United States'
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-95/086
August 1995
rxEPA
CASTNet
National Dry Deposition
Network 1990-1992 Status
Report
-------
EPA/600/R-95/086
August 1995
CASTNet
NATIONAL DRY DEPOSITION NETWORK
1990-1992 STATUS REPORT
by
Environmental Science & Engineering, Inc.
Gainesville, FL 32607
Contract No. 68-D2-0134
Project Officer
Ralph Baumgardner
Air Exposure Research Division
Research Triangle Park, NC 27711
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Floor
Chicago, IL 60604-3590 @ Printed on Recycled Paper
-------
The information in this document has been funded wholly by the U.S. Environmental
Protection Agency (EPA) under Contract No. 68-02-4451 and 68-D2-0134 to
Environmental Science & Engineering, Inc. (ESE). 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.
11
-------
Foreword
In 1986, the U.S. Environmental Protection Agency (EPA) established the National Dry
Deposition Network (NDDN) as a part of the Agency's support to the National Acid Precipitation
Assessment Program (NAPAP). The goal of the NDDN was to provide NAPAP with estimates of dry
deposition flux to use in model validation, determination of spatial patterns of dry deposition, and dry
deposition flux data to relate deposition to ecological effects. NAPAP was completed in 1990 with the
issuance of the Integrated Assessment.
Also, in 1990, Congress amended the Clean Air Act. These amendments require reduction in
emissions of sulfur and nitrogen oxides. A national monitoring network was mandated as part of the
Clean Air Act Amendments to determine the effectiveness of these future emission reductions. EPA has
established the Clean Air Status and Trends Network (CASTNet) to provide data to determine
relationships between emissions, air quality, deposition, and ecological effects. The basic tenets of
CASTNet are to monitor sensitive ecosystems, to define the spatial distribution of pollutants, to detect
and quantity trends in pollutants, to implement monitoring in cooperation with other agencies and
organizations, and to implement monitoring only to fill gaps in monitoring coverage.
In 1990, the NDDN became part of CASTNet. CASTNet is the primary source for atmospheric
data to estimate dry deposition and to provide data on rural ozone. The National Atmospheric
Deposition Program (NADP) is the primary source for data on wet deposition. CASTNet supplements the
NADP with wet deposition measurements at selected sites. The National Oceanic and Atmospheric
Administration provides intensive dry and wet deposition monitoring as part of the Atmospheric and
Integrated Research Monitoring Network (AIRMoN). The National Park Service operates an air quality
monitoring network for ozone, sulfur dioxide, and paniculate matter at a number of National Parks.
Each of the above networks contributes specific data to provide a comprehensive picture of deposition
and air quality in primarily rural areas of the United States.
This report is the first in a series of summaries of concentration and deposition data from the
CASTNet Deposition Network. This report covers the years 1990, 1991, and 1992.
ill
-------
Contents
Foreword jjj
Abstract v
Figures vi
Tables x
Abbreviations and Symbols xi
Acknowledgements xii
1. Introduction 1
2. Summary and Conclusions 3
3. Network Description and Operations 6
3.1 Network Description 6
3.2 Methods 12
3.2.1 Field Operations 12
3.2.2 Laboratory Operations 15
3.2.3 Model Calculations 17
4. Results and Discussion 19
4.1 Air Chemistry and Dry Deposition 19
4.1.1 Annual Atmospheric Concentrations 19
4.1.2 Quarterly and Annual Variability of
Atmospheric Concentrations 35
4.1.3 Ammonium Versus Sulfate 50
4.1.4 Calculated Dry Deposition 52
4.2 Wet Deposition 81
4.3 Total Deposition . . ,. 88
4.4 Ozone Concentrations 92
5. Special Studies 105
5.1 Comparability of United States and Canadian Atmospheric
Concentration Data 105
5.2 Elevational Gradients in Concentration and Fluxes 110
6. Quality Assurance Summary 117
6.1 Laboratory Accuracy and Precision 117
6.1.1 Laboratory QC for Filter Pack Measurements 118
6.1.2 Laboratory QC for Precipitation Chemistry Measurements 118
6.2 Results of Collocated Measurements 121
6.2.1 Filter Pack Data 121
6.2.2 Continuous Measurements 125
References 127
IV
-------
Abstract
The National Dry Deposition Network (NDDN) was established to provide long-term estimates of
dry acidic deposition across the continental United States. Fifty routine sites were operational from 1990
to 1992, including 41 sites in the eastern United States and 9 sites in the western United States. Each
site was equipped with sensors for continuous measurements of ozone (Og) and meteorological variables
required for estimation of dry deposition rates. Weekly average atmospheric concentrations of
paniculate sulfate (SOj), paniculate nitrate (NC£), paniculate ammonium (NHJ), sulfur dioxide (SO^,
and nitric acid (HNOa) were measured at all sites and wet deposition of acidity and related species were
measured at selected sites. Two methods development sites were installed during 1991 to evaluate:
1) comparability of United States and Canadian air quality measurements, and 2) effects of terrain on
pollutant concentration and deposition. Routine application of an inferential model for calculation of
deposition velocities and dry deposition fluxes was also begun.
Atmospheric concentration data show species-dependent variability in space and time. In
general, the highest concentrations are observed in the northeast and midwest, and these are a factor of
5 to 10 times higher than those observed in the west. Significant concentration gradients are also
observed from the northeast through upper northeast and midwest through upper midwest. Annual
average concentrations of most species decreased from 1989 to 1992 in all subregions of the network.
Dry deposition calculations for 1990 and 1991 show that SO2 and HNO3 dominate sulfur and
nitrate-nitrogen fluxes, respectively. In general, SO2 accounts for more than 75 percent of dry sulfur
deposition at eastern sites and more than 50 percent of dry sulfur deposition at western sites. HNO3
accounts for more than 90 percent of dry nitrate-nitrogen deposition at all sites. Total deposition
estimates for approximately 15 sites show that dry deposition accounts for about 20 to 50 percent of wet
plus dry sulfur deposition and 30 to 60 percent of wet plus dry nitrate-nitrogen deposition.
Data from a pair of sites located in a valley and on a nearby ridge show that elevational
gradients in concentration, deposition velocity, and flux can be significant. Reactive gas concentrations
and fluxes are 2 to 4 times higher at the ridge than in the valley, suggesting that deposition in areas of
complex terrain may be difficult to estimate.
Collocated data from the Canadian Acid Precipitation Monitoring Network (CAPMoN) and NDDN
for a site in southern Ontario indicate that annual average concentrations are within ±5 percent, except
for SO2. NDDN data for SO2 are lower than CAPMoN by about 10 percent. These results indicate that
methodological differences between networks should not interfere significantly with pattern and trend
analyses across eastern North America.
-------
Figures
1 NDDN monitoring sites-December 1992 7
2 Sensitive terrestrial ecosystems with defined or likely effects from
air pollution stress 10
3 Acidic and susceptible aquatic ecosystems 11
4 Mean annual SO*" concentrations (jig/m3) for 1990 20
5 Mean annual SO* concentrations (ng/m3) for 1991 21
6 Mean annual SO* concentrations (|ig/m3) for 1992 22
7 Mean annual NO3 concentrations (|ig/m3) for 1990 24
8 Mean annual NO3 concentrations (ng/m3) for 1991 25
9 Mean annual NO3 concentrations (jig/m3) for 1992 26
10 Mean annual NHJ concentrations (ug/m3) for 1990 27
11 Mean annual NH} concentrations (|ig/m3) for 1991 28
12 Mean annual NHJ concentrations (ug/m3) for 1992 29
13 Mean annual SO2 concentrations (|ig/m3) for 1990 30
14 Mean annual SO2 concentrations (jig/m3) for 1991 31
15 Mean annual SO2 concentrations (jig/m3) for 1992 32
16 Mean annual HNO3 concentrations (jig/m3) for 1990 34
17 Mean annual HNO3 concentrations (jig/m3) for 1991 36
18 Mean annual HNO3 concentrations (fig/in3) for 1992 37
19 Mean annual total NO3 concentrations (|ig/m3) for 1990 38
20 Mean annual total NO3 concentrations (|ig/m3) for 1991 39
21 Mean annual total NO3 concentrations (|ig/m3) for 1992 40
22 Seasonal variability of SOj (ng/m3) for selected sites 42
23 Seasonal variability of NO3 (jig/m3) for selected sites 43
24 Seasonal variability of NHJ (|ig/m3) for selected sites 44
25 Seasonal variability of SO2 (jig/m3) for selected sites 45
VI
-------
Figures (Continued)
26 Seasonal variability of HNO3 (|ig/m3) for selected sites 46
27 Seasonal variability of total NO3 (ng/m3) for selected sites 47
28 Scattergram of NHJ versus SO* (molar basis) for 1991 51
29 Calculated dry deposition for SO2 (kg-S/ha) for 1990 54
30 Calculated dry deposition for SO2 (kg-S/ha) for 1991 55
31 Calculated dry deposition for SO2 (kg-S/ha) for 1992 56
32 Calculated dry deposition for SO* (kg-S/ha) for 1990 57
33 Calculated dry deposition for SO* (kg-S/ha) for 1991 58
34 Calculated dry deposition for SO* (kg-S/ha) for 1992 59
35 Calculated dry deposition for total sulfur (kg-S/ha) for 1990 60
36 Calculated dry deposition for total sulfur (kg-S/ha) for 1991 61
37 Calculated dry deposition for total sulfur (kg-S/ha) for 1992 62
38 Percent gaseous sulfur deposition for 1991 63
39 Weekly deposition velocities for SO2 for selected sites 65
40 Weekly SO2 fluxes for selected sites 66
41 Calculated dry deposition of HNO3 (kg-N/ha) for 1990 67
42 Calculated dry deposition of HNO3 (kg-N/ha) for 1991 68
43 Calculated dry deposition of HNO3 (kg-N/ha) for 1992 69
44 Calculated dry deposition of NO3 (kg-N/ha) for 1990 71
45 Calculated dry deposition of NO3 (kg-N/ha) for 1991 72
46 Calculated dry deposition of NO3 (kg-N/ha) for 1992 73
47 Calculated dry deposition of total NO3 (kg-N/ha) for 1990 74
48 Calculated dry deposition of total NO3 (kg-N/ha) for 1991 75
49 Calculated dry deposition of total NO3 (kg-N/ha) for 1992 76
50 Percent gaseous NO3 deposition for 1991 77
51 Weekly deposition velocities for HNO3 (cm/sec) for selected sites 78
VII
-------
Figures (Continued)
52 Weekly HN03 fluxes for selected sites 79
53 Scattergrams of SO2 flux versus concentration and HNO3 flux versus
concentration for 1991 80
54 Wet deposition of SOiJ (kg-S/ha) for 1990 82
55 Wet deposition of SOj (kg-S/ha) for 1991 83
56 Wet deposition of SOj (kg-S/ha) for 1992 84
57 Wet deposition of NO3 (kg-N/ha) for 1990 85
58 Wet deposition of NO3 (kg-N/ha) for 1991 86
59 Wet deposition of NO3 (kg-N/ha) for 1992 87
60 Hourly average O3 concentrations for typical rolling terrain (Site 108), complex
terrain (Site 119), mountaintop (Site 120), semi-urban (Site 116), and western
(Sites 168 and 174) sites 97
61 W126 (ppm-hr) for eastern sites for 1990 99
62 W126 (ppm-hr) for eastern sites for 1991 100
63 W126 (ppm-hr) for eastern sites for 1992 101
64 SUM60 (ppm-hr) for eastern sites for 1990 102
65 SUM60 (ppm-hr) for eastern sites for 1991 103
66 SUM60 (ppm-hr) for eastern sites for 1992 104
67 Time series of SOj" and NO3 concentrations (ng/m3) collected by CAPMoN
and NDDN at Egbert, Ontario 107
68 Time series of NH« and SOa concentrations (fig/m3) collected by CAPMoN
and NDDN at Egbert, Ontario 108
69 Time series of HNO3 and total NO3 concentrations (ng/m3) collected by
CAPMoN and NDDN at Egbert, Ontario 109
70 Time series of SO* concentrations (ug/m3) for ridge and valley sites at
Coweeta Hydrologic Laboratory (Site 137) 112
71 Time series of SO2 concentrations (ug/m3) for ridge and valley sites at
Coweeta Hydrologic Laboratory (Site 137) 113
72 Time series of HNO3 concentrations dig/m3) for ridge and valley sites at
Coweeta Hydrologic Laboratory (Site 137) 114
Vlll
-------
Figures (Continued)
73 Scattergrams of collocated filter pack data for SOj" and SO2 for one eastern site
site (128) and one western site (163) ................................... 123
74 Scattergrams of collocated filter pack data for NOg and HNO3 for one eastern site
site (128) and one western site (163) ................................... 124
75 Scattergrams of collocated continuous data for O3, temperature, solar radiation,
-------
Tables
1 Site-Selection Criteria for CASTNet Dry Deposition Sites 8
2 Deployment History of the NDDN 9
3 NDDN Site Listing 13
4 Precision and Accuracy Objectives of NDDN Field Measurements 16
5 Precision and Accuracy Objectives for NDDN Laboratory Data 18
6 Annual Average Concentrations (|ig/m3) of SO*, NOg, and NH}, 1989-1992 48
7 Annual Average Concentrations (ng/m3) of HNO3) SO2, and Total NOg,
1989-1992 49
8 Total Sulfur and Nitrate-Nitrogen Deposition for NDDN Sites Measuring Wet
and Dry Deposition for 1990 89
9 Total Sulfur and Nitrate-Nitrogen Deposition for NDDN Sites Measuring Wet
and Dry Deposition for 1991 90
10 Total Sulfur and Nitrate-Nitrogen Deposition for NDDN Sites Measuring Wet
and Dry Deposition for 1992 91
11 Summary of NDDN Ozone Maxima for 1990 93
12 Summary of NDDN Ozone Maxima for 1991 94
13 Summary of NDDN Ozone Maxima for 1992 95
14 Comparison of 1991 Atmospheric Concentration Data for CAMPoN and
NDDN at Egbert, Ontario 110
15 Comparison of Ridge (R) and valley (V) data for Coweeta Hydrologic
Laboratory, July through December 1991 115
16 Summary of Laboratory Accuracy and Precision Data for Filter Pack Measurements,
1992 119
17 Summary of Laboratory Accuracy and Precision Data for Precipitation Chemistry,
1992 120
18 Summary of Collocated Filter Pack Data for 1992 122
-------
Acronyms and Abbreviations
AES
°C
Ca2*
CAPMoN
Cl
cm/sec
EPA
ESE
HN03
1C
ICAP
ICP-AE
K+
kg-N/ha
kg-S/ha
km
L/min
LAI
m
MAD
MAPD
MFC
Mg2*
m/sec
mmHg
Na+
NAAQS
NADP/NTN
NDDN
NH3
NH;
NH«NO3
NIST
NOAA
NO,
NO;
NO3
03
ppb
RPD
RTD
SO2
SO2
tpy
jig/m3
jim
USGS
UV
Vd
W/m2
Atmospheric Environment Service
degrees Celsius
participate calcium
Canadian Acid Deposition Monitoring Network
chloride
centimeters per second
U.S. Environmental Protection Agency
Environmental Science & Engineering, Inc.
nitric acid
ion chromatography
inductively coupled argon plasma
inductively coupled plasma-atomic emission
particulate potassium
potassium carbonate
kilogram of nitrate per hectare
kilogram of sulfate per hectare
kilometer
liters per minute
leaf area index
meter
median absolute difference
median absolute percent difference
mass flow controller
particulate magnesium
meters per second
millimeters of mercury
particulate sodium
National Ambient Air Quality Standard
National Atmospheric Deposition Program/National Trends Network
National Dry Deposition Network
ammonia
particulate ammonium
ammonium nitrate
National Institute of Standards and Technology
National Oceanic and Atmospheric Administration
nitrogen oxide
nitrite
particulate nitrate
ozone
parts per billion
relative percent difference
resistance-temperature device
sulfur dioxide
particulate sulfate
tons per year
micrograms per liter
micrograms per cubic meter
micrometer
U.S. Geological Survey
ultraviolet
deposition velocity
watts per square meter
XI
-------
Acknowledgements
The authors gratefully acknowledge the following site operators, without whose dedicated efforts
NDDN could not succeed: S. Scott (104); R. Hopkins (105); D. DeCapria (106); F. Wood (107); G.
Brooks (108); S. Nolan (109); T. Butler (110); R. Russell (111); D. Dom (112); D. Croskey (113); S.
Hammond (114); L. Chilcote (115); V. Miller (116); J. Hufman (117); R. Gubler (118); B. Jenkins (119);
S. Long (120); M. Brotzge (121); T. Chatfield (122); D. Stineman (123); F. Matt and J. Matt (124); H.
Perry and B. Perry (125); P. Hughes (126); M. Hale (127); S. Scamack (128); M. Yewell (129/131); M.
