EPA/600/R-93/065
April 1993
DRY DEPOSITION FLUX CALCULATIONS FOR THE
NATIONAL DRY DEPOSITION NETWORK
by
John F. Clarke*
Atmospheric Characterization and Modeling Division
Atmospheric Research and
Exposure Assessment Laboratory
Research Triangle Park, NC 27711
and
Eric S. Edgerton
Environmental Science & Engineering, inc.
1000 Park Forty Plaza
Durham, NC 27713
Contract No. 6S-02-4451
Project Officer
Barry E. Martin
Exposure Assessment Research Division
Atmospheric Research and Exposure Assessment Laboratory
Research Triangle Park, NC 27711
*On Assignment from the National Oceanic and
Atmospheric Administration, U.S. Department of Commerce
ATMOSPHERIC RESEARCH AND EXPOSURE .ASSESSMENT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
TECHNICAL REPORT DATA
1. REPORT MO.
EPA/600/R-93/065
2.
3' PB93- 178242
*. TITLE AND SUBTITLE
Dry Deposition Flux Calculations for the National
Dry Deposition Network
5.REPORT DATE
April 1993
(.PERFORMING ORGANIZATION CODE
7. AUTSOR(S)
John F. Clarke (1)
Eric S. Edgerton (2)
6.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION MAKE AND ADDRESS
(1) A REAL, ACMD.USEP
RTP, NC 27711
(2) Environmental Science and Technology
Durham, NC
10.PROGRAM EL2KENT NO.
N104Q/D/10
11. CONTRACT/GRANT NO.
68-02-4451
12. SPONSORING AGENCY NAME AND ADDRESS
Atmospheric Research and Exposure Assessment
Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13.TYPE OF REPORT AND PERIOD COVERED
Project Report 1/91 - 9/92
1*. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
li. ABSTRACT
The National Dry Deposition Network (NDDN) was established in 1987 to document the
magnitude, spatial variability, and trends in dry deposition of ozone and acidic
particles and gases across the United States. Currently, the network consists of
50 stations: 41 in the eastern United States and 9 in the western United States.
The NDDN will be assimilated into the Clean Air Status and Trends Network
(CASTNet)s and the number of sites is expected to increase significantly over the
next several years. Dry deposition is not measured directly in the NDDN, but is
determined by an inferential approach, i.e., dry deposition fluxes are calculated
as the product of measured ambient concentration and modeled deposition velocity.
Chemical species include 03, sulfate, nitrate, sulfur dioxide, and nitric acid.
The temporal resolution for the dry deposition calculations is hourly for O3
and weekly for the other species. This report describes the dry deposition
calculation method used in the NDDN/CASTNet program and presents dry deposition
data for the network for 1990 and 1991. Sources of uncertainty in the calculations
are discussed.
17. XEY WORES AND DOCUMENT ANALYSIS
*. DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED TERMS
c.COSATI
18. DISTRIBUTION STATQ4ENT
RELEASE TO PUBLIC
19. SECURITY CLASS (^lis R#oort)
UNCLASSIFIED
21.NO. OF PAGES
98
20. SECURITY CLASS (This
UNCLASSIFIED
22. PRICE
-------
The information in this document has been funded by the U.S.
Environmental Protection Agency. It has been subjected to
the Agency's peer and administrative review, and it has been
approved for publication as an EPA document. Mention of
trade names or commercial products does not consritute
endorsement or recommendation for use.
-------
ABSTRACT
The National Dry Deposition Network (NDDN) was established in 1987 to document the
magnitude, spatial variability, and trends in dry deposition of ozone (03) and acidic particles
and gases across the United States. Currently, the network consists of 50 stations: 41 in the
eastern United States and 9 in the western United States. The NDDN will be assimilated into
the Clean Air Status and Trends Network (CASTNet), and the number of sites is expected to
increase significantly over the next several years. Dry deposition is not measured directly in the
NDDN but is determined by an inferential approach, i.e., dry deposition fluxes are calculated as
the product of measured ambient concentration and modeled deposition velocity. Chemical
species include 03, sulfate CSO^"), nitrate (NOj), sulfur dioxide (S02), and nitric acid (HN03).
The temporal resolution for the dry deposition calculations is hourly for 03 and weekly for the
other species. This report describes the dry deposition calculation method used in the
NDDN/CASTNet program and presents dry deposition data for the network for 1990 and 1991.
Sources of uncertainty in the calculations are discussed.
iii
-------
CONTENTS
Abstract iii
Figures v
Tables vii
Acknowledgements viii
1.0 Introduction 1
2.0 Summary and Conclusions 5
3.0 NDDN Overview 7
3.1 Network Description 7
3.2 Network Operations 11
3.3 Vegetation and Surface Condition Monitoring 12
4.0 Dry Deposition Calculations 19
4.1 Inferential Model 19
4.2 Big Leaf Dry Deposition Mode! 21
4 3 Multilayer Inferential Model 22
4.4 Large Area Deposition Model 22
5.0 Dry Deposition Data 25
5.1 Sulfur Dioxide 25
5.2 Nitric Acid 30
6.0 Uncertainty of Inferential Model Calculations 42
6.1 Bias and Uncertainty from Weekly Concentration Sampling Protocol 46
6.2 Calculation of Fluxes for Non-Ideal Sites 53
6.3 Meteorology and Land-Use Parameters 55
6.4 Accuracy and Precision of NDDN Atmospheric Concentration Measurements .... 57
6.5 Within and Between Network Precision 58
6.6 Summary of Accuracy and Precision of Dry Deposition Calculations 61
6.7 Spatial Variability of Dry Deposition 63
References 67
Appendices 69
Appendix A Data Processing and Output Options 70
Appendix B 1990 NDDN Data 78
Appendix C 1991 NDDN Data 83
iv
-------
FIGURES
Number Page
1 RADM predicted dry sulfur deposition as a percent of total sulfur deposition
for 1985 3
2 NDDN monitoring sites 8
3 Average number of days between 50% leafout and 50% leaf drop as
monitored by the NDDN 15
4 Land use and land cover classification map for NDDN Site 108 (Prince
Edward, VA) 16
5 1991 mean annual S02 concentrations (jig/m3) 26
6 1991 mean annual S02 deposition velocities (cm/sec) 28
7 1991 mean annual S02 fluxes (kg/ha) 29
8 Weekly S02 concentrations for 1991 for three sites in Indiana, Pennsylvania,
and Ohio 31
9 Weekly S02 dry deposition velocities for 1991 for three sites in Indiana,
Pennsylvania, and Ohio 32
10 Weekly S02 dry deposition fluxes for 1991 for three sites in Indiana.
Pennsylvania, and Ohio 33
11 1991 mean annual HN03 concentrations (ng/m3) 34
12 1991 mean annual HN03 deposition velocities (cm/sec) 36
13 1991 mean annual HN03 fluxes (kg/ha) 37
14 Weekly HN03 concentrations for 1991 for three sites in Indiana, Pennsylvania,
and Ohio 39
15 Weekly HN03 dry deposition velocities for 1991 for three sites in Indiana,
Pennsylvania, and Ohio 40
16 Weekly HN03 dry deposition fluxes for 1991 for three sites in Indiana,
Pennsylvania, and Ohio 41
17 Time series of hourly 03 deposition velocities measured (eddy correlation)
and modeled (Big Leaf) for Pennsylvania site 44
v
-------
FIGURES (continued)
Number Page
18 Comparison of measured and modeled 03 deposition velocities for
Pennsylvania site, shown as one diumal average 44
19 Comparison eddy correlation measurements and Big Leaf calculations
for S02 Vd in Huntington Forest 45
20 Ratios of Flux2/Fluxl for indicated chemical species, averaged over all
eastern sites, by month 48
21 Standard deviation of the ratio Flux2/Fluxl by month 49
22 The average ratio Flux2/Fluxl for S02 for groups of eastern sites SO
23 Locations of NDDN sites by groupings identified on Figure 22 52
24 Ratio of concentration measurements made at ridge-top site and the
NDDN site (Site 137) located in a nearby valley 64
vi
-------
TABLES
Table Page
1 NDDN site information 9
2 Precision and accuracy objectives of NDDN field measurements 13
3 Summary of 1990 and 1991 midsummer LAI measurements for the NDDN 17
4 Estimated uncertainty for short periods (30 to 60 minutes) of direct
measurements and inferential model calculations of dry deposition
velocities for a level site with homogeneous vegetation 42
5 Estimated bias of weekly flux calculations 51
6 Change in Vj resulting from setting Ra - 0 and Wet = 0 for
approximately 1 year of data 54
7 Results of 1990 collocated filter pack sampling 59
8 Precision analysis for duplicate NDDN' sites 60
9 Summary of uncertainty of dry deposition calculations 61
10 Deposition velocities for high and low elevation sites, Coweeta, NC 65
11 Deposition velocities (cm/sec) by plant class generated by the LAD Model
for Oak Ridge, 7N 66
vii
-------
ACKNOWLEDGEMENTS
The authors would like to thank Christopher Smith and Sara Haines of Computer Sciences
Corporation, Inc. for their role in implementing the dry deposition models on the EPA computer
system and for making the sensitivity and operational model runs used in this report. Special
thanks to Renee Lucas of Environmental Science & Engineering, Inc. (ESE) for ushering this
document through production.
VI!
-------
SECTION 1.0
INTRODUCTION
Lake acidification, forest damage, and other ecological effects of air pollution have created
interest in quantifying the deposition of acidic chemical species. Acid species are deposited to
the earth's surface through wet and dry processes. Deposition of pollutants by wet processes is
relatively easy to determine through analysis of precipitation samples. Precipitation samples
have been routinely analyzed for acidic species at many sites in North America since the late
1970s [National Atmospheric Deposition Program (NADP), 1990], and spatial patterns and
temporal trends have been characterized from these data.
Deposition of pollutants by dry processes is more difficult to measure; however, several direct
approaches have been developed and applied over the past decade. One approach (referred to as
the eddy correlation method) determines the flux of a pollutant to the surface based on the
correlation of highly resolved (several cycles per second) vertical wind velocity and concentration
fluctuations. The flux is calculated as w"C, where w" and C* are the instantaneous departures of
the vertical velocity and concentration from their mean values. The over bar indicates a time
average, usually 30 minutes to an hour. This approach has been applied in a number of
short-period (about 2 weeks) research measurement programs, primarily for ozone (03) and
sulfur dioxide (S02) (Matt et al., 1987; Meyers, 1992).
A second approach for direct flux measurements (the gradient method) requires very high
precision measurements of the mean concentration and selected meteorological variables at two
heights generally within 10 meters (m) above the top of the vegetation. The flux of materiai to
the surface is proportional to the measured gradients. This approach is usually applied to
chemical species that are highly reactive with the surface, e.g., nitric acid (HN03) and are
expected to exhibit significant vertical gradients. Recent applications include HN03 flux to a
forested area (Meyers et al., 1989) and to grass (Huebert and Robert, 1985).
The eddy correlation and gradient approaches provide fluxes temporally resolved to about 30
to 60 minutes. However, both approaches require a long fetch over level terrain with uniform
vegetation. These approaches require extensive instrumentation and technical personnel
1
-------
resources and consequently are expensive to implement for routine applications. Routine
monitoring using direct measurements of dry deposition has not been practical. Those
measurements that have been made are insufficient to define spatial patterns and trends.
Air quality models are frequently used to estimate patterns, trends, and magnitudes of
atmospheric trace species, as well as their transformation and removal by physical and chemical
processes. Figure 1 shows the spatial variability of dry sulfur deposition as a percent of total
deposition as predicted by the Regional Acid Deposition Model (RADM) (Dennis, 1992).
According to the RADM, dry deposition is a significant part of the total deposition budget,
exceeding 40 percent of total deposition budget in Illinois and northeastern Ohio, and 30
percent over much of the eastern United States.
Hicks et al. (1985) proposed the dry deposition inferential model as an alternative approach
to direct measurement of dry deposition for large network, operations extending for long periods
of time. The inferential model determines deposition fluxes as the product of a measured
concentration and modeled deposition velocity (Vd). Deposition velocity is calculated based on
an understanding of the physical and chemical processes of dry deposition as described through
measured meteorological and site variables. The inferential model has been used since 1984 in
a research network (i.e., the CORE network) coordinated by the National Oceanic and
Atmospheric Administration (NOAA). Annual and seasonal dry deposition for total sulfur and
total nitrogen "nave been reported for the nine sites in the CORE nerwork (Meyers et al., 1991).
The percentage of total sulfur and nitrogen deposition attributed to dry processes ranges from 25
and 31 percent, respectively, at White Face Mountain, NY, to 59 and 57 percent, respectively, at
Argonne, II. The CORE measurements and RADM results suggest that cry deposition is a
significant component of the total sulfur and nitrogen deposition budget.
The U.S. Environmental Protection Agency (EPA) established the National Dry Deposition
Network (NDON) in 1986 as a monitoring program to characterize dry deposition patterns and
trends. The NDDN is currently operated as a component of the Clean Air and Status Trends
Nerwork (CASTNet). Ambient pollutant concentrations, meteorological conditions, and land-use
data required for the inferential model are collected at the 50 sites that comprise the network
(Edgerton and Lavery, 1990). Recently, a program was initiated to calculate dry deposition
fluxes from these data. The dry deposition calculation program is described in this report.
2
-------
Figure 1. RADM predicted dry sulfur deposition as a percent o* total
sulfur deposition tor 1985.
3
-------
Section 3.0 describes the NDDN measurements, which include concentration and meteorological
variables, and land-use/vegetation characteristics. Section 4.0 provides a brief description of
three inferential models and their application to the NDDN/CASTNet sites, focusing on the Big
Leaf inferential model currently in use. Section 5.0 presents seasonal and annual S02 and HNOa
deposition data for the NDDN for 1991. Uncertainty and precision of the dry deposition data are
discussed in Section 6.0.
Three appendices are included in this report. Appendix A discusses model input data
requirements and model output options. Appendices B and C present seasonal and annual dry
deposition data (Vds, fluxes, and concentrations) for 03, S02, HN03, sulfate (SO|"), and nitrate
(NOj) for 1990 and 1991, respectively.
-------
SECTION 2.0
SUMMARY AND CONCLUSIONS
Quantitative dry deposition information is needed to characterize spatial patterns and trends
in total deposition and to support ecological effects, modeling, and chemical budget studies.
Fluxes are required for specific sites and also for large areas [e.g., 80-kilometer (km) grids]. An
inferential method for determining dry deposition as the product of a measured concentration
and modeled Vd can provide relevant dry deposition information. EPA has initiated a program to
calculate dry deposition fluxes for the NDDN sites and to assess sources of uncertainty. The first
phase of that program has been completed. Three dry deposition inferential models developed
by NOAA have been implemented at the Atmospheric Research and Environmental Assessment
Laboratory (AREAL): the Big Leaf site-specific model which is currently used for the NDDN dry
deposition calculations, a more advanced Multilayer model, and a Large Area Deposition (LAD)
model. Dry deposition fluxes have been calculated for all NDDN sites for 1990 and 1991 using
the Big Leaf inferential model. Results are included and discussed in this report.
Ambient concentrations and dry deposition fluxes for S02 and HN03 generally reflected
emission patterns. Highest concentration and largest fluxes were observed in the industrial
regions in the eastern United States. Both concentrations and fluxes were very low at western
NDDN sites.
An initial assessment of uncertainty was made in this study. 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 40 percent, depending on the chemical
species. Dry deposition fluxes for S09 and 0- are likely to be more accurate than those for
HN03 and particulates which are strongly dependent on atmospheric turbulence. As terrain and
vegetation complexity increases, uncertainty also increases, and fluxes, especially those for HNOj
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 S02 and :IN03 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. Additional measurements and studies are being conducted at
5
-------
AREAL that could lead to recommendations to modify the sampling protocol (e.g., day and night
weekly integrated filter packs, or continuous sampling for some species). Sulfur dioxide
concentrations from the NDDN may be biased low (10 to 15 percent), which also contributed 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.
The representativeness of flux calculations was uncertain, especially in areas of complex
terrain where measurements and calculations suggested HN03 flux can change by a factor of
four over short distances. The Large Area Deposition (LAD) model will help address spatial
variability of fluxes resulting from spatial variability of vegetation and terrain features.
Calculations using the LAD model suggested that Vds will vary by a factor of two as a function of
vegetation alone. While variability of this magnitude can be expected over a large area, the
difference between site-specific flux calculations and the LAD model are not expected to be that
large since the local site usually contains many species indigenous to the larger surrounding
area. The LAD model, however, cannot account for spatial variability of concentration, especially
in complex terrain.
The Multilayer inferential model will replace the Big Leaf inferential model in the
NDDN/CASTNet program when testing and modifications are completed early in 1993. It will
incorporate advances in the science and understanding of dry deposition processes resulting from
the 10-year National Acid Precipitation Assessment Program.
A program was initiated in 1992 to help define the uncertainty of the inferential model.
Direct measurements of dry deposition fluxes will be made in 1993 and 1994 at selected NDDN
sites. The goal of the program is to define model uncertainty at each site and as a function of
meteorology and vegetation, as well as provide a database for general model refinement.
6
-------
SECTION 3.0
NDDN OVERVIEW
3.1 NETWORK DESCRIPTION
The NDDN began operating in 1987 with six sites and expanded to 50 sites by July 1989.
The locations of the NDDN monitoring sites are shown in Figure 2. As of January 1989, 41
eastern sites and 9 western sites were operational. Most of the sites were selected to be
regionally representative and, consequently, were well removed from major population centers,
transportation corridors, and point sources of pollutants. Exceptions to this include sites near
Chicago, IL (146), Washington, DC (116), Sc. Louis, MO (157), and Evinston, IN (140), which
were established to assist model evaluation studies of the transport, transformation, and
deposition of acidic species.
Relevant site information, including latitude, longitude, initial reporting date, elevation,
terrain, and land-use classification of all operational sites since the inception of the NDDN is
presented in Table 1. Terrain and land-use classification refer to a 10-km radius around the site
and are presented to convey a sense of the site setting. Rolling terrain refers to sites with
average ground slopes between 5 and 10 degrees, while complex terrain refers to sites with
ground slopes greater than 10 degrees.
For discussion purposes, sites in the eastern United States have beer, grouped subjectively
into six subregions (see Figure 2). 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 land-use characteristics and general spatial patterns of atmospheric concentration
data. There may be considerable variability within subregions, and differences between
subregions may vary as a function of chemical species.
Sites in the upper northeast subregion are exclusively rural-forested, while those in the
northeast subregion exhibit a variety of characteristics. Six northeastern sites are rural-forested,
two are rural-agricultural (Sites 106 and 128), and three are near or within the
Washingtor.-Baltimore-Philadelphia-New York City conurbation (Si:es 104, 116, and 144). The
upper midwest sites are roral-agricultural (134 and 124) or rural-forested (149). The midwest
7
-------
CO
1 u
p- --
/ / •
\ 1fi8
V„,
-r-\.
UPPER
NORTHEAST
-i .. I 'S3
t - /
m
. j ifis
16?
169
\
\/
iei
V
174
m j
- -*Lj~
\
v. /
northeast
midwest
south
central
V. f
SOUTHERN
PERIPHERY ^
<1
•V»
V J
-/
Hgurez NOON momloring sues
-------
TABLE 1. NDDN SITE INFORMATION
Initial
Site
Reporting
Lati-
Longi-
Elevation
Land
No.
Site Name
Date
tude
tude
(m)
Use
Terrain
101a
Research Triangle
01/06/87
35.91
78.88
94
Forested-
Rolling
Park, NC
Urban
102b
Oak Ridge, TN
01/06/87
35.96
84.29
341
Forested
Complex
103c
West Point, NY
01/06/87
41.35
74.05
203
Forested
Complex
104
West Point, NY
01/06/87
41.35
74.05
203
Forested
Complex
105
Whiteface
01/06/87
44.39
73.86
570
Forested
Complex
Mountain, NY
106
PSU, PA
01/06/87
40.73
77.95
378
Agricultural
Rolling
107
Parsons, WV
01/14/88
39.09
79.66
510
Forested
Complex
108
Prince Edward, VA
11/01/87
37.17
78.31
146
Forested
Rolling
109
Woodstock, NH
12/31/88
43.94
71.70
258
Forested
Complex
110
Connecticut Hill,
N.TV
09/14/87
42.40
76.65
515
Forested
Rolling
111
IN I
Speedwell, TN
07/01/89
36.47
83.83
361
Agricultural
Rolling
112
Kane Experimental
12/31/88
41.60
78.77
622
Forested
Rolling
Forest, PA
113
M.K. Goddard, PA
01/08/88
41.43
80.15
384
Forested
Rolling
114
Deer Creek State
09/30/88
39.63
83.26
265
Agricultural
Rolling
Park, OH
115
Ann Arbor, Ml
06/30/88
42.42
83.90
267
Forested
Flat
116
Beltsville, MD
12/31/88
39.03
76.82
46
Urban-
Flat
Agnc.
117
Laurel Hill State
12/10/87
40.00
79.25
615
Forested
Complex
Park, PA
118
Big Meadows, VA
06/30/88
38.52
78.44
1,073
Forested
Mountain
top
119
Cedar Creek State
11/09/87
38.88
80.85
234
Forested
Complex
Park, WV
120
Horton Station, VA
06/03/87
37.33
80.55
920
Forested
Mountain
top
121
Lilley Comett
01/19/88
37.08
82.99
335
Forested
Complex
Woods, KY
122
Oxford, OH
08/18/87
39.53
84.72
284
Agricultural
Rolling
123
Lykens, OH
09/30/88
40.92
83.00
303
Agricultural
Flat
124
Unionville, Ml
06/30/88
43.61
83.36
201
Agricultural
Flat
125
Candor, NC
09/30/90
35.26
79.84
198
Forested
Rolling
126
Cranberry, NC
12/31/88
36.11
82.04
1,219
Forested
Mountain
127
Edgar Evins State
03/22/88
36.04
85.73
302
Foresred
top
Rolling
Park, TN
128
Arendtsville, PA
06/30/88
39.92
77.31
269
Agricultural
Rolling
(continu
ed)
9
-------
TABLE 1. (continued)
Initial
Site
Reporting
Lati-
Longi-
Elevation
Land
No.
Site Name
Date
tude
tude
Cm)
Use
Terrain
129d
Perryville, KY
08/11/87
37.68
84.97
279
Agricultural
Rolling
130
Bondville, IL
02/09/88
40.05
88.37
212
Agricultural
Flat
131
MackvUle, KY
07/31/90
37.70
85.05
353
Agricultural
Rolling
133
Salamonie, IN
06/30/88
40.82
85.66
249
Agricultural
Flat
134
Perkinstown, W1
09/30/88
45.21
90.60
472
Agricultural
Rolling
135
Ashland, ME
12/31/88
46.61
68.41
235
Agricultural
Flat
137
Coweeta, NC
11/03/87
35.06
83.43
686
Forested
Complex
140
Vincennes, IN
08/05/87
38.74
87.49
134
Agricultural
Rolling
144
Washington's
12/31/88
40.30
74.87
58
Agric.-
Rolling
Crossing, NJ
Urban
146
Argonne National
07/01/87
41.70
8799
229
Urban-
Laboratory, IL
Agric.
Flat
149
Weilston, MI
06/30/88
44.22
85.82
295
Forested
Flat
150
Caddo Valley, AR
09/30/88
34.18
93.10
71
Forested
Rolling
151
Coffeeville, MS
12/31/88
34 00
89.80
134
Forested
Rolling
152
Sand Mountain, AL
12/31/88
34.29
85.97
352
Agricultural
Rolling
153
Georgia Station,
06/30/88
33.18
84.41
270
Agricultural
Rolling
156
vJA
Sumatra, FL
12/31/88
30.11
84.99
14
Forested
Flat
157
Alhambra, IL
06/30/88
38.87
89.62
164
Agricultural
Flat
161
Gothic, CO
07/01/89
38.96
106.99
2,926
Range
Complex
162
Uinta, UT
07/01/89
40.55
110.32
2,500
Range
Complex
163
Refolds Creek, ID
07/01/89
43.21
116.75
1,198
Range
Rolling
164
Saval Ranch, NV
07/01/89
41.29
115.86
1,873
Range
Rolling
165
PinedaJe, WY
12/31/88
42.93
109.79
2.388
Range
Rolling
167
Chiricahua, AZ
07/01/89
32.01
109.39
1,570
Range
Complex
168
Glacier National
12/31/88
48.51
114.00
963
Forested
Complex
Park, MT
169
Centennial, WY
07/01/89
41.31
106.15
2,579
Range
Complex
174
Grand Canyon, AZ
07/01/89
36.06
112.18
2,073
Forested
Complex
aOperation rerminared 01/02/90.
