600382094
FINAL EVALUATION OF URBAN-SCALE
PHOTOCHEMICAL AIR QUALITY SIMULATION MODELS
by
Kenneth L. Schere and Jack H. Shreffler
Meteorology and Assessment Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina. 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. Mention of trade names or commerical products does not constitute
endorsement or recommendation for use.
AFFILIATION
Dr. Shreffler and Mr. Schere are on assignment to the Meteorology and
Assessment Division, Environmental Sciences Research Laboratory, from the
National Oceanic and Atmospheric Administration, U.S. Department of Commerce.
n
-------
ABSTRACT
The research study discussed here is a continuation of previous work whose
goal was to determine the accuracy of several selected urban photochemical air
quality simulation models using data from the Regional Air Pollution Study in
St. Louis. The report summarizing the initial work on this project discussed
four models and used a sample of 10 days for testing. This work continues
testing on three of the models with an increased sample size of 20 days.
The models evaluated here are: The Photochemical Box Model (PBM) de-
veloped in-house by EPA, The Lagrangian Photochemical Model (LPM) developed by
Environmental Research and Technology, Inc., and The Urban Airshed Model (UAM)
developed by Systems Applications, Inc. Emphasis in this report is directed at
the ability of the models to reproduce the maximum 1-hour ozone concentrations
observed on 20 days selected from nearly "2 years of data. The PBM, LPM, and
UAM have been evaluated using statistical methods and graphical techniques and
all show potential as air quality management tools.
The standard deviation of the differences between observed ozone maxima
and predicted concentrations at the same place and time ranged from 0.04 to
0.06 ppm for maxima of 0.16 to 0.26 ppm. This measure of uncertainty should be
recognized by decision-makers using these models in regulatory and planning
processes.
This report covers a period from March 1981 to February 1982.
-------
CONTENTS
Abstract iii
Figures vi
Tables " ix
1. Introduction 1
2. Conclusions and Recommendations 5
3. The Data Set . . . 8
4. Photochemical Box Model 12
Overview 12
Results 13
Conclusions 21
5. Lagrangian Photochemical Model 39
Introduction 39
Method 39
Results 43
Conclusions 46
6. Urban Airshed Model 63
Introduction 63
Results 65
Conclusions .- 74
References 93
Appendices
A. PBM - Time series for remaining test days 95
B. PBM - Summary of statistical results 126
C. LPM - Time series for remaining test days 132
D. UAM - RS and RO plots for remaining test days 179
E. UAM - Contour plots for remaining test days 228
F. UAM - Summary of statistical results 237
-------
FIGURES
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
A-l
A-2
A-3
A-4
A-5
A-6
A-7
A-8
A-9
Map of the St. Louis area ~
Schematic drawing of the PBM domain
Map of the PBM domain
PBM simulation results for Day 142 of 1975
PBM simulation results for Day 178 of 1975
PBM simulation results for Day 207 of 1975
PBM simulation results for Day 212 of 1976
PBM simulation results for Day 275 of 1976
PBM frequency histograms for all data
PBM residual vs. obs. cone, plots for all data ....
LPM simulation results for Day 231 of 1975
LPM simulation results for Day 251 of 1975
LPM simulation results for Day 226 of 1976
LPM simulation results for Day 237 of 1976
LPM simulation results for Day 275 of 1976
Schematic drawing of the UAM domain
UAM simulation results for CO on Day 142 of 1975 . . .
UAM simulation results for N02 on Day 142 of 1975 . . .
UAM simulation results for 03 on Day 142 of 1975 . . .
UAM contours of max. 03 field on Day 142 of 1975 . . .
UAM simulation results for CO on Day 207 of 1975 . . .
UAM simulation results for NOg on Day 207 of 1975 . . .
UAM simulation results for 03 on Day 207 of 1975 . . .
UAM contours of max. 03 field on Day 207 of 1975 . . .
UAM simulation results for CO on Day 195 of 1976 . . .
UAM simulation results for N0£ on Day 195 of 1976 . . .
UAM simulation results for 03 on Day 195 of 1976 . . .
UAM contours of max. 03 field on Day 195 of 1976 . . .
UAM simulation results for CO on Day 275 of 1976 . . .
UAM simulation results for M02 on Day 275 of 1976 . . .
UAM simulation results for 03 on Day 275 of 1976 . . .
UAM contours of max. 03 field on Day 275 of 1976 . . .