Snider (130); D. Conrad* and J. 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); J. Bishop (156); W. Steiner and B. Steiner
(157); B. Barr (161); C. Jenson (162); J. Moubray (163); S. Kiracofe (164); C. Laster (165); B. Smith, D.
Anderson, and A. King (167); R. Ljung (168); D. Lukens (169); and P. Hays (174).
^deceased
XII
-------
Chapter 1
Introduction
Atmospheric deposition takes place via two pathways: 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 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 10 to 15 years; however, comparable information is unavailable for dry deposition rates.
This lack of information increases the uncertainty in estimates of interregional, national, and
international transport and confounds efforts to determine the overall impact of atmospheric deposition.
The direct measurement of dry deposition is not straightforward, but a number of investigations
have shown that it can be reasonably inferred by coupling air concentration data with routine
meteorological measurements (Shieh et 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
submicron particle and sulfur dioxide (SO2) 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 atmospheric species [including nitric acid (HNO3)] and to use site-specific
meteorological data. Similar 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 Environmental Science
& Engineering, Inc. (ESE) to establish and operate the National Dry Deposition Network (NDDN). The
objective of the NDDN is to obtain field data at approximately 50 sites throughout the United States to
establish patterns and trends of dry deposition. The approach adopted by the NDDN is to calculate dry
deposition using measured air pollutant concentrations and inferred deposition velocities (Vds) estimated
from meteorological, land use, and site characteristic data. The inferential model currently used for dry
deposition calculations is a multi-layer version of the Big Leaf Model developed by Hicks et al. (1985).
-------
This report summarizes results of NDDN monitoring activities from 1990 through 1992. Annual
concentration data for atmospheric sulfur and nitrogen species are presented, and temporal variability is
described. Results of dry deposition calculations for 1990, 1991, and 1992 are discussed, and the
relative contribution of gases versus aerosols are evaluated. Wet deposition data for approximately 15
NDDN sites are presented and then used, along with dry deposition calculations, to estimate total
depositions of sulfur and nitrate-nitrogen. The relative magnitude of wet and dry deposition are
discussed. Ozone concentrations and exposure statistics are presented for 1990, 1991, and 1992.
Data are also presented from two comparability studies initiated in 1991. The first of these
involves investigation of measurement biases between atmospheric sampling methods used by United
States and Canadian acid deposition trends programs. Data are presented and analyzed from a
collocated site in Ontario, Canada. The second study is an investigation of a terrain-induced bias in
concentration measurements and dry deposition estimates. Data are presented from a pair of sites in
southwestern North Carolina that are separated horizontally by about 1,000 meters (m) and in elevation
by about 350 m.
-------
Chapter 2
Summary and Conclusions
Quantitative information on dry deposition fluxes and atmospheric concentrations is needed to
evaluate patterns and trends in total deposition and to support model evaluation and ecosystem effects
studies. Fifty monitoring sites were operational from 1990 through 1992, including approximately 15
sites which measured wet deposition. Ozone (63) and meteorological data were collected continuously,
and atmospheric sulfur and nitrogen species and wet deposition (if applicable) were sampled on weekly
intervals. Dry deposition fluxes for 1990, 1991, and 1992 were calculated using the multi-layer model
developed by the National Oceanic and Atmospheric Administration (NOAA) and data collected from the
50-site network. Studies at two non-routine sites investigated the comparability of United States and
Canadian trends measurements and local variability in concentrations, Vds, and fluxes.
Annual concentration data for sulfur and nitrogen species show fairly consistent spatial patterns
from 1990 through 1992. In the eastern United States, the highest average concentrations of particulate
sulfate (SOj*) [6 to 7 micrograms per cubic meter (ug/m3)] and SO2 (15 to 20 jig/m3) occur in an area
surrounding the Ohio River Valley. Average SO*' decreases gradually toward the periphery of the
network to about 2.0 jig/m3 in northern Maine and 3 to 4 ug/m3 in Florida, Arkansas, and Wisconsin.
Average SOa exhibits much more local variability than SO$" and decreases more rapidly towards the
periphery of the network. For western sites, SOj" and SO2 concentrations are generally well below
1.0 (ig/m3 and show no strong evidence of a pattern, except for somewhat elevated values at two Arizona
sites.
Atmospheric concentration data for HNO3 and particulate nitrate (NOg) also show substantial
variability across the network; however, differences between eastern and western sites are not as great as
for SOl' and SO2. In general, concentration patterns for HNO3 and NO3 appear to be strongly influenced
by land-use and topographic features. Maximum HNO3 concentrations (3 to 4 jig/m3) are observed at
scattered sites in New York, Pennsylvania, Virginia, and Ohio, while minimum concentrations
(<1.0 |ig/m3) are observed in Maine, Arkansas, Florida, Kentucky, and North Carolina. Substantial
variability in HN03 concentrations (factors of 2 to 3) occurs over fairly short distances in and around the
southern Appalachian Mountains. Sites in this area with good fetch exhibit concentrations substantially
higher than sites with limited fetch.
Annual patterns of NOj aerosol show peak concentrations (3 to 4 jig/m3) in agricultural areas of
the midwest and minimum concentrations (<0.5 (ig/m3) in the forested northeast. Intermediate
-------
concentrations occur at sites in pastured land or near limited agricultural activities. These observations
suggest that land use (specifically agricultural activity) affects the partitioning of gas and aerosol
nitrogen species. Two potential mechanisms for this include reaction between HNO3 and either soil
particles or ammonia (NH3). The first mechanism would produce aerosol NO3 in the 1 to 5-micrometer
((im) range, and the second would produce aerosol NO3 in the
-------
Inspection of data from rural sites in and around the Appalachian Mountains indicates that
concentrations and fluxes of SO2 and HNO3 can vary by as much as factors of 2 to 5 within only a few
hundred kilometers and that this variability is unlikely to be related to local sources of SO2 and nitrogen
oxides (NOX). Data from a pair of sites located on a ridge and in a valley in the southern Appalachian
Mountains may explain this phenomenon. The ridge site exhibits significantly higher concentrations and
Vds than the valley site. Together, these factors give rise to dry deposition rates that are factors of
1.6 (for SO*") to 4.4 (for HNOa) higher at the ridge site than the valley site. These observations verify
modeling work by McMillen (1990), which suggested that deposition rates in complex terrain could be
highly variable. Ridgetop sites may provide upper limit estimates of regional deposition, while valley
and intermediate elevation sites may provide reasonable estimates of average deposition to specific
ecosystems. Further work has been initiated to determine if elevation is important elsewhere, and, if so,
how data at a single site cen be scaled over areas of interest.
Comparison of atmospheric concentration data from the United States and Canadian trends
networks [NDDN and Canadian Acid Precipitation Monitoring Network (CAPMoN)] shows that annual
averages of most species agree within 5 percent. Data for SO2 differ by about 10 percent, on average,
between networks, with NDDN concentrations almost invariably lower than those of CAPMoN. In
general, these findings suggest that data from the two networks can be readily combined for analysis of
spatial patterns and long-term trends across eastern North America. Additional work has been initiated
to elucidate the disparity in SO2 concentrations between networks and to compare United States and
Canadian approaches for estimating dry deposition fluxes.
-------
Chapter 3
Network Description and Methods
3.1 Network Description
The status of the NDDN, as of December 1992, is shown in Figure 1. Forty-one primarily rural
eastern sites and nine western sites were operational from 1990 through 1992. All eastern sites, with
the exception of two relocations, were installed and collecting data as of January 1989, and all western
sites were collecting data as of July 1989. With few exceptions, as noted below, all sites complied with
siting criteria shown in Table 1. The number of sites in the network is summarized by year in Table 2.
The current network is designed to support ecological effects research and investigation of
relationships between emissions and atmospheric concentrations/depositions. Recent assessments of
sensitivity to O3 and acid deposition have shown that large areas of potentially sensitive terrestrial and
aquatic ecosystems exist in the eastern United States and that limited areas of sensitive ecosystems exist
in the western United States (see Figures 2 and 3). These findings, coupled with the expectation of
future changes in emissions of sulfur and nitrogen oxides, underly the distribution of sites in the
network.
For presentation purposes, sites in the eastern United States have been grouped into six
subregions (see Figure 1). Subregional designations, with numbers of sites in parentheses, are as
follows: northeast (11), upper northeast (3), midwest (9), upper midwest (3), south central (11), and
southern periphery (3). Besides geographic location, site groupings were based on terrain and general
spatial patterns of atmospheric concentration data. The categorization of sites also divides the
Appalachian Mountains into three convenient geographic ranges. The northern Appalachians, including
the Adirondack Mountains of New York, the Green Mountains of Vermont, and the White Mountains of
New Hampshire fall within the upper northeast subregion. The central Appalachians, including the
Catskill Mountains of New York, and the Allegheny Mountains of West Virginia, fall within the northeast
subregion. Finally, the southern Appalachians, including the Great Smoky Mountains of North Carolina
and Tennessee, fall within the south central subregion. Each of these subregions thus includes sites in
mountainous areas, both for the purpose of monitoring deposition in sensitive areas and to elucidate
variability of deposition in complex terrain.
Sites in the upper northeast subregion are exclusively rural-forested, while those in the northeast
subregion exhibit a range of characteristics. Six northeastern sites are rural-forested, two are rural-
agricultural (106 and 128), and three are near or within the Washington-Baltimore-Philadelphia-
-------
UPPER
NORTHEAST
Legend
• Dry Deposition OVy
• Wet Deposition and Dry Deposition
NORTHEAST
SOUTHERN
PERIPHERY
SOUTH
CENTRAL
Figure 1. NDDN monitoring sites-December 1992.
-------
Table 1. Site-Selection Criteria for CASTNet Dry Deposition Sites
Potential Interferant
Minimum Acceptable
Distance (km)
SO2 or NOX Pt. Source >100 tpy
SO2 or NOX Pt. Source > 1,000 tpy
Major Industrial Complex
Town, population 1,000 - 10,000
Town, population 10,000 - 25,000
City, population 15,000-50,000
City, population > 50,000
Major highway, airport, railway
Secondary road, heavily traveled
Secondary road, lightly traveled
Feedlot operations
Intensive agricultural activities
Limited agricultural activities
Parking lot or large paved area
Building with fuel combustion
Sewage treatment plant
Forced main vent or lift station
Tree line
Complex terrain
20
40
10
5
10
20
40
2
0.5
0.2
0.5
0.5
0.1
0.2
0.2
1.0
0.2
0.1
variable
Note: km = kilometer.
Source: ESE.
-------
Table 2. Deployment History of the NDDN
Year
1988
1989
1990
1991
1992
Eastern
16
41
41
41
41
No. of Sites*
Western
0
2
9
9
9
Total
16
43
50
50
50
*Indicates number of sites in operation as of January 1.
New York City conurbation (104, 116, and 144). The upper midwest sites are rural-agricultural or
rural-forested (149). The midwest sites are rural-agricultural, except for Site 146 (suburban Chicago),
which is urban-agricultural. Although rural in character, three sites (122, 140, and 157) are influenced,
to a greater or lesser extent, by SO2 emissions from nearby point sources with annual emissions in
excess of 1,000 tons per year (tpy).
The south-central sites are either rural-forested or rural-agricultural but exhibit a wide range of
terrain characteristics. Three sites are located above 1,000 m and form a line extending from northern
Virginia to southwestern North Carolina. Site 118 is situated on a ridge of the eastern Blue Ridge
Mountains, and Sites 120 and 126 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). Two
sites (121 and 137) are located in hollows or valleys, and the other six sites in the subregion are in
rolling terrain. The distribution of sites in this subregion provides an opportunity to investigate
relationships among terrain characteristics, atmospheric concentrations, and dry deposition (see
Section 4.6). Finally, all three of the southern periphery sites are rural-forested in flat or rolling terrain.
Despite apparent similarities in land use and terrain, the western sites are not homogeneous in
character. For this reason, no attempt was made to group western sites. Site 161 (Gothic, CO) occupies
a mountain valley within the central Rocky Mountains. Site 162 is located on die foothills of the High
Uintas, the most prominent east-west mountain range in North America. Sites 163 and 164 are located
in semi-arid rangeland near the northern extreme of the Great Basin. Sites 165 and 169 represent the
-------
Acid Deposition > Ozone
Ozone > Acid Deposition
Ozone = Acid Deposition
Defined Effect
SOURCE: EPA.
Figure 2. Sensitive terrestrial ecosystems with defined or likely effects from air
pollution stress.
-------
-------
transition from the western Great Plains to the Rocky Mountains. Sites 167 and 174 are located in the
arid southwest; Site 167 is in the Sonoran Desert, while Site 174 is on the extensive and forested Kaibab
Plateau. Site 168 (near the Canadian border) alone represents the western boreal forest. Thus,
although these sites are collectively termed the western part of the network, they represent a wide range
of environments.
Site locations and descriptive information are provided in Table 3. Terrain and land-use
information refers to a 10-km radius around the site and is presented to convey a sense of the setting
within which each site operates. Site numbers are used for identification purposes only and do not
correlate with order of installation or operation.
3.2 Methods
This section provides a brief overview of NDDN methods. Step-by-step protocols and additional
details on these activities can be found in the NDDN Field Operations Manual, Laboratory Operations
Manual, and Data Management Manual (ESE, 1990a, 1990b, 1991a).
3.2.1 Field Operations
Ambient measurements for O3, SO2, SO*', NO3, HNO3, paniculate ammonium (NH}), and
meteorological variables required for dry deposition calculations were performed at each NDDN site.
Meteorological variables and O3 concentrations were recorded continuously and reported as hourly
averages consisting of a minimum of nine valid 5-minute averages. Atmospheric sampling for sulfur and
nitrogen species was integrated over weekly collection periods using a 3-stage filter pack. 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 cellulose filters. Filter packs were prepared and
shipped to the field weekly and exchanged at each site every Tuesday. Blank filter packs were collected
monthly to evaluate passive collection of particles and gases as well as contamination during shipment
and handling. At 16 sites located more than 50 km from National Atmospheric Deposition
Program/National Trends Network (NADP/NTN) sites (see Figure 1), wet deposition samples were
collected weekly (according to NADP/NTN protocols) and shipped to ESE for chemical analysis.
Filter pack sampling and 03 measurements were performed at 10 m using a tilt-down aluminum
tower (Aluma, Inc.). Filter pack flow was maintained at 1.50 liters per minute (L/min) at eastern sites
and 3.00 L/min at western sites, for standard conditions of 25 degrees Celsius (*C) and 760 millimeters
of mercury (mmHg) with a Teledyne-Hastings CST-10K mass flow controller (MFC). Wet deposition
samples were collected in precleaned polyethylene buckets using an Andersen Model APS precipitation
sampler. Buckets were placed on the sampler on Tuesday and removed, whether or not rainfall had
12
-------
Table 3. NDDN Site Listing
Site
No.
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
130
131
133
134
135
137
140
144
Site Name
West Point, NY
Whiteface
Mountain, NY
PSU, PA
Parsons, WV
Prince Edward, VA
Woodstock, NH
Connecticut Hill,
NY
Speedwell, TN
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
Horton Station, VA
Lilley Cornett
Woods, KY
Oxford, OH
Lykens, OH
Unionville, MI
Candor, NC
Cranberry, NC
Edgar Evins State
Park, TN
Arendtsville, PA
Bondville, IL
Mackville, KY
Salamonie, IN
Perkinstown, WI
Ashland, ME
Coweeta, NC
Vincennes, IN
Washington's
Crossing, NJ
Initial
Reporting
Date
01/06/87
01/06/87
01/06/87
01/14/88
11/01/87
12/31/88
09/14/87
07/01/89
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
09/30/90
12/31/88
03/22/88
06/30/88
02/09/88
07/31/90
06/30/88
09/30/88
12/31/88
11/03/87
08/05/87
12/31/88
Latitude
41.35
44.39
40.73
39.09
37.17
43.94
42.40
36.47
41.60
41.43
39.63
42.42
39.03
40.00
38.52
38.88
37.33
37.08
39.53
40.92
43.61
35.26
36.11
36.04
39.92
40.05
37.70
40.82
45.21
46.61
35.06
38.74
40.30
Longitude
74.05
73.86
77.95
79.66
78.31
71.70
76.65
83.83
78.77
80.15
83.26
83.90
76.82
79.25
78.44
80.85
80.55
82.99
84.72
83.00
83.36
79.84
82.04
85.73
77.31
88.37
85.05
85.66
90.60
68.41
83.43
87.49
74.87
Elevation
(m)
203
570
378
510
146
258
515
361
622
384
265
267
46
615
1,073
234
920
335
284
303
201
198
1,219
302
269
212
353
249
472
235
686
134
58
Primary
Land
Use
Forested
Forested
Agricultural
Forested
Forested
Forested
Forested
Agricultural
Forested
Forested
Agricultural
Forested
Urban-Agric.