'"'Operation rerminared 12/31/88.
cOperation terminated 01/30/88.
dOperation terminated 07/28/90.
10
-------
sites are rural-agricultural, except for Site 146 (suburban Chicago), which is urban-agricultural.
Although rural in character, there is some evidence that S02 concentrations at Sites 122, 140,
and 157 are influenced by nearby emissions (i.e., within 50 km).
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 along the
Appalachian Mountains from northern Virginia to western 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. Finally, each of the three southern periphery
sites is rural-forested in flat or rolling terrain.
Despite apparent similarities in land use and terrain, the western sites are not homogeneous
in character. Nearly every site is located in a distinct subregion of the west. Site 161 (Gothic,
CO) occupies a mountain valley in the central Rocky Mountains, and Site 162 is located on the
foothills of the High Uintas. Sites 163 and 164 are located :n similar surroundings near the
northern extreme of the Great Basin. Sites 165 and 169 represent the transition from the
western Great Plains to the Rocky Mountains. Sites 167 and 174 are located in the arid
southwest; however, Site 167 is in the Sonoran Desert, while Site 174 is on the forested Kaibab
Plateau. Site 168 (near the Canadian border) alone represents the western boreal forest.
3.2 NETWORK OPERATIONS
Ambient measurements of meteorological variables, 03. SO^", NOj, S02, HNOj, and
ammonium (NH^) are collected at each NDDN site. Meteorological variables and 03 are
measured as hourly averages, while the sulfur and nitrogen species are measured as weekly
averages using filter packs (Edgerton and Lavery, 1991).
Windspeed, wind direction, and standard deviation of wind direction (sigma theta) are
measured at 10 rn. temperature is measured at 2 and 9 m, ar.d relative humidity is measured at
9 m. Precipitation and solar radiation are measured on i-m platforms located outside the rain
11
-------
and sun shadows of the shelter and towers. Surface wetness is measured at approximately 6 to
12 inches above the surrounding low-lying vegetation (typically grass). Field equipment is
inspected weekly by the site operator and calibrated quarterly by ESE personnel using standards
traceable to the National Institute of Standards and Technology (NJST). Precision and accuracy
objectives for NDDN continuous measurements are shown in Table 2.
Filter pack sampling and 03 measurements are performed at 10 m using a tilt-down
aluminum tower. Filter pack flow is maintained at 1.50 liters per minute (L/min) at eastern
sites and 3.00 L/min at western sites [referenced to 25 degrees Celsius (°C) and 760 millimeters
of mercury] with mass flow controllers. Ozone is measured via ultraviolet (UV) absorbance with
a Thermo-Environmental Model 49-103 analyzer operating on the 0- to 500-parts-per-billion
(ppb) range. Site operations, laboratory analysis of filter packs, and data management activities
are performed according to standard operating procedures developed for the NDDN (ESE, 1990a,
1990b, 1991).
Duplicate sers of equipment were located at various sites in the northeast (107), southeast
(153), midwest (157), and southwest (167) to evaluate the overall network precision. The
purpose of this spatial distribution was to capture precision data across a broad range of
meteorological conditions and ambient concentrations. In late 1990, a duplicate site was
established in the northwest (163), and duplicate sers of equipment at Sites 107, 153, and 157
were moved to Sites 128, 156, and 114, respectively. Results of precision analysis are described
in Section 6.0.
3.3 VEGETATION AND SURFACE CONDITION MONITORING
Various observations are periodically made at the NDDN sites to support model calculations
of dry deposition. Site operators record surface conditions (e.g., dew, frost, snow) and
vegetation status weekly. This information is collected to determine the frequency of conditions
that could influence deposition rates for gases (especially S02) and particles. Vegetation data
are obtained to track evolution of the dominant plant canopy, from leaf emergence (or
germination) to senescence (or harvesting). Each week, the site operator estimates and reports
the percentage of maximum leaf area for the plant canopy near the site.
12
-------
TABLE 2. PRECISION AND ACCURACY OBJECTIVES OF NDDN FIELD MEASUREMENTS
Measurement
Parameter
Method
Acceptance Criteria*
Precision
Accuracy
Windspeed
Wind Direction
Sigma Theta
Relative Humidity
Solar Radiation
Precipitation
O3
Filter Pack Flow
Surface Wetness
Anemometer
Wind Vane
Wind Vane
Humidity Sensor
Pyranometer
Rain Gauge
Ambient Temperature. Thermistor
Delta Temperature Thermistor
UV Absorbar.ce
Mass Flow Controller
Conductivity Bridge
±0.5 m/sec
+5°
+ 10%
±10% (of
full scale)
±10% (of
reading
taken at
local noon)
.+ 10% (of
reading)
±0.5°C
±0.25 *C
±10% (of
reading)
+.0.15 L/min
Undefined
The greater of
±0.2 m/sec or
±5%
±5°
Undefined
±10% (of
full scale)
+ 10%
±0.05 inch*
±0.25°C
±0.25CC
±10%
±10%
Undefined
Note: °C = degrees Celsius.
L/min = liters per minute,
m/sec = meters per second.
•Field precision criteria apply to collocated instruments, and accuracy criteria apply to
calibration of instruments.
fFor target value of 0.50 inch.
13
-------
Examination of several years of data shows that the growing season length at a site (as
determined from the time of 50-percent leafout to 50-percent leaf senescence) is fairly constant
(i.e., within 7 to 10 days) from year to year. However, growing season length (as defined
previously) has a strong latitudinal dependence. As shown in Figure 3, growing season in the
eastern United States differs by nearly a factor of 2 between northern Maine (126 days) and
northern Florida (247 days). For species with significant canopy resistance to deposition,
variability in growing season length may play an important role in spatial variability of V^s and
fluxes. Growing season shows less of a latitudinal dependence among the NDDN sites in the
western United States (see Figure 3). The reasons for this are not entirely clear but may be
related to differences in elevation or effects of water stress.
Site operators also provide sketches illustrating the distribution of major plant species and
land-use classifications within 1 km of the site. These sketches are refined by digitization and
analysis of aerial photographs obtained from the U.S. Geological Survey (USGS) National
Cartographic Information Center in Reston, VA, followed by interpretation according to
procedures described by Anderson et al. (1978). Plant species information is then obtained by
surveying each of the major land-use classifications within 1 km of each site. Figure 4 illustrates
the land-use distribution for a site in south-central Virginia (Site 108, Prince Edward State
Forest).
Leaf Area Index (LAJ) is the one-sided leaf area of the plant canopy per unit area of ground.
LAI has been shown to be an important model input variable for dry deposition calculations
(McMillen, 1990). LAI was measured at each site during the summers of 1990 and 1991 using
an LAI-2000 Plant Canopy Analyzer (LI-COR, 1989). The LAI-2000 makes indirect (i.e.,
nondestructive) estimates of LAI from simultaneous measurements of light interception above
and below the canopy. Initial development and testing of the LAI-2000 by the manufacturer
focused on a variety of agricultural crops, such as soybeans and wheat. Similar approaches have
been used to measure LAI of forest canopies (Pierce and Running, 1988; Chason et al., 1990).
The objective for IAI measurements is to characterize the predominant canopies near each
NDDN site. Based on the digitized aerial photographs described earlier, transects of 25 to 100
independent measurements are made across each major land-use classification at all sites.
Information on plant species, canopy density, and height is recorded during each transect.
Results are then used to estimate species-dependent LAI values for input to the Big Leaf model.
14
-------
3 CESSSS1SCS®*
figuie
-------
ClasMification Legend
i 1 ,K! 1
Uf bun,
Kesitlentiol
; .i-c j
Urbon,
Commercial
ra
Urbon,
Industrial
! 'JjJ
u rbo n.
Inatitutional
Lyrll
Urban,
Fronjportobon
[ uo~~l
Urban,
Open and Other
Forest.
Mixed
I .ic .)
Forest,
Coniferous
i roj
forest.
Deciduous
&'~
Agriculture. C'oplnnd
!_ _AP ~]
Agriculture, Postuie
i.J^L J
Agriculture. Crove ond Nursery
! ^ ]
Agriculture. Other
[ KS J
Hanqfl.
Scrub and BiubN
f"«D
Range,
Crosslond
1 «N j
Wetland, Nor*- tof anted
!>> 1
>Vat!and, Forested
!VJ
otream
C»C)
1 ako
r«R ]
Reservoir, l)omm®d
[jifii]
Barren,
Strip Mined
l«ti
Barren,
Beach
' HS I
Barren,
Other
Symbol Legend
Road (UT)
NOON Monitoring Site
Classification by pfiolo
-------
A summary of LAI data for various canopy types is shown in Table 3. LAI values at the NDDN
sites cover a broad range. Eastern deciduous canopies have the highest LAIs (3.6 to 6.4), and
western rangeland has the lowest (0.2 to 2.2). Eastern coniferous canopies, on average, exhibit
LAIs about one unit higher and one unit lower, respectively, than western coniferous canopies
and eastern deciduous canopies. The LAI of crops is somewhat sensitive to planting density and
plant species. Results appear to converge on a value of about 2.5 for a broad range of crops.
TABLE 3. SUMMARY OF 1990 AND 1991 MIDSUMMER LAI MEASUREMENTS FOR THE NDDN
CanoDV
Number of
Sites
LAI
Class
Subclass
Mean
Standard Deviation
Range
Deciduous
Eastern
23
4.43
0.66
3.64-6.40
Northeastern
8
4.33
0.61
3.65-6.10
Southeastern
7
4.23
0.45
3.64-5.18
Midwestern
8
4.71
0.80
3.78-6.40
High Elevation
1
2.19
0.24
1.94-2.40
Western Aspen
3
2.05
0.84
1.57-3.02
Mixed
Eastern
10
4.43
0.56
3.58-5.38
Northeastern
3
4.15
0.24
3.89-4.51
Southeastern
5
4.47
0.68
3.71-5.38
Midwestern
2
4.35
0.12
4.15-4.54
Conifer
Eastern
14
3.31
0.65
2.43-4.60
Loblolly
5
3.13
0.76
2.88-4.21
White Pine
3
2.90
0.45
2.60-3.42
Hemlock
2
3.60
0.16
3.49-3.71
Western
4
2.48
0.48
2.05-3.09
Juniper-Pinon
2
1.61
0.71
1.10-2.11
Grass
Short (<12")
10
1.36
0.39
0.87-1.68
Tall (>12")
6
1.99
0.62
1.44-2.30
Mixed Grass, Weeds
6
2.42
0.66
1.43-3.20
Range
Sagebrush
4
0.43
0.17
0.25-0.64
Sage, Grass
3
1.68
0.49
1.15-2.23
Crops
Clover
A.
1.94
0.44
1.63-2.25
Soybeans
3
2.70
0.59
2.33-3.33
Soybeans (dense)
1
6.36
--
-
Corn
4
2.60
0.69
1.59-3.08
Sugar Beets
1
2.46
--
-
Melons, Tomatoes
1
1.42
--
17
-------
The accuracy of LAI estimates using the LAI-2000 was evaluated at several independently
characterized sites during the summer of 1991. Preliminary results suggest that readings agree
within _+0.5 unit with other, more direct, approaches for deciduous canopies and crop species.
Additional studies are planned to evaluate accuracy for coniferous canopies and grasses, as well
as the influence of changing lighting conditions and terrain. Results of side-by-side LAI
measurements within the NDDN (two instruments with separate operators) show that precision
and reproducibility are on the order of 2 to 5 percent for mixed forest canopies (nominal LAI
approximately 4.25).
18
-------
SECTION 4.0
DRY DEPOSITION CALCULATIONS
4.1 INFERENTIAL MODEL
Assuming that fluxes are unidirectional (i.e., air to surface), dry deposition flux can be
expressed as the linear product of ambient concentration (C) and deposition velocity (Vd):
Flux = C x Vd (1)
The influence of meteorology and vegetation on the flux of material is contained in the Vd.
The inferential model (Hicks et al., 198S) simulates Vd based on a theoretical understanding of
dry deposition processes. The procedure is anchored to empirical deposition estimates obtained
through direct measurements of dry deposition fluxes, primarily of 03 by using the eddy
correlation method. The basis for the inferential model is briefly described as follows:
The dry deposition processes are represented in the inferential model by analogy to series
and parallel resistances in an electrical circuit. Three major resistance components represent the
physical and chemical processes. The total resistance (R) for a single vegetation species can be
expressed as:
R = Ra + Rb - Rc = 1/Vd (2)
Ra is the aerodynamic resistance. It is inversely proportional to the atmosphere's ability to
transfer material downward from the planetary boundary layer to the surface layer by turbulent
processes. Ra is calculated from easily measured meteorological parameters, i.e., windspeed and
standard deviation of wind direction (sigma theta). The formula used in the calculation
procedure is:
Ra = u/u.2 (3)
where u is the w:r.dspeed, and u, is Ae local friction velocity which measures the effectiveness
of turbulence exchange between the atmosphere and the surface. In application of the
inferential model to the NDDN, u, is derived from the standard deviation of the wind direction
19
-------
(sigma theta) and u. Under well-mixed conditions (e.g., during most sunny afternoons or very
windy conditions), R_, approaches 0. Under calm, stagnant conditions (e.g., at night), material is
not mixed readily to the surface, and Ra approaches a large value.
Rb is the boundary layer resistance to vertical transport through a shallow (approximately
1 millimeter) nonrurbulent layer of air in direct contact with the surface. It depends on
aerodynamics of the surface and is inversely proportional to u, in the inferentiaJ model. Rb is
also weakly dependent on physical properties (e.g., diffusivity) of the pollutant being deposited.
Rc is the canopy, or surface uptake, resistance. It contains a number of terms (represented
as parallel resistances) that account for the direct uptake/absorption of the pollutant by leaves,
soil, and other biological receptors within and below the canopy. Rc contains parameterizations
for the type of vegetation and vegetation density, solar radiation penetration of the canopy,
wetness of the surface, etc. Rc is difficult to treat theoretically, and the system of equations for
estimating Rc is normally empirically adjusted based on direct observation of dry deposition
Huxes.
For pollutant species such as 03 and S02, the controlling component of Rc is the stomotal
resistance, which has large diurnal and seasonal variability. For highly reactive species such as
HN03, Rc is generally small regardless of season or canopy type, and Ra and Rb control
variability. Particle deposition for SO^" and NOj is primarily governed by turbulent processes
and is represented in the 3ig Leaf model as a function of u. and Ra.
Using the previously described physical and mathematical framework, several operational
models have been implemented to evaluate dry deposition in the NDDN. These are the Big Leaf
inferential model, the Multilayer inferential model, and the LAD model. The Big Leaf and
.Multilayer models are site specific. The resulting dry deposition fluxes are representative only of
a 1-km-radius area around a site. The Big Leaf model differs from the Multilayer model; the Big
Leaf treats the canopy as a single sink, and the Multilayer model applies similar calculations
through a 20-layer canopy in which model parameters are modified by attenuation processes.
The IAD model fMcMillen, 1990) estimates dry deposition over a large area (up to 100 km by
100 km) surrounding the NDDN site using spatially resolved (1 km by 1 km) vegetation and
terrain data.
20
-------
All three models were developed by the NOAA Atmospheric Transport and Diffusion Division,
Oak Ridge, TN. Currently, only the Big Leaf model is fully operational. While the Multilayer
and LAD models have been implemented in the NDDN program, these models are still
undergoing testing by both NOAA and EPA. The following sections provide some details on the
models relevant to their application in the NDDN.
4 2 BIG LEAF DRY DEPOSITION MODEL
Dry deposition calculations for the NDDN sites (reported in Section 5.0 and the appendices)
are made using a version of the Big Leaf inferential model, which has been modified to address
multiple vegetation species (including barren ground and water surfaces). Land-use/vegetation
types are quantitatively defined in a 1-km radius of the site. An important feature of the Big
Leaf model (also of the Multilayer and LAD models) is that it treats dry deposition as a function
of canopy surface area as measured by the LAI.
The meteorological variables necessary to determine Ra, Rb, and Rc 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. The windspeed and u. measured on the tower are strictly applicable for
the specific vegetation species near the tower. Windspeed and u. are extrapolated to other
vegetation species in a 1-km radius as a function of the surface roughness assigned to each
species. It is assumed that the product of u and u. is constant above all vegetation species.
Thus, R_d and Rb may differ for each plant species as a function of roughness.
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 models.
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 car. be calculated by the model.
21
-------
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, again for every hour. The site averaged Vd is
the areal-weighted average Vd over all vegetation types. The hourly values of Vd are then
averaged over a week and multiplied by the weekly integrated concentrations to produce weekly
fluxes of HN03, SO^", NOj, and S02. Ozone flux is calculated using hourly measurements of 03
and hourly values of Vd. Weekly flux calculations are considered valid if >. 70 percent of hourly
Vd values are available for that week. Flux calculations based on weekly integrated
concentrations are generally biased low. This is further discussed in Section 6.0.
4.3 MULTILAYER INFERENTIAL MODEL
NOAA has recently developed a Multilayer inferential model. This model calculates Rb and
R^ for each of the 20 layers through the canopy using a wind profile and solar radiation
attenuation algorithm unique to the plant species. The resistances are integrated through the
canopy to obtain a representative Vd, which is multiplied by the above canopy concentration to
obtain the flux. The model additionally accounts for water stress on the vegetation and
deposition to snow surfaces (not accounted for in the Big Leaf model). Several parameter
values, e.g., che soil resistance, have also been modified from those used in the Big Leaf model.
In concept the Multilayer model is a significant improvement over the Big Leaf model. The
NDDN is in the process of implementing the multilayer version and expects to start reporting Vds
and fluxes using that model when it is released by NOAA in 1993. Similar to the Big Leaf
model, the NDDN version of the Multilayer model will calculate Vds for any number of plant
species and area weight these to obtain a site-specific Vd.
4.4 LARGE AREA DEPOSITION MODEL
Vegetation characteristics and terrain may change significantly within kilometers of NDDN
sites; consequently, site-specific deposition may be unrepresentative of the larger surrounding
area. For example, dry deposition to a grass field where the NDDN equipment is located may
differ from that to a coniferous or deciduous forest more characteristic of the regional land use.
In that case it would be inappropriate ro estimate regional patterns of deposition fluxes using
site vegetation data. Selection of sites that are representative in terms of land use is often not
practical; therefore, an approach has been developed to estimate the bias of a site with respect
22
-------
to a large area surrounding that site. The approach, documented by McMillen (1990), estimates
dry deposition through the Multilayer algorithm based on land use within a 100-km by 100-km
area surrounding the site. Vegetation type is determined by associating Advanced Very High
Resolution Radiometric (AVHRR) satellite statistics with ground truth vegetation data.
The model system has two components. The first component (program EXTRACT) is a
processor for AVHRR satellite and terrain data upon which the system is based. The second
component is an algorithm to calculate statistics on dry deposition for the desired area (up to
100 km by 100 km) with 1-km by 1-km grids using the Multilayer inferential model.
The EXTRACT program processes a 1-km resolution United States database. The input is the
latitude/longitude of the site (usually NDDN site coordinates) and size of the grid (from 1 to
100 km). The outputs of EXTRACT are:
1. Vegetation class for each grid (1 km by 1 km),
2. Normalized Difference Vegetation Index (NDV1) class for each 1-km grid (NDVI is a
measure of biomass or LAI),
3. Calculated terrain complexity and slope for each grid based on the elevation of the terrain
at the four surrounding grids, and
4. NDVI and terrain complexity mean and standard deviation for the area and as a function
of vegetation class.
The basic system allows for 167 different vegetation categories. The LAD, as currently
implemented, uses nine lumped vegetation classes: corn, soybeans, wheat, grasses, northern
hardwood forest, southern hardwood forest, loblolly pine, spruce, and ponderosa/lodgepole pine.
The output from EXTRACT is input to the LAD program along with meteorological and
concentration data for the NDDN sites and composite vegetation response parameters for the
nine lumped vegetation classes.
The LAD program calculates Vds and fluxes using the Multilayer model as follows. Ra and
Rh + Rc are calculated for each of the nine vegetation classes, and values are assigned to each
1-km grid according to its vegetation class. For each grid, R. is adjusted for terrain complexity,
and Rc is adjusted for the NDVI and terrain slope (solar elevation angle) of the particular grid,
and Vd is calculated. The mean and standard deviation of Vd are obtained for the entire area
23
-------
and also for individual vegetation species. Fluxes are obtained by multiplying the grid-averaged
Vd by the concentration at the NDDN site. Averaging options are for hourly, daily, or weekly
output.
The model does not account for spatial variability of concentration and meteorology, other
than adjusting u and u, for variation in vegetation type and Rg for general terrain complexity.
These limitations probably make the approach inappropriate for short-time flux calculations (less
than a week), for grids dominated by local "sources," and for grids with substantial variability in
elevation (see Section 6.7). For weekly or longer averages and in areas of level or rolling terrain
away from major sources, the LAD model is expected to provide reasonable results. The
approach is currently used to estimate potential bias of the NDDN site relative to the larger
surrounding area.
24
-------
SECTION 5.0
DRY DEPOSITION DATA
This section describes spatial and temporal patterns of concentration, Vd, and flux for S02
and HN03 for calendar year 1991. These species were chosen for discussion because they have
been shown to account for most of the dry sulfur and nitrogen deposition in the eastern United
States (Meyers et al., 1991). Spatial plots are presented for annual averaged concentration, Vd,
and flux at all NDDN sites. Temporal (i.e., week to week) patterns are presented for three sites
in Indiana, Pennsylvania, and North Carolina. Annual and seasonal concentration, Vd, and dry
deposition flux data for S02, HN03, 03, SO*", and N03 for calendar years 1990 and 1991 were
calculated and are included in Appendices B and C, respectively.
5.1 SULFUR DIOXIDE
Mean annual concentrations for S02 during 1991 are shown in Figure 5. Seasonal values of
concentrations, Vds, and fluxes and the number of weeks making up the season are included in
Appendices B and C. Concentrations across the eastern United States ranged from
1.6 micrograms per cubic meters (jig/m3) in northern Maine (Site 135) to 18.8 ng/m3 in
southwestern Indiana (Site 140). Concentrations greater than 15 ng/m3 were limited to two
sites in western Pennsylvania, as well as isolated sites in northern Illinois, southern Indiana, and
western Ohio. The high concentrations at the Pennsylvania sites are believed to be regionally
representative, while the Illinois, Indiana, and Ohio sues appear to be influenced by local
sources. Average concentrations of 10 to 15 ng/m3 extend over a much larger area (from
southern Illinois to New Jersey) and generally correspond to the major sulfur emission regions.
Outside this area, concentrations decrease rapidly to 2 ng/m3 or less in Maine, Wisconsin,
Arkansas, and Florida.
The overall pattern of S02 concentrations is punctuated by relatively low values at two sites
in eastern Kentucky (Site 121) and western North Carolina (Site 137). Both of these sites are
located in complex terrain and exhibit S02 concentrations that are 30 to 50 percent of those
observed at neighboring sites. Nocturnal phenomena (e.g., shallow inversions) may account for
the lower concentrations at these sites. For example, a ridge-top site recorded S02
concentrations 2.4 times higher than the Coweeta site, which is located in a protective valley less
than 1,000 m away (see Section 6.7). Low emission densities for S02 in the central Appalachian
25
-------
r~r
07
35
11.8
11.6
5.9
144
•103
17.3
0.6
03
8.7 v
•63
100
7.8
0.4
8 4
10.8
02
3.B «6
09
7.0
3.2
—"vL, 19
04
c$r
We, ,991 moan a^SO; concent
-------
Mountains may also contribute to the low concentrations in eastern Kentucky and western North
Carolina.