PBM simulation results for Day 182 of 1975
PBM simulation results for Day 183 of 1975
PBM simulation results for Day 184 of 1975
PBM simulation results for Day 209 of 1975
PBM simulation results for Day 221 of 1975
PBM simulation results for Day 230 of 1975
PBM simulation results for Day 231 of 1975
PBM simulation results for Day 251 of 1975
PBM simulation results for Day 159 of 1976
11
...'.. 23
24
25
27
29
31
33
35
37
48
51
54
57
60
76
77
...... 78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
96
98
100
102
104
106
108
110
112
-------
A-10 PBM simulation results for Day 160 of 1976 114
A-ll PBM simulation results for Day 195 of 1976 116
A-12 PBM simulation results for Day 211 of 1976 118
A-13 PBM simulation results for Day 225 of 1976 120
A-14 PBM simulation results for Day 226 of 1975 122
A-15 PBM simulation results for Day 237 of 1976 124
C-l LPM simulation results for Day 142 of 1975 134
C-2 LPM simulation results for Day 178 of 1975 137
C-3 LPM simulation results for Day 182 of 1975 140
C-4 LPM simulation results for Day 183 of 1975 143
C-5 LPM simulation results for Day 184 of 1975 146
C-6 LPM simulation results for Day 207 of 1975 149
C-7 LPM simulation results for Day 209 of 1975 152
C-8 LPM simulation results for Day 221 of 1975 155
C-9 LPM simulation results for Day 230 of 1975 158
C-10 LPM simulation results for Day 159 of 1976 161
C-ll LPM simulation results for Day 160 of 1976 164
C-12 LPM simulation results for Day 195 of 1976 167
C-13 LPM simulation results for Day 211 of 1976 170
C-14 LPM simulation results for Day 212 of 1976 173
C-15 LPM simulation re'sul ts for Day 225 of 1976 ' . . 176
D-l DAM simulation results for CO on Day 178 of 1975 180
D-2 UAM simulation results for NOg on Day 178 of 1975 181
0-3 UAM simulation results for 03 on Day 178 of 1975 182
D-4 UAM simulation results for CO on Day 182 of 1975 183
D-5 UAM simulation results for N02 on Day 182 of 1975 184
D-6 UAM simulation results for 03 on Day 182 of 1975 185
D-7 UAM simulation results for CO on Day 183 of 1975 186
D-8 UAM simulation results for N02 on Day 183 of 1975 187
D-9 UAM simulation results for 03 on Day 183 of 1975 188
0-10 UAM simulation results for CO on Day 184 of 1975 189
0-11 UAM simulation results for N02 on Day 184 of 1975 190
D-12 UAM simulation results for 03 on Day 184 of 1975 191
D-13 UAM simulation results for CO on Day 209 of 1975 192
D-14 UAM simulation results for NOg on Day 209 of 1975 193
D-15 UAM simulation results for 03 on Day 209 of 1975 194
0-16 UAM simulation results for CO on Day 221 of 1975 195
D-17 UAM simulation results for HO? on Day 221 of 1975 196
0-18 UAM simulation results for 03 on Day 221 of 1975 197
D-19 UAM simulation results for CO on Day 230 of 1975 198
D-20 UAM simulation results for N02 on Day 230 of 1975 199
D-21 UAM simulation results for 03 on Day 230 of 1975 200
D-22 UAM simulation results for CO on Day 231 of 1975 201
D-23 UAM simulation results for N02 on Day 231 of 1975 202
D-24 UAM simulation results for 03 on Day 231 of 1975 203
D-25 UAM simulation results for CO on Day 251 of 1975 204
D-26 UAM simulation results for NOg on Day 251 of 1975 205
D-27 UAM simulation results for 03 on Day 251 of 1975 206
D-28 UAM simulation results for CO on Day 159 of 1976 ........ 207
D-29 UAM simulation results for M02 on Day 159 of 1976 208
D-30 UAM simulation results for 03 on Day 159 of 1976 209
-------
D-31 UAM simulation results for CO on Day 160 of 1976 210
D-32 UAM simulation results for N02 on Day 160 of 1976 211
D-33 UAM simulation results for 03 on Day 160 of 1976 212
D-34 UAM simulation results for CO on Day 211 of 1976 213
D-35 UAM simulation results for M02 on Day 211 of 1976 214
D-36 UAM simulation results for 03 on Day 211 of 1976 215
D-37 UAM simulation results for CO on Day 212 of 1976 216
D-38 UAM simulation results for N02 on Day 212 of 1976 217
D-39 UAM simulation results for 03 on Day 212 of 1976 218
D-40 UAM simulation results for CO on Day 225 of 1976 219
D-41 UAM simulation results for NOg on Day 225 of 1976 220
D-42 UAM simulation results for 03 on Day 225 oT 1976 221
D-43 UAM simulation results for CO on Day 226 of 1976 222
D-44 UAM simulation results for N02 on Day 226 of 1976 223
D-45 UAM simulation results for 03 on Day 226 of 1976 224
D-46 UAM simulation results for CO on Day 237 of 1976 ........ 225
D-47 UAM simulation results for N02 on Day 237 of 1976 226