Forested
Forested
Forested
Forested
Forested
Agricultural
Agricultural
Agricultural
Forested
Forested
Forested
Agricultural
Agricultural
Agricultural
Agricultural
Agricultural
Agricultural
Forested
Agricultural
Agric.-Urban
Terrain
Complex
Complex
Rolling
Complex
Rolling
Complex
Rolling
Rolling
Rolling
Rolling
Rolling
Flat
Flat
Complex
Mountaintop
Complex
Mountaintop
Complex
Rolling
Flat
Flat
Rolling
Mountaintop
Rolling
Rolling
Flat
Rolling
Flat
Rolling
Flat
Complex
Rolling
Rolling
13
-------
Table 3. NDDN Site Listing (Continued)
Site
No.
146
149
ISO
151
152
153
156
157
161
162
163
164
165
167
168
169
174
Site Name
Argonne National
Laboratory, IL
Wellston, MI
Caddo Valley, AR
Coffeeville, MS
Sand Mountain, AL
Georgia Station,
GA
Sumatra, PL
Alhambra, IL
Gothic, CO
Uinta, UT
Reynolds Creek, ID
Saval Ranch, NV
Pinedale, WY
Chiricahua, AZ
Glacier National
Park, MT
Centennial, WY
Grand Canyon, AZ
Initial
Reporting
Date
07/01/87
06/30/88
09/30/88
12/31/88
12/31/88
06/30/88
12/31/88
06/30/88
07/01/89
07/01/89
07/01/89
07/01/89
12/31/88
07/01/89
12/31/88
07/01/89
07/01/89
Latitude
41.70
44.22
34.18
34.00
34.29
33.18
30.11
38.87
38.96
40.55
43.21
41.29
42.93
32.01
48.51
41.31
36.06
Longitude
87.99
85.82
93.10
89.80
85.97
84.41
84.99
89.62
106.99
110.32
116.75
115.86
109.79
109.39
114.00
106.15
112.18
Elevation
Cm)
229
295
71
134
352
270
14
164
2,926
2,500
1,198
1,873
2,388
1,570
963
2,579
2,073
Land
Use
Urban-Agric.
Forested
Forested
Forested
Agricultural
Agricultural
Forested
Agricultural
Range
Range
Range
Range
Range
Range
Forested
Range
Forested
Terrain
Flat
Plat
Rolling
Rolling
Rolling
Rolling
Flat
Flat
Complex
Complex
Rolling
Rolling
Rolling
Complex
Complex
Complex
Complex
Source: BSE.
occurred, the following Tuesday. Buckets were weighed in the field, then sealed and shipped to ESE for
chemical analysis. Precipitation amount (depth) was also monitored at wet deposition sites.
Og was measured via ultraviolet (UV) absorbance with a Thermo-Environmental Model 49-103
analyzer operating on the 0- to 500-ppb range. Ambient air was drawn from the 10-m air quality tower
through a 3/8-inch TFE Teflon* sampling line. 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 parts per billion
(ppb)], and span (400 ppb) checks of the 0, analyzer were performed every third day using an internal
0, generator.
In addition to the above, various observations were periodically made at the NDDN sites to
support model calculations of dry deposition. Site operators recorded surface conditions (e.g., dew, frost,
snow) and vegetation status weekly. Vegetation status and land-use information were used to define the
14
-------
distribution and condition of plant species around each site that could influence deposition rates for
gases (especially SO^ and particles. Vegetation data were obtained to track evolution of the dominant
plant canopy, from leaf emergence (or germination) to senescence (or harvesting). Once a year, site
operators also provided information on major plant species and land-use classifications within 1.0 km of
the site. Additional land-use data was obtained by digitization and analysis of aerial photographs
obtained from the U.S. Geological Survey (USGS) National Cartographic Information Center in Reston,
VA. Photographs were interpreted according to procedures described by Anderson et aL (1978).
Leaf area index (LAI) measurements were conducted at all NDDN sites during the summers of
1991 and 1992. LAI is the one-sided leaf area of the plant canopy per unit area of ground at full leaf
emergence and has been shown to play an important role in atmosphere-canopy exchange processes
(McMillen, 1990). LAI was measured using an LAI-2000 Plant Canopy Analyzer manufactured by Li-Cor
(Lincoln, NE). The LAI-2000 makes indirect (i.e., nondestructive) estimates of LAI from simultaneous
measurements of light interception by the plant canopy at five angles of inclination (Li-Cor, 1989).
Initial development and testing of the LAI-2000 by the manufacturer focused on a variety of agricultural
crops, such as soybeans and wheat, and similar approaches have been used to measure LAI of forest
canopies (Pierce and Running, 1988; Chason et ah, 1990).
All field equipment was subjected to quarterly inspections and multipoint calibrations, using
standards traceable to the National Institute of Standards and Technology (NIST). In addition,
independent equipment audits were performed annually by Ogden Environmental and Energy Services,
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. Precision and accuracy criteria for NDDN field
measurements are shown in Table 4.
3.2.2 laboratory Operations
Filter pack samples were loaded, shipped, received, extracted, and analyzed by ESE personnel at
the Gainesville, FL laboratory. Filter packs contained three types of filters in sequence: a Teflon* filter
for collection of aerosols, a nylon filter for collection of HNO3, and dual potassium carbonate (KjCOa)
impregnated cellulose filters for collection of SO2.
Following receipt from the field, exposed filters and blanks were extracted and then analyzed for
SO* and NOj by micromembrane-suppressed ion chromatography (1C). Teflon* filter extracts were also
analyzed for NH} by the automated indophenol method using a Technicon II or TRAACS-800
Autoanalyzer system. All analyses were completed within 72 hours of filter extraction.
15
-------
Table 4. Precision and Accuracy Objectives of NDDN Field Measurements
Measurement
Parameter
Windspeed
Wind Direction
Sigma Theta
Relative Humidity
Solar Radiation
Precipitation
Ambient Temperature
Delta Temperature
03
Filter Pack Flow
Surface Wetness
Method
Anemometer
Wind Vane
Wind Vane
Hygrometer
Pyranometer
Rain Gauge
Platinum RTD
Platinum RTD
UV Absorbance
Mass Flow Controller
Conductivity Bridge
Objectives
Precision
±0.5 m/sec
±5°
±10%
+10% (of
full scale)
±10% (of
reading
taken at
local noon)
±10% (of
reading)
±0.5eC
±0.25°C
±10% (of
reading)
+.0.15 L/min
Undefined
*
Accuracy
the greater of
±0.2 m/sec or
±5%
±5°
Undefined
±10% (of
full scale)
±10%
±0.05 inch+
±0.25°C
±0.25eC
±10%
±10%
Undefined
Note: m/sec = meters per second.
RTD = resistance-temperature device.
'Precision criteria apply to collocated instruments, and accuracy criteria apply to calibration of
instruments.
fFor target value of 0.50 inch.
Source: ESE.
16
-------
Wet deposition samples were filtered and then analyzed for pH, conductivity, acidity, sodium
(Na+), potassium (K+), NHJ, calcium (Ca2+), magnesium (Mg2*), chloride (CT), nitrite (NO^), NO^ and
SO2,". Analysis of NH^ and anions was as described previously for filter pack samples. Analysis of Na+,
Mg2*, and Ca2+ was performed with a Perkin-Elmer P-2 inductively coupled argon plasma (ICAP)
emission spectrometer. Acidity was determined via titration to approximately pH 8.3, and K+ was
analyzed via atomic emission.
Results of all valid analyses were stored in the laboratory data management system.
Atmospheric concentrations were calculated (based on volume of air sampled) following validation of
hourly flow data. Atmospheric concentrations of particulate SO2,", NO3, and NHJ were calculated based
on the analysis of Teflon* filter extracts; HNO3 was calculated based on the NO3 found in nylon filter
extracts; and SO2 was calculated based on the sum of SO2," found in nylon and cellulose filter extracts.
Precision and accuracy objectives for NDDN laboratory analyses are listed in Table 5.
3.2.3 Model Calculations
Dry deposition calculations for 1990, 1991, and 1992 were made using a multi-layer version of
the NOAA inferential model, as described by Hicks et al. (1985). The model calculates fluxes, F, as the
product of measured concentrations and inferred Vds. Deposition velocity, in turn, is calculated as the
inverse sum of three separate resistances: atmospheric resistance (RJ, boundary layer resistance
and canopy resistance (R,.), as follows:
where: i = chemical species.
The three resistance terms are calculated for each chemical species and vegetation/surface type
for every hour of available meteorological input data. Hourly values of Vd are averaged over a week and
multiplied by the weekly integrated concentrations to produce weekly fluxes of HNO3, SOj, and particles.
Ozone flux is calculated using hourly measurements of 03 and hourly values of Vd. Weekly flux
calculations for all chemical species are considered valid only if .>70 percent of hourly Vd values are
available for that week. Seasonal values require at least 10 of 13 weeks of data to be considered valid.
Annual values are considered valid only if Vd and flux data for all four seasons are available.
17
-------
Table 5. Precision and Accuracy Objectives for NDDN Laboratory Data
Objectives*
Analyte
pH*
Conductivity
Acidity
NH+
Na+
K+
Mg2+
Ca2+
cr
NOj
NOj
4
Medium
W
W
W
W/F
W/F
W/F
W/F
W/F
W
W
W/F
W/F
Precision
Method (RPD)
Electrometric
Electrometric
Titrimetric
Automated colorimetry
ICP-AE
Flame atomic emission
ICP-AE
ICP-AE
1C
1C
1C
1C
12
10
15
10
10
10
10
10
5
5
5
5
Accuracy
88-
90-
85-
90-
90-
90-
90-
90-
95-
95-
95-
95-
112
110
115
110
110
110
110
110
105
105
105
105
Note: W = wet deposition samples.
F = filter pack samples.
1C = ion chromatography.
ICP-AE = inductively coupled plasma - atomic emission.
RPD = relative percent difference.
*Precision and accuracy criteria represent ±0.05 pH unit.
Source: ESE.
18
-------
Chapter 4
Results and Discussion
This section presents atmospheric concentration and dry deposition flux data for NDDN. It
illustrates spatial patterns and seasonal variations for 1990, 1991, and 1992, and describes inter-annual
variability. The comparability of United States and Canadian atmospheric concentration sampling
approaches is discussed, and sources of local variability in concentrations and fluxes are investigated.
4.1 Air Chemistry and Dry Deposition
This section presents atmospheric concentration data for SO$~, NH}, NO3, SOj, and HNO3 derived
from weekly filter pack measurements and model calculations of dry deposition for SOj", SO& HNO3, and
NO3. A detailed analysis of spatial and temporal trends is beyond the scope of this report. Therefore,
the emphasis of this section is on salient features of concentration patterns as well as season-to-season
and year-to-year variability. Spatial correlations in the data and relationships between aerosol cations
and anions are also discussed.
4.1.1 Annual Atmospheric Concentrations
Annual arithmetic mean concentrations of SOj" for 1990, 1991, and 1992 are shown in Figures 4
through 6. Annual SO$" concentrations in the eastern United States range from 7.0 (ig/m3 in central
West Virginia (Site 119) to 2.1 ng/m3 in northern Maine (Site 135). Mean values above 6.0 jig/m3 cover
much of the eastern United States, from Indiana, Ohio, and Pennsylvania to northern Alabama and
Georgia. Averages of 5.0 to 6.0 jig/m3 cover most of the region from southern New York to Illinois to
northern Florida. Only sites from northern New York to Maine, northern Michigan, Wisconsin, and
Arkansas exhibit concentrations below 4.0 |ig/m3. Strong concentration gradients (i.e., factor of 2) exist
between Pennsylvania and northern New York, Ohio and Michigan, and from Illinois into Wisconsin.
Atmospheric concentrations of SOj" across the western United States during 1990 are
substantially lower than those in the east. Mean SO* values range from 1.40 |ig/m3 in southern Arizona
(Site 167) to 0.56 |ig/m3 in northern Nevada (Site 164). Results of paired t-tests of weekly
concentrations indicate that the two sites located in the Great Basin exhibit significantly (p <0.05) lower
concentrations than other sites. Similarly, the two Arizona sites exhibit higher concentrations than the
other western sites.
Generally similar results are observed in 1991 and 1992; however, peak concentrations across
the eastern United States are somewhat lower and less widespread than in 1990. For example, the
19
-------
Figure 4. Mean annual SO* concentrations (/ig/m3) for 1990.
-------
Figure 5. Mean annual SO% concentrations (/ig/m3) for 1991.
-------
N>
Figure 6. Mean annual SO^ concentrations (jig/m3) for 1992.
-------
number of sites exhibiting mean annual concentrations >6.0 (ig/m3 was 20 in 1990, 17 in 1991, and 5
in 1992. Relatively small year-to-year changes in SO* were observed for most sites in the eastern and
western United States.
Annual NO3 concentrations exhibit much more variability than SO*' and a definite pattern of
higher concentrations in the midwest than elsewhere (see Figures 7 through 9). The lowest
concentrations in the eastern United States (i.e., <0.25 (ig/m3) are observed at forested sites in New
England and the southern Appalachian Mountains, while the highest concentrations (i.e., 2 to 3 (ig/m3)
are observed in agricultural areas of the midwest. Intermediate values (i.e., 1.0 to 2.0 ng/m3) appear to
be associated with agricultural sites anywhere in the eastern United States. Two potential mechanisms
for NO3 production include gas-phase reaction between HNO3 and NH3 and gas-particle reaction of HNO3
with soil particles. Although both of these reactions are likely to be enhanced in agricultural areas, the
spatial correlation of NHJ and NO3 concentrations suggests that the former may be more important.
Annual NO3 concentrations exhibit minimal variability across western sites. NO3 concentrations
range from 0.15 iig/m3 in the mountains of central Colorado (Site 161) to 0.41 |ig/m3 in southwestern
Idaho (Site 163). The datum for Idaho may reflect agricultural activity near Boise, which is located
approximately 60 km northeast of the site. Paired t-tests for NO3 indicate three widely spaced groups of
sites in the following states: Montana, Wyoming, and Colorado (mean concentration 0.16 (ig/m3);
Nevada, Utah, and southern Arizona (0.27 jig/m3); and Idaho and northern Arizona (0.39 (ig/m3).
Annual average values for NH} range from nearly 2.7 (ig/m3 in northern Indiana (Site 133) to
0.53 (ig/m3 in Maine (Site 135) and, in general, exhibit higher values at agricultural sites than at
forested sites (see Figures 10 through 12). Concentrations above 2.0 (ig/m3 are detected at all sites
within Illinois, Indiana, and Ohio, plus the two southernmost sites in Michigan. Additional
concentrations in this range also are found in northern Alabama, eastern Pennsylvania, and Maryland.
Average values below 1.0 jig/m3 are found only in the upper northeast and northern Florida. Average
concentrations for NH^ range from 0.23 (ig/m3 (Site 164) to 0.45 (ig/m3 (Site 167), and only these two
sites differ significantly, based on paired t-tests.
Annual values for SO2 range during 1990 from 19.5 (ig/m3 in southwestern Indiana (Site 140) to
1.9 (ig/m3 in Maine (Site 135) (see Figures 13 through 15). Concentrations of 15 (ig/m3 or greater
occur in a small area in western Pennsylvania, as well as at isolated sites in northern Illinois, southern
Indiana, and western Ohio. The Pennsylvania sites are regionally representative, but the Illinois,
Indiana, and Ohio sites appear to be influenced by local sources. A much larger area extending from
Illinois eastward to northern Virginia and southern New York exhibits concentrations in the range of 10
23
-------
to
Figure 7. Mean annual NOj concentrations (/zg/m3) for 1990.
-------
ho
Ul
Figure 8. Mean annual NOj concentrations (/ig/m3) for 1991.
-------
IsJ
Figure 9. Mean annual NO^ concentrations (/xg/m3) for 1992.
-------
tSJ
Figure 10. Mean annual NHj concentrations (/zg/m3) for 1990.
-------
N)
00
Figure 11. Mean annual NH^ concentrations (/zg/m3) for 1991.
-------
NJ
VO
Figure 12. Mean annual NHj concentrations (/xg/m3) for 1992.
-------
u>
o
Figure 13. Mean annual SO2 concentrations
for 1990.
-------
Figure 14. Mean annual S02 concentrations (jig/m3) for 1991.
-------
U)
ho
Figure 15. Mean annual SO2 concentrations (/zg/m3) for 1992.
-------
to 15 jig/m3. Sharp concentration gradients occur from Pennsylvania to northern New York, from the
lower midwest to upper midwest, and around isolated sites in Kentucky and North Carolina. Terrain
effects could account for large differences (i.e., a factor of 3) between rolling terrain sites in central
Kentucky and Tennessee, and neighboring complex terrain sites in eastern Kentucky and western North
Carolina.