Annual average S02 concentrations for the western NDDN sites ranged from 0.28 iig/m3 in
northeastern Nevada (Site 164) to 2.1 yg/m3 in southern Arizona (Site 167). With the
exception of Site 167, concentrations at western sites were less than 50 percent of the lowest
value observed in the eastern United States. The density of western sites is too low to generalize
a regional pattern; however, relatively high SOz concentrations in southern Arizona may reflect
power plant and smelter emissions in the southwest.
Annual average S02 Vds and fluxes, as calculated with the Big Leaf model, are shown in
Figures 6 and 7, respectively. The INS labels on the figures indicate that less than four valid
seasons, consisting of 9 weeks or more of data, were available for a site. Annual average S02
Vds (Figure 6) showed considerably less variability across the NDDN than S02 concentrations
(Figure 5). Among eastern sites, Vj ranged from about 0.2 to 0.4 centimeter per second
(cm/sec) but showed no evidence of a latitudinal gradient as might be expected based on
growing season information (Figure 3). This suggests that other factors compensate for any
growing season effect. Inspection of land-use information (Table 1) shows that forested sites
tend to exhibit larger Vd than agricultural sites. The higher surface roughness of the forested
sites may be a factor, since it impacts windspeed and Ra. Sulfur dioxide Vds are also strongly
modified by surface wetness. Sites with high frequency of surface wetness, due to dew
formation, tend to have higher Vds. In general, there are no obvious regional patterns to the
S°2 vd field in the eastern United States.
Results for western sites were limited, due to the data completeness requirements discussed
previously, but were generally lower than those observed in the eastern United States. Several
factors can account for the lower Vds at the western sites. Three of the four western sites
meeting data completeness requirements are in rangeland with low LAI values (i.e., less than 1)
and low surface roughness (in contrast, LAls at eastern sites ranged from 2 to 5). The fourth
site (near Grand Canyon, AZ) is sparsely forested with substantial barren rock. In addition, the
frequency of surface wetness from dew at western sites is only about 50 percent of that at
eastern sites.
27
-------
00
016 /IN
W029 0
o.3i •
\ 0.31
Note: INS indicates lewer than four valid seasons.
Figure 6 1991 mean annua! SO2 deposition velocities (cm/sec).
-------
Note:
ins «
M«>BS'e'"e"han
tour
vaW seasons- v_3
figure
7. 1991 mean
a„™alS02«"«slk9'MV
-------
Annual S02 fluxes for 1991 (Figure 7) generally mirrored the patterns of atmospheric
concentrations. Among eastern sites, fluxes ranged from 1 kilogram per hectare (kg/ha) in
southwestern North Carolina (Site 137) to about 17 kg/ha in northwestern Pennsylvania
(Site 113) and were above 5 kg/ha in a region extending from southern Illinois to New York,
New Jersey, and Virginia. Fluxes were about 1 kg/ha around the periphery of the eastern
network. As observed for the concentration data, fluxes were substantially lower at Sites 121
and 137 than at neighboring sites, reflecting the combination of relatively low Vds and
concentrations. Western sites have very small fluxes, reflecting the low concentrations and low
Vd*.
Figure 8 shows the weekly variability of S02 concentrations at a flat, mixed vegetation site in
SaJamonie, Indiana (Site 133), a primarily agricultural site in rolling terrain at a state college in
Pennsylvania (Site 106), and a forested complex terrain site in Coweeta, North Carolina
(Site 137). Concentrations were highest during fall and winter and considerably higher in
Indiana and Pennsylvania than in North Carolina.
Deposition velocities (Figure 9), in contrast, showed a much more pronounced pattern of
variability with substantially higher Vds during the growing season at all sites. These velocities
were nearly constant at low values (i.e., 0.08 to 0.15) during the winter and increased
dramatically from April through July. Of these three sites, Site 133 had the highest weekly value
of about 0.57, which occurred in May. Site 137 had the lowest summer maximum of about 0.4,
also occurring in early May.
The pattern of weekly fluxes for 1991 was similar at the Indiana and Pennsylvania sites,
showing maxima in the summer and large week-to-week variability (Figure 10). Thus, the
annual flux profile does not follow the concentration profile at these sites. S02 fluxes were low
year round at the Coweeta, NC site.
5.2 NITRIC ACID
Annual average concentrations of HN03 (see Figure 11) ranged from 0.5 jig/m3 in norrhem
Maine (Site 135) to 3.5 |ig/m3 in southeastern Pennsylvania (Site 128). Concentrations were
generally between 2 and 3 yg/m3 in the central portion of the eastern network and between 1
and 2 yg/m3 across the upper northeast, upper midwest, and southern periphery. As observed
30
-------
H-J/
bD
:><;
•-r.i uo
:iC CO
=>0 . 00
0 0
I
II
i |
~ ~ ~ Coweeta, NC (Site 137)
-• Penn State, PA (Site 106)
¦k—b—A Sclomonle, IN (Site 133)
tj-
fyrv\
P \i
W
>v
\
u
no ;
nni43
\A*A/ ^ >.. IP
J-fT'0 a 3>B'BL J RtH' Uta-n n-ET^tHrf-E.! n-B-B-€hB-B 3B-h-B e^Q-©"0
/
' V VV\'f
* A V
1 1 1—
9107B
91 1 70
Week Stdft !) otes (yyddo>
9 J ?7B
a-EJ
qpo 13
Figure 8. Weekly SO2 concentrations for 1991 lor three sites in
Indiana, Pennsylvania, andQ&&-.—' ^ jp~
-------
w
to
0 LO
n -i c
o . 4
o. do
o. 1 c
0 0 0
90
-¦fir-tf \
fW
NQ fr
\
\ K
\ ! \
O-O B Coweeta, NC (Site 137)
• • • Penn State, PA (Site 106)
-tr ix ¦& Salomonie, IN (Site 133)
f\ \ / \ A
m I *\ • ^ h'v*'* \
/ A * U \
\
rt-G =
WVYJ
i" i 1 r~
-i 1 r-
y 3078
T
~ 7 1 r—
31170
W-jp< Start Oot:L'is lyydrit)
^ I ^
91 27B
—r
9201 3
Figure 9. Weekly SO2 dry deposition velocities for 1991 for three sites in
Indiana, Pennsylvania, and Ohio.
-------
w
OJ
ky/ha
a. 7 c
:( i >o
¦lO -
0 - .< 0
o ~,o
c . z o
c 0
0 CO
A /;5,
J. i * \
~ D ~ Coweelo, NC (Site I 37)
• Penn Stole, PA (Site 106)
a
-------
06 v.
03
0.7
2.8
08
2.41
• 2.8
to.*
3.2
3.6
03
•2.6
2.2
0 5
04
3.0
2.7 N
• 2.1
23
0.5
0.8
25-
0.3
2.7
20
1.8
0.9
1.4
1;
1 A
.r~
Figure 11. 1991 mean annual HNO 3 concentrations lug/m').
-------
for S02, the sites in eastern Kentucky and western North Carolina exhibited substantially lower
concentrations than nearby sites.
Local HN03 concentrations may be influenced by terrain. Sites with good exposure generally
had higher average values than neighboring sites located in complex terrain or in small
clearings. For species with large Vds (such as HN03), the microclimate in complex environments
could result in high removal of HN03 and high local variability of concentrations.
Average annual concentrations for the nine western sites ranged from 0.28 ^g/m3 in central
Colorado to 0.86 yg/m3 in northern Arizona and, with the exception of the latter site, were
lower than concentrations in the eastern United States. Ratios of HN03 to S02 also differed
significantly between eastern and western sites. For eastern sites, the ratio was invariably less
than unity, whereas for the western sites it was on the order of unity. Given higher Vds for
HNOj than for S02, this suggests that HN03 will have higher atmospheric fluxes in the western
United States.
Annual average Vds for eastern sites (Figure 12) ranged from values less than 1 cm/sec in
West Virginia and Kentucky to values exceeding 2.5 cm/sec in western North Carolina and
Maine. Deposition velocities exceeding 2 cm/sec occurred at sites in Illinois, Indiana, West
Virginia, and Pennsylvania; however, a regional pattern was not obvious. Nitric acid Vds in the
Big Leaf inferential model are primarily driven by windspeed and turbulence. Consequently, Vds
would be expected to be higher in those regions usually experiencing high winds or high surface
roughness (i.e., the northern part of the United States and at complex sites). The low Vd in
central and southern Illinois probably reflects a transition region from moderate to lighter winds
and also very low surface roughness (the vegetation is primarily farm crops). Deposition
velocities at the western sites are as high or higher than those in the eastern United States,
reflecting higher winds and complex topography at the western sites.
Calculated HN03 fluxes (Figure 13) for 1991 ranged from 1.8 to 24 kg/ha for eastern sites
and 2.2 to 41 kg/ha for western sites. Extremely high spatial variability is evident across the
eastern states, indicating local influences of sources, meteorology, and possibly removal
processes. As suggested for S02, nocturnal inversion layers may promote depletion of HN03
within the surface layer. This effect will increase with shallower and longer-lasting inversions.
35
-------
Nole: INS indicates fewer than four valid seasons
Hgure 12. 1991 mean annual HNO3 deposition velocities (cm/sec).
-------
GJ
-a
•- ^ iO
s1 ^
,-¦( * 18 0
, JOJU^ *30 * 24«
( N, »V * »1<;0 \ 9 *•
\ J INS \ • \
\ l> \ «.7 \INS. r
\ - -~"t' * \ t «-V * r is
Vr \ " I >5"w~( j •
S 7 *6.t fW- . 18 U5-"
1 V. ^ WS 1^2" „ a3
i J\l 8 9 J
1 ._Vs5>'"*"v:/
Note
• INS indicates
• than tour
valid seasons
Hgufe
13 l9gi wean
aWUamn03^es
(Kg/ha).
-------
Weekly HN03 data for 1991 for three sites in Indiana, Pennsylvania, and North Carolina are
shown in Figures 14 through 16. The maximum weekly concentration (Figure 14) of 7.1 ng/m3
occurred at the Indiana site in April, although the highest annual average occurred at the
Pennsylvania site. The North Carolina site, for reasons discussed previously, had relatively low
values of HN03. The highest values at all sites generally occurred during the spring and
summer months. In other words, the seasonal behavior of HN03 is the opposite of that for S02.
HNOj Vds (Figure 15) exhibit only moderate season to season variability but were generally
highest during the winter and spring. Deposition velocities for HN03 are primarily driven by
wind/turbulence variables, which are normally higher during those seasons. Thus, the annual
profile of HN03 Vd is very different from that of S02. Nitric acid fluxes (Figure 16) were highest
at the Pennsylvania site and very low at the North Carolina site. The highest values at the
Indiana and Pennsylvania sites occurred during the summer months, while those for the North
Carolina site occurred in early spring. In all cases, the temporal pattern of fluxes mirrored that
of concentration.
38
-------
U1
[i ! • I'
3 i>C
7 00
5 . 0C
' . 0 0
. 0 J
,, ; )
r.o
CO
o oc
m
(V X .m
V V A
ivti y viv
v \
.H ' '•tj-'-' M
It
V
-ti u
~ ~ ~ Coweeto, NC (Site 137)
• • • Penn State, PA (Site 106)
it -ft ¦ -ft Solomonle, IN (Site 133)
A ¦
I
\
F I V
/ %
Ai\K a. A
N. / * /
\ //^v
\ i ,
- 13Q-S t:H'*
Vn^ >H3
~r— t " r
no ri/j-
q m/ii
1 ? 1 1 1 1 ( J 1 I
"1 17 2
wpp< fit art Dates (yyaad!
r t r-
91278
-i [
920 1 3
Figure 14 Weekly HNO3 concenlralions for 1991 for three sites in Indiana,
Pennsylvania, and Ohio.
-------
h i
A n h
* V V V
B-B-B Coweeto, NC (Site 137)
Penn State, PA (Site 106)
if Solomonle, IN (Site 133)
A
T .
/ \/ \f* f
\ £
V
/ U/ 4 \ '\ * * \ i k r \ T P
y f w i^y' ^ a\ Asa a V 1/ /
W ^\AA/
« \ ?n \/ %>U vVV
^ K »
H p H Lae—B-tj
y
&'• v \ F
e-s
Hr q.,
93070 9117H 91270 92013
W<->Hk Start Cafes (yydddl
Figure 15 Weekly HNO3 dry deposition velocities lor 1991 for three sites in
Indiana, Pennsylvania, and Ohio.
-------
K a /1 In
R-B-O Coweeto, NC (Site 137)
•—•—• Penn State, PA (Site 106)
* Solamonle, IN (Site 133)
* /fHac.
^n,&R-d =j >-e"
920 1 ?
¦ n n 7 n
91 27 8
week Start Lie tes (yycaal
Figure 16. Weekly I fNOs dry deposition fluxes for 1991 for three siles in
Indiana, Pennsylvania, and Ohio
-------
SECTION 6.0
UNCERTAINTY OF INFERENTIAL MODEL CALCULATIONS
Dry deposition fluxes for the NDDN were calculated as the product of a measured
concentration and a modeled Vrf (Equation 1, Section 4.0), both of which have sources of error.
This section will focus on the sources of uncertainty associated with the Big Leaf inferential
model (i.e., the determination of Vd) and its application in the NDDN.
The inferential model was developed based on relatively few direct observations of dry
deposition fluxes, primarily eddy correlation measurements of 03, and to a lesser extent, eddy
correlation measurements of S02 and SO^", and gradient measurements of HNOa. These direct
measurement techniques are difficult to implement, are usually employed in research field
studies for relatively short time periods (weeks), and do not cover a broad spectrum of surface
and meteorological conditions.
The accuracy of direct flux measurement techniques is difficult to determine empirically and
differed from site to site. Ideal site requirements are uniform vegetation on a level surface
extending 100 times the measurement height in all directions. In addition, there should not be
major plumes interacting with the measurement site, so the concentration does not change
significantly over the measurement period. For these conditions, the accuracy of the direct
methods is estimated to be about ±15 to 20 percent (see Table 4). Direct measurements are
usually more accurate during daytime (normally the period of highest fluxes) than at night.
TABLE 4. ESTIMATED UNCERTAINTY FOR SHORT PERIODS
(30 TO 60 MINUTES) OF DIRECT MEASUREMENTS AND
INFERENTIAL MODEL CALCULATIONS OF DRY DEPOSITION
VELOCITIES FOR A LEVEL SITE WITH HOMOGENEOUS VEGETATION.
Measurement
Species
Method
Uncertainty (%)
Model Uncertainty (%)
hno3
Gradient
15
40
so2
Eddy Correlation
20
30
o3
Eddy Correlation
15
25
42
-------
In genera!, comparisons of Vds for 03 determined by eddy correlation techniques and the
inferential model have shown good agreement. For example, Figure 17 shows hourly Oa Vds
measured (by eddy correlation) and modeled (Big Leaf inferential model) for a site in
Pennsylvania for 10 days in August 1987 (McMillen, 1990). The model captured both the
magnitude and diurnal variation of Vd. Averaged over a diumal cycle, the model tended to
undercalculate Vd slightly during the morning transition period and overcalculate Vd slightly
during the afternoon (Figure 18). Matt et al. (1987) found the inferential model to
undercalculate eddy correlation S02 Vds above a forested area by about 20 percent, primarily
during the afternoon. Meyers (1992) also showed that modeled and measured (eddy
correlation) SOz Vds (see Figure 19) for a site in Huntington Forest in upstate New York were in
good agreement. These and similar studies support the subjective uncertainty estimates for
inferential model calculations of Vds as shown in Table 4.
The uncertainties in Table 4 reflect experience in making direct measurements of deposition
fluxes. Direct eddy correlation measurements of 03 fluxes have been made for at least 10 years,
resulting in refinement of the technique and a good database for development and evaluation of
the inferential model. The database of S02 flux measurements is not nearly as large because
fast response S02 instruments are not as advanced as those for 03. While few gradient
measurements for HN03 have been reported, the method is believed to have small uncertainty
because of the large gradients of HN03 normally observed near the surface. The 40-percent
uncertainty for the HN03 inferential model reflects the absence of a large database with which
to calibrate the model and the high sensitivity of the model to windspeed and turbulence.
Uncertainties for SO^ and N03 are likely to be even higher than 40 percent. These numbers
apply to short period (30 to 60 minutes) measurements and calculations. They also represent
generally favorable meteorological conditions and ideal measurement of the meteorological and
vegetation variables required for the inferential model.
Routine calculation of dry deposition fluxes as done for the NDDN introduces additional
uncertainties. These result from less precise definitions of meteorology and vegetation variables
associated with a monitoring program as opposed to a research program. However, the NDDN
calculations are integrated weekly and are usually reported as seasonal and/or annual values.
The integration process should average out random uncertainties of the model input data, and
43
-------
i *
w o
£, Modeted-tooothed
L® ft
-------
3
~3"
¦Q=«nr-cj3^-
a
3 B
cF^
ob
209
210
211
Julian Date
212
Figure 19. Comparison of eddy correlation measurements (¦ ) and Big Leaf
calculations (~) for SO2 Vd in Huntington Forest.
45
-------
integrated results for routine calculations may fall near the uncertainty values in Table 4 for
sites with good exposure.
Other sources of uncertainty arise in the NDDN dry deposition database because of the
NDDN weekly concentration sampling protocol, influence of complex terrain on deposition, and
uncertainty in meteorology and vegetation data. These topics are discussed briefly in the
following sections. The precision of the database and potential near-site spatial variability of dry
deposition fluxes are also discussed.
6.1 BIAS AND UNCERTAINTY FROM WEEKLY CONCENTRATION SAMPLING PROTOCOL
As described previously, weekly average deposition fluxes for SOz, HN03, SO| , and NOj
were calculated by multiplying a weekly average Vd by the weekly integrated concentration
(Section 4.2). A potential source of bias and uncertainty results if Vd and concentration are
correlated. This is illustrated in the following equation:
168 (4)
Weekly Flux = £ Vd*C = Vd*C + Vj*C'
where: C = hourly values of concentration,
Vd = hourly value of deposition velocity, and
L = summation over a week (168 hours).
The first term on the right-hand side (Vd* C ) is the product of the weekly average Vd and
weekly average concentration (dry deposition flux as calculated in the NDDN). The second term
on the right-hand side (Vd'*C') represents the covariance (correlation) between the hourly
concentrations and Vds, which was assumed to be small and ignored in the NDDN dry deposition
calculation procedure. There is concern that the covariance term may be large under certain
conditions. Meyers and Yuen (1987), using fall and winter data for a forested site near Oak
Ridge, TN, found that the product of weekly average S02 and weekly average Vd provided a very
good estimate of the weekly flux calculated as the sum of the product of hourly concentrations
and Vds. However, Matt and Meyers (1992) suggest the correlation term to be as large as
40 percent for S02 during the summer months at the same site. This amounted to an
approximately 20-percent undercalculanon of annual S02 flux using weekly concentrations.
46
-------
For NDDN applications, 03 fluxes were calculated using hourly concentrations and Vds and
should be a reasonable approximation of the true flux. However, for sulfur and nitrogen species,
fluxes were calculated as the product of weekly integrated concentrations and weekly averaged
Vds, so the correlation term may be important.
To assess the potential impact of the correlation effect on the NDDN dry deposition
calculations, an analysis of dry deposition data was conducted with 21 months of data from
January 1987 to September 1989 when the network collected weekly integrated filter packs for
separate daytime and nighttime periods. Fluxes were calculated as:
Fluxl = Vj(day)*C(day) + Vd (nig hi) *C(n ight) (5)
Flux2 = V.(week)*C(week) (6)
The ratio Flux2/Fluxl was calculated for each week of available data, and averages and
standard deviations were calculated by species and month. A composite for the network is
shown in Figure 20. Figure 20 shows the monthly average (over all eastern sites) ratio of
Flux2/Fluxl. Standard deviations of the ratios are shown in Figure 21. For S02, SO^", and
HN03, calculating flux by Equation 6 results in underestimation of the flux calculated by
Equation 5, and the undercakulation is a function of season. The undercalculation was largest
during the summer season. The correlation between concentration and Vd (represented by the
last term in Equation 4) was significant during the summer due to low concentration at night
being associated with low Vds and high daytime concentrations being associated with high Vds.
The undercalculation is a function of species, being largest for HN03 and generally smallest for
SO^'. The standard deviations also appear to be seasonally related with maximum values
occurring during the summer months (about 0.06 for S02 and 0.14 for HNOa).
Because of the large standard deviation of Flux2/Fluxl across eastern sites (Figure 21), an
additional analysis was completed for S02. Sites with similar annual variability of the
Flux2/Fluxl ratio were grouped together using cluster and principal component analysis and are
shown on Figure 22. There is a large variation of the correlation effect across groups. Group 1
47
-------
X
D
c\i
X
ID
<
CC
/A.
f\
^ V°3
/ \
Jf
°3 D/N—77—
10
MONTH
Figure 20. Ratio of Flux2 / Fluxi for indicated chemical species, averaged
over al: eastern sites, by month.
48
-------
/°3^-o^^X
=- S02
J 4 6 8 7 8 9
MONTH
1 : 12
Figure 21. Standard deviation of the ratio Flux2 / Fluxl by month.
49
-------
1.2
*
-J
U_
GROUP 4
CM
X
ZD
GROUP 3
_i
LL
o
<
DC
V GROUP 2
GROUP 1
0.8
0.7 -\ t 1 1 ; 1 —r——-——-—-—r 1
: 2 -5 4 b 6 7 6 9 101!
MONTH
Figure 22. The average ratio Flux2 / Fluxl for SO2 for groups of eastern sites.
50
-------
shows a major effect during the summer season, while Group 4 shows little or no iridication of a
correlation between concentration and V^. In general, the majority of sites fall into groups with
slight to moderate correlation effects, and there is little indication of a regional partem. Strong
variability of groupings occurs along the spine and western flanks of the Appalachian Mountains
from central Tennessee to northern New York. In this area, a number of Group 1 sites are
located adjacent to Group 4 sites. The Group 1 sites are all located in valleys in complex terrain,
while the Group 4 sites are located on mountaintops or at relatively high elevation in rolling
terrain (Figure 23).
A more accurate value of flux is obtained from the summation term in Equation 4 using
hourly values of concentration and Vds. This was only possible for 03 in the current analysis.
Figure 20 contains a curve labeled 03 HR. This curve is the ratio of the 03 fluxes calculated
using Equation 6 to that calculated using hourly concentrations and V^s (i.e., the summation
term in Equation 4). The ratio is about 0.05 lower than the ratio calculated using day/night
averaged concentrations (curve 03 D/N). This suggests that the calculations based on
Equation 5 may underestimate the fluxes for other species as well. For lack of other
information, the 0.05 adjustment factor was also applied to S02, HN03, and SO4". A summary of
the adjusted bias averaged over all sites is shown in Table 5. The biases are recommended for
seasonal adjustments to the database reported in Section 5.0 and Appendices B and C. The bias
may be higher or lower at individual sites. The magnitude of the variability is indicated by the
standard deviation (Figure 21).
TABLE 5. ESTIMATED BIAS OF WEEKLY FLUX CALCULATIONS
Summer Winter
Species
Mean (%)
Standard Deviation
Mean (%)
Standard Deviation
°3
0
0
0
0
so2
14
6
4
2
hno3
20
14
4
2
sol
5
4
2
2
51
-------
LEGEND
• Group 1
O Group 2
D Group 3
¦ Group 4
A Insufficient Data
Figure 23 Locations ot NOON sites by groupings identilied on Figure 22
-------
Thus, a negative bias was introduced into the NDDN flux calculations for S02 and HN03,
and to a smaller extent SO^\ because of the weekly integrated sampling of these species. The
bias is generally greatest during the summer and early fall when nocturnal stratification results
in both small Vds and low concentrations near the ground. Daytime mixing generally results in
higher Vds and concentrations. The current analysis indicated the largest bias is with HNOa.
Because of the dynamics of pollutant and interacting meteorology, it is expected that the bias
would be small for high elevation sites and near major low-level sources. Further analysis is
planned to more fully understand the nature of the bias. Our goal is eventually to correct for
the bias as a function sice, season, and pollutant species.
