United States
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
EPA-600/3-90/020
August 1990
National Dry
Deposition Network
Second Annual
Progress Report
(1988)
-------
EPA/600/3-90/020
August 1990
NATIONAL DRY DEPOSITION NETWORK
SECOND ANNUAL PROGRESS REPORT (1988)
by
Eric S. Edgerton, Thomas F. Lavery
Mark G. Hodges, Jon J. Bowser
Hunter/ESE, Inc.
Gainesville, FL 32607
Contract #68-02-4451
Project Officers
Barry E. Martin and Rudolph P. Boksleitner
Exposure Assessment Research Division
Atmospheric Research and Exposure Assessment Laboratory
Research Triangle Park, North Carolina 27711
ATMOSPHERIC RESEARCH AND EXPOSURE ASSESSMENT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
NOTICE
The information in this document has been funded wholly by the United
States Environmental Protection Agency under Contract No. 68-02-4451 to
Hunter/ESE, Inc. It has been subjected to the Agency's peer and
administrative review, and it has been approved for publication as an EPA
document. Any mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
ii
-------
FOREWORD
The Atmospheric Research and Exposure Assessment Laboratory (AREAL) is
committed to performing research and development in the characterization of
air pollutant sources, sinks, transport, and transformations, in the
assessment and prediction of exposure of humans and ecosystems to
environmental pollutants, and in the development of monitoring systems and
other technologies to determine the status and trends in pollutant
concentrations and the condition of the nation's ecosystems.
As national and international concern over acid deposition has grown, so
has the need for information on spatial patterns and secular trends.
Although wet deposition monitoring has been able to provide such
information, it is realized that total acid deposition has both wet and dry
components the latter of which has not been well defined. AREAL has
supported the development of a network of dry acid deposition monitoring
sites to demonstrate the feasibility of operating such a network and to
provide a database of atmospheric dynamics and chemical concentrations that
can be transformed to dry acid deposition quantities through the
application of one or more dry deposition algorithms. Applications for the
final data may include determination of spatial and temporal trends,
evaluation of deposition models, refinement of mass balance estimates, and
support to effects studies.
This report summarizes the second year's progress of the National Dry
Deposition Network.
Gary J. Foley, Ph.D.
Director
Atmospheric Research And Exposure Assessment Laboratory
Research Triangle Park, North Carolina 27711
iii
-------
ABSTRACT
Progress in the National Dry Deposition Network during calendar year 1988
is presented. The network configuration and operating procedures for the
field, laboratory, and data management center are described, and data are
summarized. Forty-three sites were operational at the close of 1988. Each
site was equipped with sensors for measuring ozone and meteorological
parameters required by inferential dry deposition algorithms. Weekly
average day/night concentrations of sulfate, nitrate, sulfur dioxide, and
nitric acid were measured using a three-stage filter pack. Sodium,
potassium, calcium, magnesium, and ammonium were measured during half of
the year. Ambient concentration data for 18 sites operational throughout
the year showed species-dependent variability from site to site, season to
season, and day to night. Average SO^2 and HN03 were highest during summer
(10 and 3 /ig/m , respectively) and lowest during fall (4 and 1.8 JLlg/in ,
respectively); S02 and N03 were highest in winter (18 and 1.7 fJLg/m3,
respectively) and lowest in summer (10 and 0.6 Mg/ra3 respectively).
Day/night variability was generally low for SO^2 and N03, but frequently
pronounced for S02 and HN03. Approximations of annual dry deposition for
SO^2 plus S02 and N03 plus HN03 suggest that fluxes are of the same order of
magnitude as wet deposition at numerous sites. Application of one or more
dry deposition algorithms is needed to refine such estimates. Examination
of ozone data showed numerous observations above 100 ppb mainly associated
with stagnating high pressure.
iv
-------
CONTENTS
Foreward iii
Abstract iv
Figures vi
Tables ix
Abbreviations and Symbols x
Acknowledgement xi
1. Introduction 1
2. Conclusions 3
3. Network Description and Operations 5
3.1 Network Description 5
3.2 Network Operations 9
3.2.1 Field Operations 12
3.2.2 Laboratory Operations 13
3.2.3 Data Management 17
4. Results and Discussion 21
4.1 Filter Pack Measurements 21
4.1.1 Annual Concentration Data 21
4.1.2 Seasonal Concentration Data 24
4.1.3 Day Versus Night Concentration Data 44
4.2 Ozone 47
4.3 Summary ,of Meteorological Observations 61
4.4 Estimated Dry Deposition 73
4.5 Overall Data Quality 78
4.5.1 Field Data 78
4.5.2 Laboratory Data 81
References 87
-------
FIGURES
Number Page
1 Status of NDDN monitoring sites- -December 1988 ........... 6
2 Filter pack assembly ........................ 10
3 Flowchart of laboratory operations for filter pack samples ...... 15
4 Flowchart for extraction of NDDN air quality filters ........ 16
5 Flow of data through the NDDN data management system ........ 19
6 Annual average concentration of SO^2 (A) and N03 (B)
(Mg/m3) for 1988 ......................... 22
7 Annual average concentration of HN03 (A) and S02 (B)
) for 1988 ......................... 23
8 Average SO^2 (A) and N03 (B) concentrations (/ig/m3) - -first
quarter 1988 ............................ 26
9 Average HN03 (A) and S02 (B) concentrations (/ig/m3) - -first
quarter 1988 ....... . .................... 27
10 Average SO^2 (A) and N03 (B) concentrations (/ig/m3) - -second
quarter 1988 ............................ 29
11 Average HN03 (A) and S02 (B) concentrations (jig/m3) - -second
quarter 1988 ............................ 30
12 Average SO^2 (A) and N03 (B) concentrations (/ig/m3) - -third
quarter 1988 ............................ 32
13 Average HN03 (A) and S02 (B) concentrations (/Ltg/m3) - -third
quarter 1988 ............................ 33
14 Average SO^2 (A) and N03 (B) concentrations (/Hg/m3) - -fourth
quarter 1988 ............................ 35
15 Average HN03 (A) and S02 (B) concentrations (jUg/m3) - -fourth
quarter 1988 ............................ 36
vi
-------
FIGURES (continued)
Page
16 Average NH^ (A) concentrations (Mg/m3) and Na+ (B) concentrations
(ng/m3)- -fourth quarter 1988 .................... 37
17 Average K+ (A) and Ca+2 (B) concentrations (ng/m3) - -fourth quarter
1988 ................................ 38
18 Average Mg+2 concentrations (ng/m3) - -fourth quarter 1988 ...... 39
19 Network-wide concentrations of SO^/SC^ (A) and N03/HN03
(B) by season .......................... 43
20 03 daily averages and maxima for Sites 101 (A) and 102 (B)--1988 . 48
21 03 daily averages and maxima for Sites 103/4 (A) and 105 (B) --1988.49
22 03 daily averages and maxima for Sites 106 (A) and 107 (B)--1988 . 50
23 03 daily averages and maxima for Sites 108 (A) and 110 (B)--1988 . 51
24 03 daily averages and maxima for Sites 113 (A) and 117 (B)--1988 . 52
25 03 daily averages and maxima for Sites 119 (A) and 120 (B)--1988 . 53
26 03 daily averages and maxima for Sites 121 (A) and 122 (B)--1988 . 54
27 03 daily averages and maxima for Sites 129 (A) and 137 (B)--1988 . 55
28 03 daily averages and maxima for Sites 140 (A) and 146 (B)--1988 . 56
29 Maximum hourly 03 (ppb) for May 28 - June 1, 1988 (A) and
June 13 - 17, 1988 (B) ...................... 57
30 Maximum hourly 03 (ppb) for July 3 - July 9, 1988 (A) and
August 16 - 19, 1988 (B) ...................... 58
31 03 time series for Site 120 (A), Site 129 (B) , and Site 121 (C)
from July 3 - July 9, 1988 ..................... 60
32 1988 Wind rose for Site 101, Research Triangle Park, NC (A)
and Site 102, Oak Ridge, TN (B) .................. 64
33 1988 Wind rose for Site 103/4, West Point, NY (A) and
Site 105, Whiteface Mountain, NY (B) ................ 65
vii
-------
FIGURES (continued)
Number Page
34 1988 Wind rose for Site 106, State College, PA (A)
and Site 107, Parsons, WV (B) 66
35 1988 Wind rose for Site 108, Prince Edward, VA (A), and
Site 110, Connecticut Hill, NY (B) 67
36 1988 Wind rose for Site 113, M.K. Goddard, PA (A) and
Site 117, Laurel Hill, PA (B) 68
37 1988 Wind rose for Site 119, Cedar Creek, WV (A) and
Site 120, Morton Station, VA (B) 69
38 1988 Wind rose for Site 121, Lilley Cornett Woods, KY (A)
and Site 122, Oxford, OH (B) 70
39 1988 Wind rose for Site 129, Perryville, KY (A) and
Site 130, Bondville, IL (B) 71
40 1988 Wind rose for Site 137, Coweeta, NC (A) and Site 140,
Vincennes, IN (B) 72
41 Estimated dry deposition of SO^2 plus S02 (A) and NOg plus HN03 (B)
(eq/ha) 76
42 Scatter plots of collocated filter pack data from Site 103/4,
West Point, NY--second quarter 1988 86
viii
-------
TABLES
Number Page
1 NDDN sites in operation during 1988 7
2 NDDN monitoring equipment 11
3 Precision and accuracy objectives of field measurements 14
4 Precision and accuracy objectives for laboratory measurements
for NDDN 18
5 Comparison of annual average concentration at five sites
for 1987 and 1988 25
6 Aerosol ion balances for fourth quarter 1988 42
7 Day versus night concentrations of SO^ and
S02 by quarter for 1988 45
8 Day versus night concentrations of NO^ and HN03
by quarter for 1988 46
9 Quarterly average temperature, relative humidity, and
solar radiation during 1988 62
10 Quarterly average windspeed, wind direction, and sigma
theta during 1988 63
11 Estimated deposition velocities for aerosols and gases 75
12 Wet versus dry deposition at selected sites 77
13 Summary of meteorological sensor unadjusted calibrations
during 1988 79
14 Summary of 03 and filter pack unadjusted calibrations
during 1988 80
15 Summary of laboratory accuracy and precision during 1988 82
16 Precision data for collocated continuous sensors
at West Point, NY during 1988 83
17 Precision data for collocated filter pack sampling at
West Point, NY during 1988 84
IX
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
cm/sec
CVS
DAS
DMC
EEPROM
EPA
eq/ha-yr
1C
I CAP
km
L/min
m
m/sec
nun
nunHg
MFC
NAAQS
NADP/NTN
neq/m3
NDDN
ng/m3
NIST
PC
ppb
ppm
QC
ML
w/m
SYMBOLS
Br
°C
Ca+2
HN03
K+
Mg+2
Na+
NHj
NH4N03
NO;
S02
S042
centimeters per second
calibration verification standard
data acquisition system
data management center
electronically erasable programmable read-only memory
U.S. Environmental Protection Agency
equivalents per hectare per year
ion chromatography
inductively coupled argon plasma
kilometer
liters per minute
meter
meters per second
millimeter
millimeters of mercury
mass flow controller
National Ambient Air Quality Standard
National Atmospheric Deposition Program/National Trends
Network
nanoequivalents per cubic meter
National Dry Deposition Network
nanograms per cubic meter
National Institute of Standards and Technology
personal computer
parts per billion
parts per million
quality control ,'
microliter
micrograms per liter
micrograms per cubic meter
watts per square meter
- - bromine
-- degrees Celsius
-- particulate calcium
-- nitric acid
-- particulate potassium
-- particulate magnesium
-- particulate sodium
-- particulate ammonium
-- ammonium nitrate
-- particulate nitrate
- - ozone
-- sulfur dioxide
-- particulate sulfate
-------
ACKNOWLEDGEMENTS
The authors wish to thank E. Allen (University of Florida, Gainesville);
R. Baumgardner (EPA); S. Bromberg (EPA); and B. Martin (EPA) for their
thoughtful reviews of this document. We also gratefully acknowledge the
following site operators without whose dedicated efforts the NDDN could not
succeed: S. Lumpkin (101), J. Gholston (102), S. Scott (103/4), R. Prins
(105), D. DeCapria (106), F. Wood (107), G. Brooks (108), S. Nolan (109),
T. Butler (110), D. Dorn (112), D. Croskey (113), S. Hammond (114),
L. Chilcote (115), V. Miller (116), B. Hufman (117), R. Gubler (118),
J. Chisler (119), S. Long (120), M. Brotzge (121), T. Chatfield (122),
D. Stineman (123), F. Matt and J. Matt (124), P. Hughes (126), M. Hale
(127), S. Scamack (128), K. McGill (129), M. Snider (130), D. Conrad (133),
F. Emstrom (134), D. Olberding (135), R. McCollum (137), M. Lang and
T. Mouzin (140), H. Burnett (144), P. Hess (146), W. Dunn (149), D. Honnell
(150), B. Scobey, Jr. (151), P. Ruf (152), J. Melin (153), F. Loftin (156),
W. Steiner and B. Steiner (157), C. Laster (165), R. Ljung (168).
xi
-------
SECTION 1.0
INTRODUCTION
Atmospheric deposition takes place via two pathways, both of which are
thought to be significant: wet deposition and dry deposition. Wet
deposition is the result of precipitation events (rain, snow, etc.) which
remove particles and gases from the atmosphere. Dry deposition, on the
other hand, is the transfer of particles and gases to the landscape in the
absence of precipitation. Wet deposition rates of acidic species across
the United States have been well documented over the last 5 to 10 years,
but, due to measurement difficulties, comparable information is unavailable
for dry deposition rates.
Whereas the direct measurement of dry deposition can be extremely
difficult, a number of investigations have recently shown that it can be
reasonably inferred by coupling air quality data with routine
meteorological measurements (Shieh e_t al. . 1979; Hicks et al. . 1985; Meyers
and Yuen, 1987; Wesely and Lesht, 1988). Using analogies with heat and
momentum flux, Shieh et al. (1979) computed aerosol and S02 deposition
velocities as a function of land use, season, windspeed, and meteorological
stability class. Results of this calculation for the eastern United States
showed that deposition rates for both species were strongly dependent on
windspeed, solar radiation, and the condition and type of ground cover.
For example, rapidly growing vegetation was found to experience higher
deposition rates than senescent vegetation, and forests generally
experienced higher rates than short grass or snow. Wesely (1988) has
expanded this approach to calculate deposition rates for various additional
species (including HN03) and to use site-specific meteorological data.
Seminal work on the subject of dry deposition has been performed by
Hicks et al. (1985) and Hosker and Womack (1986), who developed, tested,
and deployed the first field system for inferential dry deposition
measurements.