SO2 concentrations exhibit substantial variability across the western region. Annual average
concentrations range over a factor of 6 from 0.32 (ig/m3 in central Colorado (Site 161) to 2.27 |ig/m3 in
southern Arizona (Site 167). Concentrations at Site 167 exceed the lowest values in the eastern region
and clearly reflect local emissions on both sides of the United States/Mexico border. Similarly, SO2
concentrations in northern Arizona (Site 174) may be influenced by southwestern sources. Results of
paired t-tests on weekly concentrations suggest at least three distinct site groups: Idaho, Nevada, and
•»
Colorado (mean concentrations of 0.34 jig/m3); northern Arizona (0.83 jig/m3); and southeastern
Arizona (2.27 (ig/m3). The remaining sites are not statistically different from either the lowest or
intermediate site group.
Data for 1991 and 1992 show a very similar SO2 pattern as for 1990. The domain for
concentration >.10 ng/m3 stretches from Illinois eastward to the Atlantic seaboard. The lowest
concentrations among eastern sites (i.e., >2.5 jig/m3) routinely occur in New England, Wisconsin,
Arkansas, Florida, and North Carolina.
Annual average concentrations of HNO3 for 1990 (see Figure 16) exhibit a maximum
concentration of approximately 3.5 jig/m3 in southeastern Pennsylvania (Site 128) and a minimum of 0.6
jig/m3 in northern Maine (Site 135). The majority of NDDN sites occupy a range from 2.0 to 3.0 |ig/m3.
Only five sites exhibit average concentrations of 1.0 (ig/m3 or less. Three of these sites are located in the
northern extremes of the network (i.e., Wisconsin, Maine, and New Hampshire). The other two are
located near the geographic center of the network in eastern Kentucky (Site 121) and southwestern
North Carolina (Site 137).
The overall pattern of HNO3 across the eastern United States might be influenced by terrain as
much or more than other factors (e.g., emissions). Sites with the highest average values have good
exposure (fetch), while those with the lowest values typically have poor exposure, due mainly to complex
terrain. For species with significant deposition velocities (such as HNOg), the microclimate in regions of
complex terrain produces local variability in atmospheric concentrations. This could account for the
highly variable concentrations in the vicinity of the Appalachian Mountains. Terrain effects are explored
further in Chapter 5.
33
-------
Figure 16. Mean annual HNO3 concentrations (/xg/m3) for 1990.
-------
Annual concentration values for HNO3 for 1990 vary by approximately a factor of three among
western sites. The lowest annual concentration occurs in Montana (Site 168), and the highest occurs in
northern Arizona (Site 174). Results for northern Arizona are similar to the lowest observed values in
the eastern region and may be due, in part, to the high number of visitors (during May to September
only) to Grand Canyon National Park. In contrast, the low concentrations in Montana and Colorado
could be due to the forested surroundings and high elevation of Sites 168 and 161, respectively.
Results for 1991 and 1992 are consistent with those for 1990 (see Figures 17 and 18). The
highest annual average concentration routinely occurs in southeastern Pennsylvania (Site 128), and the
lowest concentrations (<1.0 ng/m3) occur in New England and in two complex terrain sites in Kentucky
and North Carolina (121 and 137).
Results for total NO3 (i.e., NO3 plus HNOj) for 1990 show marked variability across the eastern
United States (see Figures 19 through 21). The highest concentrations (i.e., >4.0 jig/m3) occur in the
midwest, as well as in a small area from southeastern Pennsylvania to central New Jersey. The lowest
concentrations (i.e., <2.0 jig/m3) occur in the upper northeast and southern periphery, and at three
complex terrain sites in North Carolina, Kentucky, and West Virginia. As for HNOa, and to a lesser
extent S02, spatial variability of total-NO3 near the Appalachian Mountains may be due to terrain
influences. The peak concentrations in the midwest correspond well with peak NO3, suggesting that the
latter can be a significant reservoir species. Given the likely differences in deposition rates, conversion of
HNO3 to the particle phase should result in higher time-averaged concentrations.
Total NO3 concentrations for 1991 and 1992 are generally similar to those discussed for 1990
(see Figures 20 and 21). Average values above 4.0 jig/m3 occur only in the midwest and at sites in New
Jersey (144) and southeastern Pennsylvania (128). The lowest concentrations among eastern sites recur
in New England (Sites 109 and 135), Kentucky (Site 121), and North Carolina (Site 137). Among
western sites, a factor of three range in concentration persists, with the lowest annual average values in
central Colorado (Site 161) and the highest in northern Arizona (Site 174).
4.7.2 Quarterly and Annual Variability of Atmospheric Concentrations
This section describes the variability of atmospheric concentrations between seasons and
between years. For purposes of completeness and representativeness, a quarter/season has been treated
as valid if it contains 10 valid samples, and a year has been treated as valid if it contains all four valid
seasons.
35
-------
U)
Figure 17. Mean annual HNO3 concentrations (/xg/m3) for 1991.
-------
U)
Figure 18. Mean annual HNO3 concentrations (^g/m3) for 1992.
-------
U)
00
Figure 19. Mean annual total NO^ concentrations (^g/m3) for 1990.
-------
U)
vO
Figure 20. Mean annual total NOj concentrations (^ig/m3) for 1991.
-------
3.60
Figure 21. Mean annual total NOj concentrations (jig/m3) for 1992.
-------
Time-series of quarterly average concentrations for five eastern sites and three western sites are
shown in Figures 22 through 27. Eastern sites were selected to illustrate variability in the upper
northeast (Site 105), northeast (Site 106), southeast (Sites 120 and 153), and midwest (Site 122) dating
from the origin of the network. Western sites were selected to represent a transect from the United
States-Canada border to the United States-Mexico border. Despite large concentration differences,
seasonal patterns appear to be fairly similar among sites and reproducible between years. SO$", NHJ,
and HNO3 concentrations typically peak during the summer, and SO2 and NOj concentrations peak
during the winter at all sites. For example, SOj' data show an increase of a factor of about three from
winter (first quarter) to summer (third quarter) followed by a plummet to the lowest level of the year in
the fall (fourth quarter). SO2, in contrast, exhibits peak concentrations in fall and winter and
dramatically lower concentrations in spring and summer. Total-nitrate concentrations remain essentially
constant from season to season at all eastern sites, due to the mirror image seasonal variations of HN03
and NOj. Little evidence of an upward or downward trend can be seen in any of the time-series plots.
Time-series for the western sites also show pronounced seasonality. Although the data record is
brief at this point, fairly reproducible year-to-year peaks and valleys in concentrations are observed. The
southern Arizona site (167), which appears to be influenced by nearby SO2 sources, exhibits significant
variability, especially in SOj" and SO2. Sites on the periphery of the eastern United States (e.g., Maine,
Arkansas, and Florida) generally show reduced seasonal variability.
Annual average concentrations for sulfur and nitrogen species are listed in Tables 6 and 7 for all
eastern sites, all western sites, and groups of eastern sites from 1989 through 1992. Inspection of data
shows that year-to-year variability is less than 10 percent for all site groups. Despite relatively low
variability from year to year, there is evidence of a significant (10 to 20 percent) decline in
k
concentrations for all species in all regions, except the west. The data record is obviously insufficient to
tell if this is part of a secular trend or merely natural variation (e.g., meteorological variability) in
annual concentrations.
Data for SO*' show roughly a 16-percent decrease from 1989 to 1992 for 41 eastern sites and
essentially no change for nine western sites. The largest regional decreases were on die order of
20 percent and occurred in the midwest and upper midwest. Results for HNO3 and NOj show year-to-
year variability similar to SO* and apparent reductions of about 15 percent over the period of record.
Similar time series for total NOj concentrations show that year-to-year variability is not simply due to
changes in gas-particle partitioning but to lower overall concentrations of atmospheric nitrates. The
most significant change in concentration for the period occurred for SO2. In this case, average
concentrations for eastern and western sites decreased by about 25 percent from 1989 to 1991. All
41
-------
*-
NJ
H'*
eg
to
to
o
9
(u
I
o
>t>
(/>
8.
c/i
»*^*
9
t
4,
Concentration (ug/m3)
oo-i.-j.rorv)
cv-
*•*»• ^^
i
-------
u>
to
w
en
n>
81
o
O
w >
!
I
CO
IT
S 2
^
n ""*"
O> I
*^J ^^r
! 1
i i
Concentratbn (ug/m3)
p p p p p
o to 4*. b> bo
I *
8 Q?-
Qr
<*
Q?
Q?-
CD
CO
'
Concentration (ug/m3)
O —i. I\D GO
Q?-
Qr-J
Q?
Q?
Q?-
-------
f
N)
re
0)
1
z
ac
o
•t
M
re
Cu
8
Concentration (ug/m3)
oPPPPppp
-------
Cn
3
to
en
•
CD
I
i.
I.
!
1
fr
o
£
re
t
07
Qy-
07
Q?
Q}
Concentration (ug/m3)
o -»• r>o co
CD
O)
O
I
4
1
4
1
4
1
4
T
i
4
Concentration (ug/m3)
o S 8 8
o
o
-------
00
I2
1
0
Nitric Acid
f-1988-» «-1989-» 4-1990-+ 4-1991-> 4-1992-*
Site: 060105 &&A106 *HH*120 see 122
1<51 Nitric Acid
CO
1.0H
0.5
0.0
4-1989-> 4-1990-» 4-1991 -> <-1992->
Site: 909164 &&A167 ^MMC 168
Figure 26. Seasonal variability of HNO3 (/xg/m3) for selected sites.
46
-------
Total Nitrate
<-1988-» <-1989-»
Site: 960105
<-1992->
*«»~*12D aea122
2-°l Total Nitrate
1.5
1.0
0.5
0.0
«-1988-> <-1989-» <-1990-* 4-1991 -» <-1992-»
Site: 098164
Figure 27. Seasonal variability of total NOj
for selected sites.
47
-------
Table 6. Annual Average Concentrations Cng/m3) of sc£, NOj, AND NHj, 1989-1992
*»
00
soj-
Sub-
region
Northeast
Upper Northeast
Midwest
Upper Midwest
South Central
Southern Periphery
East
West
No. of
Sites
11
3
9
3
12
3
41
9
1989
6.4
3.0
6.8
4.2
6.4
5.0
6.0
0.89
1990
6.1
2.8
6.1
3.5
5.9
4.8
5.5
0.83
1991
6.2
2.6
6.0
3.4
5.7
4.0
5.4
0.81
1992
5.6
2.5
5.3
3.2
5.4
4.0
5.0
0.86
1989
1.0
0.3
2.9
1.8
0.7
0.6
1.3
0.28
NOj
1990
0.8
0.2
2.3
1.7
0.6
0.4
1.1
0.25
1991
0.7
0.2
2.4
1.6
0.6
0.5
1.1
0.28
1992
0.8
0.2
2.4
1.4
0.6
0.5
1.1
0.26
1989
2.0
0.8
2.9
1.8
1.9
1.4
2.0
0.36
NHj
1990
1.8
0.8
2.4
1.5
1.7
1.2
1.8
0.31
1991
1.9
0.7
2.4
1.5
1.7
1.2
1.8
0.31
1992
1.7
0.6
2.3
1.4
1.6
1.2
1.7
0.32
Source: ESE.
-------
Table 7. Annual Average Concentrations (|ig/m3) of HN03, SO2, and Total NOj, 1989-1992
HNO3
Sub-
region
Northeast
Upper Northeast
Midwest
Upper Midwest
South Central
Southern Periphery
East
West
No. of
Sites
11
3
9
3
12
3
41
9
1989
2.6
1.0
2.6
1.6
2.2
1.4
2.1
0.60
1990
2.6
1.0
2.4
1.4
2.1
1.4
2.1
0.50
1991
2.5
0.9
2.4
1.3
1.8
1.2
1.9
0.45
1992
2.2
0.7
2.2
1.3
1.9
1.2
1.8
0.48
1989
15.9
3.2
14.0
5.2
7.8
3.4
10.5
0.77
S02
1990
13.6
2.8
12.7
4.6
7.4
2.6
9.3
0.72
1991
13.3'
2.4
12.3
3.9
6.1
2.4
8.7
0.65
1992
11.8
1.9
10.8
3.6
t
6.0
2.6
7.9
0.55
1989
3.5
1.3
5.4
3.3
2.8
1.9
3.4
0.88
Total
1990
3.3
1.2
4.7
3.1
2.6
1.8
3.1
0.74
NOj
1991
3.2
1.1
4.7
2.8
2.4
1.7
3.0
0.72
1992
2.9
0.9
4.6
2.7
2.3
1.7
2.9
0.71
Source: ESE.
-------
regions of the network exhibited reductions in observed SO2 ranging from 40 percent in the upper
midwest to 23 percent in the midwest. Absolute changes in S02 concentrations averaged -2.5 ug/m3
across all 41 eastern sites.
The observed changes in concentrations over time appear to be real and not the result of
changes in field or analytical protocols. The only significant change in field sampling protocols occurred
in January 1990, when sample collection intervals were modified from day/night to around the clock.
This modification is not a significant factor in the apparent trend because concentration changes occur
from year to year and not as a step function. The only significant change in laboratory protocols
occurred in January 1991. This involved the addition of hydrogen peroxide to filter extracts to promote
oxidation of residual SO2. The expected impact of this new protocol is an increase in apparent SO2 by
<5 percent. In most cases, SO2 values decreased by about 5 percent from 1991 to 1992. Concentration
reductions thus appear to be the result of changes in emissions and/or transport and not the result in
changes in sampling or analytical protocol.
4.1.3 Ammonium Versus Sulfate
Data for the 50 NDDN sites suggest regional variability in aerosol speciation as well as
concentration. This is illustrated in Figure 28, which shows the relationship between NH^ and SO*
(molar basis) for each NDDN site using annual average concentrations for 1991 as an example. Similar
relationships were observed for other years. Solid and dashed lines are also plotted in this figure to
depict 1:1 and 2:1 ratios of NHJ to SO*, respectively. The 2:1 line represents completely neutralized
SO*, and the 1:1 line represents 50-percent neutralization of SO2;, assuming that only NH^ and SO2; are
present in the aerosol phase. Other aerosol species are undoubtedly present and, therefore, ratios
represent only approximate levels of neutralization.
Results show that the majority of eastern sites fall between the 50-percent and 100-percent
neutralization lines and that the western sites scatter around the 100-percent neutralization line.
Inspection of Figure 28 also shows a small number of eastern sites on or above the 100-percent
neutralization line and a small number of sites only slightly above the 50-percent neutralization line.
The highest NH^SO2" ratios all correspond to agricultural sites in the midwest and upper midwest. The
lowest ratios of NH^SO2", in contrast, all correspond to sites in predominately forested areas of the
southeast and northeast.
Although filter pack data cannot be used to quantify aerosol acidity, these results suggest broad
qualitative differences in aerosol acidity. For agricultural sites in general, there appears to be sufficient
NH} to completely neutralize SO2;. For forested sites at significant distances from agricultural activity,
50
-------
-------
observed SO2, must be balanced by other cations in addition to NHJ. The metal cations Na+, K+, Ca2+,
and Mg2"1" were measured in NDDN filter pack samples during 1989 (ESE, 1990c). Results showed that
Na+, K+, Ca2+, and Mg2* were minor aerosol constituents at all sites except those in the midwest or near
the coast CSites 104, 116, and 156). On the whole, these data show that aerosol composition varies from
region to region and suggest that aerosol acidity may be greater at predominantly forested sites then at
predominantly agricultural sites.
4.1.4 Calculated Dry Deposition
Dry deposition calculations for the NDDN sites are made using a multi-layer version of the NOAA
inferential model, which has been modified to address multiple vegetation species (see Clarke and
Edgerton, 1992). As described in Chapter 3, the meteorological variables necessary to determine R,, R,,,
and R£ are obtained from the 10-m meteorological tower at each of the sites, normally located in a
clearing over grass or other low vegetative surface. Vegetation activity parameters (minimal stomotal
resistance; light response coefficient; and optimum, minimum, and maximum growing temperatures) are
assigned to each vegetation species based on published values, primarily from plant physiology studies.
These data were provided by the authors of the inferential model.
The calculation procedure used for the NDDN is based on area weighting of dry deposition to
the individual land-use/vegetation types. Vd and flux are calculated by species and then area-weighted
within 1 km of the site to represent average values. Mixed vegetation (e.g., mixed coniferous and
deciduous forest) is disaggregated into individual species prior to calculating deposition such that
deposition to individual species can be calculated by the model.
The three resistance terms are calculated for each species and vegetation/surface type every hour
when meteorological data are available. The inverse sum of the resistances is the Vd for the specific
species and vegetation/surface type, and the site averaged Vd is the areal-weighted average Vd over all
vegetation types. Hourly values of Vd are then averaged over a week and multiplied by the weekly
integrated concentrations to produce weekly fluxes of HNO3, SO2, NOj, and SO2.