6.2 CALCULATION OF FLUXES FOR NON-IDEAL SITES
Direct measurements of deposition fluxes in complex and mountainous terrain are not
practical. This is also true for level sites with multiple vegetation species of radically different
heights. In such environments, the inferential model has had little verification. A concern in
complex environments is that of edge effects, i.e., the wind advecting pollutants into the exposed
edges of the canopy resulting in enhanced deposition. In addition, the turbulent environment in
many areas of complex terrain and vegetation may not be adequately represented by the NDDN
site measurements. For example, the wind and turbulent environment at an NDDN site located
in an open field may not be representative of the turbulent environment above a nearby
predominately forested area. The NDDN Big Leaf model adjusts wind parameters as a function
of surface roughness (Section 4.0); however, there are uncertainties in this procedure.
An indication of the magnitude of potential error associated with edge effects and complex
environments is suggested by setting Ra equal to 0 (in Equation 2) in the inferential model
computational process. is the aerodynamic resistance and a measure of the resistance to
transport to the near-surface environment of pollutants in the free atmosphere. By setting IL,
equal to 0, it is expected to provide an upperbound to the Vd.
The Big Leaf inferential model was run for five sites and approximately 1 year with Rg set to
0. Results are shown in Table 6 for 03, S02, and HN03. Ozone, S02., and HN03 Vds increased
on average 20, 27, and 93 percent, respectively. The smaller percentage increases for 03 and
S02 indicated that Vds for these species are dominated by the canopy resistance, whereas HN03
53
-------
TABLE 6. CHANGE IN Vd RESULTING FROM SETTING Ra = 0
AND WET = 0 FOR APPROXIMATELY 1 YEAR OF DATA
Site Vd (cm/sec)
No. 03 S02 HN03
121 BASE 0.19S 0.209 0.831
- 0 0.246 0.267 1.48
Wet = 0 0.207 0.208
% Change Ra = 0 26.2% 27.8% 78.1%
% Change Wet = 0 6.2% -0.5%
107 BASE 0.169 0.202 1.18
Ra = 0 0.23 0.276 2.75
Wet = 0 0.181 0.186
% Change 36.1% 36.6% 133.1%
% Change 7.1% -7.9%
151 BASE 0.187 0.242 0.879
Ra = 0 0.219 0.303 1.57
Wet = 0 0.198 0.229
% Change 17.1% 25.2% 78.6%
% Change 5.9% -5.4%
110 BASE 0.201 0.32 268
R3 = 0 0.217 0.365 a.14
Wet = 0 0.221 0 266
% Change 8.0% 14.1% 54.5%
% Change 10.0% -16.9%
120 BASE 0.171 0.313 154
Ra = 0 0.201 0.408 3.37
Wet = 0 0.2 0.209
% Change 17.5% 30.4% 118.8%
% Change 17.0% -33.2%
AVG (Ra = 0 21.0% 26.8% 92.6%
AVG (Wet = 0) 9.2% -12.8% 0.0%
STD (Ra = 0 9.5% 7.4% 28 9%
STD (Wet = 0) 4.1% 11.5% 0.0%
54
-------
Vd is dominated by turbulent processes, primarily through R^. The exercise (R., = 0) was not
valid for particles, but increases in particle Vds should be on the order of those for HN03. Thus,
at extremely complex sites, Vds and fluxes are probably undercalculated.
6.3 METEOROLOGY AND LAND-USE PARAMETERS
Dry deposition velocities are calculated in the inferential model based on meteorology and
vegetation variables. Two categories of uncertainties arise from these variables. The first is due
to errors in the measurement of meteorology and vegetation parameters, and the second is due
to misrepresentation of the role of these parameters in the deposition process. Only the first
category is described in this section. The second should be contained within the overall limits of
uncertainty of deposition model comparisons with direct observations (Table 4).
Model sensitivity to meteorology and vegetation variables was tested using a limited number
of sites and weeks of data, and this work is continuing. Temperature is an important model
variable with respect to the biological response of vegetation. Modeled Vds for S02 and 03 were
temperature dependent over ranges of temperature characteristic of diurnal and annual cycles.
However, at a given temperature, the model was reasonably insensitive to the normal range in
uncertainty in measuring the temperature (0.5°C) or near-site spatial variability (2 to 3°C).
Deposition velocities for HN03 and particles are not a function of temperature in the inferential
model. Consequently, temperature is not of concern in the uncertainty of the model. However,
when temperature data are missing due to damaged or out of calibration instruments, the
inferential model algorithm rejects all data for that period. An algorithm was developed to
substitute temperature data from a designated neighboring site. Tests of the procedure indicated
no significant difference in the dry deposition calculations using site temperature or designated
neighbor temperature.
Similarly, the model was reasonably insensitive to solar radiation within the expected
accuracy range of measurement (10 percent). Vegetation biological response, and consequently
the Vds for 03 and S02, are reasonably sensitive to solar radiation in the inferential model only
for values between 0 and 100 watts per square meter (W/m2). Deposition velocities for HN03
and particles are not directly dependent on solar radiation except as a switching indicator
between stable and unstable conditions. The designated neighbor approach also worked very
55
-------
well when solar radiation was missing at a particular site (for weekly or longer integration
periods).
Deposition velocity calculations are extremely sensitive to windspeed for HN03 and particles
and, to a smaller extent, for 03 and S02- The model is similarly sensitive to the standard
deviation of wind direction. These rwo parameters are used in the calculation of u. and
subsequently in Ra. Deposition velocities for HN03 and particles responded almost linearly to
changes in u and sigma theta. Uncertainty in wind variables can arise from different
instruments and network procedures. These include starting speed, exposure, calibration, and
maintenance. Uncertainties in the flux calculations will generally be larger under light wind
conditions where an uncertainty in windspeed will translate to a correspondingly large
percentage error in Vd for HN03 and particles.
According to the inferential model, Rc for S02 and 03 is very sensitive to surface wetness.
For example, Rc increases and decreases significantly for S02 and 03, respectively, when dew is
present on the canopy. The surface wetness sensor normally responds accurately to the presence
of dew or rain. However, the wetness measurement was of concern for reasons of
representativeness rather than accuracy. First, the sensor is located about 10 inches above the
ground and may underestimate wetness from dew in tall car.opies. Second, dew wetness was
not measured at the NDDN sites prior to 1989. To calculate dry deposition for NDDN data prior
to 1989, a predictive dew formation algorithm suggested by Wesely and Lesht (1989) has been
incorporated into the model. The algorithm predicts wetness based on windspeed and relative
humidity. It is believed that the position of the wetness sensor at the NDDN sites and the
Wesely algorithm (when applied to NDDN data) lead to underestimation of wetness. To analyze
the extreme of this effect, wetness was set to 0, and the model was run for six sites for 1 year.
The results are given in Table 6. Setting wetness to 0 (no rain or dew) increased Vd for 03 an
average of about 9 percent and decreased Vd for S02 by about 13 percent. Wetness had no
effect in the model for HN03 and particle deposition.
The effects of different instruments and procedures can be derived from several sites where
NOAA CORE and NDDN equipment are collocated. The NOAA sites are independently operated
using different meteorological instrumentation and procedures. Three weeks of meteorological
data (1 week each ir. spring, summer, and autumn) from NOAA and NDDN equipment at West
56
-------
Point, NY, were analyzed by the inferential model (the NDDN chemistry files were used in both
model runs). Calculated fluxes differed on the order of 4 percent for 03, 17 percent for S02, and
about 22 percent for HN03 and particles. The differences were primarily due to variations in
windspeed and the standard deviation of wind direction and were more pronounced under light
wind conditions. This finding is preliminary, and analysis of additional data for NOAA and
NDDN collocated sites is planned.
Model results for 03 and S02 were sensitive to the percent leafout and LAI. A 50-percent
uncertainty in percent leafout can result in a 70-percent uncertainty in Vd for an hourly period.
However, when integrated over seasonal and annual periods, uncertainty in percent leafout is
not expected to be large, and its impact on fluxes is not expected to exceed 10 to 15 percent.
6.4 ACCURACY AND PRECISION OF NDDN ATMOSPHERIC CONCENTRATION
MEASUREMENTS
The Big Leaf model calculates fluxes as the product of a measured concentration and a
calculated Vd. Uncertainties in concentration measurements therefore propagate directly into
flux calculations. This section provides estimates of precision and accuracy for NDDN
measurements of 03, S02, SO^", HN03, NOa, and NH^. It should be noted that such estimates
are generalized to the entire network, and uncertainties at individual sites may differ from those
indicated in this report.
Ozone is measured at NDDN sites via UV absorbance, which is an EPA-reference technique
traceable to an N1ST photometer housed at the EPA AREAL in Research Triangle Park, NC.
Quarterly calibrations of NDDN 03 analyzers showed them to be accurate (relative to the NIST
photometer) to within +.10 percent or +.4 ppb, whichever is greater. Precision data from
duplicate NDDN sites showed that 03 concentrations were typically reproducible within 1 to
2 ppb.
The accuracy of filter pack measurements is more difficult to specify. Of the species
measured, only S02 had a reference method, but the reference method targeted concentration
levels substantially higher than typically observed at NDDN sites. A number of field comparisons
have been conducted recently to evaluate filter pack accuracy (Dasch et ah. 1990; Sickles, 1991);
however, no study is sufficiently comprehensive to determine filter pack accuracy across the
entire NDDN. Results of these studies indicated that inrermethod agreement for SO? and HN03
-------
was generally on the order of _+5 and +,10 percent, respectively, for sites in the eastern United
States. The same studies uniformly suggested that the filter pack underestimates S02
concentrations by 10 to 15 percent. The reasons for this are unclear, but there is some evidence
that conditions of high relative humidity reduce filter pack collection efficiency for S02.
Preliminary results of a study conducted at several NDDN sites indicated that up to 10 percent of
ambient S02 may be lost at high relative humidity (i.e., >90 percent), due to penetration of S02
through the filter pack. Additional studies are ongoing to confirm the existence and magnitude
of this effect.
Methods comparisons for NOj aerosol showed that filter pack accuracy may be site specific.
In the midwest, where concentrations typically exceeded 2.0 jxg/m3, filter pack concentrations
were ^15 percent of other methods. In areas of low concentrations (i.e., <0.5 ng/m3),
agreement was only within _+20 to 30 percent. Thus, the accuracy of filter pack measurements
may be a strong function of concentration and other factors.
Chemical precision data for 1990 for four NDDN sites with duplicate samplers are shown in
Table 7. Results showed that the precision for SO^', S02, and NH^, as indicated by the median
absolute percent difference (MAPD), was on the order of 5 percent or less. Annual average
concentrations for duplicate sampiers differed by less than 2 percent for SO^" and NH^ and less
than 5 percent for S02. For HN03, MAPDs were on the order of 3 to 6 percent, and annual
concentrations of duplicate samplers differed by less than 6 percent.
Results for NOj were highly variable from site to site. MAPDs ranged from about 4 percent
for a midwestem site (mean concentration 2.3 ng/m3) to about 19 percent at a northeastern site
(mean concentration 0.6 (ig/m3). These results support the idea that the uncertainty of N03
measurements is concentration dependent within the range of the NDDN observations. An
apparent exception to this was observed at a western site (167), which exhibited the lowest N03
concentrations of the four duplicate sites. Differences in environmental conditions and aerosol
composition may promote spatial variability of precision data for NOj.
6.5 WITHIN AND BETWEEN-NETWORK PRECISION
For many applications, such as analysis of trends and spatial patterns, precision is an
important consideration. Six NDDN" sires have duplicate meteorological and concentration
measuring systems.
58
-------
TABLE 7. RESULTS OF 1990 COLLOCATED FILTER PACK SAMPLING
Total
Statistic SO$ NOs HNQ3 S02 NH$ NOj
Site 107 fn = 36)
x (Hg/m3)
7.82
0.63
2.02
11.7
1.84
2.65
Y (ng/m3)
7.85
0.54
1.91
11.2
1.84
2.45
%D
-0.4
15.4
5.6
4.4
0.3
7.8
MAPD
2.1
19.4
5.5
6.0
2.6
8.4
MAD (fig/m3)
0.11
0.10
0.10
0.62
0.04
0.21
Site 153 fn = 371
X (pg/m3)
6.97
0.40
2.39
6.95
1.85
2.73
Y (|ig/m3)
7.09
0.37
2.44
7.03
1.83
2.80
%D
1.7
7.8
-2.1
-1.1
1.1
-0.7
MAPD
4.2
16.4
5.7
4.7
4.0
6.2
MAD (ng/m3)
0.17
0.04
0.11
0.21
0.04
0.15
Site 157 fn = 391
X (jig/m3)
6.54
2.30
2.50
11.4
2.65
4.81
Y (ng/m3)
6.45
2.26
2.41
11.2
2.62
4.57
%D
1.4
1.8
3.7
1.8
1.1
2.7
MAPD
1.3
3.8
3.8
2.7
1.9
2.8
MAD (>ig/m3)
0.08
0.09
0.09
0.28
0.04
0.13
Site 167 fn = 521
X (ng/m3}
1.40
0.28
0.61
2.27
0.45
0.89
Y Cng/m3)
1.41
0.28
0.61
2.31
0.46
0.90
%D
-0.7
-0.9
-0.4
-1.7
-1.1
-0.6
MAPD
1.6
5.5
2.7
3.2
1.9
2.0
MAD (jig/m3)
0.02
0.01
0.02
0.05
0.01
0.02
Note: %D = percent difference of means = 200 x CX - Y)/(X - Y).
MAD = median absolute difference between paired weekly samples
= median (|x| - | Y|).
MAPD = median absolute percent difference between paired weekly samples
= median 200 ' X - Y
(X + Y)
n = number of valid weekly sanples above detection limit for both samplers.
X arid Y = mean concentration for primary and secondary samplers, respectively.
59
-------
Precision data for the six duplicate NDDN sites are given in Table 8. The table shows the
number of weeks of data for each site and mean absolute percent difference (and standard
deviation) for Vds, fluxes, and concentrations. These values are presented for each site and
species along with weighted (by number of weeks) averages by species over the six sites.
TABLE 8. PRECISION ANALYSIS FOR DUPLICATE NDDN SITES*
Site Number
°3
so2
hno3
sd\'
NOj
Vd
FLUX
CONC
Vd
FLUX
CONC
Vd
FLUX
CONC
Vd
FLUX
CONC
FLUX
CONC
167 25
0.009
0 023
0.024
0.013
0.052
0.054
0.014
0034
0.029
0 023
0.035
0017
0.078
0.08
156 8
0.052
0.053
0.091
0.088
0.126
0.059
0.026
0.072
0.056
0.024
0 065
0 061
0.097
0.103
114 23
0.03S
0.072
0.048
0.029
0.067
0.062
0,046
0.089
0.074
0.041
0.067
0.063
0.107
0.095
153 37
0 067
0 082
0.028
0.088
0.084
0.046
0.078
0 087
0.061
0.08
0 096
0.037
0 182
0.1S
107 33
0.034
0.056
0.038
0.033
0.0S6
0.045
0.071
0.1
0.11
0 084
0.094
0.021
0.213
0.214
157 35
0.218
0.129
0.016
0.059
0.053
0.023
0.269
0.27
0.05
0 222
0.223
0.02
0.214
0.04
W. Ave
0.057
0075
0.033
0.05
0.066
0 045
0.101
0.121
0.065
0 094
0 108
0.032
0 166
0.118
Standard Deviation
167 25
0.008
0.015
0.012
0.013
0.049
0047
0.011
0.021
0.024
0.026
0.031
0.011
0.065
0053
156 8
0.03
0.046
0 02
0.056
0.071
0 034
0027
0 051
0 054
0 021
0.036
0.033
0.13
0.112
114 23
0.049
0.049
0.007
0.038
0.053
0 054
0.051
0 065
0.072
0 047
0.056
0 055
0 066
0.075
153 37
0.076
0.096
0.05
0.09
0.09
0.04
0.077
0.08
0.038
0.1
0.1
002
0.153
0.115
107 33
0 041
0.058
0.026
0.034
0043
0.036
0.059
0.082
0.078
0 081
0.082
0016
0 159
0.134
157 35
0 124
0.127
0-13
0 033
0 038
002
0.134
0 158
0.06
0 095
0.111
0.01
0.12
0.O4
W. Ave
0.063
0 073
0.049
0.045
0.056
0.038
0.069
0.085
0.054
0.072
0.079
0.021
0.12
0087
•Precision Is equal to:
£2 i x, - JCj |/ (X~Txj)
Precision for concentration measurements was about 3 to 6 percent for all species except
NOj. Precision in calculating Vds was about 5 to 7 percent for 03 and S02 and about 10 percent
for HNO3 and particles. Precision in calculating fluxes was about 7 percent for 03 and S02,
12 percent for HN03 and SO^", and 17 percent for NOj. These values are reasonable and will
permit determination of relative spatial patterns and trends of dry deposition fluxes (i.e., without
regard to the absolutely accuracy of the data).
An analysis of between-nerwork precision is planned using five collocated NOAA CORE and
NDDN sites. These collocated sites are completely independent, using different instrumentation,
60
-------
calibration procedures, and vegetation monitoring procedures. Both groups, however, essentially
use the same inferential model.
6.6 SUMMARY OF ACCURACY AND PRECISION OF DRY DEPOSITION CALCULATIONS
Quantitative assessment of the uncertainty of dry deposition calculations is not practical at
this time. We are working toward that goal through continuing sensitivity analyses and have
initiated a program where direct measurements of deposition fluxes at selected NDDN sites will
be used to assess model performance. In the interim, however, the previous discussion/results
are suggested as a preliminary qualitative assessment of uncertainty of the database presented in
Section 5.0 and Appendices B and C. Table 9 provides a summary of these results.
TABLE 9. SUMMARY OF UNCERTAINTY OF DRY DEPOSITION CALCULATIONS.
Source 03 S02 HN03 Particles
Inferential Model 25%
Weekly Sampling
Bias None
Uncertainty None
Complex Terrain
Bias SU
Uncertainty Small
Meteorology
Wetness (Bias) SO
Wind (Uncertainty) Small
Concentration
Bias None
Uncertainty 10%
Precision
Vd 6%
Concentration 3%
Flux 8%
Note: MU - moderate under calculation.
SO = sma','. over calculation.
SU = small under calculation.
30% ^0% >40%
-5 to 15% -5 to 20% -2 to 5%
3 to 7% 5 to 14% 4 to 18%
SU MU MU
Small Moderate Moderate
SU None None
Small Moderate Moderate
-15% +10% None
10% 10% 10%
5% 10% 10%
5% 7% 3-12%
7% 12% 10-17%
61
-------
The initial assessment of uncertainty of the dry deposition velocity calculations was derived
from Table 4: 25 percent for 03, 30 percent for S02, and 40 percent or greater for HN03 and
particles. These values apply to level sites of uniform vegetation and for seasonal and annual
averaging periods. In addition, the Vds and fluxes are biased low because of the weekly
integrated concentration sampling protocol, and at complex sites, because of vegetation edge
effects and/or larger turbulent exchange than represented in the model. A conservative estimate
of bias due to the weekly sampling protocol was about 5 to 15 percent for S02 and 5 to
20 percent for HN03 for winter and summer data, respectively. Ozone should have no bias due
to this source, and SO^" and N03 showed very little bias. The bias may vary with site
characteristics, being smaller for complex and mountainous sites and those influenced by major
point sources. Hourly monitoring of S02 is planned for selected NDDN sites to examine this
issue further.
The enhanced Vds resulting from setting Ra equal to 0 (20 percent for 03, 27 percent for
S02, and 93 percent for HN03) should be considered upper limits for very complex sites. Little
is known about bias due to edge effects. For most sites and situations, Ra could not be expected
to go to 0. This is especially true at night when Ra is very high. Generally, it is suggested that
the fluxes are undercalculated, the undercalculation and uncertainty being smaller for SO, and
03 than for HN03 and particulates.
For the most part, it is assumed that uncertainties due to the measurement of meteorology
and vegetation variables are contained within the uncertainty values of Table 4. This is probably
not true for a mixture of different instruments and procedures and in complex environments,
where the measured windspeed and turbulence may not be appropriate to the physical
deposition processes as represented in the inferential model. Errors in the wind measurements
should affect HN03 and particulates more than 03 and S02. Uncertainty in the wetness sensor
suggests a small bias (undercalculation) for S02 but will have no effect on HN03 and
particulates.
Additional biases may arise from errors in filter pack measurements for SO^", N03, NH^,
HN03, and S02- Information is limited but suggests that SO^", N03, and HNO- data are
generally unbiased with respect to other (nonreferer.ee) methods. For SO,, the NDDN da:a may
62
-------
be biased low by 10 to 15 percent. Reasons for this bias are unclear, but it may be related to
collection efficiency reductions at high relative humidity.
Duplicate NDDN sites showed good precision. Similar instruments and procedures resulted in
consistent Vds and flux calculations. A similar analysis for NOAA/NDDN collocated sites will be
conducted in the future.
6.7 SPATIAL VARIABILITY OF DRY DEPOSITION
Apart from the uncertainties discussed previously, there can be significant spatial variability
in dry deposition fluxes. The Vds and fluxes reported in Section 5.0 and Appendices B and C
were areal averages representative of a 1-km radius surrounding the NDDN site. Dry deposition
may vary spatially due to variability in both concentration and Vd. To minimize these effects,
sire-selection criteria emphasize representative locations away from the direct influence of major
sources.
These siting criteria cannot always be achieved. An extreme example is Site 137 (Coweeta,
NC). This site is located in a valley about 300 m below and within 1,000 m of a ridge-top site,
specifically located to study the effects of terrain on dry deposition. Ratios of concentration
measurements from the ridge site to the NDDN site between June and December 1991 are
shown in Figure 24. Concentrations of SO^', S02, and HN03 averaged 1.2, 2.4, and 2.5 times
higher, respectively, at the ridge site. It is suggested that this is primarily the result of nocturnal
phenomena, including plume interaction with the ridge-top site, but not the valley site, and
shallow, longer lasting inversions at the valley site. Deposition velocities calculated for both
sires are shown in Table 10. '["he ratios of the average Vds for high and low sites were 1.3, 1.2,
and 1.67 for SO^", S02, and HN03, respectively, reflecting the higher levels of turbulence
normally observed at higher elevations. The Vds were generally smaller for all three species a:
the ridge site during the summer and higher at the ridge site during the fall and early winter.
Multiplying the concentration ratios by the Vd ratios indicated the fluxes to be 1.6, 2.9, and 4.2
rimes higher at the ridge site for SO^, S02.. and HN03, respectively.
63
-------
Ratio (Ridgetop Site/Routine Site)
£
5
O Sulfale
Sulfur Dioxide
A Nilric Acid
0
#
—r
£
£
o
£
1991
Figure 24. Ratio ol concentration measurements made at ridge-lop site and the
NDDN site (Site 137) located in a nearby valley.
-------
TABLE 10. DEPOSITION VELOCITIES FOR HIGH AND LOW
ELEVATION SITES, COWEETA, NC
Week
S02
hno3
Particles
Low
High
Low
High
Low
High
37
0.45
0.29
0.78
0.36
0.08
0.03
38
0.40
0.39
0.78
1.16
0.07
39
0.36
0.22
0.66
033
0.06
0.02
40
0.39
0.42
0.82
0.91
0.08
0.06
41
0.36
0.26
0.84
0.74
0.09
0.05
42
0.29
0.39
0.88
1.59
0.08
0.11
43
0.28
0.55
0.89
2.13
0.07
0.13
44
027
0.34
0.94
1.74
0.08
0.10
45
0.25
0.38
0.77
1.48
0.06
0.08
46
0.23
0.34
0.69
1.62
0.04
0.07
47
0.27
1.16
0.06
48
0.26
0.46
1.17
3.33
0.06
0.12
49
0.29
0.33
0.75
1.94
0.05
0.10
50
0.33
0.54
1.19
2.49
0.06
0.11
51
0.23
0.42
0.95
2.15
0.04
0.12
52
0.30
0.26
1.02
0.37
0.06
Average
0.31
0.37
0.89
1.49
0.07
0.08
Ratio: High/Low
1.20
1.67
1.30
The LAD model can provide information on the variability of Vd due to surface vegetation
(McMillen, 1990). Table 11 contains Vds for nine plant classes generated by the LAD model for
a summer daytime hour in Oak Ridge, TN. These values suggest the possible range in Vds due
to differences in vegetation over a large area surrounding the NDDN site. In reality, site Vds are
usually calculated based on two to five different vegetation types, which are often representative
of the larger area. At Oak Ridge, for example, the site Vrfs and fluxes were based on a
50/50 mixture of oak/hickory and loblolly pine. Since these were also the dominant species
over the iarger area (80 percent, see Table 10), the site-specific Vds and large area average Vds
should be reasonably consistent. Spatial variability of deposition fluxes also depends on spatial
variability of concentration, which is not considered in I.AD calculations.