In 1986, the U.S. Environmental Protection Agency (EPA) contracted
with Hunter/ESE to establish and operate the National Dry Deposition
Network (NDDN). The specific objective of NDDN is to obtain field data at
50 to 100 sites throughout the United States to establish patterns and
trends of dry deposition. Ultimately, dry deposition fluxes will be
calculated using measured air pollutant concentrations and inferred
deposition velocities estimated from meteorological, land use, and site
characteristic data. One or more of the inference approaches developed by
Hicks et al. (1985) and Wesely (1988) will be used to estimate dry
deposition velocities.
-------
This report describes progress on the NDDN during calendar year 1988.
The purpose of this report is to familiarize the reader with the general
approach of NDDN and the various types of data which are being produced and
reported to EPA. This report is not intended to provide a definitive
analysis of the 1988 database. It describes the network configuration and
deployment schedule as well as procedures developed for field operations,
laboratory operations, database management, and quality control (QC). An
overview of air quality and meteorological data for 1988, including QC
results, is also presented. Finally, dry deposition rates are estimated
for selected sites and compared with wet deposition rates. These estimates
were prepared using literature values for deposition velocities (not the
algorithms cited above) and, therefore, are intended only to illustrate
likely ranges of deposition.
-------
SECTION 2.0
CONCLUSIONS
A 51-station dry deposition monitoring network has been established.
Forty-two sites are located in the eastern United States, and 9 sites are
located in the western United States. Results to date indicate that data
of sufficient quality to estimate dry deposition rates are being routinely
collected. In short, a monitoring network is being deployed that will
eventually provide nationwide measurements of air quality and estimates of
dry deposition. The following major conclusions were derived from the 1988
database.
1. Day and night air quality samples were collected at all sites using a
three-stage filter pack. Results for 18 sites, operational throughout
the year, show species-dependent variability from site to site, season
to season, and day to night. Annual average concentrations of
atmospheric particulate sulfate (SO^2) exhibited peak values of 7.2 to
7.5 micrograms per cubic meter (/Ltg/m3) (in southwestern Pennsylvania
and northern West Virginia) and relatively low spatial variability
relative to other measured species. Annual sulfur dioxide (S02)
concentrations showed peak values of 21 Jig/m3 (in Pennsylvania and
Tennessee) and significant spatial variability, which was presumably
due to point source emissions and local meteorological factors.
Annual nitric acid (HN03) concentrations exhibited peak values of
3.2 to 3.5 Mg/m3 (in Kentucky and Virginia) and similar spatial
variability, as S02. Annual particulate nitrate (NO^) concentrations
exhibited peak values of 1.5 to 2.0 Mg/m3 (in Illinois and Indiana)
and distinctly higher values in the midwest than the northeast.
Examination of land use characteristics showed that the highest N03"
concentrations were found in agricultural areas, and the lowest
concentrations were found in forested areas. Comparison of 1987 and
1988 concentration data for five sites showed no significant
difference from year to year.
2. Seasonal variability was considerable for all four atmospheric
constituents. Concentrations of SO^2 and HN03 increased
deterministically at all sites and by a factor of 2-3 networkwide from
winter to summer. In contrast, concentrations of N03 and S02
decreased at all sites and by a factor of two networkwide over the
same period.
3. Analysis of the water soluble component of atmospheric particulate
ammonium (NH^), particulate sodium (Na+) , particulate potassium (K+),
particulate calcium (Ca+2), and particulate magnesium (Mg+2) during
the fourth quarter of 1988 showed that NH^ was the dominant cation at
virtually all sites, followed by Ca+2 and Mg+2. Concentrations of NH^
were appreciably higher at midwestern agricultural sites than at
northeastern forested sites, suggesting a chemical association with
N03. Calculation of ionic balances for atmospheric aerosol showed
-------
good agreement for northeastern and southeastern sites, but a general
excess of cations for the majority of midwestern sites.
4. Comparison of day and night concentration data for selected sites
indicated that S02 and HN03 are typically found at moderately to
substantially lower concentrations at night. Nocturnal depletion
increased from a site in flat terrain to one in complex terrain, but
was not observed for a mountaintop site in Virginia. Day/night
variability of SO^2 and N03 was much less than that of the reactive
gases.
5. Results of continuous ozone (03) monitoring showed that concentrations
above 100 parts per billion (ppb) were observed only during the period
of May through August and that the majority of these were associated
with four episodes of stagnating high pressure over the eastern United
States. Examination of time series data for one such episode showed
that peak concentrations were similar across a transect of sites from
central Kentucky to Virginia. Average concentrations decreased from a
mountaintop site to a rolling terrain site to a complex terrain site,
presumably due to 03 depletion within the nocturnal inversion layer.
6. Annual dry deposition rates for SO^2, N03, S02, and HN03 were estimated
for selected sites using average annual concentration data and assumed
values for deposition velocity. Results have an uncertainty of at
least 50 percent. Dry deposition for S02 and HN03 appear to be much
greater than SO^2 and N03, respectively, unless estimated deposition
velocities are greatly in error. Comparison of wet deposition and dry
deposition of various sulfur and nitrogen species indicated that the
two may be comparable over much of the eastern United States.
Application of one or more inferential methods of dry deposition is
expected to refine these estimates.
-------
SECTION 3.0
NETWORK DESCRIPTION AND OPERATIONS
3.1 NETWORK DESCRIPTION
Figure 1 shows the location and status of 51 sites comprising NDDN, as
of December 1988. Forty-one eastern sites and two western sites were in
full operation at the end of the year. Monitoring equipment was installed
and awaiting utility connection at eight sites (one eastern, seven
western), all of which were scheduled for initiation by early to mid 1989.
The single eastern site (111--Speedwell, TN) was selected as a replacement
for the other site in eastern Tennessee (102--Oak Ridge). Site 102 was
demobilized at the end of 1988 due to operational difficulties and the
proximity of large sources of S02 and NOX.
The names, locations, reporting dates, elevation, terrain, and land-
use classifications of operational NDDN sites are listed in Table 1.
Terrain and land-use information refers to a 10-kilometer (km) radius
around the site and is presented to convey a sense of the setting within
which each site operates. Note that site numbers are used for
identification purposes only and do not correlate with order of
installation or operation.
Of the sites listed in Table 1, 19 were operational during the first
calendar quarter of 1988, 28 were operational during the second quarter,
and 33 were operational during the third quarter. With the exception of
the Georgia site (153--Georgia Station), all sites established through the
first half of 1988 were located in a region bounded by the Mississippi
River to the west and by Tennessee and North Carolina to the south.
Deployment in this fashion assisted another EPA program designed to provide
field data for evaluation of one or more regional acid deposition models
(Barnes et al., 1987) . Dual sets of monitoring equipment were operated at
West Point, NY (i.e., Site 103/4) for the first through third quarters of
1988. The purpose of this effort was to obtain precision data for
independently operated sensors.
For purposes of discussion, the eastern United States sites can be
divided into three regions: midwestern, northeastern, and southeastern.
The midwest includes the states of Illinois, Indiana, Kentucky, Michigan,
Ohio, and Wisconsin and contains 14 NDDN sites. The southeast includes the
states of Arkansas, Alabama, Georgia, Florida, Mississippi, North Carolina,
and Tennessee and contains 10 sites. The northeast includes the remaining
eastern states, from Virginia to Maine, and contains 17 sites. Inspection
of land-use information in Table 1 shows marked regional differences in the
-------
CO
w
E-
i— i
CO
o
O
a
0>
JD
(D
o
o
C3)
O
Q
Q
O
V)
I
55
05
0)
-------
TABLE 1. NDDN SITES IN OPERATION DURING 1988
Site
No.
101
102
103/4
105
106
107
108
109
110
112
113
114
115
116
117
118
119
120
121
122
123
124
126
127
128
129
130
Site Name
Research Triangle
Park, NC
Oak Ridge, TN
West Point, NY
Whiteface
Mountain, NY
PSU, PA
Parsons, WV
Prince Edward, VA
Woodstock, NH
Connecticut Hill,
NY
Kane Experimental
Forest, PA
M.K. Goddard, PA
Deer Creek State
Park, OH
Ann Arbor, MI
Beltsville, MD
Laurel Hill State
Park, PA
Big Meadows, VA
Cedar Creek State
Park, WV
Morton Station, VA
LLlley Comett
Woods, KY
Oxford, CH
Lykens, OH
Unionville, MI
Cranberry, NC
Edgar Evins State
Park, TN
Arendtsville, PA
Perryville, KY
Bondville, IL
Initial
Reporting
Date
01/06/87
01/06/87
01/06/87
01/06/87
01/06/87
01/14/88
11/01/87
12/31/88
09/14/87
12/31/88
01/08/88
09/30/88
06/30/88
12/31/88
12/10/87
06/30/88
11/09/87
06/03/87
01/19/88
08/18/87
09/30/88
06/30/88
09/30/88
03/22/88
06/30/88
08/11/87
02/09/88
Latitude
35.90
35.96
41.35
44.39
40.79
39.09
37.17
43.94
42.40
41.60
41.09
39.63
42.42
39.03
40.00
38.51
38.83
37.33
37.08
39.53
40.92
43.61
36.11
36.04
39.92
37.68
40.05
Longitude
78.87
84.29
74.05
73.86
77.95
79.66
78.31
71.70
76.65
78.77
80.15
82.26
83.90
76.82
79.25
78.43
80.90
80.55
82.99
84.72
83.00
83.42
82.04
85.73
77.30
84.97
88.37
Elevation Land
(m) Use
94
341
203
622
393
505
146
250
515
618
384
265
267
46
616
1,058
274
972
335
284
296
198
1,219
302
210
279
212
Fores ted-Urban
Forested
Forested
Forested
Agricultural
Forested
Forested
Forested
Forested
Forested
Forested
Agricultural
Forested
UrbanjAgric .
Forested
Forested
Forested
Forested
Forested
Agricultural
Agricultural
Agricultural
Forested
Forested
Agricultural
Agricultural
Agricultural
Terrain
Rolling
Complex
Complex
Complex
Rolling
Complex
Rolling
Complex
Rolling
Rolling
Rolling
Rolling
Flat
Flat
Complex
Mountaintop
Complex
Mountaintop
Complex
Rolling
Flat
Flat
Mountaintop
Rolling
Rolling
Rolling
Flat
(continued)
-------
TABLE 1. (continiied)
Site
No.
133
134
135
137
140
144
146
149
150
151
152
153
156
157
165
168
Site Name
Salanonie
Reservoir, IN
Perkinstown, WI
Ashland, ME
Coweeta, NC
Vircennes, IN
Washington's
Crossing, NJ
Argonne National
Laboratory, IL
Wellston, MI
Caddo Valley, AR
Coffeeville, MS
Sand Mountain, AL
Georgia Station,
GA
Sumatra, FL
Alhambra, IL
Pinedale, WY
Glacier National
Park, MT
Initial
Reporting
Date
06/30/88
09/30/88
12/31/88
11/03/87
08/05/87
12/31/88
07/01/87
06/30/88
09/30/88
12/31/88
12/31/88
06/30/88
12/31/88
06/30/88
12/31/88
12/31/88
Latitude
40.81
45.21
46.61
35.06
38.74
40.32
41.70
44.22
34.18
34.00
34.29
33.18
30.11
38.87
42.93
48.51
Longitude
85.68
90.61
68.41
83.43
87.49
74.86
87.99
85.82
93.10
89.80
85.96
84.41
84.99
89.62
109.79
114.00
Elevation Land
(m) Use
219
522
235
686
134
72
229
292
71
134
352
266
14
164
2,388
968
Agricultural
Agricultural
Agricultural
Forested
Agricultural
Agric . -Urban
Urban- Agric.
Forested
Forested.
Forested.
Agricultural
Agricultural
Forested
Agricultural
Range
Forested
Terrain
Flat
Rolling
Flat
Complex
Rolling
Rolling
Flat
Flat
Rolling
Rolling
Rolling
Rolling
Flat
Flat
Rolling
Complex
Source: Hunter/ESE, 1989.
-------
character of sites. Midwestern sites are predominantly agricultural, while
those in the northeast and southeast are mostly forested. Terrain
characteristics also differ systematically between regions. The majority
of midwestern sites are located on flat countryside and only that in
eastern Kentucky (121--Lilley Cornett Woods) is in complex terrain.
Northeastern sites in contrast are roughly evenly divided between rolling
and complex terrain, and only one site (116--Beltsville, MD) is in flat
terrain. Three northeastern sites are located above 1,000 meters (m) and
form a line extending southward from northern Virginia to North Carolina.
Site 118 (Big Meadows, VA) is situated at the crest of the Blue Ridge;
Site 120 (Horton Station, VA) and Site 126 (Cranberry, NC) occupy the spine
of the Appalachian Mountains. Due to the unique exposure of these sites,
they have been placed in a separate terrain category (i.e., mountaintop).
Regional differences in terrain and land use, as reflected by the
network, appear to faithfully represent actual differences between the
midwest, northeast, and southeast. The midwest is indeed largely
agricultural flatland, while the northeast is largely forested. By
definition, the Appalachian Mountains and foothills are almost entirely
contained within the northeastern and southeastern regions. Many of the
sites classified as complex are located along the eastern and western
flanks of the Appalachians. As will be discussed later, these regional
differences in terrain and land use appear to exert significant influences
over the air quality and meteorology observed across the network.
3.2 NETWORK OPERATIONS
This section summarizes the field, laboratory, and data management
center (DMC) operations for the NDDN. Step-by-step protocols and
additional details of these activities can be found in the NDDN Field
Operations Manual, Laboratory Operations Manual, and Data Management
Manual.
Ambient measurements for 03, S02, SO^2, N03, HN03, windspeed, wind
direction, temperature, relative humidity, solar radiation, precipitation,
and delta temperature were performed at each NDDN site. Meteorological
parameters and 03 concentrations were recorded continuously and reported as
hourly averages. Atmospheric sampling for SO^2, N03, HN03, and S02 was
integrated over 7 day and night collection periods using a three-stage
filter pack (see Figure 2). In this approach, particles and selected gases
are collected by passing air at a controlled flow rate through a sequence
of Teflon®, nylon, and base-impregnated filters. Filter packs were
prepared and shipped to the field weekly and changed out at each site every
Tuesday. Day filter pack samples were collected over the hours of 0800 to
2000 local standard time while night samples were collected over the
remaining hours of the day. Blank filter packs were collected on a monthly
basis to evaluate passive collection of particles and gases. At sites
located more than 50 km from National Atmospheric Deposition
Program/National Trends Network (NADP/NTN) sites, wet deposition monitoring
equipment was installed. Precipitation samples will be collected weekly
and shipped to the Hunter/ESE laboratory for chemical analyses beginning in
1989. Table 2 lists the equipment installed and operated at NDDN sites.
-------
E
0)
tn
Q.
t_
0)
-------
TABLE 2. NDDN MONITORING EQUIPMENT
Item
Manufacturer
Model Number
Equipment Shelter
[8 feet (ft) by 8 ft by 10 ft with
electricity and telephone]
03 Analyzer
Meteorological System
Windspeed
Wind Direction
Temperature
Delta Temperature
Relative Humidity
Solar Radiation
Precipitation (tipping bucket)
10-Meter (m) Tower
Data Acquisition System (DAS)
Primary DAS
Backup DAS
Personal Computer (PC)
Telecommunications Modem
Printer
EKTO 8810
Thermo-Environmental 49-103
Climatronics F460
or
R.M. Young AQ Series
Universal Mfg. 4-30
Odessa SDM-3260
Odessa DSM-3260L
Various Various
Packard-Bell ' 2424
Star SD-10
Air Quality Monitoring System
10-Meter Tower
Filter Packs
Pump
Flow Controller
Solenoid Valve
Wet Deposition Equipment (selected sites)
Precipitation Collector
Rain Gauge (weighing)
Triple-Beam Balance
Aluma Tower
Savillex
Thomas
Teledyne-Hastings
Mace
Andersen
Belfort
Ohaus
AT-048
101-CA11
CST-10K
800-2334
APS
5915R-12
1119-D
Source: Hunter/ESE, 1989.