D-48 UAM simulation results for 03 on Day 237 of 1976 227
E-l UAM contours of max. 03 field on Day 178 of 1975 ....'.... 229
E-2 UAM contours of max. 03 field on Day 182 of 1975 229
E-3 UAM contours of max. 03 field on Day 183 of 1975 230
E-4 UAM contours of max. 03 field on Day 184 of 1975 230
E-5 UAM contours of max. 03 field on Day 209 of 1975 231
E-6 UAM contours of max. 03 field on Day 221 of 1975 231
E-7 UAM contours of max. 03 field on Day 230 of 1975 232
E-8 UAM contours of max. 03 field on Day 231 of 1975 232
E-9 UAM contours of max. 03 field on Day 251 of 1975. 233
E-10 UAM contours of max. 03 field on Day 159 of 1976 233
E-ll UAM contours of max. 03 field on Day 160 of 1976 -.234
E-12 UAM contours of max. 03 field on Day 211 of 1976 234
E-13 UAM contours of max. 03 field on Day 212 of 1976 235
E-14 UAM contours of max. 03 field on Day 225 of 1976 235
E-15 UAM contours of max. 03 field on Day 226 of 1976 236
E-16 UAM contours of max. 03 field on Day 237 of 1976 236
viii
-------
TABLES
Number
1 Summary of Meteorological Conditions on 20 Test Days 10
2 Predicted and Observed 03 Maxima for the PBM, 20
3 Initial Pollutant Concentrations Assigned in the LPM 42
4 Predicted and Observed 03 Maxima for the LPM 47
5 Predicted and Observed 03 Maxima for the UAM 71
6 03 Regression Statistics for the UAM . . . 73
B-l Summary of Model Performance Statistics from the PBM 128
C-l Temporal Correlation Coefficients for the LPM 133
F-l Summary of Mean Concentration Statistics from the UAM 238
F-2 Summary of Analysis of Trends Data from the UAM 243
F-3 Summary of Concentration Maxima Statistics from the UAM 246
IX
-------
SECTION 1
INTRODUCTION
This report covers a final evaluation of selected urban scale photochem-
ical models which could be applied in regulatory situations. A report (herein-
after referred to as Report I) on a preliminary evaluation was prepared by
Shreffler and Schere (1982). In the interest of brevity, material in Report I
is not repeated here unnecessarily. Referenced documentation on the models and
Report I are essential to fully understand the procedures used in the simula-
tions.
The Regional Air Pollution Study (RAPS) was conducted in the St. Louis
region over the period 1974-1977 (Schiermeier, 1978). RAPS was designed to
provide a comprehensive data set for the testing and evaluation of numerical
air quality simulation models on an urban scale. While the RAPS field measure-
ments were in progress EPA surveyed the. available, -state-of-the-art, photo-
chemical air quality simulation models, and selected 3 for evaJuation.
Under contract, the builders adapted their models to the RAPS data base and
were supplied 3 days of data for test simulations. In addition, there was
a need to look at a simple box-model approach which was not embodied in an
existing model. A box model was therefore constructed by EPA.
The following models were investigated in the evaluation program.
Photochemical Box Model (PBM) - a single cell Eulerian model constructed
by EPA.
Lagrangian Photochemical Model (LPM) - a multi-level parcel model deve-
loped by Environmental Research and Technology, Inc.
Livermore Regional Air Quality Model (LIRAQ) - a single-level Eulerian
grid model developed by Lawrence Livermore Laboratory.
-------
Urban Airshed Model (UAM) - a multi-level, Eulerian grid model developed
by Systems Applications, Inc.
LIRAQ, described by MacCracken et al . (1978), was dropped from the
evalution because of a series of technical and logistics problems detailed in
Report I. LIRAQ was unable to produce significant ozone levels for St. Louis,
but no specific, correctable error could be identified. Aside from the tech-
nical problems with the model, which are unresolved and unevaluated, LIRAQ did
not seem well suited as a general-use model easily transferable to new locales.
No conclusions can be made from this effort about the use of LIRAQ by special-
ized user groups at specific locales.
Some of the other conclusions reached from the preliminary work described
in Report I are that (a) the PBM performs best in near-stagnation conditions,
(b) the LPM needs modification of the fixed-box formulation for the air parcel
and (c) the UAM tends to consistently underpredict ozone. It was also general-
ly found that the variability between specific 1-hour predicted and observed
concentrations at a particular location tends to be substantial.