This section presents results of dry deposition calculations for sulfur and nitrogen species for
1990, 1991, and 1992. Weekly patterns of Vds and fluxes are also discussed to illustrate the effects of
canopy condition and meteorology on atmospheric deposition. The deposition of SO2" and SO2 is
expressed in units of kilograms sulfur per hectare (kg-S/ha) to facilitate comparison of gas and particle
phase deposition and, ultimately, comparison with wet deposition.
52
-------
Deposition of HNO3 and NO3 is expressed in units of kilograms nitrogen per hectare (kg-N/ha),
also for comparison purposes. Unlike sulfur deposition, however, HNO3 and NO3 cannot be considered
the primary contributors to total nitrogen. Additional potentially significant contributors include
reduced nitrogen species (e.g., NHJ and NHj), organic nitrates (e.g., peroxyacetylnitrate), and other
oxidized nitrogen species (e.g., nitrogen dioxide and nitrous acid). For this reason, deposition of HNO3
and NO3 has been discussed in terms of nitrate-nitrogen rather than total nitrogen deposition.
4.1.4.1 Sulfur Fluxes
Calculated 1990, 1991, and 1992 fluxes of SOj, SOj , and total sulfur are shown in Figures 29
through 37. Annual fluxes of SO2 for eastern sites range from about 0.7 kg-S/ha in north Florida to
approximately 9.0 kg-S/ha in northern Illinois, southern Indiana, and western Pennsylvania. The overall
pattern of SO2 deposition strongly reflects that of atmospheric concentrations. Fluxes are substantially
lower for sites in complex terrain than for neighboring sites, due to a combination of low concentration
and low Vd. Lindberg et al. (1988) reported similar variability in wet deposition between low- and high-
elevation forests of the Southern Appalachians. McMillen (1990) showed that terrain-induced variability
in meteorological parameters could have a significant impact on dry deposition estimates. Observations
from the NDDN not only confirm previous studies but show that short-range variability in concentrations
and meteorology complicate estimation of dry deposition in complex terrain (see Chapter 5).
Annual SO2 fluxes for western sites are generally well below 1.0 kg-S/ha and well below
minimal values observed for eastern sites. This reflects extremely low SO2 concentrations at most
western sites. The lone exception to this is the site in southern Arizona (Site 167). Deposition at this
site is on the order of that observed at the more remote eastern sites and is the result of elevated local
SO2 concentrations.
Calculated dry deposition for SO*' is shown in Figures 29 through 31. Deposition rates for
eastern sites range from about 0.2 kg-S/ha in northern Maine (Site 135) to about 1.0 kg-S/ha in western
North Carolina (Site 126). The three mountaintop sites in Virginia and North Carolina exhibit die
highest SOj" fluxes. This finding may be due to enhanced V^s at high elevations, since SOj
concentrations at mountaintop sites are not atypical of other eastern sites. Deposition among western
sites ranges from 0.06 kg-S/ha in northern Montana to 0.28 kg-S/ha in southern Arizona. As noted
earlier, relatively high deposition rates in southern Arizona are due to locally high concentrations
Calculated dry deposition of total sulfur (i.e., SO2 plus SOj ) exhibits a pattern which is fairly
similar to SO2l indicating that gaseous deposition dominates the dry deposition process (see Figures 35
through 37). The dominance of SO2 deposition is shown in Figure 38, which illustrates the percent of
53
-------
en
Figure 29. Calculated dry deposition for SO2 (kg-S/ha) for 1990.
-------
Ui
Figure 30. Calculated dry deposition for SO2 (kg-S/ha) for 1991.
-------
Figure 31. Calculated dry deposition for SO2 (kg-S/ha) for 1992.
-------
01
Figure 32. Calculated dry deposition for SO*' (kg-S/ha) for 1990.
-------
00
Figure 33. Calculated dry deposition for SO*" (kg-S/ha) for 1991
-------
01
V0
Figure 34. Calculated dry deposition for SO*' (kg-S/ha) for 1992.
-------
Figure 35. Calculated dry deposition for total sulfur (kg-S/ha) for 1990.
-------
Figure 36. Calculated dry deposition for total sulfur (kg-S/ha) for 1991.
-------
ON
N>
Figure 37. Calculated dry deposition for total sulfur (kg-S/ha) for 1992.
-------
Ul
Figure 38. Percent gaseous sulfur deposition for 1991.
-------
total dry sulfur deposition attributable to SO2) using 1991 data as an example. Data for eastern sites
show that SO2 accounts for 59 to 93 percent of dry sulfur deposition. In general, the highest
percentages (i.e., >85 percent) occur in an area extending from Illinois to southern New York. Not
suprisingly, the lowest percentages occur around the network periphery, where SO2 concentrations are
low. Data for western sites show lower percentages of SO2 deposition, reflecting low S02 concentrations
throughout the region. At no site, however, does SO2 account for less than 50 percent of dry sulfur
deposition.
The week-to-week variability of SO2 Vds and fluxes are shown for three sites in Figures 39 and
40. These three sites represent a range in land-use characteristics and terrain in Indiana, Pennsylvania,
and North Carolina. The Indiana site is surrounded by mixed grass and crops in flat terrain. The
Pennsylvania site is agricultural in rolling terrain and the North Carolina site is in forested complex
terrain. Weekly deposition velocities exhibit a pronounced summertime peak at all three sites, with the
highest Vds ranging from 0.4 to 0.5 centimeters per second (cm/sec). The period of peak Vd is
considerably shorter at the agricultural site in Pennsylvania than the other sites, presumably due to the
short growing season for corn and soybeans relative to native grasses (Indiana) and trees (North
Carolina). Vds are a factor of 3 to 5 lower in winter and fall than summer but still show appreciable
differences between sites. In this case, low Vds for the Pennsylvania site reflect large surface resistances
of bare soil. Relatively high Vds for the Indiana site are the consequence of high windspeeds and low
atmospheric resistance.
Weekly flux data for the three sites show a summertime peak for the Indiana and Pennsylvania
sites but little weekly variability for North Carolina site (Figure 40). The seasonal pattern of SO2
concentrations runs counter to that of Vd, and, therefore, the product of the two exhibits much less
seasonally than either one by itself.
4.1.4.2 HN03 and NO3 Fluies
Calculated annual fluxes of HNO3 for 1990, 1991, and 1992 range from 0.7 to 5.0 kg-N/ha for
eastern sites and 0.4 to 1.3 kg-N/ha for western sites (see Figures 41 through 43). Spatial variability is
high across the United States, reflecting large ranges in Vd and concentration. As noted for SO2, sites in
valleys experience more frequent and persistent nocturnal inversions than sites in rolling terrain or on
ridges. The overall effect of such inversions is to suppress vertical exchange and thus to reduce fluxes of
gases and particles to the surface. Inversions may also reduce frequency of plume impaction at sites
located in valleys.
64
-------
IS)
cm/sec
0.60 -
0.50-
0.40 -
0.30 -
0.20 -
0. 10 -
D D O Coweeto, NC (SUc 137)
Penn State, PA (Stte 106)
Solomonfe, IN (Site 133)
0.00 -
90343
^ I ^
91078
91178
Week Start Dates (yyddd)
91278
92013
Figure 39. Weekly deposition velocities for SO2 for selected sites.
-------
kg/ha
0. 70 -
0.60 -
0.50 -
0.40 -
0.30 -
0.20 -
0. 10 -
0.00 -
Coweeto, NC (Site 137)
Penn Stele, PA (Site 106)
Solomonie, IN (Site 133)
90343
9107B
91178
Week Start Dates (yyddd)
91278
92013
Figure 40. Weekly SO2 fluxes for selected sites.
-------
Figure 41. Calculated dry deposition of HNO3 (kg-N/ha) for 1990.
-------
00
Figure 42. Calculated dry deposition of HNO3 (kg-N/ha) for 1991.
-------
Figure 43. Calculated dry deposition of HNO3 (kg-N/ha) for 1992.
-------
Annual fluxes for NO3 (Figures 44 through 46) range from 0.01 kg-N/ha in northern Maine and
northern Montana (Sites 135 and 168) to 0.2 kg-N/ha in northern Illinois and northern Indiana
(Sites 133 and 146). In general, NO3 deposition is similar between eastern and western sites and,
provided the former are not located near intensive agricultural activities, and is inconsequential relative
to HNO3 deposition. This is illustrated in Figures 47 through 49, which show total nitrate-nitrogen dry
deposition, and Figure 50, which shows the percentage contribution of HNO3 to dry nitrate-nitrogen
deposition. Gaseous deposition accounts for at least 91 percent, and usually more than 95 percent, of
nitrate-nitrogen dry deposition at NDDN sites. Unlike sulfur deposition, the percentage of nitrate-
nitrogen does not appear to differ between eastern and western sites.
Weekly HNO3 Vd and flux data for three sites in Indiana, Pennsylvania, and North Carolina are
shown in Figures 51 and 52. Deposition velocities exhibit only moderate seasonality, but were generally
highest during spring and winter (i.e., periods of highest winds). The presence of actively growing
vegetation appears to exert little influence over Vds for HNO3. Fluxes of HNO3 were generally similar at
Pennsylvania and Indiana sites but much lower at the North Carolina site. This is the result of very low
concentrations and low Vd in complex terrain.
The relationship between atmospheric concentration and dry deposition of SO2 and HNO3 is
illustrated in Figure 53. Each data point represents the intersection of annual average concentration and
annual deposition for a site, using data for 1991. Despite the complexity of the dry deposition process,
data for SO2 are fairly consistent from site to site. This, in turn, suggests that variability in deposition is
largely driven by variability in concentration. This appears to hold for western and eastern sites.
Data for HNO3 exhibit markedly greater scatter from site to site. For example, a large range in
deposition is calculated for sites with annual average concentrations of 2 to 3 |ig/m3 of HNO3. This can
only occur due to variability in both Vd and concentration. Closer inspection of data in Figure 53 shows
some evidence of geographical clustering of sites. Three sites in flat terrain (i.e., 133, 130, and 157) fall
significantly below and three mountaintop sites fall significantly above the general flux-concentration
relationship. The three sites in flat terrain are located in extensive cropland areas of Illinois and Indiana
while the three mountaintop sites are in forested ecosystems of Virginia and North Carolina. This
contrast in terrain and vegetation suggests that differences in atmospheric turbulence controls much of
the site-to-site variation of HN03 deposition.
Clarke and Edgerton (1992) assessed the overall uncertainty of dry deposition calculations using
the inferential model. Seasonal and annual values of calculated fluxes for sites located in flat terrain
with uniform vegetation and not influenced by nearby sources are probably accurate within about 25 to
70
-------
Figure 44. Calculated dry deposition of NOj (kg-N/ha) for 1990.
-------
Figure 45. Calculated dry deposition of NOj (kg-N/ha) for 1991.
-------
Figure 46. Calculated dry deposition of NOj (kg-N/ha) for 1992.
-------
Figure 47. Calculated dry deposition of total NOj (kg-N/ha) for 1990.
-------
Ul
Figure 48. Calculated dry deposition of total NOj (kg-N/ha) for 1991.
-------
Figure 49. Calculated dry deposition of total NOj (kg-N/ha) for 1992.
-------
Figure 50. Percent gaseous NOj deposition for 1991.
-------
cm/sec
4 .00 -
3.00 -
2.00 -
0.00 -
90343
9107B
91178
Week Start Dates (yyddd)
ODD Coweeto, NC (Stte 137)
Penn State, PA (Site 106)
Selomonle, IN (Site 133)
91278
^ T
92013
Figure 51. Weekly deposition velocities for HNO3 (cm/sec) for selected sites.
-------
Kg/Ha
0. 90 -
0.80 -
0.70 :
0.60 -
0.50 -
0. 40 -
0.30 -
0.20 -
0. 10 -j
0.00 :
DOB Coweeto, NC (Site 137)
Penn Stole, PA (Site 106)
Solomonic, IN (Site 133)
90343
91078
91178
Week Start Dates (yyddd)
i
91278
92013
Figure 52. Weekly HNO3 fluxes for selected sites.
-------
8
6
0
126125
117
146 140
113
128
no
114
12I|B
133
T19 157
0
5 10 15
S02 (ug/m3)
20
51
4
3-
0
0
126
105
157
110
118
1 2 3
HN03 (ug/m3)
128
Figure 53. Scattergrams of SO2 flux (kg-S/ha) versus concentration and HNO3 flux
(kg-N/ha) versus concentration for 1991.
80
-------
40 percent, depending on the chemical species. Dry deposition fluxes for SO2 are likely to be more
accurate than those for HNO3 and participates which are strongly dependent on atmospheric turbulence.
As terrain and vegetation complexity increases, uncertainty also increases, and fluxes, especially those for
HNO3 and particulates, are likely to be undercalculated.
Inferential model flux calculations for the NDDN were generally biased low due to the weekly
integrated sampling protocol, especially for SO2 and HNO3 during the summer season. The bias
appeared to be site specific and caused by a correlation between Vd and concentration on hourly and
diurnal time scales. Sulfur dioxide concentrations from the network may be biased low (10 to
15 percent), which will also contribute to an undercalculation of flux for that species. Duplicate NDDN
sites showed good precision. Deposition fluxes from duplicate sets of NDDN equipment (i.e., located side
by side) were typically within 5 to 10 percent for the major species.
4.2 Wet Deposition
The primary objective of the NDDN is the calculation of dry deposition rates for various sulfur
and nitrogen species. This will permit (1) examination of patterns and trends of dry deposition over
space and time, (2) comparison with wet deposition rates, (3) calculation of total deposition rates, and
(4) estimation of mass budgets for various regions and subregions of the United States and North
America. This section presents wet deposition observations for SO* and NO3 for 16 NDDN sites in the
eastern and western United States. Ultimately, these data should be combined with other networks (e.g.,
NADP/NTN and CAPMoN) to develop continental-scale patterns of wet deposition and total deposition.
Annual wet deposition of SO*" and NO3 for the eastern NDDN sites is illustrated in Figures 54
through 59. Data are expressed in units of kg-S/ha or kg-N/ha to permit straightforward comparison
between wet and dry deposition. Although a convenient bookkeeping approach, these units may or may
not be appropriate for determining ecological effects of deposited species.
Wet deposition of SOj" at NDDN sites varied by more than an order of magnitude during all
3 years. The highest deposition (> 10 kg-S/ha) occurred in Ohio and West Virginia, and the lowest
(about 1.0 kg-S/ha) occurred in northern Utah and central Colorado. Average depositions across the
upper midwest, upper northeast, south central, and southern periphery subregions are approximately
one-half to two-thirds of peak values. The overall deposition pattern is strongly reminiscent of
atmospheric SO* concentrations.
Wet NO3 deposition for 1990 to 1992 ranged from about 4 kg-N/ha in central Ohio and West
Virginia to less than 1 kg-N/ha for the three western sites. Although deposition amounts for NO3 are
81
-------
oo
Figure 54. Wet deposition of SC£ (kg-S/ha) for 1990.
-------
oo
to
Figure 55. Wet deposition of SO^' (kg-S/ha) for 1991
-------
00
Figure 56. Wet deposition of SO* (kg-S/ha) for 1992
-------
00
Ul
Figure 57. Wet deposition of NOj (kg-N/ha) for 1990.
-------
00
Figure 58. Wet deposition of NOj (kg-N/ha) for 1991.
-------
oo
Figure 59. Wet deposition of NOj (kg-N/ha) for 1992.
-------
only about half those for SO*", they exhibit similar spatial distributions. Deposition amounts in excess of
3 kg-N/ha cover a large area from northern Wisconsin to southwestern North Carolina. Inclusion of all
wet deposition sites in the eastern United States (and Canada) could modify deposition patterns
somewhat. By the same token, overall deposition patterns may differ from year to year, due to
precipitation variability and other factors.
4.3 Total Deposition
This section provides estimates of wet plus dry deposition (i.e., total deposition) and the
percentage as dry deposition at NDDN sites which collect both wet and dry deposition data. Total sulfur
deposition is defined as the sum of wet SO* deposition plus dry SO2 and SO*" deposition, expressed in
units of kg-S/ha. Total nitrate-nitrogen deposition is defined as wet NOj deposition plus dry HNO3 and
NOj deposition, expressed as units of kg-N/ha. As discussed earlier, estimates of dry deposition are
uncertain by about ±30 to 50 percent and biased low by 5 to 20 percent, depending on chemical species
and site characteristics (Clarke and Edgerton, 1992). These uncertainties propagate directly into total
deposition estimates. The uncertainty of wet deposition data is probably on the order of 10 to
20 percent (James, 1993). It is also worth noting that total deposition, as defined above, does not
include other deposition processes, such as cloud and fog deposition. These processes are not likely to
be important at existing NDDN sites but could be significant at high elevations and in areas of frequent
or persistent fog. Caution is therefore advised in extrapolating total deposition data to locations not
represented by the NDDN.
Estimates of total sulfur deposition and total nitrate-nitrogen deposition for 1990, 1991, and
1992 are shown in Tables 8 through 10 for 16 NDDN sites. Results show that total deposition of sulfur
ranges from about 1 kg-S/ha in central Colorado (Site 161) to 20 kg-S/ha in central Ohio (Site 114).