11
-------
3.2.1 Field Operations
Each site was equipped with a shelter (complete with telephone and
100-amp electrical service), two 10-m towers, a meteorological system, an
03 and air quality monitoring system, and a data acquisition system (DAS).
One tower was used for meteorological measurements, using a Climatronics
F460 system or R.M. Young AQ system. Windspeed and wind direction were
measured at 10 m, temperature was measured at 9 m and 2 m, and relative
humidity was measured at 9 m. Precipitation and solar radiation were
measured on 1-m platforms located outside the rain and sun shadows of the
shelter and towers.
Filter pack sampling and 03 measurements were performed at 10 m using
a tilt-down aluminum tower (Aluma, Inc.). Day, night, and blank filter
packs were fitted with non-interchangeable quick connects to prevent
confusion and to reduce time for changeout in the field. Filter pack flow
was maintained at 1.50 liters per minute (L/min) [for standard conditions
of 25 degrees Celsius (°C) and 760 millimeters of mercury (mmHg)] with a
Teledyne-Hastings CST-10K mass flow controller (MFC) and recorded as hourly
averages on the data acquisition system. Switching from the day filter
pack to the nighttime filter pack was performed by a relay-activated
solenoid which was, in turn, controlled by the DAS.
03 was measured via ultra-violet absorbance with a Thermo-
Environmental Model 49-103 analyzer operating on the 0 to 500 ppb range.
Ambient air was drawn through an all-Teflon® sampling line upstream of the
03 analyzer. A 3/8-inch TFE Teflon® sample line was used to sample air
atop the 10-m air quality tower. Teflon® filters housed at the tower inlet
and the analyzer inlet prevented particle deposition within the system.
Periodic checks indicated that line losses through the inlet system were
consistently less than 3 percent. Zero, precision (60 ppb), and span
(400 ppb) checks of the 03 analyzer were performed every third day using an
internal 03 generator.
The onsite DAS consisted of a primary datalogger (Odessa 3260) , a
back-up datalogger (Odessa 3260L), an IBM-compatible personal computer
(PC), a printer, and a telephone modem. The primary datalogger was used to
acquire, average, store, and communicate readings from all continuous
sensors. It also performed a variety of control functions, such as
switching the day/night filter pack relay and activating the internal 03
generator for consistency checks. Data were stored on the primary and
backup dataloggers in electronically erasable programmable read-only memory
(EEPROM) cartridges, both of which were accessible through the onsite PC,
or remotely, through the telephone modem. Printouts of hourly averages for
the previous day were automatically generated by the onsite printer a few
seconds after midnight each day.
Site operators visited each site on Tuesdays and Fridays and in
response to equipment malfunctions or suspected problems. On Tuesdays,
filter packs were changed out, sample lines were leak tested, sensors were
12
-------
subjected to electronic and reasonableness checks, and data from the
dataloggers were downloaded to diskettes. The site operator telephoned the
NDDN operations center at Hunter/ESE following site inspections to relay
observations and problems. Data, documentation, and samples were shipped
to Hunter/ESE the day of collection. On Fridays, the site operator
performed a similar but limited check of sensors and equipment.
All field equipment was subjected to inspections and multipoint
calibrations (using standards traceable to the National Institute of
Standards and Technology) by Hunter/ESE personnel on a quarterly basis. In
addition, independent equipment audits were performed semiannually by ERG
Environmental and Energy Services, Inc. (formerly Westec, Inc.) and
randomly by EPA or its designee. Results of field calibrations were used
to assess sensor accuracy and flag, adjust, or invalidate field data.
The overall precision of field measurements was evaluated by operation
of dual sets of equipment at West Point, NY (i.e., Sites 103 and 104).
Operation of Site 103 was terminated in October 1988 and, therefore,
precision can be estimated only for the first three quarters of 1988.
Plans were initiated for establishing five collocated sites in 1989 to
provide precision information for the major geographical regions covered by
the network. Precision and accuracy criteria for NDDN field measurements
are shown in Table 3.
3.2.2 Laboratory Operations
Filter pack samples were loaded, shipped, received, extracted, and
analyzed by Hunter/ESE personnel at the Gainesville, FL laboratory. The
flow of samples through the laboratory is shown in Figure 3. Filter packs
contained three types of filters in sequence: a Teflon® filter (Gelman,
Zefluor, 1.0 /im) for collection of aerosols, a nylon filter (Gelman,
1.0 /Lim) for collection of HN03, and dual K2C03 impregnated cellulose
filters (Whatman 41) for collection of S02. All filters were subjected to
acceptance tests before they were used in the network. Filters showing
detectable levels of SO^2 or NO^ were either washed and re-tested or
returned to the manufacturer for credit.
Following receipt from the field, exposed filters were placed in
color-coded extraction bottles and extracted as shown in Figure 4.
Extracts were then analyzed for SO^2 and N03 by micromembrane suppressed
ion chromatography (1C) using a Dionex Model 4000i 1C equipped with an
Autoion 1000 Controller. The instrumental configuration included a Dionex
Automated Sampler, an AG4-A guard column, a AS4-A separator column, and a
50-microliter (ML) injection loop. Data acquisition, display, and analysis
was accomplished with a Maxima computer integration system. The 1C was
calibrated prior to each analytical run with five standards plus a blank
covering the range of 35 to 9,000 micrograms per liter (Mg/L) for N03 and
40 to 10,000 Mg/L for SO^2. Extraction and calibration solutions were
spiked with 1.0 parts per million (ppm) particulate bromine (Br~) as a
check on 1C column condition and to assist in detection of spurious sample
injections.
13
-------
TABLE 3. PRECISION AND ACCURACY OBJECTIVES OF FIELD MEASUREMENTS
Measurement
Parameter
Method
Acceptance Criteria*
Precision
Accuracy
Windspeed (vector
and scalar)
Anemometer
±0.5 ra/sec
±0.2 m/sec
Wind direction
Standard deviation
of wind direction
Relative humidity
Solar radiation
Precipitation
Ambient temperature
Delta temperature
Ozone (03)
Filter pack flow
Wind vane
Wind vane
Humidity sensor
Pyranometer
Rain gauge
Thermistor
Thermistor
03 analyzer
Mass flow meter
±5°
Undefined
±10% (of
full scale)
±10% (of
reading)
±10% (of
reading)
±0.5°C
±0.2°C
±10% (of
reading)
±0.15 L/min
±2°
Undefined
±10%
±10%
0.10 inch+
±0.25°C
±0.1°C
±10%
±10%
Note: ° — degree.
°C - degrees Celsius.
L/min = liters per minute.
m/sec — meters per second.
*Field precision criteria apply to collocated instruments; accuracy criteria
apply to calibrations and audits of instruments.
+For target value of 1.00 inch.
Source: Hunter/ESE, 1989.
14
-------
SET UP FIELD GROUPS
IN CLASS
PREPARE FILTERS IMPREGNATE
WHATMAN ONLY - 01M KfO3
RUN ACCEPTANCE TEST OF FILTERS
1C
ASSEMBLE FILTER PACKS „
WHATMAN, NYLON, TEFLON®
DATA MANAGEMENT
ASSIGN SAMPLE NO.
LOG IN SAMPLE
ENTER FIELD DATA
FINALIZE DATA BATCH
SHIP FILTER PACKS TO
SITE OPERATORS
(WEEKLY)
REVIEW
DATA
SAMPLE RECEIPT, VISUAL INSPECTION,
COMPLETE LOGSHEET <72 HOURS
DISASSEMBLE FILTER PACK
QUARTERLY REPORT
QC SUMMARY
(WHITE LABEL)
NYLON FILTER IN
60-ntL BOTTLE
(ORANGE LABEL)
WHATMAN FILTER IN
60-mL BOTTLE
(BLUE LABEL)
STORE AT 4°C
UNTIL EXTRACTION
STORE. AT 4°C
UNTIL EXTRACTION
STORE AT 4°C
UNTIL EXTRACTION
EXTRACT
25mLOFDEIONIZED
WATER PLUS 1-ppmBr
EXTRACT
25mLOF0.003NN»OH
PLUS 1-ppmBr
EXTRACT
50mLOF0.05%H202
PLUS 1-ppmBr
OC REVIEW. Ml ORANGE
STANDARD, DRIFT
CHECK, NBS RECOVERY
STORE AT 4°C. NOT LESS THAN
12 HOURS OR MORE THAN 60 HOURS
1 ALL EXTRACTS
ANALYZE BY 1C
SOfNOj
TEF
EXT
ONI
LOW*
PACTS
_Y
ANALYZE BY
COLORIMETRY
NHJ
INITIATE DATA TRANSFER
ANALYZE BY
I CAP
ANALYZE BY
ATOMIC EMISSION
K*
Figure 3. Flowchart of laboratory operations for filter pack samples
15
-------
TEFLON
I
PUCE IN 60-mL
PLASTIC BOTTLE
(WHITE LABEL)
REFRIGERATE 4° C
UNTIL EXTRACTION
ADD 25 mL OF
DEIONIZED WATER
WITH 1 ppm Br
SHAKER TABLE
FOR 15 MINUTES
ULTRASONIFY
FOR 30 MINUTES
SHAKER TABLE
FOR 15 MINUTES
LET SIT OVERNIGHT
IN REFRIGERATOR
AT4'C
NYLON
I
PLACE IN 60-mL
PLASTIC BOTTLE
(ORANGE LABEL)
REFRIGERATE 4° C
UNTIL EXTRACTION
ADD 25 mL OF
0.003N NaOH
WITH 1 ppm Br
SHAKER TABLE
FOR 15 MINUTES
ULTRASONIFY
FOR 30 MINUTES
SHAKER TABLE
FOR 15 MINUTES
LET SIT OVERNIGHT
IN REFRIGERATOR
AT4°C
WHATMAN
I
PLACE IN 60-mL
PLASTIC BOTTLE
(BLUE LABEL)
REFRIGERATE 4° C
UNTIL EXTRACTION
ADD 50 mL OF
0.05% H 202
WITH 1 ppm Br
SHAKER TABLE
FOR 15 MINUTES
ULTRASONIFY
FOR 30 MINUTES
SHAKER TABLE
FOR 15 MINUTES
LET SIT OVERNIGHT
IN REFRIGERATOR
AT4°C
Figure 4. Flowchart for extraction of NDDN air quality filters
16
-------
Beginning in the third quarter of 1988, Teflon® filter extracts were
analyzed routinely for NHj, Na+, K+, Mg+2, and Ca+2. Analysis of Na+,
Mg+2, and Ca+2 was performed with a Perkin-Elmer P-2 inductively coupled
argon plasma (ICAP) emission spectrometer. Analysis of NH^ was by the
automated indophenol method using a Technicon II or TRAACS-800 AutoAnalyzer
system. Analysis of K+ was via atomic emission on a Perkin-Elmer 5100
atomic absorption spectrophotometer.
Various QC samples were routinely analyzed to track the accuracy and
precision of laboratory data. NIST SRM No. 2694-11 (simulated rainwater)
was analyzed immediately after instrument calibration to monitor accuracy.
Recoveries within ±5 percent of certified values were required for analyses
to continue. A control sample was analyzed after every 10 environmental
samples to monitor within-run precision. For the analytical batch to be
accepted, the maximum difference between the first analysis and each
subsequent analysis was limited to ±5 percent. Ten percent blind
replicates were also analyzed to monitor between-run precision. Due to the
potential for change in analyte concentration between analyses, replicate
samples were not used as a control for acceptance of batches. Finally, one
unknown sample provided by the EPA was analyzed with each batch. The
unknown consisted of filter media spiked with salt solutions containing
SC>42 and NO^. These were carried through the entire extraction and
analysis procedure established for Teflon®, nylon, and K2C03-impregnated
cellulose filters. Precision and accuracy objectives for NDDN laboratory
analyses are listed in Table 4.
Results of all valid analyses were stored in units of micrograms per
filter ()Hg/filter) in the laboratory data management system.
Concentrations of all species were then calculated (based on volume of air
sampled) following validation of hourly flow data. Atmospheric
concentrations of particulate SO^2, NOj, NH^, Na+, K+, Mg+2, and Ca+2 were
calculated based on the analysis of Teflon® filter extracts; HN03 was
calculated based on the NO^ found in nylon filter extracts; and S02 was
calculated based on the sum of SO^2 found in nylon and cellulose filter
extracts. Since Teflon® filters were extracted in deionized water, it
should be noted that reported concentrations of particulate species refer
to the water soluble component only.
3.2.3 Data Management
Data management center activities consisted of three major operations:
data acquisition, validation, and transmittal to EPA. These are depicted
in Figure 5. The data acquisition process stressed multiple levels of
redundancy in order to minimize data loss. The primary mode of data
acquisition from the field was via telephone modem. Each site was
automatically polled on a daily basis using a PC and software developed by
Odessa Engineering, Inc. The polling software permits recovery of hourly
data and status files, power failure logs, and automated calibration
results from the previous 7 days. The program also maintains
synchronization of the network by checking the clock within each DAS and
correcting the time if it deviates from expectation by more than 2 minutes.
If daily polling resulted in incomplete data capture from any site, then
17
-------
TABLE 4. PRECISION AND ACCURACY OBJECTIVES KR LABCRATCRY MEASUREMENTS
FOR NDDN
Acceptance Criteria
Analyte
Anmonium (NHl)
Calciun (Ca+2)
Magnesiun (Mg+2)
Sodiun (Na+)
Potassium (1C1")
Sulfate (SC>42)
Nitrate (N0£)
Sanple
Type
Filter Extract
Filter Extract
Filter Extract
Filter Extract
Filter Extract
Filter Extract
Filter Extract
tethod
Technicon AutoAnalyzer
Inductively coupled
plasma --atonic emission
Inductively coupled
plasma --atomic emission
Inductively coupled
plasma- -atomic emission
Inductively coupled
plasma- -atomic emission
Ion chromatography
Ion chromatography
Precision*
(percent)
10
10
10
10
10
5
5
Accuracy"1"
(percent)
90 to 110
90 to 110
90 to 110
90 to 110
90 to 110
95 to 105
95 to 105
*Determined from midlevel standard and initial calibration curve.