65
-------
TABLE 11. DEPOSITION VELOCITIES (cm/sec) BY PLANT CLASS
GENERATED BY THE LAD MODEL FOR OAK RIDGE, TN.
Plant Percent S02 Vd 03 Vd HN03 Vd Particles Vd
Com
9
0.56
0.41
2.52
0.34
Soybeans
8
0.76
0.69
2.07
0.3
Wheat
1.6
0.80
0.73
2.07
0.3
Grass
0.5
0.87
0.83
1.76
0.28
Oak, hickory
39
1.14
1.08
4.30
0.44
Maple, birch, beech
0
0.90
0.79
4.08
0.43
Loblolly pine
38
0.82
0.70
3.90
0.42
Spruce
0
0.57
0.39
4.53
0.42
Pond, pine
0
0.38
0.16
3.78
0.42
66
-------
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.
Dasch, J.M., Cadle, S.H., Kennedy, K.G., and Mulawa, P.A. 1989. Comparison of Annular
Denuders and Filter Packs for Atmospheric Sampling. Atmos. Environ., 23(12), 2775-2782.
Dennis, R.L. 1992. Personal Communication with J.F. Clarke, Atmospheric Sciences Modeling
Division, Air Resources Laboratory. National Oceanic and Atmospheric Administration.
Research Triangle Park, NC.
Edgerton, E.S. and Lavery, T.F. 1990. National Dry Deposition Network Fourth Annual Progress
Report (1990). Prepared for the U.S. Environmental Protection Agency. Contract
No. 88-02-4551, Environmental Science 8c Engineering, Inc., Gainesville, FL.
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
(N'DDN) Laboratory Operations Manual. Prepared for the U.S. Environmental Protection
Agency. Contract No. 68-02-4451. Gainesville, FL.
Environmental Science & Engineering, Inc. (ESE). 1991. National Dry Deposition Network
(NDDN) Data Management Manual. Prepared for the U.S. Environmental Protection Agency.
Contract No. 68-02-4451. Gainesville, FL.
Hicks, B.B., Baldocchi, D.D., Hosker, R.P., Hutcherson. B.A., Matt, D.R., McMillen, R.T., and
Satterfield, L.C. 1985. On the Use of Monitored Air Concentrations to Infer Dry Deposition.
National Oceanic and Atmospheric Administration (NOAA) Technical Memorandum
ERLARL-141, 65 pp.
Huebert, B.J. and Robert, C.H. 1985. The Dry Deposition of Nitnc Acid to Grass. J. Geop'ny.
Res. 90(01), 2085-2090.
U-COR, Inc. 1989. LAI-2000 Plant Canopy Analyzer-Technical Information. U-COR, Inc., 4421
Supenor Street, Lincoln, NE.
Matt, D.R., McMillen, R.T., Womack, J.D., and Hicks, B.B. 1987. A Comparison of Estimated
and Measured S02 Deposition Velocities. Water, Air, and Soil Pollution, 36,331-347-
Matt, DR. and Meyers, T.P. 1992. On the Use of the Inferential Technique to Estimate Dry
Deposirion of S02- Submitted for Publication in Atmospheric Environment.
67
-------
McMiilen, R.T. 1990. Estimating the Spatial Variability of Trace Gas Deposition Velocities.
National Oceanic and Atmospheric Administration (NOAA) Technical Memorandum
ERLARL-181, 37 pp.
Meyers, T.P. 1992. Dry Deposition Measurements at Huntington Forest. National Oceanic and
Atmospheric Administration (NOAA], ATDD. Unpublished Manuscript.
Meyers, T.P., Hicks, B.B., Hosker, R.P., Womack, J.D., and Satterfield, L.C. 1991. Dry Deposition
Inferential Measurement Technique-II. Seasonal and Annual Deposition Rates of Sulfur and
Nitrate. Atmosphere Environment 25A, 10, pp. 2361-2370.
Meyers, T.P., Huebert, J., and Hicks, B.B. 1989. HN03 Deposition to a Deciduous Forest.
Boundary-Layer Met. 49, 395-410.
Meyers, T.P. and Yuen, T.S. 1987. An Assessment of Averaging Strategies Associated with
Day/Night Sampling of Dry-Deposition Fluxes of S02 and 03. J. Geophysic Res., 92,
6705-6712.
National Atmospheric Deposition Program (NADP)/National Trends Network. 1990. NADP/NTN
Annual Data Summary, Precipitation Chemistry in the United States, 1989. Available from
the National Resource Ecology Laboratory, Colorado State University. Fort Collins, CO.
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.
Sickles, J.E., Hodson, L.L., McClenny, W.A., Paur, R.J., Elstead, T.G., Mulik, J.D., Anlauf, K.G.,
Weibe, H.A., Mackay, G.I., Schiif, H.I., and Bubacz, D.K. 1990. Field Comparison Methods
for the Measurement of Gaseous and Particulate Contributors to Acid Dry Deposition. Atmos.
Environ., 24(A)(1), 155-164.
Wesely, M.L. and Lesht, B.M. 1989. Comparison of RADM Dry Deposition Algorithm with a
Site-Specific Method for Inferring Dry Deposition. Water, Air, and Soil Pollution, 44,
273-293.
68
-------
APPENDICES
Data processing for the NDDN Big Leaf model is summarized in Appendix A along with
model output options. Examples of the vegetation input files are presented and briefly
explained.
Dry deposition data for 1990 and 1991 are given in Appendices B and C, respectively. The
data are presented by site and by season, and an annual value is given when all four seasons
contain at least 9 weeks of data (data completeness criteria and definition of seasons are given
in Appendix A). The first column specifies the averaging period: winter, spring, summer,
autumn, and annual. The second column contains the site number (see Table 1 for name and
location). The third column contains the average seasonal (or annual) 03 deposition velocity
(cm/sec), and the fourth contains the number of valid weeks of meteorological data making up
the season (a valid week requires 70 percent or more of the hourly data). Seasonal data are
assumed valid if they were developed from greater than 8 weeks of data, and annual data are
reported only if there are four valid seasons. Columns 5 and 6 contain 03 dry deposition flux
data (kg/ha) and the number of valid weeks (the number of weeks may differ from those in
Column 4 if hourly concentration data are missing for periods when the meteorological data files
were complete). The flux data in Column 5 have been aggregated to 13 weeks. Columns 7 and
8 contain seasonal average 03 concentration data (ppb) and the number of weeks making up the
average.
The S02 data have the same format as described for 03. The number of weeks making up
the seasonal Vd is not given since it is the same as given in Column 4 for 03. The number of
weeks associated with the flux and concentration data may differ from those for 03. The
concentration data in Column 13 are in units of ng/m3 and are referenced to 25°C and
1 atmosphere. The format for HN03 and SC>4 deposition velocities, fluxes, and concentrations is
the same as for S02- The number of weeks making up each season is not given since they are
the same as for S02. The deposition velocities for S04 are assumed to apply to N03.
69
-------
APPENDIX A
DATA PROCESSING AND OUTPUT OPTIONS
The three inferential models [i.e., Big Leaf, Multilayer, and Large Area Deposition CLAD)]
require meteorological, chemical, and vegetation input data. The meteorological and chemical
input files are the same for all of die models: windspeed, standard deviation of wind direction,
temperature, solar radiation, relative humidity, precipitation, and surface wetness are required
hourly. Of the chemical species, 03 is required hourly and the other species are read in as
integrated weekly values.
A meteorological/chemical preprocessor prepares the data for application in the inferential
model. In addition, the preprocessor checks for completeness of input data and, for some
missing meteorological variables (e.g., temperature and solar radiation), it substitutes data from
an appropriate neighboring site. If surface wetness is missing, an option is provided to calculate
dew wetness. The processed meteorological and chemical data are input in the model(s) along
with three vegetation files, which are described in the following paragraphs.
A.1 VEGETATION FILES
There are three vegetation input files: PLANT.DAT, STAT10N.DAT, and LEAF.DAT. All three
files are identically input to both the Big Leaf and Multilayer models. The LAD uses different
input files; howe\'er, the information passed to the model is the same.
PLANT.DAT-This file is shown in Table A-l. It contains physical and biological parameters
necessary to calculate Rc. There are 35 different plant species represented, including 16 species
of deciduous trees, 12 species of conifer, and 7 species of grasses and crops. For each species,
the file contains the name; minimum stomatal resistance; light response coefficient; optimal,
majtimum, and minimum plant response temperature; an index (1 to 3) specifying the vertical
profile of biomass for that species; and the height of the species (the latter two parameters are
only used in the Multilayer model). The information contained in this file were either received
from NOAA or extrapolated to similar plant types. The file can be easily updated when new
information becomes available.
70
-------
STATION.DAT--The first page of the STAT10N.DAT file is shown in Table A-2. This file is
organized by site. The first line for each site contains the site abbreviation and number, name,
latitude, longitude, time zone, and the number of plant types designated for the site. For each
plant type, a plant number is given (e.g., 11 corresponds to 'GRASS' in Table A-l), the percent
cover of that species at the site (within a 1-kilometer radius), maximum summer LAI, winter LAI,
and surface roughness length in centimeters. The percent coverage and plant type are based on
aerial photographs and physical identification at the site. LAI measurements are based on
measurements at each site, and the surface roughness is estimated from height and type of
vegetation (Section 3.0). These files are updated as vegetation changes. The file structure is
identical for the Big Leaf and Multilayer models. A different file format is required for the LAD
model, but it will not be discussed in this appendix.
L£AF.DAT--The first page of the LEAF.DAT file is shown on Table A-3. It is organized by site
and the first line for each site contains the site abbreviation and number, and the number of
periods in the year the site is divided into for the reporting of percent green leaf data.
Subsequent lines each contain beginning and ending dates, and a percent green leaf associated
with that period for each plant species contained in the STATION-DAT file. The files are not
specific to a given year but may be in the future. Values in the file vary significantly from site to
site. See Figure 3, for example, where leafout duration at northern sites may be 2 months or
more shorter than at southern sites. This specific format is used in the Big Leaf and Multilayer
models. In the LAD model, however, percent green leaf is modulated only as a function of
species and not by site.
A.2 OUTPUT OPTIONS
The data in the Appendices B and C represent the standard output format for dry deposition
data. Both seasonal and annual dry deposition velocities (Vds), fluxes, and ambient
concentrations are routinely output for 03, S02, HN03, SO|", and NOj. The column labeled (N)
gives the number of weeks of valid data that make up the seasonal average (there are
13 possible weeks in each season). Annual values are not given unless all four seasons contain
9 or more weeks of valid data. A week is considered valid if 70 percent or more of VjS are
calculated from the hourly meteorological files, and if concentration data are available.
71
-------
Deposition velocities and concentrations are averaged, and fluxes are aggregated up to
13-week seasons regardless of the number of weeks of valid data. For example, if the calculated
flux is based only on 7 weeks of valid data, the flux reported is the calculated value multiplied
by 13/7; the number of valid weeks, however, is reported as 7. The user should carefully
consider the number of weeks going into the seasonal and annual values when using the data.
Deposition velocities are reported as centimeters per second (cm/sec); fluxes as kilograms per
hectare (kg/ha) per time period (weekly, seasonal, or annual), and concentrations as
micrograms per cubic meter (^g/m3) except for 03 where it is reported in parts-per-billion
(ppb) units.
The seasons are defined as weeks starting within the following periods:
7
Winter: 22rDecember-21 March
Spring: 22 March-21 June
Summer: 22 June-21 September
Autumn: 22 September-21 December
Other model results routinely archived are the weekly values for Vds, fluxes, and
concentrations for sites and weeks meeting the 70-percent minimum data capture requirement.
Special output options include hourly Vds for ail chemicai species, hourly 03 fluxes, and Vds and
fluxes for individual vegetation species contained in the STATION.DAT file for the specified site.
However, these require addition runs of the model and some modification of the code.
72
-------
TABLE A-l. PLANT.DAT FILE FOR BIG LEAF AND MULTILAYER MODELS
'SPRUCE,r,225,40,9,35,-5,2,23
"POND/LODGP PINE, 1',500,40,25,40,5,3,23
•LOBLOLLY PINE,1',200,55,25,40,5,3,23
•WHITE OAK, 1',100,50,25,45,5,2,23
'CHEST/N.RED OAK, 1', 100,40,25,45,5,2,23
'MAPLE, 1',100,50,25,45,5,2,23
•WHITE BIRCH,2',300,40,25,40,5,2,23
'MAIZE',250,65,25,45,5,1,2.5
•WHEAT, 1', 100,25,25,40,5,1,1
'SOYBEAN1,100,50,25,45,10,1,1
'GRASS',50,20,25,45,5,1,0.5
'BLUE GRASS', 150,50,30,40,5,1,0.2
'SUGAR MAPLE,2',100,50,25,45,5,2,20
'BEECH, l',l00,50,25,40,5,2,23
•YELLOW BIRCH*,300,40,25,40,5,2,23
"WHITE ASH,100,40,25,40,5,2,23
'HEMLOCK*,225,10,25,35,-5,2,23
"YELLOW POPLAR*',150,40,25,40,5,2,23
'GUM*',150,40,25,40,5,2,20
•APPLE,PEACH,PEAR*', 150,40,25,40,5,2,10
'BLACK LOCUST*', 150,40,25,40,5,2,23
VIRGINIA PINE',2',200,55,25,40,5,3,23
'ALFALFA*',50,20,25,45,5,1,1
'RED PINE*',200,55,25,40,5,3,23
•SOUTHERN RED OAK*,2',100,40,25,40,5,2,23
'SOUTHERN YELLOW PINE',2',200,55,25.40,5,3,23
-WHITE PINE*',225,40,25,35,-5,3,23
'SUBALPINE FIR",225,25,9,35,-5,3,20
'SAGEBRUSH*',100,20,25,45,5,1,1
'JUNIPER*',225,25,9,35,-5,3,10
"VELVET ASH*',100,40,25,40,5,2.23
'EMORY OAK*'. 100,25,25,45,5,2,23
'.ARIZONA CYPRESS*',225,25,25,45,5„2;23
(Continued)
73
-------
TABLE A-l. (Continued)
'PINON PINE*',225,25,9,35,-5,3,20
'WATER*', 1,1,-99,-99,-99,1,0.1
'ROCK*', 1,1,-99,-99,-99,1,0.1
•ASPEN*',200,30,25,35,5,2,10
74
-------
TABLE A-2.
DATA FOR THE FIRST FOUR SITES IN THE STATION.DAT FILE
'ALH1577ALHAMBRA,IL',38.869,89.625,6,4
8,30,5.5,0.0,20
9,30,3.0,0.0,10
10,30,3.5,0.0,5
11,10,2.5,0.0,5
'ANA115VANN ARBOR,MI',42.417,83.902,6,4
4,30,4.5,0.0,100
5,30,4.5,0.0,100
11,28,2.5,0.0,20
27,12,3.6,2.7,100
'ANL1467ARGONNE NATL LAB,IL',41.701,87.995,6,3
4,25,4.5,0.0,100
11,50,2.5,0.0,5
13,25,4.5,0.0,100
'ARE 128','ARE NDTSVILLE,PA',39.923,77.307,5,4
4,15,4.3,0.0,100
5,15,4.3,0.0,100
11,20,2.5,0.0,20
20,50,3.5,0.0,75
73
-------
TABLE A-3. DATA FOR THE FIRST FOUR SITES IN THE LEAF.DAT FILE
'ALH157.11
1,90,25,0,0,0
91,104,25,0,0,0
105,134,50,0,25,25
135,151,75,50,50,50
152,181,100,75,75,75
182,243,100,100,100,100
244,259,80,75,75,75
260,273,60,25,25,25
274,289,40,0,0,0
290,304,25,0,0,0