Both eastern and western regions exhibit substantial ranges in depositions which is largely due to
variability in SOZ concentrations (see Section 4.1.4). The contribution of dry sulfur deposition to total
sulfur deposition ranges from less than 25 percent for sites in Wisconsin (134), northern Florida (156),
and central Colorado (161) to more than 40 percent for sites in Virginia (108), Ohio (114), and
Pennsylvania (128).
Among eastern sites, it appears that the percentage of dry sulfur deposition decreases from the center of
the region to the periphery (e.g., Florida and Wisconsin).
Estimates of total nitrate-nitrogen deposition range from about 1 kg-N/ha in central Colorado
(Site 161) to about 8 kg-N/ha in northern Ohio (Site 123) and eastern Pennsylvania (Site 128) and are
fairly consistent from 1990 to 1991. As a rule, differences between eastern and western sites are not as
88
-------
Table 8. Total Sulfur and Nitrate-Nitrogen Deposition for NDDN Sites Measuring Wet and Dry
Deposition for 1990
Site
108
111
114
115
119
123
124
125
126
128
134
156
157
161
162
167
Total Sulfur
kg/ha
8.3
13.5
19.7
11.4
,17.3
17.6
9.1
INS
11.9
17.2
6.9
5.4
12.5
1.4
INS
3.5
CS)
%Dry
46
29
42
37
31
36
33
INS
29
44
21
24
39
20
INS
43
Nitrate-Nitrogen
kg/ha
4.6
5.2
7.7
6.6
6.1
8.0
4.5
INS
6.9
8.5
4.3
3.0
4.6
1.4
INS
2.3
fN)
%Dry
60
43
45
46
27
43
38
INS
56
50
31
48
35
41
INS
55
Note: INS = insufficient data.
Total sulfur deposition = wet SOj -S + dry SO^ -S + dry SO2-S.
Nitrate-nitrogen deposition = wet NOg-N + dry NOg-N + dry HNO3-N.
Source: ESE.
89
-------
Table 9. Total Sulfur and Nitrate-Nitrogen Deposition for NDDN Sites Measuring Wet and Dry
Deposition for 1991
Site
108
111
114
115
119
123
124
125
126
128
134
156
157
161
162
167
Total
kg/ha
9.5
13.0
15.8
11.3
15.3
13.0
10.0
7.9
11.5
16.1
6.4
6.8
11.8
1.0
1.3
3.1
Sulfur fS)
%Dry
39
25
43
35
31
46
28
35
25
51
18
19
40
21
25
43
Nitrate-Nitrogen
kg/ha
4.9
4.9
7.8
6.4
5.6
7.0
5.4
4.7
6.1
8.0
4.1
3.5
4.1
1.1
1.7
1.9
fisn
%Dry
50
34
47
41
24
51
28
48
49
56
26
30
38
39
48
57
Note: Total sulfur deposition = wet SO^-S + dry SO^-S + dry SO2-S.
Nitrate-nitrogen deposition = wet NOg-N + dry NOj-N + dry HNO3-N.
Source: ESE.
90
-------
Table 10. Total Sulfur and Nitrate-Nitrogen Deposition for NDDN Sites Measuring Wet and Dry
Deposition for 1992
Site
108
110
111
114
115
119
123
124
125
126
128
134
156
157
161
162
167
Total Sulfur
kg/ha
9.2
15.7
9.7
14.3
10.7
15.3
16.0
9.6
8.7
12.8
15.5
6.2
7.3
11.1
1.0
1.2
2.8
fS)
%Dry
41
40
35
41
34
31
37
26
34
23
44
22
25
39
23
27
40
Nitrate-Nitrogen
kg/ha
4.7
8.7
4.0
6.7
6.7
5.4
7.5
5.0
4.8
6.3
7.2
3.9
3.3
3.5
1.0
1.6
1.8
(N)
%Dry
55
51
47
48
48
24
48
30
52
51
48
40
43
41
49
57
60
Note: Total sulfur deposition = wet SOj -S + dry SOj -S + dry SO2-S.
Nitrate-nitrogen deposition = wet NOj-N + dry NOj-N + dry HNO3-N.
Source: ESE.
91
-------
great as for sulfur deposition. Dry deposition of nitrate-nitrogen ranges from approximately 30 percent,
or less, of total deposition in central West Virginia (Site 119) to 50 percent, or more, in central Virginia
(Site 108), eastern Pennsylvania (Site 128), and southern Arizona (Site 167). As discussed earlier, much
of the variability in nitrate-nitrogen deposition appears to be due to the effects of complex terrain in and
around the Appalachian Mountains. Terrain effects are discussed further in Chapter 5.
Comparison of results for sulfur and nitrate-nitrogen show that dry deposition generally
contributes a greater percentage of the latter. This may be explained, in part, by substantially higher
(factor of 4) Vd for HNO3 than for SO2. Another reason for this could involve network design and site
selection. The network is designed to provide regionally representative deposition data, and, therefore,
sites are selected which are outside the immediate influence of pollution sources, including sources of
SO2. The total sulfur deposition and the percentage dry deposition are likely to decrease with increasing
distance from sources of SO2. As a secondary pollutant, HNO3 is much less sensitive to proximity of
precursors than SO2.
4.4 Ozone
As described in Section 3.0, continuous O3 concentrations were monitored throughout the period
at a height of 10 m at all NDDN sites. Annual averages, valid observations, and peak observed
concentrations for 1990, 1991, and 1992 are summarized in Tables 11 through 13. Annual averages
among eastern sites ranged from about 25 to 50 ppb for all three years. The highest annual averages
(i.e., 45 to 50 ppb) occur at mountaintop sites along the Blue Ridge and Appalachian Mountains, while
the lowest annual averages (i.e., 20 to 25 ppb) occur in sites located in sharp valleys (e.g., Sites 119 and
121) and in semiurban areas (e.g., Sites 116 and 146). Hourly average concentrations equal to or
greater than the NAAQS were relatively rare for all years. In 1990, eight sites exhibited one or more
hourly values .>120 ppb, and only 15 days of exceedances occurred at the 48 sites operational
throughout the year. During 1991 and 1992, 9 sites and 5 sites recorded O3 values .>120 ppb,
respectively, and there were a total of 14 days and 7 days with exceedances, respectively.
For the 3-year period, as a whole, average O3 concentrations fell by approximately 3 ppb over
the eastern United States. Reductions were not consistent from site to site or year to year. Southeastern
sites exhibited about a 2-ppb change from 1990 to 1991 and again for 1991 to 1992. Northeastern sites
exhibited a 2-ppb increase from 1990 to 1991 and then a 7-ppb decrease from 1991 to 1992. This year-
to-year variability is similar to that reported earlier for SO*, SO2) and HNO3; however, insufficient data
are available to determine whether the similarity is simply fortuitous or the result of linkage between
chemical and meteorological processes.
92
-------
Table 11. Summary of NDDN Ozone Maxima for 1990
Highest
Site
No.
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129/131
130
133
134
135
137
HO
144
146
149
150
151
152
153
156
157
161
162
163
164
165
167
168
169
174
Valid
-------
Table 12. Summary of NDDN Ozone Maxima for 1991
Highest
Site
No.
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
130
131
133
134
135
137
140
144
146
149
150
151
152
153
156
157
161
162
163
164
165
167
168
169
174
Valid
(n)
8638
8605
8518
8663
8244
8632
8685
8550
8624
8524
8399
8297
8288
8483
8345
8614
8563
8595
8636
8532
8314
8599
8576
8503
8604
7916
8319
8627
8594
8602
8569
8664
8634
8338
8545
8641
8226
8301
8643
7837
8606
8317
8390
8386
7852
7768
7878
8642
7605
8566
Percent
valid
99
98
97
99
94
99
99
98
98
97
96
95
95
97
95
98
98
98
99
97
95
98
98
97
98
90
95
98
98
98
98
99
99
95
98
99
94
95
99
89
98
95
96
96
90
89
90
99
87
98
Annual
Average (ppb)
29.3
37.0
33.4
30.4
33.2
30.7
39.5
28.6
37.6
34.0
32.4
31.5
26.6
31.0
46.0
26.0
44.0
21.7
32.9
34.2
33.7
35.6
44.9
35.6
38.3
31.0
36.4
33.3
33.4
31.6
26.5
30.0
30.5
24.5
33.6
24.7
35.7
33.6
34.7
30.4
30.6
45.3
46.2
38.1
43.1
47.0
41.9
23.9
48.0
45.9
Value
(Ppb)
149
122
117
99
108
98
109
108
112
116
107
101
156
118
114
111
108
96
113
120
111
118
95
98
122
145
109
107
79
99
86
112
143
132
106
87
110
99
123
90
117
86
88
73
91
76
83
62
81
81
Date
07/19
06/27
07/18
08/03
06/15
05/24
06/28
08/03
06/27
07/18
06/10
05/16
07/20
08/30
06/21
07/22
10/30
08/03
06/21
06/14
06/14
10/31
04/18
06/03
07/19
06/14
08/03
06/20
05/10
08/17
04/18
06/19
07/19
06/01
07/20
08/01
07/27
07/15
09/13
05/09
08/03
05/15
05/18
06/06
07/19
05/09
05/23
05/23
05/20
05/18
Hour
16
01
15
13
17
16
03
14
16
12
16
17
16
17
21
14
20
14
15
15
19
16
20
15
17
13
15
17
19
23
17
16
15
16
14
15
15
16
17
12
13
18
18
07
17
24
20
13
14
16
Second Highest
Value
(ppb)
114
111
116
98
103
88
105
100
108
115
106
101
137
110
100
111
99
90
113
115
103
102
92
93
121
116
100
106
79
90
86
111
141
118
105
86
101
95
118
85
112
83
82
73
88
74
82
61
79
79
Date
06/15
06/10
07/17
06/09
08/17
06/26
06/27
08/17
05/16
07/17
06/21
05/20
07/19
06/10
05/16
08/08
05/16
05/15
07/18
06/03
05/22
07/22
03/21
09/14
06/28
07/01
08/01
08/16
08/27
08/18
05/03
08/24
05/23
06/20
05/15
06/18
08/12
06/05
08/05
08/21
08/16
05/02
05/15
08/22
08/05
05/19
05/24
05/20
06/14
05/10
Hour
16
20
17
18
15
15
15
16
19
14
15
18
16
16
20
15
01
14
17
19
18
17
14
16
15
23
14
16
16
01
18
17
16
18
19
15
18
17
18
14
14
19
10
18
06
10
09
15
15
14
Third Highest
Value
(ppb)
113
109
110
95
99
86
99
96
105
113
104
101
132
106
98
106
97
88
112
113
101
86
90
89
117
116
95
99
74
88
82
104
135
117
105
85
88
92
115
83
110
78
78
67
86
74
79
61
78
79
Date
07/17
07/19
07/19
06/27
09/16
06/27
07/17
07/23
06/28
06/27
08/07
06/25
06/21
07/18
06/10
06/10
05/15
10/31
07/20
05/15
05/21
08/08
11/01
08/04
06/21
07/02
08/02
07/18
05/14
06/11
03/26
06/09
08/08
06/18
07/19
08/24
08/14
09/15
06/29
10/12
09/12
05/13
06/11
05/23
08/04
05/20
05/04
06/03
07/14
09/07
Hour
18
23
12
18
16
14
11
14
02
16
16
15
14
14
13
12
24
16
20
19
17
16
11
14
17
02
•13
18
18
03
17
17
18
17
21
17
15
14
16
15
17
11
18
19
23
14
17
19
19
14
Source: ESE.
94
-------
Table 13. Summary of NODN Ozone Maxima for 1992
Highest
Site
No.
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
130
131
133
134
135
137
140
144
146
149
150
151
152
153
156
157
161
162
163
164
165
167
168
169
174
Valid
8304
8595
8675
8405
8509
8536
8715
8543
8635
8538
8553
8383
8043
8658
8056
8604
8631
8503
8687
8666
8549
8602
8638
8655
8566
8503
8595
8695
8351
8642
8592
8614
8486
8631
8608
8604
8608
8365
8687
8480
8471
8250
8569
8425
7592
8360
8101
8498
8139
8144
Percent
valid
95
98
99
96
97
97
99
97
98
97
97
95
92
99
92
98
98
97
99
99
97
98
98
99
98
97
98
99
95
98
98
98
97
98
98
98
98
95
99
97
96
94
98
96
86
95
92
97
93
93
Annual
Average (ppb)
25.0
34.1
30.4
27.4
30.9
29.7
35.5
26.7
32.8
28.5
29.0
27.9
23.7
26.6
39.5
22.6
40.3
19.4
29.1
31.5
30.1
35.8
42.7
30.7
32.0
26.3
35.3
29.9
32.3
31.5
26.9
27.3
26.1
21.5
31.3
24.2
34.2
32.5
32.9
30.8
29.7
43.5
45.2
37.8
42.1
46.0
40.1
20.9
47.2
46.0
Value
(PPb)
144
115
104
89
123
100
102
96
106
110
98
126
118
96
99
91
92
100
100
104
98
102
97
99
105
98
97
103
95
93
93
103
139
131
96
88
90
101
117
88
106
73
78
73
79
72
76
77
74
81
Date
05/23
05/23
06/30
05/23
07/14
05/23
05/22
05/12
06/29
06/30
05/12
05/21
07/12
05/23
05/23
05/23
05/23
05/12
06/16
05/22
08/23
07/09
05/23
05/09
05/23
06/30
05/10
06/30
06/04
05/23
05/10
06/29
08/24
06/13
05/22
04/09
05/30
05/11
05/23
03/27
08/18
04/30
05/27
05/05
12/06
04/30
05/18
05/17
04/30
04/28
Hour
18
03
16
14
18
10
23
14
19
20
16
17
15
16
14
15
17
13
19
17
17
16
19
17
18
13
15
17
13
16
13
18
16
16
22
14
18
16
17
20
16
14
15
20
07
10
15
16
16
04
Second Highest
Value
(PPb)
112
111
102
81
97
94
98
90
97
101
97
116
117
93
96
90
85
87
98
103
97
102
87
88
95
90
87
100
89
89
85
99
125
99
92
88
88
97
107
84
104
72
76
72
74
72
76
76
73
79
Date
08/27
05/22
05/23
05/13
07/21
06/14
05/23
05/11
06/30
06/29
06/06
06/30
07/20
05/22
05/22
07/29
05/12
05/11
05/16
06/12
06/17
07/13
05/10
05/10
07/20
06/29
08/10
06/16
06/13
05/21
05/12
05/16
06/30
06/29
05/20
10/05
05/09
05/15
05/24
06/22
06/30
05/11
06/05
05/16
07/16
05/06
06/03
05/22
10/02
06/28
Hour
15
22
13
14
16
13
01
18
11
18
17
18
16
17
14
16
13
12
15
18
16
16
17
15
17
13
15
17
14
24
12
15
16
14
18
15
17
18
16
18
13
16
17
18
23
13
24
16
22
08
Third Highest
Value
(PPb)
107
109
98
79
93
89
98
90
92
101
94
110
113
92
95
85
82
85
98
99
96
101
85
88
94
89
86
95
83
87
83
96
122
99
92
84
84
97
106
82
97
71
75
71
72
70
71
69
71
78
Date
08/24
06/13
06/13
07/30
05/23
05/30
06/30
06/23
05/23
07/01
06/15
07/01
07/01
05/13
07/02
06/15
05/22
05/10
06/29
05/11
05/11
05/23
06/24
06/19
05/11
05/09
05/23
06/12
05/10
05/22
04/09
06/30
06/29
07/01
06/03
05/09
05/31
06/29
06/24
03/01
06/29
04/29
04/27
07/04
06/09
08/21
04/29
04/11
08/20
06/26
Hour
15
24
16
13
17
16
11
16
17
14
15
17
15
15
22
14
17
16
19
19
18
15
08
20
17
14
15
17
18
01
18
16
16
13
17
13
14
14
14
18
16
16
19
18
13
17
15
16
17
11
Source: ESE.
95
-------
Data for western sites showed average O3 concentrations ranging from 25 ppb in northern
Montana (Site 168) to 47 ppb in northern Arizona (Site 174). In general, while average concentrations
exceeded those in the east, peak concentrations were almost invariably less than those observed across
the eastern part of the network. The average elevation for the western sites is well above the highest
elevation of any eastern site. Relatively high annual averages among western sites are consistent with
similarly high averages for the eastern sites located on mountaintops.
Differences in diurnal O3 cycles appear to offer partial explanation for the differences in
concentrations between sites on mountaintops, in valleys, and near urban areas. Hour-by-hour average
O3 concentrations for 1990, for a rolling terrain site (108), a complex terrain site (119), a mountaintop
site (120), a semiurban site (116), and two western sites (168 and 164) are plotted in Figure 60.