-(-Determined from reference traceable to the National Institute of Standards and
Technology (MIST).
Source: Hunter/ESE, 1989.
18
-------
CONTINUOUS DATA
I
p
—
4INTOUT
^
DSH-3260
DSM-3260L
DAILY REPORT
AND
EXCEPTION
REPORT,
RAW DATABASE
HOURLY AVERAGES
Figure 5. Flow of data through the NDDN data management system
19
-------
diskettes of data from the primary and backup DAS were read into the
database management system. If the database was still incomplete, then
missing data were entered manually from site printouts. Each datum was
automatically given a source flag that could be used to trace its mode of
entry into the system (i.e., modem, primary DAS, backup DAS, or printout).
Data validation consisted of thorough review of operator logs, onsite
reasonableness checks, results of field calibrations and audits and a
variety of parameter-specific range and consistency checks. In addition,
diurnal patterns for numerous parameters and a variety of inter-parameter
relationships were examined for reasonableness. Solar radiation data, for
example, were used to check for time shift errors in the database (caused
by power failure), while significant rainfall events were used to check the
response of relative humidity and temperature sensors (the former should
approach 100 percent, and delta temperature should approach zero).
Following validation of data for a calendar quarter, flag counts, and
parameter averages were calculated. The database, flag counts, and
averages were then loaded onto a 9-track tape and submitted to the NDDN
Project Officer. To verify the data transfer, flag counts and parameter
averages were independently generated by EPA and echoed back to Hunter/ESE.
Data sets were subjected to a variety of reasonableness and consistency
checks by EPA before final acceptance into the NDDN database.
20
-------
SECTION 4.0
RESULTS AND DISCUSSION
4.1 FILTER PACK MEASUREMENTS
4.1.1 Annual Concentration Data
Annual arithmetic average concentrations of SO^2, N03, HN03, and S02
are shown in Figures 6 and 7 for 18 sites in operation throughout 1988.
Aerosol SO^2 concentrations range from 7 . 5 Mg/m3 in northern West Virginia
(Site 107) to 3.4 Mg/m3 in northern New York (Site 105). Concentrations
above 7.0 Hg/m3 occur in a band extending from southwestern Pennsylvania to
Virginia and also in central Kentucky. Concentrations of 5.0 to 5.9 /Llg/m
extend around virtually the entire periphery of the network from northern
Illinois to southwestern North Carolina to northern New York. A strong
concentration gradient (i.e., factor of 2) occurs between the local maximum
in West Virginia and northern New York. Weaker gradients extend
southeastward and northwestward from the locus of concentrations above
7.0
Aerosol N03 concentrations exhibit much more variability than SO^ and
a definite pattern of higher concentrations in the midwest than elsewhere.
The lowest concentrations (<0.3 /Ltg/m3) are observed in forested sites in
the Adirondack and Appalachian Mountains (i.e., Site 105 and Site 137),
while the highest concentrations (>2.0 /ig/m ) are observed in urban and
agricultural areas of the midwest. The only northeastern site to exhibit
annual N03 concentrations in excess of 1.0 Hg/m3 was located on a research
farm operated by the Pennsylvania State University (Site 106) . This
observation supports the notion of a linkage between agricultural activity
and N03 concentrations. Two potential mechanisms for N03 production
include gas phase reaction of HN03 and NH3 and gas -particle reaction of
HN03 with soil particles. Both of these reactions are likely to be
enhanced in agricultural areas, relative to forested areas.
Annual averages for HN03 exhibit maximum concentrations around
3.5 Mg/ro3 in central Kentucky and southwestern Virginia and minimum
concentrations of 1 . 1 Jlg/m3 i-n eastern Kentucky and southwestern North
Carolina. The majority of concentrations networkwide fall in the range of
2.5 to 3.0 Mg/m3. The overall pattern for HN03 might be influenced by
terrain as much or more than other factors (e.g., emissions). Sites with
the highest average values typically have good exposure to wind flow
(fetch) while those with the lowest values typically have poor exposure due
mainly to complex terrain. For a species, such as HN03, with a relatively
21
-------
B
Figure 6. Annual average concentration of S042 (A) and NO3 (B) OL/g/m3) for 1988
22
-------
B
Figure 7. Annual average concentration of HN03 (A) and S02 (B)
for 1988
23
-------
high deposition velocity, the microclimate in regions of complex terrain
could induce small-scale variability in atmospheric concentrations.
Results for S02 show maximum annual concentrations around 20 Mg/ra3 in
western Pennsylvania, eastern Tennessee, and northern Illinois, and minimum
concentrations below 5.0 Hg/m3 in southwestern North Carolina, northern New
York, and eastern Kentucky. Large concentration gradients (i.e., factor of
3 or more) occur within the states of New York and Kentucky and across the
southern Appalachians from Tennessee to North Carolina. Other than the
relatively high values in eastern Tennessee, the overall pattern is
reminiscent of that for HN03.
Average concentrations for 1987 and 1988 are compared for five sites
in Table 5. Results show highly consistent concentrations from year to
year. SO^2 values differ by up to 1.0 Jig/m3, or approximately 14 percent
at Site 102, but are not significantly different for any site at the 95-
percent confidence level. Annual averages for N03 differ by no more than
0.1 jUg/m3, which is on the order of magnitude of the field precision of the
sampling/analysis protocol currently in use (see Section 4.5). Similar
results are obtained for HN03 and S02. In general, the data presented in
Table 5 indicate no evidence of significant inter-annual variability (based
on two years of data).
4.1.2 Seasonal Concentration Data
4.1.2.1 First Quarter 1988
Average concentrations of SO^2, N03, HN03, and S02 for first quarter of
1988 are shown in Figures 8 and 9. Results show unique patterns for each
species, but in all cases the lowest values are observed at the northern
and southern extremes of the Appalachian Mountains (i.e., Site 105 and
Site 137). For SO^2, quarterly averages range from 2.5 Mg/m3 at Site 105
to 4.9 |ig/m at Site 106 and are slightly higher in the eastern and
southeastern sites than the midwestern sites. Two strong concentration
gradients occur between central Pennsylvania and upstate New York and
between eastern Tennessee and southwestern North Carolina.
Aerosol N03 concentrations range from 0.3 /ig/m3 (Sites 105 and 137) to
4.1 /ig/m3 (Site 146) and show characteristic differences between the
midwest (i.e., Onio,. Indiana, and Illinois) and the northeast. Midwestern
concentrations of N03 are uniformly above 2.0 Mg/m3 (mean 3.1 /Jg/m3) , while
all other sites except central Pennsylvania are below 2.0 Mg/m3 (mean
0.8 /ig/m3) . This pattern may be related to higher levels of dust and/or
ammonia in the agricultural midwest than in the forested northeast.
HNOo data for the quarter range from 1.2 Mg/m3 at Site 146 to
3.5 Mg/m3 at Site 120. The lowest values (i.e., below 2.0 Mg/ra3) are
observed in northern New York, western North Carolina, and Illinois and
highest values (i.e., above 3.0 /Lig/m3) in central New York, Kentucky, and
southwestern Virginia. Due to its high deposition velocity, the spatial
distribution of HN03 is likely more sensitive to elevation and exposure of
sites than any of the other measured parameters. This might explain what
24
-------
TABLE 5. OCMPARISCN OF ANNUAL AVERAGE GCNCENIEATICNS
AT FIVE SITES FOR 1987 AND 1988*
Site
101 Research Triangle Park, NC
102 Oak Ridge, TN
104 West Point, NY
105 Whiteface Mountain, NY
106 Perm State University, PA
Year
1987
1988
1987
1988
1987
1988
1987
1988
1987
1988
so;2
6.0 ±3.4
5.9 ±3.7
7. 6 ±6.1
6.6 ±5.4
5.4 ±2. 8
5.4 ±3.0
3.0 ±1.7
3.4 ±1.9
7.2 ±3.2
6.7 ±3.3
™*
0.63 ±0.38
0.64 ±0.37
0.58 ±0.36
0.51 ±0.34
0.62 ±0.27
0.63 ±0.28
0.29 ±0.15
0.27 ±0.13
1.5 ±0.62
1.4 ±0.59
HN03
2.6 ±1.1
2.7 ±1.2
2.7 ±0.83
2.4 ±0.88
2.7 ±0.96
2.8 ±0.87
1.5 ±0.75
1.6 ±0.84
3.0 ±1.4
2.8 ±1.2
so,
8.1 ±5. 6
8.8 ±4.9
20 ±9. 7
21 ±9.4
12 ±6.7
13 ±6.6
3.7 ±1.6
4.2 ±1.8
17 ±8.8
16 ±8.7
^Values shown indicate arithmetic average and standard deviation.
Source: Hunter/ESE, 1989.
25
-------
Figure 8. Average SO;2 (A) and NO^ (B) concentrations
first quarter 1988
26
-------
Figure 9. Average HN03 (A) and S02 (B) concentrations
first quarter 1988
27
-------
appear to be somewhat depressed concentrations at sites located in the
complex terrain of central West Virginia and southwestern North Carolina.
Relatively low HN03 in the midwest versus the northeast could be related to
gas scavenging by alkaline dust particles or to reactions with atmospheric
NH3, both of which might be present at higher concentrations in
agricultural areas than in forested areas.
The S02 pattern for the period shows considerable spatial variability.
Quarterly average values range from 4.2 Mg/m3 in southwestern North
Carolina to 28 /ig/m3 in southwestern Pennsylvania. A pronounced gradient
occurs across New York state and a moderate gradient occurs from the
midwest to Virginia and North Carolina. Results for eastern Tennessee
(Site 102-Oak Ridge, TN) are significantly higher than those for western
North Carolina and central Kentucky, suggesting local influences. Indeed,
Site 102 is ringed by a large number of S02 point sources. Data from
Site 127 in central Tennessee should assist in evaluating the magnitude of
local influences in the Oak Ridge area.
4.1.2.2 Second Quarter 1988
Average concentration data for SO^2, N03 , HN03, and S02 for second
quarter 1988 are shown in Figures 10 and 11. As was observed for the first
quarter, lowest overall concentrations are typically found in upstate New
York (Site 105) and southwestern North Carolina (Site 137). SO^2 values
range from 9.3 Mg/m3 at Horton Station, VA (Site 120), to 3.0 Mg/m3 at
Whiteface Mountain, NY (Site 105). The highest SO^2 concentrations for the
quarter (i.e., >8.0 )Llg/m3) were found in a band of sites from eastern
Tennessee to southwestern Pennsylvania. From this band, concentrations
drop off gradually to the east and northwest but dramatically to the
northeast (i.e., through Pennsylvania and New York state). As expected,
the overall S0<|2 pattern is relatively smooth and suggests a slight
easterly displacement of peak concentrations with respect to the major
source areas of S02.
Aerosol N03 concentrations range from 0.24 jig/m3 (Site 137) to
2.0 /ig/m3 (Site 146) and show characteristic differences between the
midwest and northeast. Midwestern concentrations are generally at or above
1.0 /Ltg/m3, whereas northeastern concentrations are usually well below
1.0 Mg/m . As mentioned previously, this pattern appears to reflect
differences in land use between the two regions. The midwestern sites are
typically located in or near areas of heavy agriculture, whereas the
northeastern sites are typically located in rural-forested settings. The
principal exception to this is Site 106, which occupies an agricultural
research farm of the Pennsylvania State University, and, perhaps
coincidentally, is the only site in the northeast with N03 concentrations
above 1.0 /Ltg/m3. Thus, agricultural activity appears to correlate with
aerosol N03 concentrations.
Quarterly average HN03 concentrations range from 1.5 jUg/™3 (Site 105
and 137) to 4.2 Mg/m (Sites 122 and 129) and exhibit a tendency for
slightly higher concentrations in the midwest than elsewhere. As expected
for a secondary air pollutant, the overall pattern is relatively free of
28
-------
Figure 10. Average SO^2 (A) and NOj (B) concentrations
second quarter 1988
29
-------
Figure 11. Average HN03 (A) and S02 (B) concentrations (jug/m3)-
second quarter 1988
30
-------
sharp concentration gradients. Average concentrations for West Virginia
and eastern Kentucky, however, are slightly to markedly lower than for
surrounding sites. This suggests that HN03, with its high deposition
velocity, may be strongly influenced by microclimate and/or terrain. In
areas of extremely complex terrain (such as eastern Kentucky and West
Virginia) , it is conceivable that valleys and ridges are both
representative and that differences in microclimate will cause sharp
gradients in the ambient concentration of rapidly depositing species.
S02 concentrations range from 2.3 Mg/n>3 (Site 105) to 23 Mg/m3
(Site 102) and are generally above 10 Mg/ro3 from Indiana eastward through
Ohio, Kentucky, and Pennsylvania and below 10 Mg/m3 elsewhere.
Persistently high concentrations in eastern Tennessee and northern Illinois
(Sites 102 and 146) are almost certainly the result of local emissions.
Results for the eastern Kentucky site (121) and the southwestern North
Carolina site (137) show marked concentration depressions, indicating that
S02, like HN03, may be sensitive to microclimate.
4.1.2.3 Third Quarter 1988
Average concentration data for SO^2, N03, HN03, and S02 during the
third quarter 1988 are depicted in Figures 12 and 13. Average SO^2 values
range from 13.3 Mg/m3 at Parsons, WV (Site 107) to 4.7 Hg/m3 at Wellston,
MI (Site 149). The highest SO^2 concentrations (i.e., >10 Mg/m3) for the
quarter were found in a band of sites extending from southwestern Indiana
to central Pennsylvania and the mountains of Virginia. From this band,
concentrations drop off gradually to the east and northwest and rapidly to
the northeast (i.e., through Pennsylvania and New York state). Relatively
strong gradients in concentration occur in Michigan and New York state. In
both cases, concentrations decrease by 30 to 40 percent from south to north
over a distance of a few hundred kilometers.
Aerosol N03 concentrations range from less than 0.2 /Ltg/m (Sites 137,
102, and 119) to 2.0 Mg/m3 (Site 146) and show characteristic differences
between the midwest and northeast. Midwestern concentrations, except for
Site 149, are above 0.75 /Ltg/m3, while northeastern concentrations, except
for Site 128, are invariably below 0.75
Quarterly average HN03 concentrations range from 0.9 Mg/ro3 (Site 137)
to 4.7 Mg/m3 (Site 128). Concentrations above 3.0 Mg/m3 occur in all
states except Tennessee, Georgia, and West Virginia. As for SO^2,
concentration gradients are found in Michigan and New York state. Unlike
SO^2, sharp gradients also are found in areas of complex terrain extending
from southwestern North Carolina to northern West Virginia. Sites 137,
121, 120, and 107 show slightly to markedly depressed concentrations with
respect to neighboring sites.