305,366,25,0,0,0
'ANA115M1
1,90,25,0,0,75,100
91,110,25,0,0,75,100
111,130,40,25,25,85,100
131,151,60,50,50,90,100
152,166,80,75,75,100,100
167,250,100,100,100,100,100
251,266,80,75,75,90,100
267,273,60,50,50,80,100
274,280,40,25,25,75,100
281,304,25,0,0,75,100
305,366,25,0,0,75,100
'ANL146',12
1,75,25,0,0
76,105,25,0,0
106,120,50,20,25
121,135,75,40,50
136,151,100,60,75
152,165,100,80,100
166,259,100,100,100
260,273,80.75,75
(Continued)
76
-------
TABLE A-3. (Continued)
274,289,60,50,50
290,304,40,0,0
305,320,25,0,0
321,366,25,0,0
'ARE 128*, 11
1,75,25,0,0,0
76,91,25,0,0,25
92,105, SO,25,25,25
106,120,75,50,50,50
121,135,100,75,75,75
136,260,100,100,100,100
261,275,100,75,75,80
276,290,80,50,50,60
291,304,60,25,25,40
305,320,30,0,0,0
321,366,25,0,0,0
77
-------
APPENDIX B
1990 NOON DATA
OZONE
S02
HN03
S04
N03
SEA
SITE
Vd
N
FLUX
N
CONC
N
Vd
FLUX
CONC
N
Vd
FLUX
CONC
Vd
FLUX
CONC
FLUX
CONC
WIN
157
0.06
13
2.52
13
25.12
13
0.07
0.69
13.52
13
0.80
0.56
0.86
0.06
0.14
3.37
0.17
3.94
SPR
157
0.16
13
11.50
13
38.32
13
0.19
1.31
8.61
13
1.03
1.98
2.41
0.08
0.36
5.77
0.11
1.94
SUM
157
0.28
12
23.46
11
41.74
12
0.32
3.14
12.05
13
0.94
3.14
4.14
0.10
0.79
10.38
0.08
1.02
AUT
157
0.06
6
2.36
6
25.86
12
0.06
0.70
13.34
12
0.72
0.71
1.39
0.05
0.13
3.28
0.11
2.65
ANN
32.76
11.88
2.20
5.70
2.39
WIN
115
0.07
13
2.87
13
24.76
13
0.21
1.71
10.70
13
2.01
2.41
1.55
0.10
0.27
3.47
0.24
3.21
SPR
115
0.20
13
14.80
13
39.81
13
0.38
1.72
5.89
13
2.07
4.20
2.71
0.15
0.59
4.87
0.16
1.29
SUM
115
0.31
13
22.91
13
37.03
13
0.50
2.12
5.36
13
1.57
3.82
3.15
0.12
0.73
7.94
0.05
0.48
AUT
115
0.09
12
3.41
12
22.52
12
0.23
1.53
8.74
12
1.95
2.34
1.57
0.10
0.24
3.05
0.17
2.20
ANN
0.17
43.99
31.03
0.33
7.09
7.67
1.90
12.78
2.25
0.12
1.82
4.83
0.61
1.80
WIN
146
0.07
10
1.91
10
18.42
12
0.09
1.33
17.36
12
2.11
1.36
0.82
0.10
0.28
3.64
0.36
4.46
SPR
146
0.26
11
16.37
11
30.29
13
0.34
3.99
14.79
13
2.46
4.08
2.41
0.17
0.73
5.62
0.39
2.86
SUM
146
0.42
11
28.77
11
31.05
11
0.60
7.39
15.60
11
1.97
6.28
4.09
0.15
1.22
10.79
0.16
1.47
AUT
146
0.13
12
5.38
12
17.46
12
0.18
2.38
17.66
11
2.23
1.89
1.07
0.11
0.27
3.34
0.29
3.76
ANN
0.22
52.43
24.30
0.30
15.08
16.35
2.19
13.60
2.10
0.13
2.51
5.85
1.21
3.14
WIN
128
0.08
11
3.24
11
24.57
13
0.09
1.40
19.68
13
2.00
4.19
2.79
0.12
0.40
4.44
0.22
2.70
SPR
128
0.36
8
29.30
8
44.67
13
0.46
3.09
10.03
13
2.48
7.23
3.81
0.18
1.00
6.71
0.20
1.55
SUM
128
0.35
13
29.41
13
46.05
13
0.45
3.63
10.38
13
1.91
6.61
4.41
0.14
1.22
11.03
0.06
0.50
AUT
128
0.14
12
7.54
12
28.52
12
0.16
1.85
17.18
12
2.22
5.15
3.07
0.13
0.42
4.26
0.16
1.68
ANN
35.95
14.32
3.52
6.61
1.61
WIN
135
0.C8
13
4.39
13
32.92
13
0.12
0.37
4.17
13
2.31
1.51
0.84
0.11
0.20
2.26
0.02
0.26
SPR
135
0.20
13
12.85
13
36.86
13
0.33
0.31
1.29
13
2.44
1.16
0.59
0.17
0.35
2.55
0.03
0.19
SUM
135
C.35
10
17.03
10
26.06
13
0.47
0.26
0.68
13
1.86
0.68
0.43
0.15
0.25
2.12
0.01
0.12
AUT
135
0.09
2
4.12
2
26.55
12
0.20
0.42
1.45
12
2.88
1.82
0.56
0.11
0.13
1.66
0.02
0.23
ANN
30.59
1.90
0.61
2.15
0.20
WIN
116
0.10
13
3.37
13
17.88
13
0.17
2.23
16.71
13
2.10
3.15
1.91
0.12
0.41
4.38
0.16
1.74
SPR
116
0.23
13
17.22
9
31.92
9
0.35
2.45
9.29
13
2.25
5.80
3.36
0.17
0.88
6.53
0.08
0.63
SLM
116
0.34
2
31.74
2
33.77
13
0.43
4.43
8.22
13
2.08
7.68
3.80
0.19
1.58
10.45
0.07
0.38
A~T
116
0.14
11
5.04
11
16.25
11
0.21
2.15
13.76
12
1.89
2.55
2.04
0.12
0.32
4.09
0.08
0.83
AN»J
24.96
12.CO
2.78
6.36
0.90
w;n
130
0.06
12
2.13
9
23.35
9
0.36
0.42
9.C8
12
0.81
0.65
1.01
0.05
0.14
3.45
0.16
3.96
SPR
130
C.05
5
2.86
4
38.70
12
0.05
0.33
7.39
13
0.72
1.40
3.23
0.05
0.21
5.75
0.12
2.02
SUM
130
0
37.40
13
7.81
13
3.85
10.25
AUT
130
0.07
12
2.98
12
22.31
12
0.07
0.52
9.12
12
0.96
1.03
1.26
0.06
0.17
3.34
0.15
3.41
ANN
30.44
8.35
2.34
5.70
w I N
150
0.10
13
4.73
13
26.58
13
0.19
0.37
2.44
13
1.29
1.11
1.07
0.08
0.15
2.37
0.C4
0.59
SPR
150
0.27
13
16.46
13
30.29
13
0.40
0.46
1.50
13
1.29
1.42
1.42
0.11
0.34
4.14
0.02
0.25
78
-------
APPENDIX B
1990 NDDN DATA
(continued)
SUM
150
0.29
13
20.26
13
30.68
13
0.40
0.67
2.16
13
1.09
1.50
1.75
0.10
0.48
6.10
0.02
0.20
AUT
150
0.13
12
6.90
12
23.11
12
0.21
0.29
1.80
12
1.00
0.91
1.20
0.07
0.17
3.10
0.02
0.34
ANN
0.20
48. 34
27.67
0.30
1.79
1.98
1.17
4.94
1.36
0.09
1.14
3.93
0.09
0.35
WIN
119
0.07
13
2.41
13
21.40
13
0.09
0.78
12.14
13
1.21
1.41
1.48
0.06
0.19
3.48
0.02
0.32
SPR
119
0.26
12
17.84
12
31.70
13
0.34
2.28
8.67
13
1.23
1.65
1.68
0.09
0.49
6.74
0.01
0.15
SUM
119
0.33
13
22.09
13
29.13
13
0.37
2.08
6.91
12
0.82
1.14
1.79
0.07
0.70
13.39
0.00
0.06
AUT
119
0.11
12
5.09
12
19.08
12
0.15
1.09
11.70
12
1.02
1.08
1.35
0.06
0.22
4.60
0.01
0.26
ANN
0.19
47.43
25.33
0.24
6.22
9.86
1.07
5.28
1.58
0.07
1.60
7.05
0.04
0.20
WIN
167
0.16
13
10.28
12
39.17
12
0.20
0.33
2.07
13
2.27
0.66
0.37
0.16
0.12
0.92
0.03
0.22
SPR
167
0.24
13
18.45
13
46.14
13
0.25
0.36
1.85
13
2.82
1.54
0.69
0.26
0.25
1.20
0.09
0.45
SUM
167
0.22
13
15.41
13
41.37
13
0.27
0.32
1.52
13
2.43
1.79
0.94
0.18
0.31
2.17
0.03
0.24
AUT
167
0.12
12
7.75
12
38.38
12
0.13
0.35
3.73
12
2.34
0.85
0.45
0.16
0.17
1.37
0.02
0.20
ANN
0.19
51.89
41.27
0.21
1.36
2.29
2.47
4.83
0.61
0.19
0.85
1.42
0.18
0.28
U1N
169
0.08
10
5.87
10
45.79
12
0.12
0.07
0.78
12
3.37
0.98
0.43
0.22
0.11
0.72
0.02
0.13
SPR
169
0
50.35
12
0.49
12
0.55
0.99
SUM
169
0
47.16
13
0.59
13
0.69
1.10
AUT
169
0
42.75
12
0.71
12
0.41
0.65
ANN
46.51
0.64
0.52
0.87
U1N
137
0.08
12
4.50
11
29.24
12
0.11
0.23
2.83
13
1.07
0.74
0.90
0.07
0.14
2.39
0.01
0.24
SPR
137
0.22
13
16.89
13
38.70
13
0.26
0.38
1.94
13
1.07
1.01
1.19
0.09
0.40
5.41
0.01
0.13
SUM
137
0.25
13
15.73
13
27.18
13
0.34
0.26
0.94
13
0.80
0.55
0.86
0.07
0.55
9.53
0.00
0.06
AUT
137
0.11
9
5.92
9
25.50
11
0.15
0.23
2.04
11
0.94
0.64
0.79
0.07
0.19
3.45
0.01
0.15
ANN
0.17
43.04
30.16
0.22
1.10
1.94
0.97
2.94
0.94
0.08
1.27
5.20
0.03
0.15
WIN
110
0.07
13
3.50
11
29.53
11
0.09
1.22
16.86
13
1.96
4.55
3.02
0.10
0.26
3.55
0.06
0.72
SPR
110
0.24
13
18.06
13
44.71
13
0.35
2.17
8.20
13
2.03
4.04
2.51
0.14
0.58
5.21
0.05
0.45
SUM
110
0.35
13
24.59
13
41.93
13
0.52
3.37
8.13
13
1.68
3.32
2.55
0.12
0.70
7.62
0.03
C.27
AUT
110
0.10
12
5.13
12
29.37
12
0.15
1.73
15.33
12
1.96
4.56
3.01
0.10
0.30
3.91
0.05
0.68
ANN
0.19
51.29
36.39
0.28
8.48
12.13
1.91
16.47
2.77
0.11
1.85
5.07
0.19
0.53
WIN
151
0.11
12
6.58
12
34.57
12
0.16
0.49
3.97
12
1.34
1.40
1.37
0.08
0.18
2.94
0.04
0.69
SPR
151
0.24
13
16.76
13
40.59
13
0.29
0.50
2,28
13
1.15
1.47
1.65
0.09
0.36
5.19
0.02
0.29
SUM
151
0.25
12
19.61
12
45.12
13
0.29
0.70
3.09
13
0.36
1.21
1.83
0.07
0.52
9.32
0.01
0.16
AjT
151
0.11
12
7.55
12
37.56
12
0.13
0.51
4.70
10
0.91
1.22
1.70
0.06
0.19
3.77
0.02
0.53
ANN
0.18
50.50
39.46
0.22
2.21
3.51
1.06
5.30
1.64
0.07
1.24
5.31
0.10
0.42
WIN
114
0.07
1
2.08
1
24.26
13
0.21
1.81
12.43
13
1.75
2.96
1.57
0.08
0.22
3.40
0.10
2.42
SPR
114
0.27
12
22.49
12
43.07
13
0.50
3.75
9.86
13
2.01
4.79
3.07
0.15
0.77
6.38
0.18
1.66
SL'M
114
0.35
8
26.28
8
40.37
13
0.54
6.01
13.32
13
1.63
5.78
4.69
0.13
1.20
13.29
0.07
0.54
AUT
114
0.13
11
6.46
11
23.74
11
0.28
2.82
13.93
12
1.91
3.05
2.20
0.11
0.35
4.11
0.17
1.98
ANN
32.86
12.39
2.88
6.80
1.65
WIN
127
0.07
13
3.86
13
32.51
13
0.19
1.29
8.51
13
1.21
1.75
1.88
0.C7
0.17
3.15
0.05
0.84
SPR
127
0.21
13
16.21
13
43.31
13
0.41
1.68
5.42
•3
1.13
1.91
2.16
0.09
C. 43
6.46
0.02
0.28
79
-------
APPENDIX B
1990 NDON DATA
(continued)
SUM 127 0.20 12 16.19 12 a.16 13 0.37 1.34 4.43 13 0.76 1.16 1.98
AUT 127 0.11 12 6.62 12 32.17 12 0.25 1.40 7.87 12 1.08 1.66 1.95
ANN 0.15 42.88 38.04 0.31 5.71 6.56 1.04 6.48 1.99
0.04 0.34 11.61 0.00 0.11
0.07 0.24 4.59 0.03 0.51
0.07 1.17 6.45 0.10 0.44
WIN
153
0.11
12
6.66
12
35.23
12
0.20
1.27
7.84
13
2.07
2.88
1.84
0.13
0.32
3.21
0.08
0.76
SPR
153
0.28
13
23.33
11
46.22
11
0.44
2.30
6.93
13
2.03
4.04
2.54
0.17
0.88
6.51
0.05
0.32
SUM
153
0.29
13
28.01
13
49.72
13
0.45
2.24
6.39
13
1.74
3.84
2.85
0.14
1.28
12.00
0.01
0.12
AUT
153
0.16
12
10.63
12
34.77
12
0.24
1.70
9.69
12
1.72
3.11
2.40
0.12
0.41
4.59
0.04
0.46
ANN
0.21
68.63
41.49
0.33
7.52
7.71
1.89
13.87
2.41
0.14
2.88
6.58
0.18
0.42
WIN 174 0.14 13 9.61 13 44.24 13
SPR 174 0.16 13 13.00 13 51.60 13
SUM 174 0.15 13 11.06 13 48.83 13
AUT 174 0.14 12 9.39 12 44.65 12
ANN 0.15 43.06 47.33
0.12 0.09 0.94 13
0.15 0.06 0.53 13
0.14 0.06 0.59 13
0.11 0.11 1.34 12
0.13 0.33 0.85
1.42 0.80 0.72
1.87 1.22 0.83
1.59 1.48 1.20
1.30 0.88 0.88
1.55 4.37 0.91
0.14 0.08 0.72 0.03 0.27
0.25 0.20 1.03 0.12 0.55
0.19 0.21 1.43 0.06 0.40
0.14 0.10 0.89 0.03 0.23
0.18 0.59 1.02 0.24 0.36
WIN
168
0.08
5
3.37
5
28.21
12
0.16
0.09
0.99
13
1.01
0.28
0.32
0.05
0.04
0.85
0.01
0.26
SPR
168
0.15
13
7.82
13
28.41
13
0.23
0.09
0.51
13
1.19
0.31
0.34
0.09
0.06
0.77
0.01
0.08
SUH
168
0.20
12
9.82
12
23.09
13
0.26
0.11
0.54
13
0.93
0.27
0.38
0.08
0.05
0.76
0.01
0.09
AUT
168
0.08
9
2.63
9
22.19
12
0.16
0.07
0.62
12
0.84
0.16
0.24
0.04
0.01
0.56
0.01
0.23
ANN
25.48
0.67
0.32
0.74
0.17
U1N
161
0.07
13
5.30
9
47.07
9
0.08
0.02
0.34
13
2.48
0.62
0.32
0.15
0.07
0.61
0.01
0.11
SPR
161
0.18
13
14.49
13
50.44
13
0.21
0.06
0.31
13
2.69
0.90
0.43
0.20
O.K
0.90
0.05
0.27
SUM
161
0.27
13
18.93
13
39.29
13
0.35
0.08
0.30
13
2.43
0.77
0.39
0.18
0.15
1.05
0.02
0.15
AUT
161
0.10
11
6.22
11
40.50
11
0.12
0.03
0.29
11
2.33
0.42
0.23
0.13
0.06
0.55
0.01
0.08
ANN
0.16
44.95
44.33
0.19
0.18
0.31
2.48
2.71
0.34
0.17
0.42
0.78
0.09
0.15
WIN
112
0.07
13
3.01
8
27.76
8
0.08
1.54
23.62
11
1.59
3.27
2.60
0.08
0.23
3.79
0.04
0.61
SPR
112
0.20
13
15.73
12
43.65
12
0.29
2.04
9.16
12
1.50
2.68
2.20
0.11
0.48
5.70
0.02
0.21
SUH
112
0.25
13
19.03
12
39.38
12
0.40
2.50
7.75
13
0.88
1.45
2.13
0.07
0.48
9.53
0.01
0.11
AUT
112
0.10
12
5.20
12
29.54
12
0.16
2.73
22.40
12
1.37
2.71
2.59
0.07
0.24
4.56
0.03
0.55
ANN
0.15
42.97
35.08
0.23
8.82
15.73
1.34
10.12
2.38
0.08
1.44
5.90
0.10
0.37
WIN
121
0.08
13
3.23
13
21.30
13
0.09
0.35
4.93
13
1.05
0.46
0.57
0.06
0.16
3.08
0.04
0.81
SPR
121
0.28
13
16.70
13
27.31
13
0.31
0.60
2.60
13
0.94
0.76
1.02
0.08
0.32
5.38
0.02
0.34
SUM
121
0.30
13
16.71
13
23.00
13
0.31
0.45
1.82
13
0.62
0.47
0.96
0.06
0.46
10.57
0.01
0.17
AUT
121
0.13
12
5.70
12
17.63
12
0.13
0,28
3.03
12
0.70
0.40
0.74
0.05
0.16
3.95
0.01
0.33
ANN
0.20
42.34
22.31
0.21
1.68
3.10
0.83
2.09
0.82
0.06
1.10
5.75
0.08
0.41
hlN 117 0.08 13 3.39 13 23.48 13
SPR 117 0.27 13 20.30 13 36.98 13
SUM 117 0.32 13 23.77 13 32.79 13
AUT 117 0.13 12 6.44 12 22.79 12
ANN 0.20 53.91 29.01
0.12 2.33 25.22 13
0.36 3.45 12.49 13
0.39 3.23 10.35 13
0.17 2.63 22.40 12
0.26 11.65 17.62
1.40 2.50 2.28
1.53 2.98 2.49
1.06 2.10 2.52
1.43 2.15 1.89
1.36 9.73 2.30
0.07 0.22 3.76 0.03 0.44
0.11 0.55 6.38 0.02 0.26
0.08 0.72 11.21 0.01 0.10
0.08 0.28 4.45 0.03 0.42
0.09 1.77 6.45 C.08 0.31
WIN 123 0.08 13 3.14 13 25.65 13
SPR "23 0.23 13 17.24 13 41.27 13
0.22 2.06 12.40 13
0.41 2.57 8.51 13
2.15 2.33 1.41
2.00 4.50 2.91
0.11 0.31 3.62 0.3' 3.63
0.15 0.72 6.00 0.24 2.27
80
-------
APPENOIX B
1990 NDDN DATA
(continued)
SUH
123
0.32
13
24.15
10
39.70
10
0.51
3.32
8.21
13
1.57
4.79
3.97
0.13
1.02
10.67
0.08
0.81
AUT
123
0.15
12
7.89
11
24.74
11
0.34
3.58
13.30
12
1.99
2.82
1.81
0.12
0.40
4.12
0.30
3.38
ANN
0.19
52.43
32.84
0.37
11.53
10.61
1.92
14.45
2.53
0.13
2.45
6.10
0.94
2.52
UIN
131
0
SPR
131
0
SUM
131
0
47.24
8
10.33
8
3.94
13.69
AUT
131
0
33.11
11
15.54
9
2.83
4.63
ANN
UIN
113
0.07
10
2.48
10
24.09
13
0.19
4.17 26.70
13
1.64
2.81
2.34
0.08
0.25
4.09
0.10
1.58
SPR
113
0.19
10
14.63
9
39.60
10
0.41
3.94
11.56
12
1.72
3.89
2.79
0.14
0.63
5.56
0.07
0.55
SUH
113
0.26
12
19.33
12
37.39
13
0.52
3.96
9.68
13
1.41
3.52
3.27
0.10
0.79
9.99
0.02
0.20
AUT
113
0.10
12
4.53
12
24.55
12
0.24
3.93
21.94
12
1.55
2.86
2.38
0.09
0.32
4.37
0.08
1.26
ANN
0.16
40.97
31.41
0.34
16.00
17.47
1.58
13.07
2.70
0.10
1.98
6.00
0.27
0.90
WIN
122
0.07
13
2.40
13
22.68
13
0.07
1.21
22.38
13
1.34
1.74
1.69
0.09
0.24
3.73
0.17
2.55
SPR
122
0.23
12
17.54
12
39.87
13
0.31
2.92
12.96
13
1.40
3.74
3.47
0.12
0.61
6.48
0.14
1.52
SUM
122
0.35
9
30.12
9
42.71
13
0.38
3.38
12.10
13
0.96
3.23
4.19
0.10
0.96
12.68
0.04
0.49
AUT
122
0
24.09
12
22.19
12
2.49
4.61
2.00
ANN
32.34
17.41
2.96
6.88
1.64
WIN
107
0.08
13
3.49
13
25.28
13
0.10
1.29
17.05
13
2.48
3.01
1.55
0.11
0.35
3.86
0.10
1.13
SPR
107
0.31
10
22.89
10
38.07
13
0.44
2.97
10.49
13
2.55
3.86
2.05
0.17
0.86
6.89
0.08
0.51
SUH
107
0.41
13
30.92
13
35.10
13
0.55
3.16
7.23
13
1.92
3.33
2.22
0.14
1.28
11.98
0.03
0.24
AUT
107
0.14
12
6.85
12
22.44
12
0.18
1.87
16.18
11
2.09
2.52
1.52
0.11
0.42
4.82
0.06
0.65
ANN
0.24
64.15
30.22
0.32
9.29
12.74
2.26
12.72
1.84
0.13
2.91
6.89
0.27
0.63
WIN
129
0.07
13
3.32
13
30.60
13
0.08
0.72
11.72
13
1.51
2.46
2.12
0.09
0.24
3.53
0.11
1.58
SPR
129
0.27
13
20.95
13
44.08
13
0.30
1.88
8.92
13
1.57
3.79
3.13
0.12
0.63
6.43
0.07
0.65
SJM
129
0.36
4
32.78
4
51.32
4
0.37
2.22
7.47
4
1.44
3.88
3.41
0.13
1.05
9.79
0.05
0.43
AUT
C
ANN
UIN
108
0.09
13
4.09
13
28.06
13
0.11
0.79
9.10
13
1.56
2.81
2.36
0.11
0.29
3.42
0.05
0.51
SPR
108
0.38
9
29.12
8
42.39
12
0.46
1.77
5.31
13
1.85
3.67
2.52
0.16
0.91
6.51
0.03
0.23
SUM
108
0.39
9
27.08
8
38.18
12
0.44
1.25
3.81
13
1.61
3.22
2.54
0.14
1.01
9.55
0.01
0.10
ALT
108
0.15
12
8.92
12
29.57
12
0.21
1.45
10.51
12
1.61
2.88
2.31
0.11
0.39
4.22
0.04
0.41
ANN
0.25
69.21
34.55
0.30
5.26
7.18
1.66
12.58
2.43
0.13
2.60
5.93
0.12
0.31
UIN
165
0.06
13
3.12
12
31.20
12
0.07
0.03
0.50
12
1.34
0.28
0.29
0.07
0.03
0.54
0.01
0.11
SPR
165
0.13
13
9.80
13
49.51
13
0.13
0.04
0.38
13
1.95
0.60
0.39
0.17
0.11
0.83
0.03
0.22
SUM
165
0.27
13
21.26
13
47.82
13
0.28
0.13
0.60
13
1.86
1.06
0.73
0.17
0.12
0.92
0.02
0.18
AUT
165
0.08
12
5.60
12
42.41
12
0.09
0.03
0.44
12
1.60
0.34
0.28
0.10
0.04
0.55
0.01
0.15
ANN
0.14
39.79
42.74
0.14
0.23
0.48
1.69
2.28
0.42
0.13
0.31
0.71
0.07
0.17
UIN
126
o.oa
13
5.31
13
42.04
13
0.09
0.32
4.74
13
3.11
4.51
1.89
0.16
0.33
2.59
0.04
0.29
SPR
126
3.28
12
24.78
12
53.85
12
0.34
0.99
4.18
13
2.82
4.95
2.29
0.19
0.95
6.55
0.04
0.27
81
-------
APPENDIX B
1990 NDON OATA
(conti nued)
SUM
126
0.41
10
35.60
10
51.82
10
0.47
1.17
3.38
13
2.06
3.46
2.26
0.14
1.33
12.48
0.01
0.13
AUT
126
0.13
12
9.18
12
42.41
12
0.16
0.57
5.62
12
3.08
4.90
1.95
0.17
0.47
3.72
0.04
0.26
ANN
0.22
74.87
47.53
0.26
3.05
4.48
2.77
17.82
2.10
0.17
3.08
6.34
0.13
0.24
WIN
134
0.07
13
3.05
13
29.96
13
0.10
0.30
3.67
13
1.76
0.77
0.56
0.11
0.18
2.18
0.22
2.80
SPR
134
0.15
13
9.61
13
41.02
13
0.22
0.26
1.63
13
1.91
1.76
1.23
0.16
0.35
2.93
0.13
1.02
SUM
134
0.32
13
17.37
13
32.33
13
0.41
0.44
1.37
13
1.40
1.12
1.02
0.12
0.29
3.21
0.04
0.43
AUT
134
0.10
12
4.16
12
26.36
12
0.16
0.31
2.69
12
1.66
0.78
0.61
0.10
0.13
1.80
0.16
2.45
ANN
0.16
