 six week periods  starting in mid-August
 1988, mid-October 1988 and mid-January
 1989. After examination of the data pro-
 duced in the total Acid MODES program
 the network managers proposed some
 changes so that  sampling  was  not re-
 sumed  until July 1989. After that time op-
 eration was continuous until May 31,1990
 with the exception of one  week before
 January 1, 1990. The interruption unfortu-
 nately eliminated the collection of samples
 in the spring of 1989 and limited sampling
 in the spring of 1990 to two months.
   The sampling schedule was further com-
 plicated  by the closing of  some of the
 original sites and  starting new  sites, as
 described in the network description sec-
 tion. Only 17 sites operated from the start
 to the end of the program. Specific dates
 that each site operated are indicated in
 Table 1.

 Analytical Procedures
   After  the samples  were  delivered  to
 EPA, those  with  invalid  quality  control
 codes were removed  and the remainder
 were  analyzed  by  wavelength dispersive
 x-ray fluorescence (WDXRF) spectrometry.
 The WDXRF  spectrometer, a Siemens
 MRS-3 system  with one scanning mono-
 chromator and 13 fixed monochromators,
 gives analytical capability for elements from
 sodium through bismuth. The fixed mono-
 chromators were tuned  to one element
 and dedicated to that element;  all other
 elements were  analyzed sequentially by
 the scanning monochromator. A chromium
 target x-ray tube operated at 54 kv and 44
 ma was used to excite the samples. The
 characteristic radiation from each element
 was defracted by a crystal and its inten-
 sity measured  by a detector  for each
 monochromator. The type  of crystal and
 detector employed was optimally selected
for the characteristic wavelength of each
 element. Spectrometer operation and data