S02 concentrations range from 1.1 Mg/m3 (Site 137) to 19.8
(Site 146) and are generally above 10 Mg/m3 from southern Illinois eastward
through Ohio and Pennsylvania. Concentrations around the periphery of the
network (i.e., northern New York, northern Michigan, and central Georgia)
are on the order of 2 to 4 Mg/m3- Relatively high concentrations in
31
-------
na — Not Available
Figure 12.
Average SO;2 (A) and NOj (B) concentrations (jug/m3)-
third quarter 1988
32
-------
na — Not Available
Figure 13. Average HN03 (A) and S02 (B) concentrations
third quarter 1988
33
-------
eastern Tennessee, northern Illinois, and southwestern Indiana (Sites 102,
140, and 146) appear to reflect local emissions. Relatively low
concentrations for sites in eastern Kentucky and southwestern North
Carolina (121 and 137) suggest that S02 exposure may be strongly influenced
by terrain, as well as remoteness from sources.
4.1.2.4 Fourth Quarter 1988
Average concentrations for SO^2, NOj, NH^, Na+, K+, Ca+2, Mg"*"2, HN03,
and S02 for the fourth quarter 1988 are depicted in Figures 14 through 18.
Average SO^2 concentrations for the quarter range from 2.3 /ig/m3 at
Site 105 (Whiteface Mountain, NY) to 4.0 Mg/m3 at Sites 101 and 102
(Research Triangle Park, NC; and Oak Ridge, TN), and show remarkably
constant values from the Mississippi River to the Atlantic seaboard.
Significant gradients occur in latitudinal transects through the states of
New York and Michigan where values decrease by about 30 percent from north
to south. Relatively low concentrations at Sites 121 and 137 (both
averaged 2.9 fJig/m3 during the quarter) and, to a lesser extent, Site 119
(3.2 Mg/ni3) , could be terrain related. Results for Site 150 (Caddo Valley,
AR), the most westerly site in operation for the entire quarter could
reflect the transition from expectedly low concentrations in the great
plains and far west to high concentrations in the east.
N03 concentrations varied by more than an order of magnitude across
the network and ranged from 0.26 Mg/m3 at Site 137 (Coweeta, NC) to
3.4 jUg/m3 at Site 133 (Salamonie Reservoir, IN). In general,
concentrations are above 2.0 MgA>3 in the midwestern states of Illinois,
Wisconsin, Indiana, Ohio, and Michigan (except for the extreme northern
site) and well below 1.0 /Llg/m3 nearly everywhere else. The states of
Kentucky and Pennsylvania exhibited intermediate concentrations in the
range of 1.0 to 2.0 /ig/m3, perhaps reflecting transport of aerosol N03 from
the midwest.
Quarterly NH^ concentrations ranged from 0.66 /Ltg/m3 at Site 105 to
2.1 Mg/ra3 at Site 133 and were slightly higher in the midwest than
elsewhere. The overall pattern of NH^ concentrations was similar to that
for SO^ , perhaps reflecting a chemical association between the two (i.e.,
ammonium sulfates). For most sites in the northeast, there is sufficient
NH4 (on an equivalent basis) to completely account for ambient S042
concentrations. In the midwest, there is excess NH4 with respect to S042,
which suggests that some ammonium nitrate (NH4N03) could be present in the
atmosphere.
Results for Na+ range from 54 nanograms per cubic meter nanograms per
cubic meter (ng/m3) at Site 105 to 250 ng/m3 at Site 101 and suggest marine
influences along the Atlantic coast from New York to Georgia. Although no
site is within 50 km of the coast, the four closest sites (Sites 101, 104,
108, and 153) average 188 ng/m3, whereas the remaining sites average
(87 ng/m3). Relatively high Na+ concentrations at Site 146 (160 ng/m3)
almost certainly reflect urban influences of the Chicago metropolitan area
(e.g., road salt). Na+ is frequently used to determine marine-derived S042
in precipitation and aerosols, based on the ratio of Na+ to S042 in bulk
34
-------
B
Figure 14. Average SO^2 (A) and NOj (B) concentrations (/L/g/rrr5)--
fourth quarter 1988
35
-------
B
Figure 15. Average HN03 (A) and S02 (B) concentrations (jug/m3)-
fourth quarter 1988
36
-------
B
Figure 16. Average NH£ (A) concentrations (|L/g/m3) and
Na* (B) concentrations (ng/m3)--fourth quarter 1988
37
-------
Figure 17. Average K* (A) and Ca+2 (B) concentrations (ng/m3)-
fourth quarter 1988
38
-------
Figure 18. Average Mg+2 concentrations (ng/m3)-fourth quarter 1988
39
-------
seawater (i.e., 0.248) and the assumption that all Na+ in the sample is of
marine origin. Calculations for fourth quarter 1988 indicate that marine
sources account for well below 5 percent of the aerosol SO^2 observed
across the NDDN.
Quarterly average K+ concentrations show minimal variability across
the network. Values range from 60 ng/m3 at Site 105 to 210 ng/m at
Site 133, but are below 100 ng/m3 at only two sites in New York state.
There is little evidence of regional variability and, perhaps surprisingly,
no evidence of urban influence at sites such as those located outside of
Chicago (Site 146) or New York (Site 104).
Ca+2 data cover a wide range of concentrations, from 75 ng/m3 at
Site 105 to 2,300 ng/m3 at Site 121. Extremely high values at the latter
site are the result of coal mining traffic in the area, the impact of which
is currently being investigated. In general, Ca+2 results show
considerable site-to-site variability superimposed upon regional
variability. The majority of sites in Ohio, Illinois, Indiana, and
Michigan exhibit concentrations above 300 ng/m3, whereas the remaining
sites are typically well below this value. Relatively high concentrations
(for northeastern sites) at Sites 107 and 120 appear to reflect; nearby
construction activities (now completed). Concentrations at Site 128 may be
the result of local agricultural activity.
The pattern for Mg"1"2 is similar to that for Ca+2 and shows a
combination of local and regional variability. Concentrations range from
17 ng/m3 at Site 105 to 320 ng/m3 at Site 146 and are almost invariably
above 50 ng/m in the midwest but below this figure elsewhere. This
pattern presumably reflects differences in regional land use between the
midwest and for other regions (i.e., agricultural versus forested).
Locally high concentration of Mg+2 in eastern Kentucky (Site 121) appears
to be due to coal mining activities.
Quarterly-average HN03 concentrations range from 0.6 /^g/m3 at Site 121
to 2.9 Mg/m3 at Sites 118 and 120. As for most other parameters, there is
strong evidence of regional variability. In this case, the northeast
exhibits generally higher concentrations than the midwest, and the
southwest exhibits intermediate concentrations. A band of concentrations
on the order of 2.0 Mg/n>3 and above runs nearly north-south from New York
to North Carolina With the exception of Site 129 in central Kentucky and
102 in eastern Tennessee, concentrations outside of this band are less than
1.5 /Ig/m3. Clearly the most dramatic feature of the HN03 distribution is
the gradient between Sites 120 and 121. Concentration differs by nearly a
factor of 5 at these sites, which are separated by only about 250 km.
These two sites also present extremes in terms of local terrain. Site 120
is perched upon a mountain ridge, whereas Site 121 is nestled in a valley.
Results for S02 also show marked variability, both on a regional and
local scale. Concentrations range from 3.1 Mg/m3 at Site 137 to 23.7 Mg/m3
at Site 117. The highest concentrations occur throughout Pennsylvania and
in northern West Virginia, southern Indiana, and northern Illinois.
Relatively low concentrations (i.e., <5.0 jLtg/m3) occur around the periphery
40
-------
of the network (Arkansas, Wisconsin, the northernmost sites in New York and
Michigan) and in mountainous areas of eastern Kentucky and southeastern
North Carolina. In general, the overall pattern appears similar to the
geographical distribution of sources, with occasional punctuation due to
possible terrain effects.
The analysis of NH4 and base metal cations on Teflon* filters permits
evaluation of various ionic relationships in aerosol samples. Ion balances
for fourth quarter 1988 are shown in Table 6 for representative sites in
the northeast, midwest, and southeast. Units of charge have been used to
permit direct comparisons of ionic concentrations across species. Results
show that S042 is the dominant anion at all sites but that the degree of
dominance is less in the midwest than elsewhere. In the northeast and
midwest, S042 typically accounts for 80 to 90 percent and 60 to 80 percent
of measured anions, respectively. NH4 is the dominant cation at all sites,
except Site 121, and typically accounts for 70 to 80 percent of total
cations. In the northeast, low concentrations of N03 suggest that NH4 is
predominantly associated with S042 (or an unmeasured anion). In the
midwest, in contrast, higher N03 concentrations and a frequent excess of
NH4 over S042 strongly suggest the presence of NH4N03. Whether this species
is actually present in the atmosphere or formed as an artifact of the
sampling technique is not known. Results for Ca+2 indicate that it is the
second most abundant cation at virtually all sites and that it is more
abundant in the agricultural midwest than elsewhere. Concentrations of
Na+, K+, and Mg+2 are of only marginal importance to the overall ion
balance. Calculations of ionic ratios based on total measured cations and
anions show that reasonable balances (i.e., ±15 percent) are attained for
most northeastern and southeastern sites. This suggests that the major
ionic species are currently measured and that other ions (e.g., H+ and Cl")
play a minor role in the ionic composition of these aerosol samples. Since
the database for cations is limited, it is unknown whether this is true at
other times of the year. As mentioned earlier, it is also uncertain
whether these results represent true atmospheric conditions or artifact
reactions on Teflon® filters. Ionic balances for midwestern sites show
persistent and frequently large cation excesses. These excesses correlate
strongly with Ca+2 concentrations, suggesting that an unmeasured anion
(e.g., HC03 or C032) is associated with this species.
Network-wide average concentrations of S042, N03, HN03, and S02 are
shown in Figure 19, by season, for the 18 sites in operation throughout
1988. Results for S042 and S02 exhibit dramatic differences in seasonal
variability. S042 concentrations increase by nearly a factor of three from
winter (i.e., January to March) through summer (i.e., July to September),
then plummet during the fall to the lowest levels of the year. Clearly,
the transition from summer to fall represents a major change in the
atmospheric potential for S042 production. Average S02 concentrations, in
contrast, decrease markedly from winter to summer, then return to
intermediate levels in the fall.
Seasonal concentrations for N03 and HN03 also show significant
variability (see Figure 19). HN03 follows a similar, but less dramatic
pattern as S04 . N03 concentrations, in contrast, more than double from
41
-------
*iH
4J
a
•i-i
00
i
00
§r-i Q ft op so i—i vo oo co ON in m vo ^ m vo r>
ON cr> r> oo oo os r- o in m co r>- oo ON oo oo oo
o o o o o o o o o" o o o o o o o o o
r^OOvincyi-j-oo r^cMr-iCNr^Q mgcsiovo
r~oor^r~r~-r-r-. p^r--r~-incNoo r-.oor-.r~r>.
o o o o o o o o o o o o o o o o o o
r- ON i—i <—4 oo in
OO ON i-l O ON O
O O t-i i-i O r-J
i—i in m oo fi t—i ON CM CM Q
1-1 <-i CM m m 1-1 ONONi-Jo
i-li-ICMi-l OOi-li-li-l
QN
CM
i—i oo r> i— i—
oo i—i ON vo in oo oo^jvovo
CSlCMvTr-i'~l r-l i—I
mcOvfronin i—) -j- -j-
-------
CONCENTRATION (ug/m3)
20
WINTER SPRING SUMMER
SEASON
CONCENTRATION (ug/m3)
3.5
WINTER
SPRING SUMMER
SEASON
S02
S04—
FALL
N03-
HN03
B
FALL
Figure 19.
Network-wide concentrations of S0^2/S02 (A) and
NO'/HNOj (B) by season
43
-------
summer to winter, perhaps due to the temperature-dependent equilibrium
between aerosol NH4N03 and gas phase HN03 and NH3.
4.1.3 Day Versus Night Concentration Data
Day and night filter pack samples were collected during 1988 to
determine whether diurnal variations existed which could affect the
uncertainty of dry deposition estimates. Results are summarized here for
four sites that are representative of the major terrain classifications
within the network. Data from these sites exemplify the day/night
variability observed throughout the network.
Day/night concentrations of 804 and S02 for sites in central Illinois
(Site 130), central Kentucky (Site 129), eastern Kentucky (Site 121), and
southwestern Virginia (Site 120) are compared by quarter in Table 7.
Results for S042 show differences of less than 20 percent between day and
night for all sites and all times of the year. Data for Site 130, which is
located in prime agricultural flatland, show a consistent decrease in
concentration of approximately 10 percent from day to night. For Site 129
in rolling terrain, day/night concentrations are virtually identical during
quarters one and four and show a 10-percent decrease during the middle two
quarters of the year. Day/night concentrations at Site 121 (complex
terrain) are similar during first and fourth quarters and show nocturnal
depressions of 10 and 20 percent, respectively, during second and third
quarters. Site 120 (mountaintop) is unique in that day/night
concentrations are invariant throughout the year.
The diurnal variability of S02 is much more pronounced than that of
S042 and apparently more dependent upon terrain and season. Nocturnal
concentrations at Site 130 are 10, 25, and 35 percent lower than daytime
values during second, third, and fourth quarters, respectively. At
Site 129, nocturnal concentrations are reduced by up to 25 percent during
quarters one and four and by 33 percent during quarters two and three.
Site 121 shows dramatic day/night differences. Night concentrations are
reduced by 40 to 50 percent during first and fourth quarters and show
nearly total depletion during quarters two and three. As was observed for
S042, Site 120 shows no evidence of significant day/night variability.
Examination of S02 data across Sites 129, 121, and 120 suggests that
nocturnal depletion could explain some, but not all, of the annual and
seasonal spatial variability described previously.
Results for NOj show unique patterns of day/night variability (see
Table 8). For Sites 129 and 120, differences between day and night are
small but generally suggest slightly higher concentrations at night than
during the day. At Site 121, differences are also small but suggest
slightly higher concentrations during the day. Results for Site 130 show
substantially higher nighttime concentrations during all three seasons that
the site was in operation. Although these data suggest nocturnal
production of N03 aerosol at certain sites, it should be noted that
artifact reactions on filter media could also account for these patterns.
For example, HN03 may be scavenged by wetted Teflon® filters at high
relative humidity, periods of which occur frequently at night.