34.19
32.42
0.22
1.31
2.34
1.68
4.43
0.86
0.12
0.95
2.53
0.55
1.68
WIN
106
0.07
13
2.72
13
24.74
13
0.07
1.37
23.99 13
2.13
4.23
2.56
0.10
0.34
4.27
0.14
1.70
SPR
106
0.16
13
11.90
13
43.48
13
0.21
1.43
9.07
13
2.21
4.95
2.87
0.15
0.75
6.31
0.09
0.83
SUM
106
0.23
10
17.29
10
37.87
13
0.34
1.83
7.69
13
1.84
4.27
3.09
0.13
1.01
10.03
0.04
0.35
AUT
106
0.08
11
3.85
11
26.54
12
0.09
1.29
18.96
12
2.17
4.13
2.38
0.12
0.41
4.36
0.11
1.24
ANN
0.13
35.77
33.16
0.18
5.92
14.93
2.09
17.58
2.73
0.13
2.50
6.24
0.38
1.03
WIN
163
0.07
11
3.57
10
35.57
12
0.07
0.02
0.34
13
2.03
0.35
0.23
0.12
0.04
0.47
0.04
0.57
SPR
163
0.10
13
6.31
12
40.84
12
0.10
0.03
0.31
13
2.08
0.65
0.41
0.19
0.09
0.62
0.04
0.26
SUM
163
0.15
12
10.86
12
43.82
13
0.15
0.05
0.46
13
2.03
1.16
0.72
0.19
0.10
0.71
0.05
0.33
AUT
163
0.08
12
4.23
12
34.79
12
0.08
0.02
0.31
12
1.86
0.37
0.26
0.12
0.04
0.46
0.03
0.36
ANN
0.10
24.97
38.76
0.10
0.12
0.36
2.00
2.53
0.41
0.16
0.28
0.57
0.16
0.38
WIN 133 0.08 12 3.20 12 25.46 12 0.16 1.48 11.86 12
SPR 133 0.26 13 20.08 13 41.14 13 0.37 2.23 7.82 13
SUM 133 0.36 12 27.25 12 39.98 12 0.47 2.77 7.58 12
AUT 133 0.13 11 6.40 11 25.42 11 0.20 1.79 12.51 10
ANN 0.21 56.93 33.00 0.30 8.26 9.94
2.54
2.10
1.06
0.12
0.34
3.77
0.39
4.34
2.65
5.63
2.76
0.16
0.78
5.95
0.31
2.66
2.03
5.78
3.68
0.14
1.09
10.43
0.10
0.96
2.24
2.38
1.37
0.10
0.31
3.96
0.31
4.23
2.37
15.89
2.22
0.13
2.52
6.03
1.11
3.05
WIN 164 0.07 13 4.22 12 39.93 12 0.07 0.02 0.34 13 1.42 0.40 0.38 0.09 0.03 0.49 0.02 0.25
SPR 164 0.21 13 15.82 10 44.24 11 0.21 0.04 0.26 13 1.99 0.60 C.39 0.19 0.10 0.64 0.04 0.24
SUM 164 0.39 13 30.03 12 46.38 12 0.38 0.13 0.43 12 2.03 1.11 0.68 0.19 0.10 0.66 0.05 0.31
AUT 164 0.11 12 6.12 6 37.58 6 0.11 0.03 0.26 12 1.57 0.52 0.42 0.11 0.04 0.42 0.02 0.21
ANN 0.19 0.19 0.22 0.32 1.75 2.63 0.47 0.15 0.27 0.55 0.12 0.25
WIN 118 0.08 13 4.56 12 38.58 12 0.09 0.69 10.35 13
SPR 118 0.19 13 16.17 13 53.63 13 0.24 1.33 8.25 12
SUM 118 0.25 10 23.10 9 54.25 10 0.33 2.13 7.82 11
AUT 118 0.09 9 5.83 9 39.93 12 0.10 1.35 15.04 12
ANN 0.15 49.65 46.60 0.19 5.51 10.37
2.20
4.30
2.58
0.13
0.31
3.08
0.06
0.57
2.11
5.13
3.19
0.16
0.86
6.92
0.03
0.26
1.55
3.89
3.32
0.12
1.07
11.44
0.01
0.11
1.94
4.90
2.99
0.11
0.34
3.76
0.04
0.44
1.95
18.22
3.02
0.13
2.58
6.30
0.14
0.35
WIN
152
0.07
13
3.72
13
32.77
13
0.07
0.49
8.42
13
1.85
1.98
1.37
0.11
0.29
3.32
0.19
2.19
SPR
152
0.17
13
14.00
13
44.76
13
0.28
1.21
5.54
13
1.85
3.73
2.59
0.15
0.78
6.66
0.10
0.85
SUM
152
0.24
13
22.71
13
50.62
13
0.37
1.81
6.49
13
1.66
4.19
3.20
0.15
1.44
12.C4
0.05
0.45
AUT
152
0.11
12
6.97
12
33.41
12
0.13
0.92
9.97
12
1.69
2.62
2.03
0.12
0.44
4.64
0.18
2.00
ANN
0.15
47.41
40.39
0.21
4.43
7.61
1.76
12.53
2.30
0.13
2.95
6.67
0.52
1.37
WIN
111
0.C9
9
4.06
9
25.13
13
O.M
0.88
11.38
12
2.28
2.82
1.58
0.12
0.37
4.13
0.13
1.53
SPR
111
0.20
9
13.95
8
38.47
12
0.34
1.52
5.53
13
2.52
4.20
2.12
0.17
0.73
6.35
0.07
0.43
82
-------
APPENDIX B
1990 NDDN
DATA
(continued)
SUM
111
0.29
7
23.61
7
37.67
12
0.49
1.55
3.72
12
1.91
3.03
2.09
0.14
1.05
10.49
0.03
0.21
AUT
111
0.11
11
4.92
11
21.53
11
0.14
0.77
7.92
11
1.82
2.05
1.45
0.11
0.37
4.33
0.08
0.99
ANN
30.70
7.14
1.81
6.33
0.79
UIN
156
0.15
10
8.66
7
32.15
10
0.29
0.60
2.57
12
2.27
1.39
0.86
0.13
0.30
3.01
0.09
0.82
SPR
156
0.28
9
20.06
9
37.72
10
0.45
0.72
1.98
11
2.35
2.48
1.46
0.19
0.77
5.62
0.07
0.40
SUM
156
0.27
9
17.03
9
31.86
13
0.37
0.60
2.16
12
1.90
2.27
1.52
0.15
0.87
7.54
0.03
0.22
AUT
156
0.10
6
5.07
6
29.16
11
0.12
0.37
3.29
12
2.00
2.23
1.27
0.14
0.45
3.93
0.03
0.35
ANN
32.72
2.50
1.28
5.03
0.45
WIN
162
0.07
12
5.12
12
44.46
12
0.12
0.05
0.51
12
1.97
0.63
0.42
0.12
0.05
0.59
0.03
0.32
SPR
162
0.20
13
15.78
13
49.95
13
0.24
0.07
0.39
13
2.96
1.18
0.51
0.26
0.18
0.91
0.06
0.30
SUM
162
0.32
12
24.55
12
45.53
13
0.37
0.16
0.53
13
2.57
1.63
0.79
0.21
0.19
1.12
0.04
0.20
AUT
162
40.12
12
0.48
12
0.38
0.63
ANN
45.02
0.48
0.53
0.81
WIN
124
0.07
13
3.21
13
29.30
13
0.10
0.78
9.27
13
1.30
1.36
1.36
0.07
0.18
3.28
0.19
3.60
SPR
124
0.15
11
10.73
11
42.01
13
0.21
0.88
5.33
13
1.94
3.14
1.95
0.14
0.49
4.46
0.24
2.53
SUM
124
0.24
13
17.26
13
38.37
13
0.39
1.23
3.97
13
1.62
3.35
2.70
0.12
0.59
6.53
0.10
1.08
AUT
124
0.08
12
3.97
12
28.30
12
0.12
0.68
7.36
12
1.67
1.69
1.27
0.09
0.20
2.90
0.20
3.02
ANN
0.14
35.18
34.50
0.21
3.56
6.48
1.63
9.54
1.82
0.10
1.45
4.29
0.72
2.56
WIN
140
0.08
13
2.90
13
23.42
13
0.08
1.45
23.29
13
2.09
2.52
1.57
0.11
0.32
3.87
0.25
2.94
SPR
140
0.24
13
16.60
13
37.39
13
0.26
3.14
15.54
13
2.23
4.31
2.53
0.15
0.78
6.49
0.22
1.89
SUM
140
0.34
13
25.43
13
38.00
13
0.35
4.29
15.62
13
1.71
4.85
3.65
0.13
1.35
12.91
0.14
1.28
AUT
140
0.13
12
6.29
12
23.47
12
0.14
2.41
23.23
12
1.83
2.57
1.81
0.10
0.37
4.39
0.19
2.53
ANN
0.19
51.22
30.57
0.21
11.30
19.42
1.97
14.24
2.39
0.12
2.82
6.92
0.80
2.16
WIN
120
0.07
13
4.18
12
36.83
12
0.08
0.66
10.03
13
1.74
3.99
2.93
0.10
0.26
3.40
0.04
0.57
SPS
120
0.25
13
20.90
13
51.65
13
0.40
2.22
7.59
13
1.79
4.43
3.15
0.13
0.73
7.02
0.04
0.39
SUM
*.20
0.33
13
29.19
13
54.34
13
0.64
3.48
6.76
13
1.54
4.25
3.48
0.12
1.29
13.29
0.02
0.25
AUT
120
0.12
12
8.28
12
38.97
12
0.23
2.12
13.70
12
1.77
4.30
3.12
0.11
0.37
4.30
0.06
0.69
ANN
0.19
62.56
45.45
0.34
8.47
9.52
1.71
16.96
3.17
0.11
2.65
7.CO
0.16
0.48
WIN 149 0.08 13 3.54 13 29.52 13
SPR 149 0.20 13 14.36 13 41.41 13
SUM 149 0.24 13 14.65 13 31.97 13
AUT 149 0.09 11 3.83 11 26.90 12
ANN 0.15 36.37 32.45
0.13 0.72 7.04 13
0.28 0.85 4.04 13
0.32 0.76 3.00 13
0.13 0.45 5.40 12
0.22 2.79 4.87
1.54 1.85
1.44 1.96
0.87 1.16
1.34 1.13
1.30 6.09
1.53
1.77
1.74
1.17
1.55
0.07
0.10
0.07
0.06
0.08
0.16 2.71
0.32 3.95
0.29 5.89
0.10 2.32
0.86 3.72
0.09
0.05
0.01
0.06
0.21
1.77
0.56
0.20
1.60
1.C3
WIN
105
0.07
13
3.91
12
36.88
12
0.08
0.42
7.13
13
1.84
2.50
1.69
0.10
0.18
2.35
0.02
0.28
S=*
105
0.12
13
8.91
13
43.97
13
0.18
0.31
2.55
13
1.68
1.92
1.41
0.11
0.36
3.99
0.02
0.19
SUM
105
0.23
11
13.54
11
34.59
13
0.34
0.47
1.68
13
1.17
1.19
1.28
0.08
0.29
4.24
0.01
0.15
AUT
105
0.07
12
3.55
12
32.08
12
0.07
0.26
4.59
12
1.72
2.13
1.51
0.08
0.15
2.38
0.02
0.32
ANN
0.12
29.91
36.88
0.17
1.46
3.99
1.60
7.73
1.47
0.09
0.98
3.24
0.07
0.24
WIN ',34 0.07 12 2.50 12 22.67 13 0.14 2.04 19.32 13 1.30 2.37 2.34
SPR 104 0.14 9 9.14 9 35.78 9 0.25 1.89 9.71 9 1.44 3.16 2.83
0.09 0.28 4.05 0.09 1.33
0.12 0.42 4.57 0.04 0.37
83
-------
APPENDIX B
1990 NODN DATA
(continued)
SUM
104
0.26
12 17.69
12 31.19
12
0.36
2.17
7.15
12
0.96
1.88
2.47
0.09
0.51
7.36
0.01
0.16
AUT
104
0.09
12 3.83
12 21.60
12
0.17
1.43
11.41
12
1.26
1.87
2.00
0.09
0.25
3.62
0.04
0.58
ANN
0.14
33.16
27.81
0.23
7.52
11.90
1.24
9.28
2.41
0.10
1.45
4.90
0.17
0.61
WIN
144
o.oa
13 2.52
13 19.09
13
0.09
1.37
19.70
13
2.07
3.55
2.17
0.11
0.42
4.74
0.22
2.61
SPR
144
0.32
9 24.10
9 37.09
13
0.44
2.47
8.21
13
1.98
6.33
3.98
0.15
0.66
5.67
0.06
0.69
SUM
144
0.38
3 31.31
3 37.14
13
0.47
2.44
6.68
13
1.55
6.37
4.46
0.13
1.07
9.13
0.02
0.26
AUT
144
0.14
10 7.08
10 21.10
12
0.17
1.38
14.02
12
2.24
3.95
2.54
0.15
0.45
4.20
0.12
1.32
ANN
28.61
12.15
3.29
5.94
1.22
UIN
109
0
30.65
12
4.70
12
1.09
2.37
SPR
109
0
36.42
13
1.61
13
1.08
3.82
SUM
109
0.29
8 15.74
8 26.70
13
0.41
0.28
0.98
12
0.96
0.54
0.69
0.08
0.24
3.84
0.01
0.10
AUT
109
0
25.78
12
2.31
12
0.77
2.24
ANN
29.89
2.40
0.91
3.07
84
-------
APPENDIX C
1991 NOON DATA
OZONE
S02
HN03
S04
N03
SEA
SITE
Vd N FLUX
N
CONC
N
Vd
FLUX
CONC
N
Vd
FLUX
CONC
Vd
FLUX
CONC
FLUX
CONC
UIN
157
0.06 10 2.63
10
25.46
13
0.06
0.65
16.88
13
0.75
0.76
1.78
0.05
0.15
4.15
0.16
3.83
SPR
157
0.18 13 13.02
13
37.17
13
0.19
1.39
9.15
13
0.98
2.30
3.02
0.09
0.50
6.80
0.11
1.72
SUM
157
0.28 13 22.12
13
39.68
13
0.30
2.41
10.14
13
0.94
2.30
3.10
0.10
0.68
9.00
0.07
0.94
AUT
157
0.07 12 2.95
11
20.65
13
0.08
0.70
11.79
13
0.85
0.72
1.02
0.06
0.15
3.26
0.11
2.88
ANN
0.15 40.72
30.74
0.16
5.15
11.99
0.88
6.08
2.23
0.07
1.48
5.80
0.45
2.34
UIN
115
0.07
13,
3.07
13
26.35
13
0.22
1.91
11.50
13
2.00
2.42
1.57
0.10
0.28
3.51
0.28
3.44
SPR
115
0.22
12
18.33
12
43.03
12
0.41
1.79
5.70
12
1.98
4.11
2.86
0.15
0.66
5.87
0.13
1.07
SUM
115
0.32
13
23.34
12
37.03
12
0.48
1.59
4.27
13
1.40
2.62
2.37
0.12
0.64
6.52
0.06
0.60
AUT
115
0.09
12
3.09
11
19.97
11
0.23
1.21
6.88
13
1.96
1.67
1.04
0.09
0.21
2.97
0.18
2.69
ANN
0.18
47.83
31.60
0.34
6.50
7.09
1.83
10.82
1.96
0.12
1.80
4.72
0.64
1.95
U2N
146
0.07
12
2.14
12
18.36
13
0.11
1.92
20.84
13
2.03
1.69
1.07
0.09
0.29
4.10
0.35
5.01
SPR
146
0.27
13
19.63
13
32.01
13
0.35
4.26
16.07
13
1.78
4.23
3.38
0.13
0.61
6.52
0.25
2.69
SUM
146
0.49
11
38.38
9
35.00
11
0.64
8.65
17.21
13
2.06
6.19
3.71
0.17
1.17
7.86
0.21
1.50
AUT
146
0.12
10
3.59
10
14.92
11
0.15
1.75
14.97
12
2.42
1.46
0.80
0.12
0.32
3.45
0.33
3.49
ANN
0.24
63.74
25.07
0.31
16.59
17.27
2.07
13.57
2.24
0.13
2.39
5.48
1.15
3.17
UIN
128
0.08
13
3.06
13
24.98
13
0.09
1.49
20.57
13
2.31
5.01
2.80
0.13
0.38
3.82
0.14
1.32
SPR
128
0.30
13
25.97
13
48.71
13
0.34
2.94
11.02
13
2.10
6.68
4.11
0.16
1.04
7.93
0.15
1.23
SUM
128
0.39
13
37.05
13
55.02
13
0.47
4.37
11.74
13
2.11
8.01
4.75
0.17
1.60
11.60
0.10
0.75
AUT
128
0.14
12
6.47
12
24.27
13
0.18
1.82
14.25
13
2.04
4.28
2.69
0.11
0.41
4.47
0.16
1.92
ANN
0.22
72.54
38.25
0.27
10.62
14.40
2.14
23.98
3.59
0.14
3.43
6.96
0.55
1.31
WIN
135
0.08
12
4.60
12
34.26
13
0.12
0.30
3.31
13
2.44
1.42
0.71
0.12
0.19
2.05
0.03
0.31
SPR
135
0.25
12
15.55
12
38.24
12
0.34
0.22
0.96
13
2.94
0.91
0.39
0.21
0.25
1.51
0.03
0.20
SUM
135
0.36
13
18.20
13
29.29
13
0.53
0.30
0.72
13
2.52
0.87
0.43
0.19
0.33
2.16
0.02
0.16
AUT
135
0.11
12
4.39
12
24.09
12
0.25
0.27
1.50
12
2.57
1.10
0.54
0.10
0.16
1.98
0.02
0.24
ANN
0.20
42.74
31.47
0.31
1.08
1.62
2.62
4.30
0.52
0.15
0.92
1.93
0.10
0.23
UIN
116
0.10
13
3.33
13
17.19
13
0.16
2.14
17.32
12
2.10
2.63
1.59
0.13
0.39
3.95
0.14
1.45
SPR
116
0.25
12
19.98
12
35.06
13
0.32
1.72
7.28
13
2.01
4.61
2.95
0.16
0.82
6.33
0.09
0.74
SUM
116
0.31
12
28.89
11
39.31
12
0.41
2.64
7.98
13
1.83
5.68
3.92
0.15
1.33
10.74
0.04
0.37
AUT
116
0.15
12
4.48
10
13.69
10
0.24
2.15
12.29
12
1.77
2.22
1.61
0.11
0.35
4.19
0.09
1.14
ANN
0.20
56.69
26.31
0.28
8.65
11.22
1.93
15.13
2.52
0.14
2.89
6.30
0.36
0.93
UIN
130
0.06
13
2.28
13
24.63
13
0.06
0.48
10.38
11
0.30
0.91
1.39
0.05
0.16
3.69
0.18
4.21
SPR
130
0.12
12
9.89
10
39.99
11
0.12
0.64
7.08
13
1.27
3.25
3.54
0.10
0.54
6.98
0.15
1.88
SUM
130
0.24
13
19.86
13
00
o
in
13
0.25
1.42
7.26
13
1.15
3.11
3.46
0.10
0.71
8.84
0.07
0.90
AUT
130
0.08
10
3.25
9
20.52
10
0.08
0.39
6.44
10
0.96
0.83
1.03
0.07
0.16
3.11
0.12
2.65
ANN
0.12
35.28
32.56
0.13
2.94
7.79
1.04
8.10
2.36
0.08
1.57
5.66
0.52
2.41
UIN
150
0.09
13
4.19
13
24.15
13
0.17
0.33
2.47
13
1.18
1.28
1.39
0.07
0.18
3.18
0.03
0.55
SPR
150
0.25
12
14.85
12
27.91
13
0.36
G.51
1.71
13
1.17
1.20
1.39
0.09
0.30
4.35
0.02
C.28
SUM
'SO
0.27
13
18.23
13
26.60
13
0.35
C .48
1.70
13
0.88
1.08
".54
0.08
0.34
5.40
0.01
C.17
AUT
'50
0.12
6
5.22
6
20.33
13
0.20
0.23
1.75
13
1.16
0.67
1.08
0.07
0.10
2.63
0.02
0.45
85
-------
APPENDIX C
1991 NDDN DATA
(continued)
ANN
24.75
1.91
1.35
3.89
WIN
119
0.07
13
2.22
13
19.78
13
0.08
0.89
15.17
13
1.09
1.33
1.55
0.06
0.16
3.72
SPR
119
0.27
13
19.35
13
33.28
13
0.31
1.58
6.66
13
0.95
1.18
1.60
0.08
0.43
7.27
SUM
119
0.31
13
22.96
13
32.79
13
0.34
1.77
6.69
13
0.71
0.93
1.66
0.06
0.57
11.92
AUT
119
0.11
13.
4.77
13
18.88
13
0.14
0.98
11.41
13
0.98
0.94
1.23
0.05
0.16
3.84
ANN
0.19
49.29
26.18
0.22
5.21
9.98
0.93
4.38
1.51
0.06
1.32
6.69
WIN
125
0.09
13
4.81
13
30.84
13
0.11
0.59
6.95
13
1.61
2.56
2.06
0.10
0.27
3.32
SPR
125
0.33
13
25.90
13
41.88
13
0.46
0.78
2.32
13
1.61
2.32
1.83
0.14
0.61
5.72
SUM
125
0.38
13
27.62
13
37.79
13
0.51
1.01
2.53
13
1.34
2.10
1.96
0.12
0.71
7.93
AUT
125
0.15
13
8.26
13
31.76
13
0.19
0.72
6.59
13
1.47
2.32
2.03
0.11
0.32
3.91
ANN
0.24
66.59
35.57
0.32
3.11
4.60
1.51
9.31
1.97
0.12
1.92
5.22
WIN
167
0.16
11
10.49
11
40.30
11
0.21
0.36
2.35
12
2.38
0.66
0.35
0.16
0.12
1.01
SPR
167
0.24
12
20.22
12
51.28
12
0.25
0.35
1.80
13
2.68
1.16
0.53
0.25
0.24
1.26
SUM
167
0.22
12
13.95
7
37.40
7
0.26
0.26
1.23
13
2.21
1.39
0.81
0.16
0.25
1.97
AUT
167
0.12
12
7.87
12
38.14
12
0.13
0.28
2.89
13
2.33
0.86
0.45
0.16
0.17
1.27
ANN
0.19
52.53
41.78
0.21
1.26
2.07
2.40
4.08
0.54
0.18
0.79
1.38
WIN
169
0.09
10
6.33
10
46.85
13
0.12
0.05
0.59
13
3.28
0.85
0.33
0.22
0.09
0.51
SPR
169
0.16
12
14.15
12
53.96
13
0.25
0.07
0.42
13
2.71
1.08
0.51
0.24
0.20
1.05
SUM
169
0.24
13
20.01
9
46.12
9
0.33
0.13
0.49
13
2.57
1.31
0.67
0.22
0.16
0.96
AUT
169
0.10
13
5.44
9
44.76
9
0.11
0.06
0.59
13
3.74
1.13
0.39
0.23
0.11
0.59
ANN
0.15
45.93
47.92
0.20
0.31
0.52
3.08
4.37
0.48
0.23
0.56
0.78
WIN
137
0.09
13
4.34
13
28.89
13
0.11
0.21
2.35
13
1.32
0.91
0.84
0.09
0.17
2.29
S?R
137
0.20
13
12.09
13
29.44
13
0.30
0
1.11
13
1
0.78
0.76
0.10
0.33
4.41
SUM
137
0.25
13
13.50
12
22.53
12
0.34
0.16
0.61
13
0.80
0.41
0.65
0.07
0.45
7.85
AUT
137
0.11
12
5.94
12
25.88
13
0.14
0.18
1.90
13
0.92
0.56
0.79
0.07
0.17
3.07
ANN
3.16
35.87
26.69
0.22
0.77
1.49
1.06
2.66
0.76
0.08
1.12
4.41
WIN
110
0.07
13
3.40
13
30.14
13
0.10
1.18
15.95
13
1.94
4.19
2.79
0.C9
0.23
3.13
SPR
110
0.27
13
21.84
13
48.81
13
0.36
2.40
8.34
13
1.86
3.74
2.61
0.14
0.58
5.22
SUM
110
0.40
13
32.87
13
50.33
13
0.51
3.57
8.94
13
1.64
3.91
3.05
0.12
0.81
8.36
AUT
110
0.10
12
5.05
12
28.38
12
0.16
1.61
13.99
12
1.86
3.97
2.72
0.09
0.31
4.01
ANN
0.21
63.15
39.42
0.28
8.76
11.81
1.82
15.81
2.79
0.11
1.93
5.18
WIN
151
0.09
12
4.66
12
29.45
12
0.13
0.47
4.53
13
0.72
0.80
1.55
0.04
0.10
3.35
SPR
151
0.26
12
17.09
12
38.01
13
0.37
0.41
1.57
12
1.41
1.23
1.11
0.10
0.31
3.63
SUM
151
0.27
12
20.71
12
41.82
12
0.37
0.78
2.68
11
1.12
1.67
1.89
0.10
0.64
8.17
AUT
151
0.11
12
6.12
12
32.73
13
0.14
0.48
4.20
13
1.28
1.03
1.10
0.08
0.20
2.98
ANN
0.18
48.59
35.50
0.25
2.15
3.24
1.13
4.73
1.41
0.08
1.24
4.53
WIN
114
0.08
12
3.15
12
23.43
13
0.26
3.08
15.45
13
2.20
3.56
2.09
0. "2
0.36
3.93
S»R
114
0.31
1C
24.69
10
41.18
13
0.52
4.65
11.41
13
1.81
5.34
3.85
0.15
0.89
8.06
SUM
114
0.36
13
30.11
13
42.73
13
0.58
4.75
10.33
13
1.64
5.52
4.38
0.13
1.12
10.90
Aur
114
0.14
10
6.55
10
22.73
11
0.32
2.97
11.88
13
2.06
3.25
1.37
0.12
0.43
4.24
0.36
0.02 0.31
0.01 0.15
0.01 0.11
0.02 0.48
0.05 0.26
0.06 0.81
0.04 0.38
0.03 0.29
0.08 0.90
0.20 0.60
0.03 0.22
0.09 0.46
0.04 0.25
0.03 0.22
0.18 0.29
0.02 0.12
0.06 0.36
0.03 0.19
0.02 0.10
0.14 0.19
0.01 0.18
0.01 0.17
0.00 0.08
0.01 0.22
0.04 0.16
0.06 0.78
0.05 0.46
0.04 0.39
0.04 0.62
0.19 0.56
0.03 0.74
0.03 0.39
0.03 0.39
0.07 1.03
0.15 0.64
0.19 1.95
0.16 1.23
0.07 0.61
0.25 2.68
86
-------
APPENDIX C
1991 NDDN DATA
(conti nued)
ANN
0.22
64.51
32.52
0.42
15.45
12.26
1.93