The PBM, LPM and UAM were again tested extensively for this study and
corrections of obvious errors or deficiencies were made. However, no effort
was made to adjust or tune the model predictions to observed concentration
values. The prevailing philosophy behind the evaluation effort was to use the
models in an off-the-shelf mode, much as they eventually would be applied by a
user in a regulatory situation. However, great care was taken in preparation
of data sets and model executions. Although model assumptions vary, effort was
made to use data in similar manners in all models. Data preparation and actual
execution of the models was accomplished solely by the authors at EPA., The
goal of the evaluation was to provide a fair and objective determination of the
accuracy of a set of photochemical models when tested in an operational mode
against a comprehensive urban data base.
Photochemical air quality models are designed to predict ozone concentra-
tions and thereby provide a means to assess beforehand the effects of changing
emission levels in an urban area. Emission changes may result from addition of
-------
new sources or controls on existing sources designed to reach ozone levels
defined by the National Ambient Air Quality Standards (MAAQS). In either case,
knowledge of the effects of emissions changes would allow prudent economic
decisions on plant siting or control equipment use. The numerical air quality
model synthesizes knowledge about the emissions, transport and transformation
of pollutants into a workable computational framework, but the central problem
•
is how to be reasonably certain that the model is correct in its projections.
The only reasonable method of determining accuracy is to test the model
against an extensive observational data base. Providing such a data base was
the purpose behind the RAPS. Since the level of the ozone maximum for each day
is of paramount importance relative to the MAAQS,- the comparison between its
observed value and the model prediction at the same time and place would be of
central interest. The method of evaluation consists of selecting a set of
test days, executing model simulations and computing residual concentrations
(observed-predicted). The specific outcome of the evaluation is a presentation
of information on residuals under the given circumstances. Conclusions about
model acceptability require further assumptions and judgements.
The residual reflects a composite effect of input errors, input uncertain-
ties and model errors. Errors, if not compensating, will tend to produce a
non-zero average residual over many cases. If the average residual is nearly
zero, we might be willing to accept that the model is appropriately formulated
for predicting absolute levels of high ozone in the area where the observations
were taken. The case-by-case variation could be ascribed to uncertainties in
the input. The acceptance of the model for absolute prediction leads rather
naturally to accepting that it will respond to emissions changes in the same
manner as the real-world system. Finally, we are likely to conclude that the
model is transferable to another location, provided the meteorological and
chemical modules are judged sufficiently flexible to adapt to the new condi-
tions. This line of reasoning underlies the decision-making process attendant
to model use.
This report details the performance of 3 completed models using the
RAPS data base. Input information and model components were carefully checked
-------
before proceeding. Effort was directed at ascertaining that individual modules
were doing what they were purported to do, but not to evaluate their accuracy
separately. Improvements are perhaps possible in some of the modules, but this
is left to the future. This report indicates the accuracy of the finished
models used in an operational mode with the best information available at
present.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
This report presents the second, and final, evaluation of urban-scale
photochemical air quality simulation models by EPA's Meteorology and Assessment
Division. The 3 models included in the analysis'span a wide range in complexi-
ty and sophistication but are all based on numerical solutions to mass-conser-
vative equations. They are selected from the general categories of box,
trajectory (Lagrangian) and grid (Eulerian) models. The first evaluation,
described in Report I, gave detailed results from the model runs on 10 days
selected for simulation from the RAPS data base. The sample size for this
evaluation is 20 days; an additional 10 days were combined with the original
sample. Some changes in the model algorithms, including the data preproces-
sors, from the time of the original testing on the 10-day sample made it
necessary to perform the original simulations over. Emphasis in the model
performance evaluations is placed on ozone, although -results for other pollu-
tant species are also discussed. Conclusions on the performance of- specific
models did not change appreciably from those listed in Report I although our
confidence in them has increased because of the consistency in results obtained
from both studies.
The PBM predictions for maximum 03 for the average of the monitoring
stations within the model domain were generally on the high side. The average
03 residual implied a 23% overprediction over all test days. However for
the 5 stagnation-type days where the maximum observed 03 occurred within the
PBM domain, the average overprediction of 8* was considerably better than for
the entire sample. Only a slight tendency towards overprediction was indicated
for the LPM. The biases of the residuals were relatively small, 11% of the
average observations at Level-1 and only 2.5% at Level-3. The standard devia-
tions of the LPM residuals were the highest among the three models tested, but
nearly halved from the initial tests described in Report I. The large variance
in the residuals might be expected since the LPM generates a prediction which
likely is the most specific to a particular place and time- Model predictions
-------
for maximum 03 by the UAM in a specific sense (at the same time and location
as the observed maximum) were consistently low for all evaluations with an
average 32" underprediction over the sample. Originally it was thought that
spurious numerical diffusion in the advection component of the model was
responsible for the low UAM predictions. However the underprediction problem
still persisted after the original advection algorithm was replaced with one
containing much less inherent numerical diffusion. If the time and location of
the model predictions are not constrained to ^be the same as those for the
maximum observed 03, the average 'model bias for the 20 days implied a 4%
overprediction. This excellent agreement might suggest that the uncertainty in
specifying a wind field for a grid model like the UAM could lead to large
apparent errors in the model results.