44
-------
TABLE 7. DAY VERSUS NIGHT OCNCENIBATICNS OF SO;2 AND
BY QUARTER FCR 1988
Site
130
129
121
120
Terrain
Classification
Flat
Rolling
Ccnplex
Mountaintop
Sample
Period
Day
Nigit
Day
Ni^it
Day
Nigit
Day
Ni^t
so;2 S02
1
N/0
N/0
4.5
4.4
3.9
3.7
4.1
4.2
2
6.9
6.3
11.0
9.8
8.2
7.6
9.2
9.6
3
9.2
8.5
12.4
11.1
10.4
8.6
11.4
11.0
4
3.6
3.2
3.6
3.5
3.0
2.8
3.7
3.7
1
N/0
N/0
21.2
18.4
9.2
5.4
15.2
lb.7
2
10
9
17
11
7
1
8
9
.2
.1
.1
.5
.8
.1
.9
.5
3
10.6
8.1
11.0
7.4
4.7
0.4
8.4
7.6
4
10.2
6.7
13.7
10.3
5.2
2.6
13.0
13.4
Note: 1, 2, 3, and 4 refer to first, second, third, and fourth quarters, respectively.
N/0 ** not operational.
•j
Source: Hunter/ESE, 1989.
45
-------
TABLE 8. UK VERSUS NIGHT OCNCENIBATICNS OF ND^ AND WO,
BY QUARTER HK 1988
Site
130
129
121
120
Terrain Sanple
Classification Period
Flat Day
Nigit
Rolling Day
Nigit
Conplex Day
Nigit
Mountaintop Day
Night
1
N/0
N/0
1.0
1.3
1.2
1.1
0.7
1.0
N
2
1.1
1.5
0.7
0.8
0.5
0.4
0.6
0.8
"V
3
1.
1.
0.
0.
0.
0.
0.
0.
1
6
5
6
3
2
4
3
4
2.4
3.3
1.4
1.5
0.8
0.6
0.7
0.9
1
N/0
N/0
3.6
3.1
1.4
0.6
3.2
3.9
m
2
5.7
3.0
6.3
3.4
2.6
0.6
4.2
4.2
03
3
5.
2.
5.
2.
2.
0.
4.
3.
4
4
2
6
2
3
1
1
4
1.3
0.7
2.2
1.6
0.8
0.7
2.6
3.0
Note: 1, 2, 3, and 4 refer to first, second, third, and fourth quarters, respectively.
N/0 - not operational.
Source: Hunter/ESE, 1989.
46
-------
HN03 data show significant nocturnal depletion for most seasons at
Sites 121, 129, and 130. As with S02, HN03 nearly vanishes during the
night at Site 121, especially during second and third quarters. Reductions
at Sites 129 and 130 are close to 50 percent over the same period. In
contrast, results for Site 120 exhibit only small decreases, if any, during
second and third quarters and small but statistically significant increases
during first and fourth quarters. As noted for N03, artifact reactions may
confound interpretation of HN03 data. For example, nocturnal depletion of
HN03 and enhancement of N03 at Site 130 could be explained, in part, by
phase changes on the Teflon* filter. At the remaining sites, however,
calculations of total N03 (i.e., HN03 plus N03) support observations of
significant nocturnal depletion during the second and third quarters.
4.2 OZONE
Daily average and daily maximum 03 concentrations for sites that were
operational throughout 1988 are shown in Figures 20 through 28. Each
figure also shows a horizontal line at 120 ppb, which represents the
National Ambient Air Quality Standard (NAAQS) for 03. In general, results
for all sites reflect the classical seasonal cycle of 03 production.
Maximum concentrations are observed during the period June through
September and minimum concentrations are observed in November, December,
and January. Seasonal variability of this nature has been observed
previously over widespread areas of the eastern United States (Altshuller,
1987; Meagher et al., 1987).
Inspection of daily 03 averages for 1988 shows that there is
considerable fine structure in the overall annual pattern. 03
concentrations at nearly all sites rise and decay with a period of
approximately 5 to 10 days, which is probably meteorologically driven.
Examination of daily maxima shows that approximately 75 pencent of the 1988
03 values greater than or equal to 100 ppb occurred during four such
oscillations: May 28 through June 1; June 13 through June 17; July 3
through July 9; and August 16 through August 19. Maximum hourly 03
concentrations observed during these four periods are shown for each site
in Figures 29 and 30.
Results for the period May 28 through June 1 show maximum 03 values
that frequently differ between sites by little more than the instrumental
uncertainty. Maximum concentrations range from 81 ppb in southwestern
North Carolina (Site 137) to 204 ppb in southeastern New York (Site 103/4)
but are between 110 and 125 ppb in all states except Kentucky and
Tennessee.
Data for the period June 13 through June 17 also exhibit the highest
and lowest concentrations at Sites 103/4 and 137, respectively. The region
of highest concentrations (i.e., >110 ppb) is somewhat contracted, relative
to the previous episode, and there is considerable patchiness along the
flanks of the Appalachian Mountains. For many sites, the highest hourly
average concentration of the year was observed during the period July 3
through July 9. Maximum hourly concentrations for the episode ranged from
47
-------
240
220
° 200
0
ieo
p
p
B 140-
120
100-
80-
60-
40-
20-
NAAQSFORO,
_ /
0-
OUAN 01FEB 01HAR 01 APS 01MAY 01JUN 01JUL 01AU6 01SEP 010CT 01NOV 010EC 01JAN
DATE
240
220
200'
E
. 160
P
P
B 140-
120
100-
80-
60
40
20
0
NAAQSFORO.
OUAN 01FEB 01MAH 01APR 01MAY 01JUN 01JUL 01AU6 01SEP 010CT 01NQV 01DEC OUAN
DATE
Figure 20. 0, daily averages and maxima for Sites 101 (A) and 102 (B)~1988
48
-------
240-
220-
20° '
E
160
P
P
B 140-
120
NAAQSFORO,
100-
80-
60-
40-
20-
Q-\
I
01 JAN
I I 1 I 1 I T [ I T \ [^
01FEB 01MAR OlAPH 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01NOV 01DEC OlJAN
DATE
240-
220-
160-
140-
120
100-
80-
60-
40-
20-
B
NAAQSFORO,
o
~i
OlJAN
01FEB 01MAH OlAPfl 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01NOV 01DEC OlJAN
DATE
Figure 21. 03 daily averages and maxima for Sites 103/4 (A) and 105 (B)--1988
49
-------
240
220
200
180
160
140
120
100-
80
60
40-
20
NAAQSFORO,
T 1 1 1 1 1 1 1 1 1 T
01JAN 01FEB 01MAR 01APR 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01NOV
DATE
240
220
200
180
160
140
120
100
80-
60-
40-
20
0-
01DEC 01JAN
NAAQS FOR 0,
B
01JAN 01FEB 01MAR 01APB 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01NOV 01DEC 01 JAN
DATE
Figure 22. 03 daily averages and maxima for Sites 106 (A) and 107 (B)--1988
50
-------
240
220-
20°-
E
. 160-
p
p
B 140-
120
NAAQSFORO,
100-
80-
60
40-
20-
0-
'• '\
II
i 1 1 1 1 1 1 1 1 1 1 1 r
01JAN 01FEB 01MAH 01APR 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01NOV 01DEC 01JAN
DATE
240-
220-
160
120
100
80-
60-
40-
20-
0-
NAAQSFORO,
B
1 1 1 1 1 1 1 1 1 1 1 r
01JAN 01FEB 01MAR OlAPfl 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01NOV OlDEiC 01JAN
DATE
Figure 23. 03 daily averages and maxima for Sites 108 (A) and 110 (B)--1988
51
-------
240
220
200
180-
160
140
120
100-
80
60
40-
20-
0-
NAAQSFORO,
OUAN 01FEB 01MAR 01APR 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01NOV 01DEC 01JAN
DATE
240 '
220
200-
180-
160
140-
120
100-
80-
60-
40
20-
0
NAAQSFORO,
B
OUAN 01FES 01MAH 01APR 01MAY 01JUN 01JUL 01AUS 01SEP 010CT 01NOV 01DEC OUAN
DATE
Figure 24. 03 daily averages and maxima for Sites 113 (A) and 117 (B)-1988
52
-------
240-
220-
o
°1
E
. 160-
P
P
B 140-
iao
100-
80
60-
40-
20-
NAAQS FOR 0,
—i 1 1 1 1 1 1 1 : 1 1 r
01FEB 01MAR 01APR 01MAY OUUN 01JUL 01AUG 01SEP 010CT 01NOV 01DEC OlJAN
DATE
240-
220-
0 20° "I
0180
E
. 160
P
P
B 140
120
100-
ao •
60-
40-
20 -
0
NAAQSFORO,
OlJAN 01FEB 01MAR 01APH 01MAY OUUN 01JUL 01AUG 01SEP 010CT 01NOV 01DEC OlJAN
DATE
Figure 25. 03 daily averages and maxima for Sites 119 (A) and 120 (B)--1988
53
-------
240
220-
0
0180
E
. 160
P
P
8 140
120
NAAQS FOR 0,
100-
80
60
40
20 :
0-
01JAN 01FEB 01MAR 01APR 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01MOV 01DEC 01JAN
DATE
240
220
0
0180
E
. 160
P
P
B 140
120
100
80
60-
40-
20-
0-
NAAQS FOR 0,
B
1 1 1 1 1 1 1 1 1 1 1 1 r
OUAN 01FEB 01MAH 01APR '01MAY 01JUN 01JUL 01AU6 01SEP 010CT 01NOV 01DEC OlJAN
DATE
Figure 26. 03 daily averages and maxima for Sites 121 (A) and 122 (B)--1988
54
-------
240-|
220
0 200-1
0190
E
. 160
P
B 140-j
120
100-
80-
60-
40
20 -)
NAAQS FOR 0,
01JAN 01FEB OlMAfi 01APR 01MAY 01JUN 01JUL OlAUG 01SEP 010CT 01NOV 01DEC OUAN
DATE
240
220-
200-
180-
N
E
. 160-1
P
a 140-]
120
100-
80-
60-
40
20-1
NAAQS FOR 0,
OUAN 01FEB 01MAR 01APH 01MAY 01JUN 01JUL 01AUS 01SEP 010CT 01NOV 01DEC OUAN
DATE
Figure 27. O3 daily averages and maxima for Sites 129 (A) and 137 (B)--1988
55
-------
240-
220
20° '
E
. 160-
P
P
B 140-
120
100-
80-
60 '
40-
20 :
NAAQS FOR 0,
01JAN 01FEB 01MAR 01APH 01MAY 01JUN 01JUL 01AUG 01SEP 010CT 01NOV 01DEC 01JAN
DATE
240
220
0 cuu
OiBO
E
. 160
P
P
B 140
120
100
80
60-
40-
20-
0
NAAQS FOR 0,
OUAN 01FEB OlMAfl 01APB 01MAY 01JUN 01JUL 01AU6 01SEP 010CT 01NOV 01DEC 01JAN
DATE
Figure 28. 03 daily averages and maxima for Sites 140 (A) and 146 (B)~1988
56
-------
Figure 29. Maximum hourly 03 (ppb) for May 28 - June 1, 1988 (A) and
June 13-June 19, 1988 (B)
57
-------
Figure 30. Maximum hourly 03 (ppb) for July 3 - July 9, 1988 (A) and
August 16 - August 19, 1988 (B)
58
-------
197 ppb in southeastern New York (Site 104) to 101 ppb in extreme northern
New York. A region of concentrations in excess of 125 ppb extended from
northern Indiana to New York and southward (excluding Kentucky) to
Tennessee and North Carolina. As with previous episodes, the two sites in
Illinois appear to be along the fringe of the region of elevated
concentrations.
The episode of August 16 through 19 shows a fairly dramatic
latitudinal gradient in maximum concentrations. Maximum hourly values were
well below 100 ppb across New York, Michigan, and northern Illinois,
roughly 100 ppb across central Pennsylvania, northern Indiana, and central
Illinois, and well above 100 ppb (except for one site in West Virginia) at
all remaining sites. Relatively low concentrations in New York, Michigan,
and Illinois appear to be due to a stationary front, which brought
widespread cloudiness and scattered rainfall to the northern extremes of
the network.
Synoptic meteorological conditions during the four 03 episodes were
generally very similar. A stagnating high pressure system covered the
region, bringing high temperatures, clear skies, and light and variable
winds. Precipitation was nearly absent, except for isolated thunderstorms
in the southeast. In most cases, episodes were terminated by the incursion
of a cold front and associated precipitation.
Although peak concentrations of 03 were reasonably constant from site
to site during the July 3 through 9 episode, the same does not apply for
time-averaged values. Figure 31 shows the diurnal evaluation of 03 for
three sites extending from central Kentucky to southwestern Virginia. The
central Kentucky site (Site 129) is located in an open meadow in rolling
terrain; the eastern Kentucky site (Site 121) is located in a valley in
complex terrain; and the Virginia site (Site 120) is located on a
mountaintop at an elevation of 972 m. Time series data for*these sites
show marked differences in the temporal behavior of 03. The mountaintop
site shows a day-to-day increase in 03 concentrations but little evidence
of day/night variability. The open meadow site shows significant decay of
03 from mid-day highs to early morning lows. For the days shown, the early
morning low is in the range of 40 to 60 ppb and roughly 50 percent of the
previous peak. The complex terrain site shows an extremely pronounced
day/night pattern. Daytime maxima increase steadily throughout the episode
while nighttime minima show nearly complete 03 depletion on all but the
last day.
One explanation for this behavior involves the relationship between
sampler location and the nocturnal inversion layer. The mountaintop site
presumably sits above the inversion layer and, therefore, is always in
contact with a large reservoir for 03. The rolling terrain and complex
terrain site lies below the inversion layer, within which 03 is subject to
a variety of depletion processes. Longer lasting and/or shallower
inversions could result in more rapid decay of 03 in complex terrain than
in rolling terrain. The end result of this day/night variability is a
gradient in integrated exposure (i.e., average concentrations) from
mountaintop to rolling terrain to complex terrain. Although there is no
59
-------
300 0
160 0 •
120 0
a 80 o
40 0
0 0
0123456
DAY
200 0T
160 0
1200-
80 0
40 0
0.0
B
234567
DAY
200.0
160.0
M
40 0
0 0
123456
DAY
Figure 31. 03 time series for Site 120 (A), Site 129 (B), and Site 121 (C)
for July 3 - July 9, 1988
60
-------
continuous data to support this, the same terrain effect might also account
for the S02 and HN03 pattern described previously for these sites (see
Figure 7).
4.3 SUMMARY OF METEOROLOGICAL OBSERVATIONS
Meteorological data are gathered at each site to satisfy input
requirements for one or more dry deposition algorithms. These algorithms
produce hourly deposition velocities (based on field observations) which
are then averaged over weekly periods and merged with concentration data to
calculate fluxes. The hourly meteorological data for 1988 have been
submitted to EPA and are far too voluminous to evaluate in this report.
The purpose of this section is to present quarterly averages for
meteorological parameters and annual wind roses for the network. The
sources of data for this section are 18 sites, which were operational
throughout 1988.