Results are typical of all 3 years and show generally consistent differences between types of sites for the
annual period. The rolling terrain site exhibits moderate day/night variability, with nocturnal minima
typically on the order of 50 to 60 percent of the day maxima. Day maxima show a fairly broad plateau
between the hours of 1400 through 1700. Hourly averages for the complex terrain site shows markedly
different behavior. In this case, nocturnal minima are much less than half the daily maxima. In
addition, periods of maximum concentration are appreciably shorter than at rolling terrain sites, and
average maximum concentrations are slightly lower than at rolling terrain sites. The mountaintop site
exhibits a unique diurnal pattern, or lack of one. For this, and other mountaintop sites, there is no
distinct period of maximum or minimum concentration. Rather, hourly average values consistently
remain within a few parts per billion of hourly maxima observed at the other sites. Results for the
semiurban site show markedly depressed nocturnal values and slightly depressed daytime values, relative
to rolling terrain sites.
One explanation for the observed behavior involves the relationship between sampler location
and the nocturnal inversion layer. The mountaintop sites presumably sit above the nocturnal inversion
layer and, therefore, are always in contact with a large reservoir for O3. The rolling terrain and complex
terrain sites are situated below the inversion layer, within which O3 is subject to a variety of depletion
processes. Longer lasting and/or shallower inversions could result in rapid decay of O3 in complex
terrain. The result of this day/night variability is a gradient in average concentrations from mountaintop
to rolling terrain to complex terrain. For the semiurban sites, broad nocturnal minima could be the
result of destruction by nitric oxide (NO), while sharp day maxima could be due to enhanced
photochemical production in the presence of NO, and natural or manmade volatile organic gases.
As mentioned previously, relatively few O3 concentrations greater than or equal to 120 ppb were
observed during the 3-year period. Other measures of integrated O3 exposure or concentration, however,
96
-------
VO
60
50
^40
o.
at
820
0
10
0
| Sifes: «MK)8 OOO 1 1
.0
'•..
-••o
>0°°0oo0
00001111
246802461
Hour
60
50
"S*
d.
05
820
o
10
0
9j 60
50
* . ^40
_ -0
' • °-
* • • -2*0
O Q)
o o on
o _. N ^U
°00 0
10
n
1222
3024
| Sites: •••168 000174
,00000000000000000
.*••*••
•
•
•
* * • • • • *
| Sites: •••116 OOO120 |
0 0 0 o
0°000 ° ° • • • °00
°°oO°.- -.
• •
• •
00001 1 1 1 1222
246802468024
Hour
a
°oo0oo
0
•
"••.
000011111222
246802468024
Hour
Figure 60. Hourly average O3 concentrations for typical rolling terrain
(Site 108), complex terrain (Site 119), mountaintop (Site 120), semi-
urban (Site 116), and western (Sites 168 and 174) sites.
-------
suggest marked differences from year to year. The sigmoidally weighted O3 concentration CW126) is a
statistic proposed by Lefohn and Runeckles (1987) as a tool for examining O3 damage to forests and
crops. This S-shaped function weights O3 concentrations in a manner that emphasizes high values (e.g.,
>80 ppb) and deemphasizes low values (e.g., <30 ppb). Thus, concentrations believed to be more
harmful to crops are given greater weight in the averaging scheme than low concentrations.
Calculations of W126 for 1990 through 1992 are presented in Figures 61 through 63.
Calculated W126 values for 1990 range from less than 20 ppm-hr in northern Maine and
Wisconsin to approximately 100 ppm-hr in northern Virginia. Not suprisingly, sites with the highest
values include mountaintop locations along the spine of the central and southern Appalachian
Mountains. These sites exhibit limited diurnal variations and would be expected to experience among
the highest O3 exposures among rural sites. Typical values elsewhere are on the order of 50 to
70 ppm-hr, except for two sites in northern Alabama and northern Georgia, which exceeded 90 ppm-hr.
Reasons for these isolated high values are unknown.
W126 data for 1991 are generally similar to 1990, except for slightly higher values in
Pennsylvania and New York and dramatically lower (50 percent) values at the two sites in Georgia and
Alabama. Data for 1992 show substantially lower W126 values for a large number of sites compared to
1990 and 1991. Peak values in 1992 were on the order of 60 ppm-hr for the three mountaintop sites in
Virginia and North Carolina. These findings show that year-to-year variability in exposure statistics can
be substantial, even where measures of central tendency (e.g., averages) and extremes (e.g.,
concentrations >.120 ppb) show limited variability.
Similar variability is observed in other statistics used to estimate O3 exposure. The SUM60
statistic, for example, has been related to O3 damage in agricultural crops and has been suggested as a
potential model for a secondary O3 standard (Lefohn et al., 1988). By definition, SUM60 is the sum of
all O3 concentrations .>60 ppb over a 3-month period (typically June through August) which is equated
to the growing season for many crops. SUM60 calculations for eastern NDDN sites for 1990 through
1992 are displayed in Figures 64 through 66. The overall pattern of SUM60 for 1990 is very similar to
that for W126. The highest exposures occur along the Appalachian Mountains from Pennsylvania to
Georgia. The lowest exposures occur in Maine, Wisconsin, and eastern Kentucky. Generally similar
patterns and somewhat to significantly lower values are observed in 1991 and 1992.
98
-------
Figure 61. W126 (ppm-hr) for eastern sites for 1990.
99
-------
Figure 62. W126 (ppm-hr) for eastern sites for 1991.
100
-------
Figure 63. W126 (ppm-hr) for eastern sites for 1992.
101
-------
Figure 64. SUM60 (ppm-hr) for eastern sites for 1990.
102
-------
Figure 65. SUM60 (ppm-hr) for eastern sites for 1991.
103
-------
Figure 66. SUM60 (ppm-hr) for eastern sites for 1992.
104
-------
Chapters
Special Studies
Two special studies were conducted during 1990 to 1992 to provide information on the
comparability and representativeness of NDDN atmospheric concentration data. The first of these studies
involved side-by-side collection of air samples using the NDDN and CAPMoN filter pack systems. The
purpose of this experiment was to provide data on the comparability of United States and Canadian air
quality trends databases. Information of this nature will prove useful in establishing uncertainty limits
for spatial patterns across eastern North America.
The second study involved simultaneous collection of filter pack and meteorological data at two
closely spaced sites in the southern Appalachians. One site was located on a mid-elevation ridge and the
other (Site 137) was located about 1.0 km away in a valley approximately 300 m below the ridge. The
objective of this study was to determine if concentrations and deposition velocities of acidic species were
affected by complex terrain. This chapter presents initial findings from these two studies.
5.1 Comparability of United States and Canadian Atmospheric Concentration Data
Late in 1990, NDDN filter pack sampling was initiated at Egbert; Ontario, a Canadian acid
deposition trends site. The objective of this sampling effort was to establish a database for tracking the
comparability of data that will be used by the two countries to evaluate trends in atmospheric
concentration and dry deposition. Although both NDDN and CAPMoN use filter pack approaches for
measuring atmospheric concentrations, several notable differences exist between programs. First, NDDN
samples weekly, and CAPMoN samples daily. NDDN samples are thus exposed to a broader range of
environmental conditions (e.g., temperature and relative humidity) over the 7-day sampling interval and
may, therefore, be more susceptible to various artifact reactions. Second, NDDN and CAPMoN samples
operate at substantially different flow rates (i.e., 1.5 L/min versus 25 I/min). Despite a shorter
collection interval, the CAPMoN filter pack samples about 2.5 times the amount of air than the NDDN
filter pack. This difference raises the question of gas-particle interactions, due to build-up of material on
filters and other effects. Finally, differences exist in laboratory and data management operations. Both
laboratories use 1C for analysis of filter extracts; however, quality control protocols differ, and analytical
accuracy is traceable to different standards. Data management procedures mandate slightly different
approaches for calculating sample volume and, in the case of CAPMoN, blank correction of filter loadings
prior to calculation of atmospheric concentrations.
105
-------
Validated 1991 CAPMoN data for Egbert were received from Mr. Robert Vet of the Canadian
Atmospheric Environment Service (AES) in the form of daily concentrations referenced to 273 degrees
Kelvin (K) and 1 atmosphere pressure. These data were aggregated into weekly composite
concentrations, corresponding to the start/end dates of NDDN samples, and then scaled to a reference
temperature of 298 K (i.e., the NDDN reference temperature). The weekly aggregated CAPMoN data
were then used, for purposes of this comparison, only if seven valid daily samples were available to be
composited. In all, 47 pairs of valid weekly samples met selection criteria.
Results of side-by-side collection of CAPMoN and NDDN samples are shown in Figures 67
through 69 and summarized in Table 14. Data for SO^" exhibit a very high correlation coefficient
(r2 = 0.99) and show no evidence of bias, either for the year as a whole or for extended periods (i.e.,
about 3 months) between equipment calibrations. Over all samples, mean concentrations differ by less
than 1 percent. Median absolute differences (MADs) and median absolute percent differences (MAPDs)
are 0.1 (ig/m3 and 2.7 percent, respectively. These data indicate that no substantial biases occur in the
collection and analysis of samples and manipulation of data by the two networks. This is not surprising
since atmospheric SO^" is relatively straightforward to measure; however, it shows that the two networks
were in good control of measurement systems during 1991.
Results for NH^ are generally similar to SO^~, while those for NO3 and HNO3 exhibit
considerable scatter. Annual average concentrations of NO3 and HNO3 are well within 5 percent, but
correlation coefficients are only 0.86 and 0.90, respectively. Inspection of time series shows extended
periods during which one sampler is higher than the other. In most cases, the pattern for NOj runs
counter to that for HNO3. For example, the NDDN appears to underestimate HNO3 and overestimate
NO3 during the spring and summer months. The opposite is true during the fall and winter. Large
MAPDs compared to overall percent differences for NO3 and HNO3 suggest that there may be
compensating effects which cancel out if averaging periods are sufficiently long. This is supported, to
some extent, by total NO3 data. In general, results for total NO3 show that the two samplers collect the
essentially same amount of material, even though they may partition the particle and gas phases
somewhat differently.
Data for SO2 exhibit a distinct and persistent bias between samplers. The overall mean
concentrations differ by 10 percent and the MAPD (a measure of the typical difference between samples)
is nearly 15 percent. Differences cannot be explained by errors in sample flow rates or analytical biases
because these would also be evident for SO^" and perhaps other species. McNaughton and Bowne
(1993) recently reviewed numerous intercomparisons of atmospheric sampling methods. In general,
results showed that filter pack methods were biased low relative to annular denuder and continuous
106
-------
o
**j9,i~
n w i
zS. V-
fD
w 3
i-t- Cu
" 2
3 2
Is
o
%;-
*%/-
V
%
^-
Concentration (ug/m3)
ro ^ o> oo
Concentration (ug/m3)
0=^±
8:
I
***,.
n
f?
§
w <
-------
o
oo
£
ON
00
s rt
H
n
a s
w
v_/
o
n>
o
IT)
Concentration (ug/m3)
00
IV)
Concentration (ug/m3)
O -*. N> CO 4*
CXI
to
O
z
-------
TJ
o
VO
28.
§,8
- §
f
n
O
5s
o
Concentration (ug/m3)
00
ro
Concentration (ug/m3)
o ->• r>o co ^
en
o
u
-------
Table 14. Comparison of 1991 Atmospheric Concentration Data for CAPMoN and NDDN at Egbert,
Ontario
Statistic
Mean NDDN (jig/m3)
Mean CAPMoN (jig/m3)
MAD (iig/m3)
MAPD (%)
%D
Slope
Intercept (jig/m3)
r2
so|-
4.08
4.09
0.12
2.7
< 1
0.96
0.19
0.99
NOj
2.47
2.42
0.26
13.2
2.0
0.98
0.00
0.86
NHj
1.80
1.80
0.08
5.4
< 1
0.98
0.03
0.96
Species
S02
5.46
6.05
0.65
14.8
10.3
1.04
0.38
0.96
HNO3
1.43
1.38
0.12
10.1
3.6
0.72
0.35
0.90
Total NOg
3.94
3.79
0.29
7.8
3.9
1.06
-0.36
0.97
Note: MAD = median absolute difference between sample pairs.
MAPD = median absolute percent difference between sample pairs.
% D = percent difference between mean concentrations.
methods and that the magnitude of the bias varied inversely to flow rate through the filter pack. Further
studies are needed to elucidate the mechanism for such biases.
In summary, results for 1991 show excellent comparability between NDDN and CAPMoN for all
species, except SO2, when averaged over the year. Seasonal and possibly temperature-dependent biases
occur for NOj and HNO3, but not for total NOg, suggesting that the two samplers observe slightly
different phase distributions for certain species. Based on these results, CAPMoN and NDDN data are
generally comparable, except for SO2, and it would appear that data sets can be merged for analysis of
annual concentration patterns across eastern North America. For SO2, it may be necessary to adjust
results for the observed bias to avoid a transboundary step function.
5.2 Elevational Gradients in Concentration and Fluxes
Data presented earlier show that spatial variability of concentrations and deposition is relatively
high in and around the Appalachian Mountains, especially for the reactive gases SO2 and HNO3. This
110
-------
variability is of concern because it complicates estimation of deposition to sensitive and protected
ecosystems as well as to mountainous areas in general.
In mid-1991, an auxiliary monitoring site was established at the Coweeta Hydrologic Laboratory
in southwestern North Carolina. The purpose of this site was to examine short-range variations in
concentration and dry deposition, specifically, the effect of elevation. Coweeta Hydrologic Laboratory
was selected for this study because of its remoteness from pollutant sources and because of the
accessibility to a ridgetop site.
Equipment was installed at an elevation of 850 m on a ridge located about 1.5 km northwest of
the routine NDDN site (elevation 550 m). Weekly filter packs and hourly meteorological data were
collected from mid-June through the end of the year. An ozone monitor was not installed because an
environmentally controlled shelter was not available.
Results of the 1991 study are shown in Figures 70 through 72 and summarized in Table 15.
Weekly concentration data for SO*' exhibit a very strong correlation between the ridge and valley sites,
with a definite tendency for higher concentrations at the ridge site. The strong correlation between sites
suggests that the two are generally exposed to the same air mass. The time series of SO2; concentrations
also shows some evidence of seasonality. Differences between ridge and valley are pronounced through
October, after which they agree essentially within the precision of the measurements. The transition
coincides with leaf senescence and leaf drop at Coweeta, suggesting that ridge/valley differences may be
the result of canopy interactions.
Data for SO2 and HNO3 also show substantial ridge/valley differences. For these species,
ridgetop concentrations are frequently 2 to 3 times higher than valley concentrations. Correlations
between the two sites are lower than for SO2; but still highly significant. Differences between ridge and
valley concentrations also persist through the end of the year (i.e., beyond leaf drop).
Ratios of ridge/valley concentrations of SO2, SO2, and HNO3 over the 6%-month study were 1.3,
2.4, and 2.6, respectively (see Table 15), and, in general, increase in order of calculated Vd. This
behavior may indicate that a variety of factors, including dry deposition, contribute to ridge/valley
differences.
Calculated Vds for the two sites show moderate enhancements for SO* and SO2 and a dramatic
enhancement for HNO3 at the ridgetop. This observation reflects species-dependent sensitivity to
windspeed, since this is the only model input variable to differ significantly from valley to ridge. As
111
-------
ZII
& ri
CM
">
3.
~
S
g
3.
O.
SB
g.
i
4*
VI
I
Concentration (ug/m3)
en
-------
en
f
re1
^4
O H
li
re c/>
5> n
a. o
0
3
o S
I
I
Q
Concentration (ug/m3)
ro
oo
QlS:
CQ CD
-------
T!
f
%
vj
to
R ri
PI
cj re
09 V)
&
3
R-8
r* 3
Is
II
5.8
i-»
rb TO
-'
3.
G.
1
Concentration (ug/m3)
-j. ro co 4^. en 05
Q
*
*
*
(D CD
-------
Table 15. Comparison of Ridge (R) and Valley (V) Data for Coweeta Hydrologic Laboratory, July
through December 1991
Concentration Cue/m3)
R
V
R/V
Deposition Velocity fcm/sec)
R
V
R/V
Flux Pcg/hal as S or N
R
V
R/V
SO*
6.04
4.75
1.3
0.08
0.07
1.1
0.25
0.17
1.4
Species
S02
3.44
1.41
2.4
0.37
0.31
1.2
1.00
0.34
2.9
HNO3
1.90
0.73
2.6
1.49
0.89
1.7
1.96
0.45
4.4
discussed previously, Vd for HNO3 is very sensitive to windspeed because surface resistances are assumed
to be negligible. Other species, such as SO^ and SO2, have significant surface resistances and, therefore,
are much less sensitive to windspeed effects.