The choice of which particular model to use in a specific application
involves not only the accuracy of the model but also the resources required to
operate it. The models test'ed here have resource requirements correlated with
their level of complexity. In terms of man-months needed to set up a single
day simulation and computer time expended (minutes of CPU on a UNIVAC 1100/82)
the approximate requirements are:
PBM 0.15 man-month 1 minute CPU
LPM 0.20 man-month 10 minutes CPU
UAM 0.50 man-month 110 minutes CPU
The 3 models discussed in this report have all shown themselves to be
acceptable tools for analysis of urban ozone air quality. The specific config-
uration of an application along with the quantity and quality of related data
and resources available to the user must all be considered in the final selec-
tion of a model. For an indication of average 03 air quality in an urban
area under stagnation conditions or as a screening method for a more complex
model the PBM is appropriate. The choice of a trajectory model, such as the
LPM, or a grid model, like the UAM, might well be decided by resource require-
ments or by the number of proposed simulations. - In any event, the user of any
of these models must have a strong scientific background and exercise extreme
care in implementing the air quality simulations. For an air quality analyst
making regulatory decisions based, in part, on model results it should be
-------
understood that substantial variability can exist between a single 1-hour
predicted and observed concentration at a particular location.
These models are now being made available to EPA's Office of Air Quality
Planning and Standards for further statistical and sensitivity testing and
ultimately for use in their regulatory decision-making process. Because model
development is an evolving area of research it is very likely that subsequent
"improved" versions of the models tested here will become available. A per-
formance test with benchmark results, as described and tabulated in this
report, now exists for future use in urban air quality model comparisons with
any subsequent versions of the models.
-------
SECTION 3
THE DATA SET
As an integral part of the RAPS a network of 25 surface stations was
established in and around the St. Louis region. These stations, shown in
Figure 1, comprise the- Regional Air Monitoring System (RAMS). The RAMS
stations continually monitored various meteorological variables as well as
ambient concentrations of pollutant gases. Schiermeier .(1978) provides
details on the instrumentation. In addition to the RAMS, upper air balloons
were released each hour from urban and rural sites to provide wind profiles
for modeling purposes.
An essential element in the data set for an air quality modeling exercise
is a complete source emissions inventory for all gaseous and particulate
species over the domain of interest. Littman (1979) summarizes the comprehen-
sive data gathering and archiving of the RAPS emissions inventory. The inven-
tory is divided into point source and area source classes with points defined
as sources emitting at least 0.01 percent of the total emissions of a pollutant
for the whole air.quality control region. Line source emissions are included
in the area source class. The area source emissions data are subdivided onto a
horizontal grid of 1989 cells with variable spatial resolution, dependent upon
the emissions source density. Grid resolution varies from 1 km in the more
dense areas to 10 km in the sparse areas. The temporal resolution of the
emission inventory is one hour. Criteria pollutants included in the inventory
are TSP (total suspended particles), S0£ (sulfur dioxide), NOX (nitrogen
oxides), THC (total hydrocarbons), and CO (carbon monoxide). Furthermore,
total hydrocarbons are broken down into the component classes of non-reactives,
olefins, paraffins, aldehydes, and aromatics. Examples of the gridded spatial
distribution of selected pollutants are shown in Report I.
-------
This report presents the results of model simulations for 20 individ-
ual days chosen from the RAPS data base. These days, 11 from 1975 and 9
from 1976, account for some of the higher 03 measurements observed in the
RAPS surface monitoring network. Maximum hour-average single station 03
values all exceeded 0.16 ppm on the 20 days. Table 1 lists the basic char-
acteristics of the days: wind speed and direction, temperature, and solar
radiation computed as averages over 0700-1359 CST. Also listed are the maximum
mixing heights for each day as determined from the upper air sounding data.
Higher 03 levels generally occur when a prevailing high atmospheric pressure
system exists over the area with little associated cloud cover, and represent
the situations conducive to production of 03 from locally generated precursor
emissions. Both the dates and Julian day numbers for the 20 days are listed in
Table 1 and either form may be used throughout this report to reference a
particular day.