Quarterly average values for temperature, relative humidity, and solar
radiation are shown in Table 9. Sites have been organized by region (i.e.,
southeast, midwest, and northeast) in order to facilitate identification of
large-scale patterns, if they exist. Results generally conform to the
expected seasonal patterns of temperature, relative humidity, and solar
radiation. Comparison of values across regions, however, show that the
midwest was relatively hot, dry, and sunny, especially during the second
and third quarters. This apparently reflects the extreme drought, which
parched much of the midwest during 1988. Data for second quarter 1988
indicate that relative humidity and solar radiation were as much as
20 percent lower and 20 percent higher, respectively, in the midwest than
the northeast.
Data for windspeed, wind direction, and sigma theta (the standard
deviation of wind direction) are listed by quarter in Table 10. Results
for windspeed and wind direction show appreciable seasonality across all
regions. In general, windspeeds decrease markedly from first through third
quarters, then pick up again during fourth quarter. Wind direction backs
from approximately southwest to south during the first half of the year and
then veers to southwest through the second half of the year. Dramatic
differences between sites are apparent for windspeed and sigma theta. As a
rule, these differences correlate well with terrain classification.
Windspeeds are highest and sigma theta lowest in flat or rolling terrain,
such as at Site 130,-while the converse is true for sites in complex
terrain such as Site 121. Based on the 1988 data set, representative sigma
theta values for flat, rolling, and complex terrain sites are on the order
of 10 to 20 degrees, 20 to 30 degrees, and 40 to 55 degrees, respectively.
Windspeed and sigma theta are important variables for characterization of
gas phase resistance in a variety of dry deposition algorithms (Hicks
et al. . 1988; Wesley and Lesht, 1988). Terrain-dependent values for these
parameters may, therefore, produce appreciable differences in deposition
velocity between sites.
Annual plots of windspeed and wind direction are shown in Figures 32
through 40. As discussed earlier, there are similarities among sites in
61
-------
r-l r-4 i-i
R§RR
rn.— inr^fr.— i
CMCMCMCMi-ICM
&3%
i-l CM R
>— i r"> in
t
T
a
ro
r*-vovo
O>CT\
a
(2
oo
in
ino
OOONVO
m
o-i
i
c
OOr-IOOCSICNi-IO> VO 00 -* P^ i-H r-l t-l r-lfOin
fHi-lr-ICNtH.-ICMSRr-l CNCMCMCMCSICN CMCSIr-l
i— t i— I i-< i— I r-1 t— 1 i— ( H t— 1
o\ooOooco
o> \o in
t— 1 i— 1 <— i
^-1 in C*1 r-l i-l CM
i i i i I i
1-1 m o
r- o -* r
a\ o n
~d- vo ro
a
8
-s
.a
ON
oo
a
4
cS
I
62
-------
c»Scsl»-l<3-CMnCMt-l
i-- CM f> CM Q O vj in m
,-3 CM i-l in r-( Z CM in
ro m eo 1-1 i—I
-
K
D
o i— i in CM o
CMroOr-5inO
CNJCNICMCMCMi— ICMr-li— ICM CMCMCMCMCMCM
inoooroinO"—(ooooin ^-i-d-ooin-j- 01—^^3
r-(i—i i—i ro m r-(CMOi—i c*i inrocMCMOrO i—I i—I o
VI
£
4_>
^
in
O oo CM
in vo co 1-1 O ~* r-i O
o
-------
HIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
20 X
10 X
5 X
0 X
Hind Speed
13 X call
5.0
MIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
20 X
10 X
5 X
0 X
36 X call
Mind Speed
(nps)
0.5
2.0
3.5
5.0
B
Figure 32. 1988 Wind rose for Site 101, Research Triangle Park, NC (A)
and Site 102, Oak Ridge, TN (B)
64
-------
HIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
20 X
10 X
5 X
0 X
32 X »!•
Mind Speed
(np<)
0.5
2.0
3.5
5.0
40 X
20 X
10 X
0 X
HIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
28 X call
Mind Speed
2.0
3.5
5.0
Figure 33. 1988 Wind rose for Site 103/4, West Point, NY (A) and
Site 105, Whiteface Mountain, NY (B)
65
-------
HIND ROSE ANALYSIS FOR 01/01/86 TO 12/31/88
N
10 X
5 X
0 X
Mind Speed
tos)
0.5
2.0
3.5
5.0
20 X call
MIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
40 X
20 X
10 X
0 X
Mind Speed
(ma)
0.5
2.0
3.5
5.0
IB X call
Figure 34. 1988 Wind rose for Site 106, State College, PA (A) and
Site 107, Parsons, WV (B)
66
-------
MIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
20 X
10 X
5 %
0 X
Hind Speed
(•pa)
0.5
2.0
3.5
5.0
14 X call
HIND HOSE ANALYSIS FOR 01/01/88 TO 12731/88
N
20 X
10 X
5 X
0 X
Mind Speed
0.5
7 X call
Figure 35. 1988 Wind rose for Site 108, Prince Edward, VA (A) and
Site 110, Connecticut Hill, NY (B)
67
-------
MIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
8 X
4 I
2 X
0 X
12 X »!•
Mind SpMd
0.5
2.0
3.5
5.0
HINO ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
20
10 X
5 X
0 X
42 % call
Wind Speed
(nps)
0.5
2.0
3.5
5.0
B
Figure 36. 1988 Wind rose for Site 113, M.K. Goddard, PA (A) and
Site 117, Laurel Hill, PA (B)
68
-------
MIND HOSE ANALYSIS FOR 01/01/86 TO 12/31/88
N
20 X
10 %
5 X
0 X
38 X call
Mind Spaed
(apa)
0.5
2.0
3.5
5.0
MIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
40 I
20 I
10 X
0 X
6 X C«lB
Mind Speed
taps)
0.5
2.0
3.5
5.0
Figure 37. 1988 Wind rose for Site 119, Cedar Creek, WV (A) and
Site 120, Morton Station, VA (B)
69
-------
WIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
8 S
4 S
2 I
0 S
64 I call
MIND ROSE ANALYSIS FOfl 01/01/88 TO 12/31/88
N
Kind SpMd
0.5
20 X
10 X
5 X
0 X
Hind Speed
Imps)
0.5
2.0
3.5
5 0
12 X call
B
Figure 38. 1988 Wind rose for Site 121, Lilley Cornett Woods, KY (A) and
Site 122, Oxford, OH (B)
70
-------
KINO ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
20 X
10 X
5 X
0 X
Mind Speed
(•ps)
0.5
2.0
3.5
5.0
8 X call
MIM) ROSE ANALYSIS FOR 02/01/88 TO 12/31/88
N
20 I
10 I
5 X
0 I
Mind Speed
tos)
0.5
2.0
3.5
5.0
B
0 X call
Figure 39. 1988 Wind rose for Site 129, Perryville, KY (A) and
Site 130, Bondville, IL (B)
71
-------
KINO ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
20 X
10 X
5 X
0 X
Hind Speed
(•ps)
0.5
2.0
3.5
5.0
52 X call
MIND ROSE ANALYSIS FOR 01/01/88 TO 12/31/88
N
20 X
10 I
5 X
0 X
Wind Speed
(*>«)
0.5
2.0
3.5
5.0
11 X call
Figure 40. 1988 Wind rose for Site 137, Coweeta, NC (A) and
Site 140, Vincennes, IN (B)
72
-------
the same terrain classification and distinct differences between
classifications. Results for the site in central Illinois (flat terrain)
indicate a complete absence of calm periods and a nearly isotropic
distribution of wind direction. Winds in all four speed categories are
recorded in all 16 direction categories. Directional frequencies are
greater than 10 percent only for one category (south-southwest) and less
than 5 percent only for three categories (centered on east).
Wind characteristics in sites of rolling terrain are typified by data
from central Kentucky (Site 129) and central Virginia (Site 108). Calm
periods were recorded 8 percent and 14 percent of the time for these sites
respectively, and there is a definite persistence of winds in the quadrant
from south to west. The highest windspeeds are clearly associated with
directional categories with high persistence.
Results for sites in complex terrain exhibit a high percentage of calm
periods as well as fairly pronounced persistence in one or more directional
categories. Results for the two sites in West Virginia (Sites 107 and 119)
show distinctly bimodal distributions of wind direction. This is probably
the result of terrain-induced circulation in the mountain valleys of the
area. Data for the site in eastern Kentucky (Site 121) show the virtual
absence of winds in the three highest speed classifications and the north-
northwest, north, and north-northwest directional classifications.
Results for the mountaintop site in southwestern Virginia show
relatively few calm periods but very high persistence. Winds are generally
strong and persistent from the west-northwest and light and infrequent from
the northeast, southwest, and adjoining directions. Winds in the north-
northwest, northwest, and southeast directional categories account for more
than 60 percent of the observations.
In general, the wind rose data support systematic differences in
ventilation between sites in flat or rolling terrain and those in complex
terrain. The latter thus appear to be meteorologically isolated from
neighboring sites. This could account in part for the large spatial
variability of S02 and HN03 along the flanks of the Appalachian Mountains.
4.4 ESTIMATED DRY DEPOSITION
The objective of NDDN, among other things, is to produce long-term
data on patterns and'trends of dry deposition across the continental United
States. Ultimately, dry deposition rates will be calculated using one or
more algorithms that couple land use or vegetation data, meteorological
data, and air quality data to produce weekly, seasonal, and annual fluxes.
This section presents estimation of dry deposition rates for 1988 using
assumed values for annual average deposition velocity taken from the
literature and annual average concentration data from 18 NDDN sites
operational throughout 1988. Results are intended only to illustrate a
possible range of dry deposition rates across the network. Results are
specifically not intended to be an approximation of, or substitute for,
deposition estimates based on the inferential approaches described by
Hicks et al. (1985) and Wesely (1988).
73
-------
The estimated deposition velocities used in these calculations are
listed in Table 11. Tabulated values reflect a variety of theoretical
estimates, experimental tests, and modeling results. Wu and Davidson
(1988) used the resistance model of Hicks e_t al. (1985) to estimate weekly
and annual deposition velocities of S02, SO^2, and HN03 for three sites
during 1986. Ranges of annual average deposition velocity for Oak Ridge,
TN; Penn State University, PA; and Whiteface Mountain, NY (all current NDDN
sites) were 0.07 to 0.13 centimeters per second (cm/sec) for SO^2, 0.18 to
0.33 cm/sec for S02, and 0.96 to 1.9 cm/sec for HN03. Sheih et al. (1979)
and Wesely and Lesht (1988) have developed computer routines for estimating
deposition velocity as a function of land use, season, and meteorological
stability class. Wesely and Lesht (1988) recently compared their model
with a site-specific inferential technique similar to that which may be
used for the NDDN (Hicks e_t al., 1988). Calculated deposition velocities
were consistent with data shown in Table 11 and, on average, within ±30 to
40 percent of deposition velocities provided by the inferential technique.
Little information is available on the deposition velocity for atmospheric
N03. By analogy, the same range of deposition velocities was used for N03
and SC>4 . However, differences in particle size and reactivity could
result in substantial differences between these species.
Estimates of annual dry deposition for SO^2 plus S02 and for N03 plus
HN03 are shown in Figure 41. Values shown represent the mid-point of the
estimated ranges of deposition and, therefore, are uncertain by at least
±50 percent. Estimated dry deposition for SO^2 plus S02 appears to be
highest in southwestern Pennsylvania (Site 117) and eastern Tennessee
(Site 102) and lowest in southwestern North Carolina (Site 137), northern
New York (Site 105), and eastern Kentucky (Site 126). Not surprisingly,
this pattern reflects the annual average S02 concentration, since in no
case did SO^2 contribute more than 30 percent of the estimated dry
deposition at any site. Relatively high deposition in northern Illinois,
southern Indiana, and eastern Tennessee may be the result of local S02
emissions. Similarly high values in western Pennsylvania and West
Virginia, on the other hand, seem to reflect more widespread sources.
Estimated dry deposition of N03 plus HN03 shows similar values over
much of the northeast and midwest but considerable variability in the
vicinity of the Appalachian Mountains. Deposition at the mountaintop site
in Virginia (Site 120) appears to be approximately three times that at
complex terrain sites in eastern Kentucky and North Carolina (Site 121 and
Site 137). The overall pattern is almost an exact transformation of annual
HN03 concentration due to large differences in concentration and deposition
velocity between HN03 and N03. At no site was N03 responsible for more
than 15 percent of the estimated dry deposition of N03 plus HN03, and only
at Site 146 (Argonne, IL) was N03 responsible for more than 10 percent.
Wet and dry deposition data from a variety of NDDN sites are shown in
Table 12. Wet deposition values (averaged over the 5-year period 1983 to
1987) were obtained from the National Atmospheric Deposition Program for
all sites except Ithaca, NY. Data for this sit;e, at the same time period,
were obtained from quarterly data reports produced by the MAP3S program
74
-------
TABLE 11. ESTIMATED DEPOSITION VELOCITIES TOR AEROSOLS AND GASES
Species
Vd
(cm/s)
Reference
0.1 - 0.2 Sehmel, 1980; Sheih et al., 1979; Voldner et al.,
1986; Wesely and Lesht, 1988; Wu and Davidson,
1988
0.1 - 0.2 Assured to be the same as SO^2
0.2 - 0.4 Sehmel, 1980; Sheih et al., 1979; Cadle gt al.,
1987; Wesely and Lesht, 1988; Wu and Davidson,
1988
1.0 - 2.0 Heubert, 1983; Wesely and Lesht, 1988; Wu and
Davidson, 1988
S02
Note: cm/s = centimeter per second
Vd = deposition velocity.
Source: Hunter/ESE, 1989.
75
-------
Figure 41.
Estimated dry deposition of SO^2 plus S02 (A) and NOj plus
HN03 (B) (eg/ha) for 1988
76
-------
TABLE 12. WET VERSUS DRY DEPOSITION AT SELECTED SITES
Deposition (ea/ha-vr)
Site
105 Whiteface, NY
110 Connecticut Hill,
NY
106 Penn State, PA
107 Parsons, WV
122 Oxford, OH
130 Bondville, IL
129 Perryville, KY
101 Research Triangle
Park, NC
102 Oak Ridge, TN
Wet*
390
490
720
850
560
550
470
340
600
so;2
Dry+
100-200
310-630
360-730
370-730
360-730
240-470
330-670
210-420
450-910
Wet
230
310
350
360
220
220
190
160
200
NOj
Dry**
100-200
190-380
180-370
170-340
200-400
180-340
180-350
140-290
160-330
Note: eq/ha-yr = equivalents per hectare-year.
*Source of wet deposition data is National Atmospheric Deposition Program
(IR-7)/National Trends Network, 1988.
+Sum of estimated dry deposition for SO;2 aerosol and S02.
**Sum of estimated dry deposition for NO^ aerosol and HN03.
Source: Hunter/ESE, 1989.