17.67
3.05
0.13
2.79
6.78
0.66
1.62
U1M
127
O.OS
13
3.56
13
27.98
13
0.18
1.31
9.54
13
1.27
1.90
1.94
0.07
0.19
3.71
0.03
0.61
SPR
127
0.22
13
15.65
13
41.09
13
0.42
1.12
3.45
13
1.07
1.35
1.61
0.08
0.39
6.27
0.02
0.32
SUN
127
0.26
12
20.61
12
43.17
12
0.42
1.07
3.29
12
0.84
1.18
1.79
0.07
0.61
10.58
0.01
0.17
AUT
127
0.11
12
5.86
12
30.87
13
0.24
1.60
8.85
13
1.17
1.68
1.80
0.07
0.22
3.83
0.05
0.85
ANN
0.16
45.68
35.78
0.31
5.10
6.28
1.09
6.12
1.7V
0.07
1.41
6.10
0.12
0.49
WIN
153
0.10
13
5.51
13
30.52
13
0.19
1.22
8.45
13
2.28
2.86
1.61
0.13
0.34
3.57
0.06
0.61
SPR
153
0.23
9
16.81
9
38.53
11
0.39
1.40
4.32
11
1.88
2.38
1.56
0.14
0.58
4.90
0.04
0.33
SUM
153
0.21
11
16.40
11
37.77
13
0.45
1.58
4.39
13
1.52
2.30
1.98
0.12
0.74
8.70
0.02
0.19
AUT
153
0.16
13
10.19
13
33.18
13
0.22
1.35
8.45
13
1.85
2.49
1.78
0.13
0.39
3.86
0.08
0.81
ANN
0.18
48.92
35.00
0.31
5.55
6.40
1.88
10.03
1.73
0.13
2.05
5.26
0.20
0.49
WIN
174
0.13
13
8.07
13
40.00
13
0.12
0.13
1.31
13
1.51
0.86
0.75
0.12
0.06
0.65
0.02
0.16
SPR
174
0.17
12
15.02
12
56.10
12
0.14
0.05
0.47
13
1.87
1.03
0.69
0.27
0.21
1.04
0.12
0.55
SUM
174
0.15
13
10.79
13
46.96
13
0.15
0.07
0.60
13
1.54
1.42
1.18
0.18
0.18
1.24
0.05
0.31
AUT
174
0.13
12
8.59
12
41.61
12
0.11
0.10
1.20
13
1.29
0.81
0.82
0.13
0.10
0.95
0.03
0.20
ANN
0.14
42.46
46.17
0.13
0.35
0.90
1.55
4.12
0.86
0.17
0.54
0.97
0.20
0.31
WIN
168
0.07
13
2.82
13
24.56
13
0.12
0.11
1.19
12
0.69
0.13
0.26
0.03
0.02
0.69
0.01
0.35
SPR
168
0.16
12
10.07
11
31.31
13
0.22
0.07
0.42
12
1.19
0.28
0.32
0.09
0.07
0.99
0.01
0.13
SUM
168
0.19
13
9.34
13
21.29
13
0.24
0.08
0.42
13
0.87
0.21
0.31
0.07
0.04
0.65
0.01
0.12
AUT
168
0.11
3
5.65
2
18.98
13
0.17
0.10
0.83
13
0.92
0.19
0.30
0.04
0.03
0.92
0.01
0.33
ANN
24.04
0.72
0.30
0.81
0.23
WIN
161
0.07
13
5.21
13
48.04
13
0.07
0.02
0.28
13
2.12
0.42
0.24
0.12
0.04
0.39
0.01
0.07
SPR
161
0.07
1
6.14
1
54.17
13
0.08
0.01
0.28
13
2.33
0.40
0.29
0.15
0.C9
0.89
0.01
0.31
SUM
161
0.22
12
14.81
12
37.14
13
0.30
0.03
0.16
13
2.11
0.49
0.30
0.15
0.10
0.87
0.02
0.15
AUT
161
0.17
6
11.21
6
41.14
10
0.18
0.06
0.29
13
2.49
0.66
0.29
0.17
0.C9
0.54
0.02
0.09
ANN
45.12
0.25
0.28
0.67
0.16
WIN
112
0.07
12
3.05
12
29.11
13
0.09
1.59
22.91
13
1.59
3.24
2.55
0.07
0.20
3.69
0.03
0.54
SPR
112
0.23
13
18.91
13
47.83
13
0.28
1.44
8.11
13
1.44
2.48
2.21
0.11
0.51
6.29
0.02
0.27
SUM
112
0.32
9
27.70
9
47.87
13
0.44
3.19
9.00
13
1.07
2.04
2.40
0.08
0.66
9.44
0.01
0.17
AUT
112
0.09
13
4.18
13
25.51
13
0.16
2.03
17.50
13
1.53
2.47
2.06
0.07
0.23
3.85
0.03
0.56
ANN
0.18
53.82
37.58
0.24
8.25
14.38
1.41
10.22
2.31
0.08
1.60
5.82
0.10
0.39
WIN
121
0.07
13
2.80
13
20.13
13
0.08
0.29
4.51
13
0.88
C.50
0.69
0.05
0.14
3.25
0.02
0.56
SPR
121
0.26
13
15.67
12
26.41
13
0.23
0.42
1.90
13
0.81
0.59
0.97
0.07
0.36
7.16
0.02
0.28
SUM
121
0.30
13
16.31
13
21.88
13
0.31
0.31
1.27
13
0.62
0.36
0.73
0.05
0.38
9.17
0.01
0.18
AUT
121
0.12
13
5.71
13
19.14
13
0.13
0.32
3.61
13
0.71
0.37
0.69
0.05
0.14
3.72
0.03
0.87
ANN
0.19
40.49
21.89
0.20
1.33
2.82
0.76
1.82
0.77
0.06
1.01
5.S3
0.07
0.47
WIN
117
0.08
13
3.27
11
23.85
11
0.13
3.01
28.93
13
1.59
2.85
2.28
0.07
0.22
3.85
0.03
0.40
SPR
117
0.30
13
24.03
13
39.01
13
0.34
3.46
13.26
13
1.33
2.63
2.62
0.11
0.66
8.26
0.02
0.24
SUM
117
0.35
13
29.79
13
40.42
13
0.38
4.36
14.57
13
0.94
2.22
3.02
0.08
0.88
13.51
0.01
0.13
AUT
117
0.12
13
5.90
13
21.50
13
0.17
2.52
20.46
13
1.31
1.71
1.66
0.06
0.21
3.79
0.02
0.44
87
-------
APPENDIX C
1991 NDON DATA
(continued)
ANN
0.21
62.98
31.20
0.26
13.35
19.31
1.29
9.41
2.40
0.08
1.96
7.35
0.08
0.30
WIN
123
0.0 7
13
2.74
13
23.06
13
0.21
2.24
14.28
13
2.05
2.81
1.77
0.11
0.32
3.97
0.29
3.43
SPR
123
0.26
13
23.78
12
44.68
12
0.45
3.54
9.69
13
2.03
5.17
3.49
0.16
0.89
7.19
0.22
1.74
SUM
123
0.36
12
32.38
12
47.66
13
0.54
2.96
7.15
13
1.61
4.38
3.50
0.15
1.02
8.90
0.10
0.86
AUT
123
0.14
12
6.80
12
23.81
12
0.36
2.77
10.08
12
2.11
2.60
1.61
0.11
0.34
3.86
0.25
3.15
ANN
0.21
65.69
34.80
0.39
11.51
10.30
1.95
14.96
2.59
0.13
2.57
5.98
0.86
2.30
WIN
131
0.08
9
3.53
9
27.43
13
0.09
1.14
15.55
13
1.77
3.55
2.70
0.11
0.32
4.07
0.11
1.07
SPR
131
0.34
1
24.35
1
42.59
13
0.38
1.08
7.04
13
1.40
2.24
2.54
0.12
0.67
7.17
0.10
0.77
SUM
131
0.36
7
30.13
7
47.86
10
0.35
2.97
8.44
13
1.20
3.43
3.40
0.11
1.02
10.96
0.06
0.52
AUT
131
0.17
12
9.70
12
29.52
13
0.19
1.64
12.14
13
1.67
2.69
2.20
0.11
0.34
4.19
0.13
1.60
ANN
36.85
10.79
2.71
6.60
0.99
WIN
113
0.07
13
2.47
13
21.36
13
0.21
3.81
23.56
13
1.64
2.86
2.26
0.09
0.26
3.88
0.10
1.48
SPR
113
0.22
13
17.94
12
44.84
12
0.43
3.83
11.91
13
1.67
3.83
3.07
0.14
0.72
6.94
0.07
0.60
SUM
113
0.32
13
28.67
13
47.20
13
0.56
5.11
11.77
13
1.45
3.81
3.37
0.13
1.00
9.79
0.03
0.33
AUT
113
0.10
13
4.16
13
22.95
13
0.25
4.09
20.85
13
1.67
2.55
1.97
0.09
0.28
4.04
0.09
1.42
ANN
0.18
53.24
34.09
0.36
16.83
17.02
1.61
13.04
2.67
0.11
2.26
6.16
0.30
0.96
WIN
122
0
0
22.75
13
21.20
13
2.07
4.41
2.41
SPR
122
0.26
12
20.85
12
41.50
13
0.32
2.72
10.99
13
1.27
3.67
3.71
0.12
0.76
8.23
0.13
1.39
SUM
122
0.34
13
29.94
13
45.98
13
0.40
3.20
10.31
13
1.03
2.92
3.67
0.10
0.81
10.19
0.06
0.71
AUT
122
0.10
13
4.23
13
21.85
13
0.12
1.56
19.31
13
1.28
1.67
1.71
0.08
0.28
4.26
0.16
2.57
ANN
33.02
15.45
2.79
6.77
1.77
WIM
107
0.08
13
3.25
13
24.91
13
0.10
1.35
17.92
13
2.52
2.62
1.31
0.11
0.32
3.56
0.08
0.94
SPR
107
0.33
13
25.55
13
40.02
13
0.37
1.80
6.79
13
2.35
3.06
1.69
0.16
0.94
7.74
0.05
0.37
SUM
107
0.45
13
35.14
13
36.69
13
0.56
2.90
6.63
13
1.91
2.94
1.95
0.14
1.26
11.53
0.03
0.24
AUT
107
0.14
13
6.40
13
20.42
13
0.18
1.24
10.16
11
2.19
2.04
1.19
0.10
0.32
3.77
0.04
0.49
ANN
0.25
70.34
30.51
0.30
7.30
10.38
2.24
10.67
1.54
0.13
2.84
6.65
0.20
0.51
WIN
108
0.08
13
3.68
13
28.06
13
0.11
0.95
10.61
10
1.76
2.93
2.15
0.09
0.21
3.39
0.03
0.42
SPR
108
0.29
8
23.13
7
40.77
12
0.36
1.33
4.15
13
1.67
3.11
2.27
0.13
0.71
6.69
0.04
0.35
SUM
108
0.35
7
24.76
7
38.59
12
0.37
0.93
3.11
13
1.35
2.19
2.05
0.12
0.80
8.74
0.01
0.12
AUT
108
0.13
13
7.13
13
27.23
13
0.15
0.81
7.37
10
1.62
2.33
1.91
0.11
0.33
3.86
0.05
0.53
ANN
33.66
6.31
2.10
5.67
0.36
U1N
126
0.08
13
4.73
13
38.60
13
0.10
0.41
5.63
13
3.09
4.28
1.73
0.16
0.33
2.41
0.04
0.26
SPR
126
0.23
12
19.05
12
53.73
13
0.32
0.50
2.31
13
2.24
3.07
1.72
0.16
0.62
5.63
0.03
0.24
SUM
126
0.38
13
32.13
13
49.60
13
0.59
1.10
2.37
13
1.91
2.64
1.74
0.14
1.10
10.26
0.02
0.17
AUI
126
0.12
12
8.70
12
40.71
13
0.14
C.53
5.03
13
2.84
3.82
1.77
0.16
0.44
3.44
0.07
0.57
ANN
0.20
64.61
44.91
0.29
2.53
3.84
2.52
13.81
1.74
0.15
2.48
5.44
0.16
0.31
win
165
0.06
12
4.60
12
SPR
'65
0.12
11
9.14
8
SUM
•65
0.27
10
20.90
10
AJT
'65
0.09
5
5.78
5
0.07
0.02
o
o
12
1.33
0.25
0.25
0.08
0.03
0.48
C.01
0.17
0.13
0.03
0.33
10
2.15
0.49
0.33
0.16
0.13
0.80
0.04
0.23
3.27
0.C9
0.40
12
2.10
0.99
0.56
O.'S
0.13
0.81
C.03
0.18
0.10
0.04
0.37
12
1.67
0.26
0.24
0.10
0.33
0.61
0.00
0.13
88
-------
APPENDIX C
1991 NODN DATA
(conti nued)
ANN 47.50 0.3a 0.35 0.68 0.19
UIN
134
0.06
11
3.12
11
33.10
13
0.09
0.26
3.73
10
1.50
0.73
0.68
0.09
0.16
2.37
0.24
3.29
SPR
134
0.26
4
17.53
4
41.29
13
0.32
0.26
1.63
13
1.27
0.98
1.36
0.12
0.28
3.10
0.04
0.64
SUM
134
0.30
12
17.81
12
32.89
13
0.40
0.43
1.26
13
1.40
1.05
0.89
0.12
0.38
3.60
0.05
0.50
AUT
134
0.09
13
3.61
13
25.49
13
0.16
0.21
1.81
13
1.57
0.55
0.45
0.08
0.09
1.52
0.11
1.98
ANN
33.19
2.11
0.85
2.65
1.60
UIN
106
0.07
13
2.62
13
24.33
13
0.07
1.25
22.06
13
2.36
4.14
2.23
0.12
0.34
3.85
0.13
1.45
SPR
106
0.18
13
14.41
13
43.06
13
0.22
1.57
9.49
13
2.37
5.45
3.04
0.18
1.03
7.49
0.12
0.84
SUM
106
0.27
12
23.69
12
43.93
13
0.38
3.49
11.26
13
1.75
5.35
3.77
0.14
1.41
12.03
0.05
0.46
AUT
106
0.08
13
2.60
12
20.82
12
0.09
1.09
16.79
13
1.89
3.01
2.04
0.10
0.36
4.34
0.12
1.60
ANN
0.15
43.32
33.04
0.19
7.41
14.90
2.09
17.95
2.77
0.13
3.13
6.93
0.42
1.09
WIN
163
0.07
13
3.59
13
35.59
13
0.06
0.02
0.37
13
1.68
0.39
0.29
0.11
0.05
0.74
0.07
1.21
SPR
163
0.10
13
6.80
12
43.46
12
0.10
0.02
0.20
13
2.16
0.54
0.33
0.19
0.11
0.70
0.04
0.24
SUM
163
0.15
13
10.59
13
42.05
13
0.16
0.04
0.31
13
1.85
0.88
0.60
0.17
0.08
0.57
0.03
0.24
AUT
163
0.08
12
4.14
12
32.43
12
0.09
0.03
0.40
12
1.61
0.35
0.30
0.10
0.06
0.77
0.03
0.51
ANN
0.10
25.14
38.38
0.10
0.10
0.32
1.82
2.16
0.38
0.14
0.29
0.70
0.17
0.55
WIN 133 0.08 13 3.10 13 25.36 13
SPR 133 0.28 13 21.85 13 42.74 13
SUM 133 0.41 13 31.92 13 43.69 13
AUT 133 0.12 12 5.36 12 22.76 12
ANN 0.22 62.23 33.64
0.15 1.64 14.49 13
0.35 1.71 6.34 13
0.48 2.44 6.51 12
0.20 1.20 8.71 10
0.29 6.98 9.01
2.26 2.37 1.36
2.40 6.13 3.53
2.03 5.14 3.19
2.48 2.03 1.10
2.29 15.68 2.30
0.10 0.33 4.37
0.16 0.86 6.82
0.16 1.02 7.82
0.11 0.28 3.45
0.13 2.49 5.62
0.37 4.87
0.21 1.74
0.12 0.93
0.30 3.67
1.00 2.80
WIN
164
0.07
13
4.33
10
40.72
10
0.07
0.02
0.29
13
1.40
0.34
0.33
0.09
0.03
0.38
0.02
0.34
SPR
164
0.08
13
5.83
13
47.96
13
0.10
0.02
0.21
13
2.21
0.55
0.31
0.21
0.12
0.76
0.05
0.27
SUM
164
C.38
13
29.55
10
47.31
10
0.37
0.08
0.28
13
2.16
1.18
0.69
0.20
0.10
0.62
0.04
0.24
AUT
164
C. 12
12
7.45
12
36.74
12
0.12
0.03
0.32
12
1.60
0.50
0.42
0.11
0.C5
0.55
0.02
0.23
ANN
0.16
47.15
43.18
0.16
0.15
0.28
1.84
2.58
0.44
0.15
0.29
0.58
0.12
0.27
UIN
118
0.07
13
4.02
13
35.66
13
0.09
0.82
12.12
13
2.11
3.75
2.31
0.12
0.26
2.65
0.04
0.44
SPR
118
0.19
9
16.97
9
56.03
10
0.25
1.19
6.69
13
1.89
4.69
3.14
0.14
0.75
7.20
0.C2
0.23
SUM
118
0.28
13
25.29
13
56.73
13
0.37
2.11
7.41
13
1.48
3.48
3.01
0.12
1.03
11.19
0.02
0.16
AUT
118
0.09
13
5.56
12
37.30
13
0.11
0.74
8.75
13
1.94
3.37
2.32
0.11
0.33
3.62
0.05
0.56
ANN
0.16
51.83
46.43
0.20
4.86
8.74
1.86
15.28
2.70
0.12
2.37
6.17
0.13
0.35
UIN
152
0.07
13
O
O
13
26.95
13
0.07
0.45
7.83
13
1.91
1.31
0.87
0.11
0.31
3.84
0.20
2.53
SPR
152
0.17
12
12.02
10
37.42
11
0.23
1.12
5.66
13
1.66
2.22
1.77
0.13
0.57
5.82
0.13
1.31
SUM
152
0.24
13
17.93
13
40.41
13
0.34
1.28
4.86
13
1.41
2.85
2.60
0.12
0.94
10.07
0.06
0.58
AJT
152
0.10
13
5.27
12
29.60
12
0.12
0.89
9.83
13
1.68
2.07
1.60
0.12
0.40
4.24
0.20
2.14
ANN
0.15
38.23
33.60
0.19
3.74
7.05
1.66
8.45
1.71
0.12
2.23
5.99
0.59
1.64
WIN
111
0.08
12
3.53
12
23.14
13
0.10
0.75
9.29
13
2.17
2.50
1.39
0.11
0.34
3.79
SPR
111
0.24
10
15.77
10
34.63
13
0.30
0.82
3.52
12
1.96
2.19
1.49
0.14
0.55
5.80
SUH
111
0.34
13
23.98
13
33.74
13
0.38
0.86
2.97
12
1.62
2.30
1.81
0.12
1.02
10.88
AUT
111
0.12
13
5.77
12
23.08
12
0.13
0.76
8.46
13
1.88
1.96
1.35
0.11
0.37
4.04
0.09 1.03
0.C6 0.51
0.02 0.25
0.12 1.34
89
-------
APPENDIX C
1991 NOON DATA
(continued)
ANN
0.19 49.06 28.65
0.23 3.18 6.06
1.91 8.96 1.51 0.12 2.28 6.13 0.29 0.78
UIN
156
0.15
12
8.47
12
32.31
13
0.27
0.45
2.10
13
2.21
1.49
0.85
0.14
0.36
3.37
0.07
0.60
SPR
156
0.25
12
14.43
12
30.77
13
0.43
0.41
1.18
13
2.24
1.44
0.83
0.16
0.48
3.67
0.09
0.71
SUM
156
0.22
13
12.56
8 27.59
a
0.41
0.37
1.14
13
1.77
1.15
0.83
0.13
0.47
4.66
0.03
0.24
AUT
156
0.10
13
5.63
13
30.83
13
0.13
0.27
2.92
13
2.08
1.74
1.13
0.13
0.38
3.57
0.06
0.59
ANN
0.18
0.31
1.50
1.84
2.07
5.83
0.91
0.14
1.68
3.82
0.25
0.54
UIN
162
0.07
12
5.20
12
46.65
13
0.11
0.04
0.40
13
2.05
0.58
0.37
0.13
0.05
0.48
0.02
0.25
SPR
162
0.18
13
15.67
13
54.51
13
0.22
0.07
0.39
13
2.92
1.03
0.45
0.23
0.18
1.01
0.07
0.38
SUM
162
0.29
11
22.08
11
45.03
11
0.35
0.10
0.35
11
2.64
1.20
0.56
0.21
0.14
0.87
0.03
0.16
AUT
162
0.09
13
5.78
13
39.22
13
0.10
0.03
0.34
13
1.88
0.62
0.43
0.12
0.06
0.70
0.02
0.29
ANN
0.16
48.73
46.35
0.19
0.24
0.37
2.37
3.42
0.45
0.17
0.43
0.77
0.14
0.27
UIN
124
0.07
11
3.35
11
32.34
13
0.11
0.74
8.90
13
1.37
1.17
1.12
0.07
0.17
3.12
0.21
3.67
SPR
124
0.14
13
10.42
13 40.82
13
0.19
0.68
4.56
13
1.77
2.96
2.19
0.12
0.49
5.05
0.18
1.96
SUM
124
0.27
13
18.88
13
39.06
13
0.38
1.07
3.77
13
1.70
2.66
1.94
0.14
0.61
5.50
0.12
1.15
AUT
124
0.08
9
2.65
7
21.17
10
0.13
0.61
6.32
13
1.68
1.36
1.01
0.08
0.18
2.74
0.18
2.60
ANN
0.14
35.30
33.35
0.20
3.10
5.89
1.63
8.14
1.57
0.10
1.45
4.10
0.69
2.35
UIN HO 0.08 11 3.02 11 22.75 13 0.08 1.49 24.35 13 2.11 3.58 2.29 0.10 0.36 4.65 0.24 2.73
SPR 140 0.23 12 16.03 12 37.28 13 0.24 3.26 16.64 13 2.12 3.90 2.58 0.15 0.94 8.52 0.29 2.46
SUM 140 0.38 5 31.66 5 40.08 13 0.38 4.47 15.46 13 1.90 4.15 2.99 0.17 1.44 10.90 0.24 1.54
AUT 140 0.13 13 4.68 12 20.81 13 0.13 1.77 18.38 13 2.05 2.43 1.53 0.11 0.34 4.12 0.21 2.65
ANN 30.23 18.76 2.35 7.05 2.35
UIN
120
0.07
10
3.60
10
33.38
13
0.09
0.76
11.19
13
1.76
3.20
2.42
0.09
0.24
3.08
0.03
0.41
SPR
120
0.25
13
20.71
13
52.30
13
0.40
1.68
5.72
13
1.78
3.93
2.85
0.12
0.74
7.54
0.04
0.40
SUM
120
0.36
13
31.02
13
54.00
13
0.63
2.87
5.93
12
1.51
3.22
2.71
0.12
1.16
11.96
0.04
0.38
AUT
120
0.13
13
7.79
12
36.21
12
0.21
1.60
10.70
13
1.67
3.11
2.36
0.10
0.35
4.21
0.14
1.67
ANN
0.20
63.12
43.97
0.33
6.92
8.39
1.68
13.46
2.59
0.11
2.49
6.70
0.24
0.72
UIN 149
SPR 149
SUM 149
AUT 149
ANN
0.29 12 18.63 12 35.61 13 0.39 0.71 2.52 13 1.17 1.16 1.40
0.11 5 4.53 5 23.65 12 0.22 0.31 3.33 12 1.33 0.58 0.93
0.09 0.34 5.07 0.02 0.22
0.07 0.08 2.19 0.05 1.21
WIN 105 0.07 13 3.46 13 32.94 13 0.C8 0.37 5.83 13 1.72 2.28 1.57 0.09 0.14 2.06 0.02 0.32
SPR 105 0.14 13 10.33 13 45.15 13 0.17 0.27 2.32 13 1.63 1.78 1.34 0.11 0.29 3.20 0.03 0.28
SUH 105 0.27 13 18.26 13 40.13 13 0.35 0.63 2.24 12 1.30 1.73 1.54 0.09 0.39 4.91 0.01 0.18
AUT 105 0.07 11 3.16 11 29.58 12 0.08 0.20 3.76 10 1.65 1.67 1.33 0.07 0.17 2.66 0.01 0.27
ANN 0.14 35.22 36.95 0.17 1.47 3.54 1.57 7.46 1.45 0.09 0.99 3.21 0.08 0.26
UIN
104
0.06
12
2.35
12
22.14
12
0.14
1.84
17.05
12
1.27
2.15
2.17
0.09
0.23
3.31
0.05
0.64
SPR
104
0.20
13
15.26
13
37.62
13
0.30
2.08
9.02
13
1.27
2.55
2.54
0.12
0.49
5.05
0.03
0.34
SUH
104
0.28
12
20.73
12
38.72
13
0.37
2.11
7.92
13
0.96
2.13
3.01
0.09
0.47
7.57
0.02
0.21
AUT
104
0.08
13
2.77
13
18.50
13
0.17
1.52
12.46
13
1.21
1.79
1.91
0.07
0.20
3.49
0.04
0.64
90
-------
APPENDIX C
1991 NOON DATA
(continued)
ANN
0.16
41.10
29.25
0.24
7.55
11.61
1.18
8.61
2.41
0.09
1.39
4.86
0.13
0.46
U1N
144
0.07
13
2.59
12
20.89
12
0.08
1.02
16.41
13
2.29
3.64
2.09
0.13
0.37
3.73
0.14
1.34
SPR
144
0.31
13
25.81
13
41.58
13
0.42
2.29
7.33
13
2.11
6.45
4.03
0.17
0.86
6.39
0.08
0.58
SUM
144
0.35
13
31.89
13
43.25
13
0.53
2.66
6.35
13
1.64
5.84
4.59
0.14
0.97
9.03
0.03
0.29
AUT
144
0.11
13
4.41
12
17.05
12
0.15
1.21
12.47
13
2.02
3.01
1.99
0.11
0.35
4.08
0.12
1.43
ANN
0.21
64.70
30.69
0.30
7.17
10.64
2.02
18.94
3.18
0.14
2.54
5.81
0.37
0.91
UIN
109
31.23
13
3.84
13
0.64
2.14
0.30
SPR
109
0.13
12
B.38
12
37.68
13
0.16
0.12
1.23
13
0.81
0.46
0.71
0.07
0.13
2.38
0.01
0.14
SUM
109
0.22
13
12.30
13
28.93
13
0.27
0.18
0.79
13
0.58
0.30
0.64
0.05
0.17
4.08
0.00
0.10
AUT
109
0.08
12
3.44
12
24.41
12
0.12
0.14
1.75
12
0.76
0.42
0.72
0.04
0.07
2.22
0.00
0.15
91
------- |