-------
TABLE 1. A SUMMARY OF WIND SPEED (WS), WIND DIRECTION (WD), TEMPERATURE,
SOLAR RADIATION (ALL WAVELENGTHS), AND MAXIMUM AFTERNOON MIXING
HEIGHT (Mh) FOR THE 20 DAYS EXAMINED.*
Date
Day
(Julian)
WS
(m/s)
WD
(deg)
Temp
Solar
(ly/min)
Max MH
(m)
—
5/22/75
6/27/75
7/01/75
7/02/75
7/03/75
7/26/75
7/28/75
8/09/75
8/18/75
8/19/75
9/08/75
6/07/76
6/08/76
7/13/76
7/29/76
7/30/76
8/12/76
8/13/76
8/24/76
10/01/76
. 142
178
182
183
184
207
209
221
230
231
251
159
160
195
211
212
225
226
237
275
1.1
0.4
1.4
1.4
1.8
1.0
2.0
0.4
1.6
1.3
1.8
1.0
1.3
2.3
0.3
1.7
2.3
1.1
1.3
0.6
224
245
70
15
324
139
18
88
167
168
181
129
284
145
251 -
205
253
273
110
222
29
29
29
30
30
26
30
26
27
28
25
25
27
28
26
30
29
30
28
22
1.12
0.96
0.99
0.92
0.85
0.98
0.98
0.98
0.96
0.95
0.89
1.06
1.01
1.02
0.53
0.82
0.70
0.86
0.82
. 0.78
1504
1822
. 2606
2488
1875
1477
1909
1195
1488
1052
1797
1972
1772
1853
1706
1304
730
1427
2124
527
*Meteorological variables (except MH) are network averages over
the period 0700-1359 CST.
10
-------
Figure 1. The St. Louis area with locations of the
RAPS surface stations.
11
-------
SECTION 4
PHOTOCHEMICAL BOX MODEL
INTRODUCTION
The Photochemical Box Model (PBM), a single cell Eulerian air quality
simulation model, simulates the transport and chemical transformation of air
pollutants in smog-prone urban atmospheres. The model's domain is set in a
variable volume, well-mixed reacting cell where the physical and chemical
processes responsible for 03 generation by HC and NOX precursors are mathemat-
ically created. These processes include the transport and dispersion of
pollutant species through the cell, the injection of primary precursor species
by emission sources, and the chemical transformation of the reactive species
into intermediate and secondary products. They are schematically illustrated
in Figure 2.
To apply the model to the St. Louis RAPS data base, the horizontal length
scale of the cell was set at 20 km and the vertical scale was time dependent,
proportional to the depth of the mixed layer. The model domain was chosen such
that it included most major emission sources on either side of the Mississippi
River. Source emissions were assumed to be distributed uniformly across the
surface face of the cell. Twelve of the RAMS surface monitoring stations as
well as one upper air sounding location were located within the cell bound-
aries. This configuration of the model domain is illustrated in Figure 3.
The steps involved in the method for performing a PBM simulation are
described in the first evaluation report by Shreffler and Schere (1982),
Report I. There are 3 data preprocessors that must be exercised prior to the
model simulation itself. They access the RAPS data base and retrieve the
appropriate parameters required by the PBM.
12
-------
The framework of the model itself has not changed from the time of Report
I to the present. There was a change made, however, in the method of choosing
the upwind HC boundary condition for the model. Originally, if the HC data
were invalid from the chosen upwind monitoring locations at a particular hour
an alternate algorithm for obtaining these data was invoked. It was a ratio
technique based on the average non-methane hydrocarbon (NMHC) to CO fraction
over the model simulation period. In this way the boundary condition for HC
could be obtained indirectly from the more reliable CO observations. It was
found however, that this method occasionally gave spuriously large HC values at
the upwind boundary. In a simulation with significant advection through the
model domain this had the potential for creating an overly reactive mix of
photochemical pollutant species in the box that would lead to an overprediction
of 03. In the simulations presented here, if a particular HC background
concentration was invalid, a value of 0.05 ppmC for NMHC was substituted.
This is a low-level background value for this quantity that borders on the
threshold limits of the monitoring instrumentation. In practice this change
caused only small differences in most of the simulation results, although in
some cases it had a noticeable effect. Background concentrations of species
other than 03 were generally very small, while those of 03 were often a
significant fraction of the total 03 burden in the model domain.
PBM simulations, were conducted for each of the 20 test days listed in
Table 1, including a repetition of the original 10 days described in Report I.
Although results for 03 predictions on all days are summarized, detailed
descriptions are presented for only 5 days. Simulation results for pollu-
tants CO, N02, and 03 for Julian days 142, 178, and 207 of 1975 and 212 and
275 of 1976 are emphasized.