77
-------
(M.T. Dana, personal communication, 1989). Again, the data tabulated for
dry deposition reflect likely ranges for 1988 only. Results for sulfur
deposition suggest that wet deposition is the dominant process in northern
New York (Site 105), but that at most other sites, wet and dry deposition
could be similar, especially if the upper limit for the deposition velocity
of S02 is approached. Results for nitrogen suggest regional differences in
the comparability of wet and dry deposition. For the northeast, dry
deposition would be comparable to wet deposition only if the upper limit
deposition velocity for HN03 is attained. Data for the midwest and
southeast, in contrast, suggest that wet and dry deposition are comparable
even if the lower limit deposition velocity for HN03 is attained.
4.5 OVERALL DATA QUALITY
As described in Section 2, extensive QC measures were undertaken to
assure and document the overall quality of the NDDN database. This section
provides a brief summary of NDDN QC data for 1988. Additional details on
field and laboratory QC checks can be found in the quarterly data reports
listed in the bibliography.
4.5.1 Field Data
Results of meteorological equipment calibrations performed during 1988
are summarized by quarter in Table 13. In general, calibration data show
the majority of sensors were operating within accuracy requirements for the
network. Temperature and relative humidity sensors typically responded
within a few tenths and a few hundredths of a degree Celcius of NIST-
traceable standards. Relative humidity, solar radiation, and rainfall
amount responded within a few percent of calibration standards. Windspeed
and wind direction were generally within ±0.2 m/sec and ±2 degrees of
target values. Inspection of quarter to quarter variability for
calibration data shows that all sensors, except relative humidity,
responded consistently throughout the year. Data quality for relative
humidity improved markedly from first quarter through fourth quarter due to
adoption of a new sensor and the use of calibration salts in the field for
adjustment of sensors.
Calibration data for continuous 03 analyzers and MFC are presented in
Table 14. 03 and MFC calibrations were performed through the entire
ambient air inlet to compensate for small (<3 percent) line losses and
pressure drops within the system, respectively. In general, results for 03
show that the instrument employed is highly stable and that it functioned
with acceptable accuracy during the year. Aggregate errors for the network
were within ±2 percent duripg each quarter, as represented by sensor versus
transfer standard slopes, and errors for individual analyzers were
invariably better than ±7 percent. Although occasional equipment failures
occurred, no 03 data were invalidated for lack of compliance with accuracy
requirements.
MFCs also showed excellent stability, the errors tabulated in Table 10
indicate little instrumental drift between calibrations (i.e., from quarter
78
-------
oo
00
o
S
C/3
6
K
C
&
I
2
o
mo voo ^-oo
sa
vo
o
OO CMVO
oo oo
93
9?,
§ §
*r~l *r4
VO
.3
00 CM
83
9?;
a
i
42 «, K §• .
m Jj M CT>
O o>
-------
TABLE 14. SUMMARY OF 0, AND MASS FLOW CONTROLLER UNADJUSTED CALIBRATIONS DURING 198i
Quarter
First
Mean
Standard Deviation
Second
Mean
Standard Deviation
Third
Mean
Standard Deviation
Fourth
Mean
Standard Deviation
C
Slope
1.
±0.
0.
±0.
0.
±0.
0.
±0.
002
012
993
013
988
024
995
019
)3
Intercept
0.
±0.
0.
±0.
0.
±1.
-0.
±1.
88
75
14
76
60
15
17
47
Filter
Flow (%
0.
±3.
0.
±2.
-I.
±3.
-0.
±2.
Pack
error)*
6
9
7
2
1
5
9
2
Note: Slope and intercept are regression coefficients for calibration
versus a NIST-traceable 03 transfer standard.
*Error is calculated relative to the target flow of 1.50 Lpm.
Source: Hunter/ESE, 1989.
80
-------
to quarter). Thus, the impact of uncertainties in flow on concentration
calculations is expected to be minimal.
4.5.2 Laboratory Data
Laboratory accuracy and precision data from 1988 are summarized in
Table 15. As described earlier, an NIST reference (2694-11) is analyzed
once per analytical batch, a calibration verification standard (CVS) is
analyzed after every tenth environmental sample, and 10-percent blind
replicates are analyzed on different dates. Replicate analyses for Teflon®
filters were discontinued in third quarter 1988 due to insufficient volumes
after cation analyses.
Results for the NIST-reference material show that for each filter type
and for each quarter instrumental error was less than or equal to
2.5 percent. During most quarters, it was much less than this, however.
For example, mean recoveries during fourth quarter ranged from 100.5
±1.3 percent for nylon SO^2 to 99.8 ±1.1 percent for Teflon* NOj. CVS
recoveries show similar results and indicate that instrumental drift was
almost invariably less than 5 percent for each analytical run.
Results of replicate analyses show that, on average, the between-run
precision for all analytes is less than 5 percent. However, there is
considerable scatter in the replicate analyses, as reflected in the
standard deviations for each analyte. This scatter is due primarily to the
selection of field and laboratory blanks in the blind replication process.
Blanks comprised more than 15 percent of the replicate analyses performed
during 1988. For these samples, even a small increase or decrease in
concentration propagates into a large percent difference between
replicates. This is especially true if one analysis was less than the
instrumental detection limit and the other was at or above the detection
limit. In general, mass recoveries from replicate analyses differed by
1,0 /Ltg/filter or less for each analyte. Given a typical sample volume of
7.5 m , this translates into a mean laboratory precision of approximately
0.15 Mg/m3 for SO^2, NOg, and HN03 and 0.2 Mg/m3 for S02.
Laboratory accuracy and precision should not be confused with overall
sampling accuracy and precision. The accuracy of the NDDN filter pack
sampling approach is unknown and its precision should be determined based
on collocated field sampling, rather than laboratory replicates.
Nevertheless, the results in Table 15 show that laboratory operations
conform with accuracy and precision requirements for this project.
Results of collocated field monitoring during 1988 are shown in
Table 16 (continuous sensors) and Table 17 (filter packs). Temperature and
delta temperature data indicate that, on average, measurements are within
0.1"C and that only rarely do differences exceed ±0.2°C, the manufacturer's
specification for the equipment. Relative humidity data exceeded the
precision goal of ±10 percent during first quarter 1988 but showed marked
improvement in subsequent quarters. Although both sensors passed
calibration checks late in the first quarter, the collocated sensor was
found to have a dirty inlet. This inlet caused the collocated sensor to
81
-------
B
1
g
E
I
•gf
fcr
isr
isr
to r oo
s
(N
O CM
so in
0°
CM
i/l
o
O CM
vo a,, ".
si§o §§2
•* "^
* «* vo m co m
§ si ^ si si ^i
CM r-- '' so CM
co
. oo
,
si si o si
° CO 00 °°. O 00
s0
% & O g> 85 O 8 g> ^ 8 85 CM'
O
t—i o
??| T
so CM » c">
lfi.S
IN
frfr
^S
82
-------
13
,1
^
X~N
cT'fe
i
.SI
w *
'1
H
13
sj
^o.
B §
13 i
H
oo
S
oo
oo,
9?;
•H
4J
H
fc
c^j
ao
CM
a\ in
1-1 •*
o
oo
CO
m CM
88
o
'H
• n-t
S <~l
3 . 3 g.«a
Qj QJ S I
35 oS »H )
LJ Li -—j m
Sb bbrH 4J U
•8-8^88
,
a
Q)
o
a
83
-------
TABLE 17. PRECISION DATA FOR COLLOCATED FILTER PACK SAMPLING AT WEST
POINT. NY DURING 1988
Analvte
Quarter
First
Second
Third
Statistic
X (Mg/m3)
D (Mg/m3)
PD (%)
MAPD (%)
X
~D
PD
MAPD
X
~D
PD
MAPD
so;2
4.
-0.
-2.
2.
5.
-0.
-1.
2.
8.
-0.
-1.
4.
06
11
7
3
74
11
8
8
81
11
2
6
N03
1.
-0.
-1.
10.
0.
0.
16.
27.
0.
0.
17.
30.
09
02
8
1
48
08
7
9
23
04
4
9
HNO
2.
0.
4.
5.
2.
0.
5.
5.
3.
0.
22.
18.
3
52
10
0
0
94
15
1
4
23
71
0
3
S02
19.
0.
1.
3.
10.
0.
0.
2.
8.
-0.
-0.
4.
0
2
1
7
1
06
6
4
39
05
6
8
Fourth
NP
NP
NP
NP
NP
Note: X. = mean concentration for both samplers during the quarter.
D — mean difference between samplers during the quarter.
PD - percent differences of the quarterly means.
MAPD = mean absolute percent difference between weekly samples.
NP = not performed.
Source: Hunter/ESE, 1989.
84
-------
take longer than the primary sensor to dry out after a period of high
relative humidity (i.e., >95 percent). After cleaning, the two sensors
responded much more coherently, as reflected in the data for second and
third quarter. Collocated O3 data show that differences between analyzers
were on average within 1 ppb and rarely exceeded 2 ppb.
Data for solar radiation, windspeed, and wind direction meet precision
goals; however, data for wind direction show much more variability than can
be explained by the expected measurement errors. Although there is no
precision goal for sigma theta, results for this parameter also show more
scatter than was initially expected. One explanation for these
observations is terrain effects. The site is located in a narrow valley
within an area of complex terrain. In contrast, recent data for a
collocated site in rolling terrain (Site 153--Georgia Station, GA) exhibit
precision of 1.0 degrees, or less, for both wind direction and sigma theta.
Results of the collocated filter pack sampling at West Point, NY show
that precision data for SO^2 and S02 invariably meet the project objective
of ±5 percent. Percent differences based on quarterly average concentra-
tions are well below 5.0 (typically below 2.0), while mean absolute percent
differences based on weekly samples range from 2.3 to 4.8. There is no
evidence of a relationship between concentration and precision, as
reflected by relatively constant percent differences from quarter to
quarter.
Results for N03 show that differences between quarterly average
concentrations are less than 0.1 /ig/m3, but that percent differences for a
quarter or for individual samples can be quite large. As illustrated in
Figure 42, N03 data show appreciable scatter over the range of observed
concentrations and generally much more scatter than the other chemical
parameters. HN03 data show concentration differences with ±0.2 MS/11"3 and
percent differences of 5.4 or less for the first and second quarters but a
significant deterioration in precision during the third quarter. Clearly,
differences in N03 cannot explain differences in HN03 during third quarter
and, therefore, the total N03 collected by the two samples differed
appreciably. These differences rule out some but not all of the artifact
reactions that can occur within the filter pack system. The occurrence of
an undetected leak or a systematic breakthrough of HN03 in one of the
systems is currently being investigated. In any event, collocated data at
the same site fos previous quarters and other sites for subsequent quarters
suggest that the third quarter 1988 precision data for HN03 are anomalous.
85
-------
\
^s;
125
s-g
«3
o
III
a
iis
c
o
to
c
IO
t
3
0-?
III
f
• c o
*o »•-
0) °
*§
10
c.
- S8S
!ls
T3 C7
*S
O §
= o
o ®
o «p
0>
(0 Z
l_ Q_
"
0
eg
I
CD
86
-------
REFERENCES
Altshuller, A.P. 1987. Estimation of Natural Background of Ozone Present
at Surface Rural Locations. Journal of the Air Pollution Control
Association, 37(12):1409-1417.
Barnes, H.M., Hansen, D.A., and Lusis, M. 1987. Regional Field Study
Design for Evaluation of Eulerian Acidic Deposition Models. Presented
at 80th Air Pollution Control Association Conference, June 21-26,
1987. New York, NY.
Cadle, S.H., Dasch, J.M., andMulawa, P.A. 1985. Atmospheric
Concentrations and the Deposition Velocity to Snow of Nitric Acid,
Sulfur Dioxide and Various Particulate Species. Atmos. Envir.,
19:1819-1827.
Hicks, B.B., Baldocchi, D.D., Hosker, R,P., Jr., Hutchison, B.A., McMillen,
R.T., and Sdtterfield, L.C. 1985. On the Use of Monitored Air
Concentrations to Infer Dry Deposition (1985), NOAA Technical
Memorandum ERL ARL-141. 66 pp.
Hicks, B.B., Baldocchi, D.D., Meyers, T.P., Hosker, R.P., Jr., and
Matt, D.R. 1988. A Preliminary Multiple Resistance Routine for
Deriving Dry Deposition Velocities from Measured Quantities. Water,
Air, and Soil Pollut., 36:311-330.
Hosker, R.P., Jr. and Womack, J.D. 1986. Simple Meteorological and
Chemical Filter Pack Monitoring System for Estimating Dry Deposition
of Gaseous Pollutants. In: Proceedings of the 5th Annual National
Symposium on Recent Advances in the Measurement of Air Pollutants,
Raleigh, NC, 14-16 May 1985, EPA/600/8-85/029 (available from NTIS,
Springfield, VA), pp. 23-29.
Meagher, J.J., Lee, N.R., Valente, R.J., and Parkhurst, W.J. 1987. Rural
Ozone in the Southeastern United States. Atmos. Envir., 21:605.
Meyers, T.P. and Yuen, T.S. 1987. An Assessment of Averaging Strategies
Associated with Day/Night Sampling of Dry-Deposition Fluxes of S02 and
03. J. Geophys. Res, 92:6705-6712.
National Atmospheric Deposition Program (IR-Z)/National Trends Network.
1988. NADP/NTN Coordination Office, Natural Resource Ecology
Laboratory, Colorado State University, Fort Collins, CO.
Sehmel, G.A. 1980. Particle and Gas Dry Deposition: A Review. Atmos.
Envir., 14:983-1011.
Sheih, C.M., Wesely, M.L., and Hicks, B.B. 1979. Estimated Dry Deposition
Velocities of Sulfur Over the Eastern United States and Surrounding
Waters. Atmos. Envir., 13:1361-1368.
87
-------
Sheih, C.M., Wesely, M.L., and Walcek, C.J. 1986. A Dry Deposition Module
for Regional Acid Deposition, U.S. Environmental Protection Agency
Report, EPA/600/3-86/037 (available as PB86218104 from NTIS,
Springfield, VA) 63 pp.
Wesely, M.L. 1988. Improved Parameterizations for Surface Resistance to
Gaseous Dry Deposition in Regional-Scale Numerical Models, U.S.
Environmental Protection Agency Report (available from NTIS,
Springfield, VA) (in press).
Wesely, M.L., Cook, D.R., Hart, R.L., and Spear, R.E. 1985. Measurements
and Parameterization of Particulate Sulfur-Over Grass. J. Geophys.
Res., 90:2131-2143.
Wesely, M.L. and Lesht, B.M. n.d. Comparison of the RADM Dry Deposition
Module with Site-Specific Routines for Inferring Dry Deposition.
Prepared for: Environmental Monitoring Systems Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency (EPA),
Research Triangle Park, NC.
Wu, Y.L. and Davidson, C.I. 1989. Estimating Dry Deposition of S02, HN03,
and S04: The Inconsequence of Separate Daytime and Nighttime
Sampling. Prepared for: Office of Research and Development, U.S.
Environmental Protection Agency, Washington, DC.
if U.S. GOVERNMENT PRINTING OFFICE: 19*0 - 74K-IS9/20463
88
------- |