-------
Table 1. Locations and Operation Dates for Fine Paniculate Network Stations
Site
Number
301
302
303
305
306
307
308
310
312
314
315
316
318
319
320
321
322
323
324
326
327
328
329
333
334
335
337
340
344
346
357
359
362
367
368
369
370
373
374
381
383
406
422
459
Location
PKtsboro, NC
Wartburg, TN
West Point, NY
Whiteface Mountain, NY
•State College, PA
Parsons.WV
Prince Edward SF, VA
Danby. NY
Kane Forest, PA
Deer Creek Park, OH
Pickney, Ml
Beltsville, MD
Tanner's Ridge, VA
Cedar Creek SP, WV
Mountain Lake, VA
Lilley Cornett Woods, KY
'Oxford, OH
Brokensword, OH
Unionville, Ml
Roaring Creek, NC
Edgar Evins SP, TN
Arendtsville, PA
Perryville, KY
Salamonie Lake, IN
Perkinstown, Wl
Ashland, ME
Coweeta (Otto), NC
Vincennes, IN
Washington Crossing, NJ
University Park, IL
Piseco, NY
*Plainview, IL
Hawthorne, PA
Renovo, PA
Williamsport, PA
Wirt, NY
Little Marsh, PA
Brackney, PA
North Orwell, PA
Eddyvjlle, KY
New Concord, KY
*State College, PA
•Oxford, OH
*Plainvjew, IL
Latitude
35.67
36.08
41.35
44.38
40.78
39.10
37.17
42.35
41.60
39.42
42.42
39.03
38.52
38.88
37.37
37.09
39.53
40.92
43.63
36.11
36.04
39.92
37.68
40.80
^45.20
46.62
35.05
38.78
40.32
41.45
43.45
39.08
41.03
41.37
41.17
42.15
41.90
41.94
41.90
37.07
36.53
40.78
39.53
39.08
Longitude
79.23
84.54
74.05
73.85
77.93
79.66
78.31
76.49
78.77
83.22
83.90
76.82
78.48
80.85
80.52
82.99
84.72
83.00
83.38
82.05
85.73
77.31
84.97
85.60
90.60
68.41
83.43
87.49
74.87
87.72
74.52
89.95
79.27
77.53
76.92
78.11
77.44
75.92
76.28
88.03
88.09
77.93
84.72
89.95
Operating
Start
09/12/88
02/12/90
08/15/88
10/13/88
08/15/88
01/22/90
10/13/88
08/15/88
08/15/88'
08/23/88
01/21/90
1 1/07/88
08/26/88
09/01/88
09/12/88
09/07/88
01/22/90
08/15/88
08/22/88
01/22/90
09/01/88
09/02/88
09/02/88
08/15/88
08/16/88
11/06/89
08/31/88
08/15/88
11/14/88
08/22/88
01/25/90
08/15/88
08/15/88
08/30/88
09/12/88
01/15/90
08/15/88
01/15/90
03/02/90
01/08/90
01/11/90
08/15/88
02/14/90
08/15/88
Dates"
Stop
12/08/89
05/31/90
12/08/89
05/31/90
05/31/90
05/31/90
05/31/90
12/08/89
05/31/90
05/31/90
05/31/90
12/03/89
09/18/88
05/31/90
05/27/90
05/31/90
05/31/90
05/31/90
05/31/90
05/31/90
12/08/89
12/08/89
05/31/90
12/08/89
12/08/89
12/08/89
05/31/90
05/31/90
12/08/89
12/08/89
05/31/90
12/08/89
05/31/90
05/31/90
12/10/89
05/31/90
05/31/90
05/31/90
05/31/90
05/31/90
05/31/90
05/31/90
05/31/90
12/08/89
      Co-located sites
      All sites were shut down during these periods:
     2 weeks     09/26/88 - 10/09/88
     8 weeks     11/21/88-01/15/89
processing were done by dedicated labo-
ratory computers.
  For the first three sampling sessions,
33  elements were determined for all
samples. A survey of these results led to
a decision to reduce the analyses to these
11  elements for the last sampling  ses-
sions: Mg, Al, Si, P, S, Cl, K, Ca, Fe, Zn
and Cd,
18 weeks
1 week
02/27/89 - 07/02/89
12/25/89 - 12/31/89
 Results and Discussion
   During the period from mid-August 1988
 through May 1990, fine particle sampling
 was conducted at 45 sites. The network
 was not operating continuously during the
 entire period, but samples were collected
 for a total of 65 weeks. Only 17 of the
sites  were operating at all times during
the network operation.  After the removal
of invalid samples,  12,589 filters (includ-
ing blanks) from 14 sampling sessions
were  analyzed for elemental content by
WDXRF spectrometry. Duplicate analyses
were  made by reanalyzing at a later date
13% of the filters from the first three ses-
sions and  5% of the samples in the re-
maining sessions.  Repeatability was ex-
cellent, greater than 90% for the six most
abundant  elements. Chlorine  analyses
were  consistently lower in the second de-
termination, apparently because of loss
by  volatilization, either direct  or as the
result of chemical reaction.
  The average fine particulate mass value
for  all sites and all collection dates was
15.5 ng/m3. The XRF  analysis gave an
average sulfur concentration of 1.9 u.g/m3,
or a mean sulfur content of 12.1% for all
10,927 samples. If the gravimetric factors
for  SO,2'  (3;00),  NH.HSO, (3.59), or
(NH4)2SO4  (4.13) are applied, the follow-
ing  equivalent.percentages would be ob-
tained: 36% SO.2', 43% NH.HSO., or 47%
(NH4)2S04.
  The seasonal averages for all sites of
the fine paniculate mass and sulfur con-
centrations are summarized  in Tables 2
and 3 for each of the seven quarters in
which sampling was made. The concen-
trations of fine particulate mass and par-
ticulate sulfur were highest during the two
summers and lowest during tiie winter and
fall  seasons.
  The increase  of fine particulate sulfur
concentrations in the summer is consis-
tent with results of  previous studies and
has been attributed to  increased use of
electrical power for air  conditioning de-
mands  and to more intense solar radia-
tion, which converts SO2 to sulfate aero-
sols.  The increase  in the percentage of
sulfur is demonstrated in Table 4, where
the percent of sulfur in  the total fine par-
ticulate is summarized for the seasons at
the  co-located sites  306 and 406 and for
the  average concentrations  from all 44
sites.

Conclusions and
Recommendations
  A data base of  fine particulate mass
and 11  chemical element concentrations
has been produced  by  daily sampling of
ambient air aerosol at 44 sites selected
from the Acid MODES network at intermit-
tent periods from July 1988 to May 1990.
The data base was generated from 10,952
samples that were  collected during 65
weeks and seven different quarters (sea-
sons). Analysis for elemental composition
was accomplished by wavelength disper-