RESULTS
Among the 20 selected days for model simulations from the RAPS data base
are 5 in which the maximum observed afternoon 03 concentration occurred
within the PBM domain. These were days when stagnation conditions existed with
light and variable winds near the surface. Very little of the primary emis-
sions moved out of the modeled area. During this type of episode, a single
13
-------
cell box model would be appropriate for air quality simulation, since most of
the pollutant mass stays within the model domain. On the other modeled days,
the maximum 03 level occurred farther downwind, outside of the PBM's boundar-
ies. Simulations were also conducted on these days even though maximum 63
measurements within the model domain were somewhat lower than those downwind.
Results are described in detail for those 5 days where the maximum ob-
served 03 concentration occurred within the model domain. Report I high-
lighted 2 of these days as well as a case of stronger advection. -The latter
case was used to demonstrate the 'inertia1 effect of numerical box models that
can lead to problems where large temporal concentration gradients exist. The
interested reader is referred to the earlier report for details.
The PBM simulated concentrations represent spatial averages over the
volume of the box. The appropriate observed concentration with which to
compare predictions is the average over all 12 monitoring sites within the box.
For the 5 days detailed here this concentration for 03 exceeded the 1-hour-
National Ambient Air Quality Standard (NAAQS) of 0.12 ppm at least once during
each simulation period. These observed average concentrations are represented
by the open circles in the concentration time series described below~. The
observed range of concentration values among the 12 sites within the PBM domain
is another feature to consider. The highest and lowest measurements are
described by the dashed lines in the time series. A coarse measure of model
performance is made by observing whether the PBM predicted concentrations
(shown as solid lines), fall within the range of observed concentrations
(between the dashed lines) and preferably as close as possible to the average
observed concentrations (shown as circles).
The first simulation day considered is Day 142-75. Synoptic meteorolog-
ical features on this day included an expansive high-pressure system centered
over western Virginia with a general windflow from the southwest in the mid-
Mississippi Valley area. Wind speeds in St. Louis, however, were light and on
the order of 1.0 m s~*. Skies were completely clear. The maximum afternoon
hour-average 03 measurement occurred at noon at site 101 in downtown St.
Louis. According to the convention for time referencing employed here, the
1200 hour-average refers to the average concentration from 1200 to 1300 h.
14
-------
Figure 4a shows the time series for CO on Day 142. The model predictions
fall within the 'envelope1 of observed concentrations and actually track the
average observed values quite well. Carbon monoxide is included in the model
evaluation analysis because it reacts very slowly in the atmosphere and may be
considered virtually inert for the time scales discussed here. As such, any
problems with the emissions or the advective and dispersive portions of the PBM
would be reflected in the CO predictions, whereas the complex chemical inter-
actions of a more reactive pollutant could mask problems with transport or
emissions.
Figure 4b shows the time series for N02 on the same day. These model
predictions are consistently greater than the average observations throughout
the simulation but generally fall within the concentration envelope. Figure
4c displays the 63 time series. PBM predictions track the average observa-
tions through mid-morning and then diverge near noon when observed 03 levels
begin to decrease while the predictions peak one hour later at 1300 CST. The
model-predicted peak 03 concentration reached 0.148 ppm compared with the
average observed peak of 0.137 ppm, an 8* overprediction. This is somewhat
better than the 28% overprediction cited in Report I for Day 142. The model
predictions fall mostly within the concentration envelope except for the last
few hours of the simulation. Background concentrations of 03 on Day 142 were
on the order of 0.06- ppm.
The synoptic-scale meteorological features on Day 178-75 show the eastern
third of the U.S. covered by a large high-pressure system. Wind speeds were
light and directions variable near the surface as well as at upper levels of
the atmosphere. In St. Louis surface winds were generally less than 1.0 m
s~l. Skies were hazy most of the day with few clouds reported. Solar insola-
tion levels were high throughout the simulation period. The maximum hour-aver-
age 03 measurement occurred at site 112 in the western part of the model
domain at 1400 CST. Background concentrations of 03 on Day 178 were approxi-
mately 0.08 ppm.
The time series for CO on this day is presented in Figure 5a. The PBM
underpredicts CO consistently throughout the period although most predictions
15
-------
do fall within the concentration envelope. The most serious underpredictions
occur in the earliest hours of simulation where the differences observed among
the stations themselves are greatest. The large departures at the upper edge
of the concentration envelope at that time illustrate this. These high values
tend to bias the average within the model domain to the high side.