Estimated fluxes for SO*, SO2, and HNO3 exhibit ridge/valley ratios of 1.4, 2.9, and 4.4, respectively,
indicating that strong deposition gradients occur over short distances in areas of complex terrain. These
findings support recent investigations by Lovett and Likens (1991), who have compared several
approaches for estimating atmospheric fluxes of sulfur and nitrogen to the Hubbard Brook Experimental
Forest. Results of these studies indicate that estimates of dry sulfur deposition can vary widely. For
example, flux estimates based on watershed mass balances were 2 to 3 times higher than estimates
based on model calculations. Input for the model calculations came from NDDN measurements at a
valley site, similar to Coweeta, located at the base of the watershed. Model output for the NDDN site
was then assumed to represent the entire watershed. Although much less direct than the observations
115
-------
at Coweeta, these findings suggest that elevational gradients might be widespread phenomena. If so, the
issue of site representativeness should include some consideration for terrain complexity. Additional
work is needed to determine scaling factors, so measurements at individual sites can be applied to large
areas of complex terrain. Certain aspects of this work are currently under development in the Large
Area Deposition model (McMillen, 1990).
116
-------
Chapter 6
Quality Assurance Summary
As described in Chapter 3, extensive quality control checks have been implemented, both in the field
and the laboratory, to ensure that data quality consistently meets program requirements. These checks
include frequent calibration of equipment with NIST-traceable standards (when available) to determine
accuracy and collocation of equipment at numerous sites to determine overall measurement precision.
This chapter summarizes accuracy and precision data for the CASTNet deposition network. Results are
presented for 1992 (a typical year) to exemplify data quality routinely provided by the network or an
annual timeframe. Readers interested in quality control data for different years, or for time periods of
less than 1 year, are referred to periodic data reports and summaries provide to EPA (e.g., ESE, 1990).
6.1 Laboratory Accuracy and Precision
The accuracy of laboratory measurements is determined by frequent analysis of independent
standards of known concentration. These standards are typically obtained directly from NIST or another
authoritative source (e.g., EPA). In practice, standards are analyzed by ESE at the beginning and end of
every laboratory batch (roughly 50 to 100 samples). The initial analysis is used to verify the analytical
calibration curve, while the final analysis is used to check for instrumental drift. Both results must be
within a specified range for the analytical batch to be accepted.
Laboratory precision is determined by the replicate analysis of at least one sample per laboratory
batch. In this case, a sample is randomly selected and re-analyzed at the end of the batch. With this
approach, precision data incorporate sample-to-sample variability and instrumental drift across the
observed range of sample concentrations.
It should be noted that laboratory accuracy represents only a portion (and frequently a small
portion) of the total uncertainty surrounding precipitation chemistry and air quality measurements.
Currently, no standard exists for directly assessing the accuracy of filter pack data and precipitation
chemistry data. Rather, accuracy must be inferred on the basis of standard laboratory measurements,
round-robin measurements of split samples and/or direct comparison with methods of known accuracy.
Results presented in this section represent only the first of these approaches. Work has been initiated to
obtain accuracy information with the other two approaches.
As with accuracy, laboratory precision represents only part of the overall precision of environmental
measurements. Unlike accuracy, however, there is an authoritative, or at least consensus, approach for
117
-------
determining the overall precision of precipitation chemistry and air quality data. This approach
(typically referred to as collocation) involves side-by-side collection of samples with independent systems.
Results of collocated sampling at a number of sites are presented in Section 6.2.
6.1.1 Laboratory QC for Filter Pack Measurements
Laboratory accuracy and precision data for 1992 filter pack measurements (filter extracts) are shown
in Table 16. In general, accuracies are well within the ±5 percent criterion established for filter pack
data. Concentrations of NOj ion exhibit a slight (about 1 percent) negative bias in both Teflon® and
nylon analyses, while SO^~ ion exhibits a slight (2 percent) positive bias in nylon and Whatman analyses.
These biases are unlikely to have a major effect on analysis of spatial patterns.
Precision data for filter pack analyses easily meet acceptance criteria for the program. Mean
replicate differences for all analytes were less than 2 percent and for Teflon-SO^" and nylon-NO^ were
less than 1 percent.
6.7.2 Laboratory QC for Precipitation Chemistry Measurements
Precision and accuracy data for laboratory analyses of ionic species in precipitation are summarized
in Table 17. As for filter pack measurements, results easily fall within acceptable ranges for most
analytes. Average recovery of anions (SO^", NO^, and Cl") was within ±2 percent. Interestingly, NO^
ion exhibits a similar negative bias as was observed in filter pack measurements. This likely reflects a
small, but persistent, systematic difference between calibration standards used in the ESE laboratory and
NIST references used to determine accuracy.
Cation measurements generally show a wider range of recoveries than anion measurements, but still
fall within the acceptance range of ±10 percent. Unfortunately, the species with the lowest overall
accuracy (i.e., largest absolute departure from 100 percent recovery) is the most important cation, H+
In this case, there is a positive bias of about 7 percent. This corresponds to a measurement error of
about 0.03 pH unit, which is only slightly larger than the uncertainty in the standards used to calibrate
the instrument.
Precision data for precipitation chemistry measurements also meet acceptance criteria for all species.
The precision of anion measurements (especially SO^" and NOp is generally greater than that for cation
measurements because anion concentrations are rarely, if ever, near the analytical detection limit.
Cation concentrations, except for H+ and NH^, are frequently within a factor of two of analytical
detection limits.
118
-------
Table 16. Summary of Laboratory Accuracy and Precision Data for Filter Pack Measurements, 1992
Filter Parameter
Teflon® SO£ SO^'
NOj
NHj
Nylon
NOj
Whatman SO^"
Accuracy
Mean
100.1
98.6
100.7
101.4
98.8
102.2
{% Recovery)*
Std. Dev.
1.2
1.3
2.1
1.7
1.6
1.0
Precision
Mean
0.4
1.4
1.6
1.4
0.8
1.7
f% Difference) +
Std. Dev.
0.7
2.8
3.2
2.4
1.0
3.7
*Accuracy based on analyses of NIST reference solutions.
+Precision based on within-run replicate analyses of filter extracts.
Source: ESE.
-------
Table 17. Summary of Laboratory Accuracy and Precision Data for Precipitation Chemistry
Measurements, 1992
Parameter
SO*
NOj
a
H+
NH;
Na+
K+
Ca2+
Mg*+
Accuracy
Mean
100.5
98.2
101.0
107.1
100.2
97.8
103.2
103.6
103.3
(% Recovery)*
Std. Dev.
1.0
0.9
2.0
3.0
2.0
3.7
3.6
4.2
3.4
Precision
Mean
0.6
0.5
3.1
3.8
2.8
5.3
3.6
2.0
1.7
C% Difference)*
Std. Dev.
0.6
0.4
4.5
5.2
2.8
7.1
4.4
2.6
2.6
*Accuracy based on analysis of NIST simulated rainwater.
+Precision based on within-run replicate analyses of rainwater samples.
120
-------
6.2 Results of Collocated Measurements
As mentioned previously, collocated measurements are a broadly accepted approach for
determining the overall precision of atmospheric measurements. Collocated filter pack measurements
and continuous ozone and meteorology measurements were conducted at a number of sites from 1990
through 1992. This section presents results for filter pack sampling at two sites (one eastern and one
western) and continuous measurements for one site. Collocated precipitation chemistry measurements
were not conducted during the reporting period, but have been subsequently initiated..
6.2.1 Filter Fade Data
Collocated filter pack data for 1992 are summarized in Table 18 and displayed in Figures 73 and
74. Inspection of summary statistics shows that annual average concentrations are nearly identical for
all species. For the eastern-United States sites, average SO^", NO3, NH^, and SO2 agree within ±2
percent, while HN03 and total NO3 agree within about ±6 percent. Median absolute differences (MADs)
reflect typical differences between a pair of samples and are less than 0.1 fig/m3 for SO^", NO3, and NH^
and less than 0.5 (ig/m3 for HNO3 and total NO3. In general, the collocated precision for sulfur species
(i.e., SO^" and SO2) and NH^ is better than that for the nitrate-nitrogen species, when expressed on a
percentage basis (i.e., MAPD). This observation may reflect artifact reactions and interconversions of
NO3 and HNO3 in the filter pack.
Data for Site 163, located in the western United States, also show good precision, despite very
low concentrations of all species. Median absolute differences between samples are j<0.02 jig/m3, and
median absolute percent differences (MAPDs) are less than 8 percent for all species. These findings
show that overall measurement precision does not vary greatly over the observed range of concentrations
in the eastern and western United States. As expected, they also show that laboratory precision is
typically a small component of the overall precision of atmospheric measurements.
Scattergrams and regression statistics for individual sample pairs show that agreement between
samplers is generally acceptable. Linear regression statistics indicate essentially 1:1 agreement for most
species and equally good agreement across the entire range of measurements. One exception to this is
HN03 for Site 128 (Arendtsville, PA), which shows a definite bias of about 0.2 |ig/m3 between samplers.
This cannot be attributed to flow differences, since other species (especially SO^") do not show the same
effect. Work is underway to investigate reasons for this behavior.
121
-------
Table 18. Summary of Collocated Filter Pack Data for 1992
Site 128 (n = 52)
X ((ig/m3)
Y (iig/m3)
D (%)
MAD (jig/m3)
MAPD (%)
Site 163 fn = 471
6.07
6.12
-0.8
0.06
1.3
Parameter
Statistic
SQ2- NO;
NH+
SO2
Total
HN03 NOj
1.81
1.83
-1.1
0.08
6.2
2.42
2.42
0.6
0.04
1.7
12.7
12.9
-2.0
0.43
4.6
2.83
3.01
-6.2
0.21
8.1
4.60
4.79
-4.1
0.22
5.6
X (ug/m3)
Y (ng/m3)
D (%)
MAD (u.g/m3)
MAPD (%)
0.72
0.74
-2.7
0.01
1.6
0.57
0.61
-6.8
0.02
7.3
0.36
0.38
-5.4
0.01
2.8
0.24
0.24
-3.2
0.01
6.9
0.39
0.39
-1.6
0.02
5.0
0.96
0.99
-3.1
0.02
3.4
Note: D = percent difference of quarterly means = 200 x (X - Y)/(X -t- Y).
MAD = median absolute difference.
MAPD = median absolute percent difference.
X = quarterly mean for primary sampler.
Y = quarterly mean for collocated sampler.
Source: ESE.
122
-------
16
12
oo
CM
-£
575
8
0
Sulfate (ug/m3)
— One-To-One Relationship
— Linear Regression
Slope
Y-lntercept
R-Squared
8
Site 128
12
0.98
0.17
1.00
16
1.5,
1.0
ro
to
CM
(
0.5
0.0
Sulfate (ug/m3)
0.0
— One—To—One Relationship
— Linear Regression
Slope
Y- Intercept
R-Squared
0.5
1.0
Site 163
1.01
0.00
0.98
1.5
40
30
CO
CM
CM 20
<75
10
Sulfur Dioxide (ug/m3)
10
•— One-To-One Relationship
— Linear Regresiion
Slope
Y-lntercept
R-Squared
0.98
0.45
0.99
20
Site 1 28
30
40
0.8
0.6
10
to
-------
10
8
oo 6
«
Si *
2
0
Nltratt (ug/m3)
•— One-To-One Relationship
— Linear Regression
Slop*
Y-lnfercept
R-Squared
1.02
-0.02
0.99
4 6
Site 128
10
ro
to
£
575
Nitrate (ug/m3)
One—To-One Relationship
— Linear Regression
Slope
Y-lnfercept
R-Squared
2
Site 163
1.03
o.oo
1.00
NJ
00
CM
C4
£
57»
Nitric Acid (ug/m3)
1.2,
One-To-One Relationship
— Linear Regression
Slope
Y-ln
-------
6.2.2 Continuous Measurements
Results of continuous (i.e., hourly) collocated measurements for O3, temperature, solar radiation,
and windspeed are illustrated in Figure 75. Data displayed in Figure 75 represent maxima daily
values for each day during 1992. Maximum values were chosen because these are typically of
greatest importance in determining ecological effects and deposition rates. As for filter pack data,
these results indicate overall precision which is well within CASTNet requirements. Results also
show that well-maintained instruments exhibit very little drift between periodic calibrations.
125
-------
eo
CM
120
100
80
60
40
20
0
Ozone(ppb)
Dally Maxima
"•• One-To-One Relationship
— Linear Regression
Slope
Y-lntereept
R-Squored
20 40 60 80
Site 128
1.03
-1.69
1.00
100 120
•— One-To-One Relationship
— Linear Regression
Slope
Y-lntercept
R-Squared
10
Site 1 28
20
t.oo
O.OS
1.00
30
NJ
ON
Dally Maxima
eo
CM
1000
800
600
400
200
0
(wa»ts/m2)
•— One-To-One Relationship
— Linear Regression
Slope
Y-lntercept
R-Squared
1.00
-0.26
1.00
200 400 600 800
Site 128
1000 1200
20
15
oo
CM
-------
References
Anderson, J.R., Hardy, E.E., Roach, J.T., and Witmer, R.E. 1978. A Land Use and Land Cover
Classification System for Use with Remote Sensor Data. U.S. Geological Survey Circular 671.
Geological Survey Professional Paper 964.
Chason, J., Huston, M., and Baldocchi, D. 1990. A Comparison of Direct and Indirect Methods for
Estimating Forest Canopy Leaf Area. Environmental Sciences Division. Oak Ridge National
Laboratory. Oak Ridge, TN.
Clark, J.F., and Edgerton, E.S. 1992. Dry Deposition Flux Calculations for the National Dry Deposition
Network. Prepared for the U.S. Environmental Protection Agency, Atmospheric Research and
Exposure Assessment Laboratory. Research Triangle Park, NC.
Environmental Science & Engineering, Inc. (ESE). 1990a. National Dry Deposition Network (NDDN)
Field Operations Manual. Prepared for U.S. Environmental Protection Agency (EPA). Contract
No. 68-02-4451. Gainesville, FL.
Environmental Science & Engineering, Inc. (ESE). 1990b. National Dry Deposition Network (NDDN)
Laboratory Operations Manual. Prepared for U.S. Environmental Protection Agency (EPA).
Contract No. 68-02-4451. Gainesville, FL.
Environmental Science & Engineering, Inc. (ESE). 1991a. National Dry Deposition Network (NDDN)
Data Management Manual. Prepared for U.S. Environmental Protection Agency (EPA). Contract
No. 68-02-4451. Gainesville, FL.
Hicks, B.B., Baldocchi, D.D., Hosker, R.P., Jr., Hutchison, B.A., McMillen, R.T., and Satterfield, L.C.
1985. On the Use of Monitored Air Concentrations to Infer Dry Deposition (1985), NOAA
Technical Memorandum ERL ARL-141. 66 pp.
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.
James, K.O.W. 1993. 1991 Quality Assurance Report, NADP/NTN Deposition Monitoring. Office of
Atmospheric Chemistry. Illinois State Water Survey. Champaign, IL.
Lefohn, A.S. and Runeckles, V.C. 1987. Establishing a Standard to Protect Vegetation-Ozone
Exposure/Dose Considerations. Atmos. Environ., 21:561-568.
Lefohn, A.S., Laurence, J.A., and Kohut, R.J. 1988. A Comparison of Indices that Describe the Relation
Between Exposure to Ozone and reduction in the Yield of Agricultural Crops. Atmos. Environ.,
22:1229-1240.
Li-Cor, Inc. 1989. LAI-2000 Plant Canopy Analyzer-Technical Information. LI-COR, Inc., 4421 Superior
Street, Lincoln, NE.
Lindberg, S.E., Silsbee, D., Schaefer, D.A., Owens, J.G., and Petty, W. 1988. A Comparison of
Atmospheric Exposure Conditions at Low- and High-Elevation Forests in Southern Appalachian
Mountain Range, pp. 321-344. In: M.H. Unsworth and D. Fowler, eds. Acid deposition at High
Elevation Sites, NATO ASI Series C:Vol.252, Mumer Publ., Boston.
127
-------
References (Continued)
McMillen, R.T. 1990. Estimating the Spatial Variability of Trace Gas Deposition Velocities. NOAA Tech.
Memo. ERL ARL-181, 37 pp.
Meyers, T.P. and Yuen, T.S. 1987. An Assessment of Averaging Strategies Associated with Day/Night
Sampling of Dry-Deposition Fluxes of SO2 and O3. J. Geophysic Res., 92, 6705-6712.
Pierce, L.L. and Running, S.W. 1988. Rapid Estimation of Coniferous Forest Leaf Area Index Using a
Portable Integrating Radiometer. Ecology 69(6) 1762-1767.
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.
Wesely, M.L. and Lesht, B.M. 1988. Comparison of RADM Dry Deposition Algorithm with a Site-Specific
Method for Inferring Dry Deposition. Water, Air, and Soil Pollution, 44, 273-293.
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Floor
Chicago, IL 60604-3590
128 *u-s- GOVERNMENT PRINTING OFFICE: 1995 - 650-006/22068
------- |