-------
Table 2. Fine Particulate Mass Seasonal A verages for All Sites
SEASON NO. OF SAMPLES MEAN STD. DEV.
(ng/m3) (ng/m3)
Summer 1988 401
Fall 1988 661
Winter 1989 985
Summer 1989 2127
FaM 1989 1803
Winter 1990 1920
Spring 1990 1478
Table 3. Sulfur Seasonal Averages for All Sites
SEASON NO. OF SAMPLES
Summer 1988 461
Foil 1988 929
Winter 1989 1149
Summer 1989 2311
Fall 1989 2180
Winter 1990 2175
Spring 1990 1646
20.04 11.47
12.46 5.51
14.89 7.05
24.87 15.52
16.57 9.59
13.79 6.69
14.95 7.14
MEAN STD. DEV.
(Hg/m3) (ng/m3)
2.16 1.60
1.15 0.70
1.30 0.74
3.12 2.75
1.52 1.32
1.37 0.77
2.03 1.25
Table 4. Seasonal Variation of Sulfur ConcentoaSon Averages at the Pennsylvania Co-heated Sites fag/trf)
SEASON SITE 306 SITE 406
Mean Std. Dev.
Summer 1988 2.29 1.05
Fall 1988 1.20 0.64
Winter 1989 1.50 0.54
Summer 1989 3.21 2.55
Fall 1989 1.79 1.31
Winter 1990 1.53 0.75
Spring 1990 2.08 1.42
sive x-ray fluorescence spectrometry
(WDXRF) on the total of 12,589 filters
(including field blanks) for the following 1 1
elements, which were the most consis-
tently present: Mg, Al, Si, P, S, Cl, K, Ca,
Fe, Zn, and Cd. The continuity and com-
pleteness of the FPN study was affected
Mean Std. Dev.
2.20 1.32
1.31 0.67
1.48 0.62
3.15 2.44
1.82 1.40
1.61 0.81
2.12 1.49
by budget and operational problems in
the parent Acid MODES network, which
caused interruptions in the sampling
schedule, the closing of some of the origi-
nal sites, and the moving of sites later in
the program.
The original form of the data base that
was generated by the collection, gravi-
metric, and analytical procedures consisted
of 384 files with about 17.5 MB of com-
puter data. This large data base includes
for each of the 12,589 samples the weight,
volume sampled, filter collection area, qual-
ity control codes, and other collection data.
Each sample also contains the analytical
data for each of the 1 1 elements, includ-
ing concentrations, detection limits, and
random error estimates.
A special software package was devel-
oped by ManTech Environmental to com-
press the large original data base to a
more transportable size and to develop
an interface for the user to conveniently
retrieve selected information. Programs
were written in Pascal language to pro-
duce a file that could be read directly into
a personal computer for further data analy-
sis.
The FPN data base produced by this
field study is a valuable contribution to the
understanding of the climatology of
nonurban fine particulate aerosols and their
contribution to urban particulate loadings.
Cursory examination of the data indicate
high variability of the daily concentration.
Future efforts should assess the relation-
ship between meteorological conditions
and site concentrations to better under-
stand the variations. A recommended ap-
proach is to classify samples by meteoro-
logical data. This might be done by at-
tempting to correlate concentration levels
with meteorological parameters such as
surface wind direction, synoptic weather
classifications (warm, cold, occluded, or
stationary fronts) and by back-trajectory
calculations. Initially, the concentration
observations that are within distinct com-
mon wind directions may be compared.
However, incoming air from the same di-
rection can arrive through different trajec-
tories. Attempts should be made to corre-
late observations with trajectory-defined
sources to promote an understanding of
how the observed variations are related to
fluctuation in meteorology.
•&U.S. GOVERNMENT PRINTING OFFICE: 1992 - 648-080/40131

-------

-------

-------

-------
 Ft. Bennett (also the EPA Project Officer) andL. Stockburger are with the
     Atmospheric Research and Exposure Assessment Laboratory, Research
     Triangle Park, NC 27711.
 The complete report, entitled "A Regional Fine Particle Field Study: Data Base
     and Initial Results," (Order No. PB92-106939/AS; Cost: $19.00, subject to
     change) will be available only from:
         National Technical Information Service
         5285 Port Royal Road
         Springfield, VA 22161
         Telephone: 703-487-4650
 The EPA Project Off her can be contacted at:
         Atmospheric Research and Exposure Assessment Laboratory
         U.S. Environmental Protection Agency
         Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati, OH 45268
     BULK RATE
POSTAGE & FEES PAID
 EPA PERMIT NO. G-35
Official Business
Penalty for Private Use $300
EPA/800/S3-91/065

-------