The NOg time series in Figure 5b shows the predictions to match the
observations throughout the period except for the, observed peak at 0700 CST. A
phenomenon may be occurring here similar to that which biased the CO peak
toward higher values. Nitrogen dioxide, however, is predominantly a secondary
pollutant and should therefore not demonstrate fluctuations ' as large as a
primary pollutant like CO. Predicted values for 03 for Day 178 are seen in
Figure 5c. They lag behind the observations during the morning hours and peak
two hours later in the afternoon. The rapid morning rise of the 03 observa-
tions may be due to a rather large entrainment rate of 03 aloft. If the as-
sumed rate of rise of the mixed layer in the PBM is too slow for this particu-
lar day the predictions would tend to lag the observations as they do here.
The model-predicted peak of 0.173 ppm was 8.8* greater than the observed value
of 0.159 ppm. This is an improvement over the 38* overprediction indicated in
Report I.
On Day 207-75 .a high-pressure system had settled into the northeastern
quarter of the U.S. after moving rapidly southeast from the western Canadian
provinces. Winds in the St. Louis area quickly dropped to near calm following
a frontal passage the previous day. Skies were clear throughout the day. The
maximum hour-average observed 03 concentration occurred at site 113 at 1400
CST, in the northwestern portion of the PBM domain. The background value of
03 on this day was generally about 0.07 ppm.
Model predictions of CO shown in Figure 6a agree quite well with observa-
tions for Day 207. Again there are several large positive anomalies along the
upper edge of the CO observed concentration envelope, similar to those occurr-
ences on other modeled days. In Figure 6b the model-predicted N02 concentra-
tions peak at the same hour and follow the same trend as the observations in
the time series shown, although the predicted values are consistently greater
16
-------
than those observed throughout the simulation period. With one exception they
remain within the N0£ observed concentration envelope.
Model-predicted 03 values on Day 207 rise more rapidly than the observa-
tions would suggest in Figure 6c. The average observed 03 peak is a sharp
one occurring at 1400 CST at 0.140 ppm, while the predicted peak is more
plateau-like with a high point at 1500 CST of 0.145 ppm, a 3.6% overprediction.
Possible reasons for the model's more rapid generation of 03 during the
morning hours include an excessive entrainment rate from aloft or an overly
reactive assumption concerning the mix of organic hydrocarbon species on this
day.
A weak stationary front separating 2 similar zones of high pressure was
situated' near St. Louis on Day 212-76. The front' was aligned mostly in an
east-west direction and had drifted north of the area during the previous day
enabling the southern-most air mass to dominate. Winds at St. Louis were
mostly from the south and southwest at speeds of 1-2 m s~l during the simulation
period. A layer of high clouds persisted throughout much of the day and
decreased the available insolation by 10 to 20%. The maximum hour-average
03 measurement occurred at noon at site 108, located in the northeastern
corner of the model domain. The background 03 concentration on this day was
0.10 ppm.
The time series for model-predicted concentrations of CO and N02 for Day
212 are shown in Figures 7a and b, respectively. In both cases the predictions
track the observations well through most of the period except for the early
hours when the peak observed concentration is underpredicted. The 03 time
series in Figure 7c shows good agreement between domain-averaged observations
and model predictions. The model-predicted concentrations rise more quickly in
the early hours and decline more slowly in the later hours of the simulation
than the observations. The average observed 03 peak occurred at noon at
0.127 ppm and the maximum estimated concentration was 0.121 ppm at 1400 CST, a
4.7% underprediction. It appears that less than 20% of the observed 03 was
produced locally on Day 212 in St. Louis since the background levels were only
slightly less than the highest afternoon values.
17
-------
Day 275-76 is considered to be the most extreme case of stagnation during
the 1975-76 RAPS project. The synoptic meteorological conditions revealed a
high-pressure region over the Texas-Arkansas area with northwest flow aloft
over St. Louis. Near the surface however, winds were very light and variable
in direction with little transport through the area. Skies were clear all day
and a strong subsidence inversion over the city held the mixed depth to a
maximum of just over 500 m above ground during the afternoon. The greatest
hour-average 03 measurement occurred at site 102, just north of downtown, at
1400 CST. Background 03 levels on Day 275 were approximately 0.06 ppm.
The time series for CO on this day is presented in Figure 8a. Overall
concentration levels substantially exceed those on the other modeling days
presented here. Report I discusses the implications of attempting to apply the
PBM to periods of extreme atmospheric stability as occurred at the very begin-
ning and end of the simulation period on Day 275. The net effect is a substan-
tial underprediction of the peak observed concentration during the stable
period. This can be seen in the CO time series. Both N02 and 03 concen-
tration levels, seen in Figures 8b and c, are also considerably greater than
those on the other days. The model-predicted N02 concentrations follow the
average observed concentrations well and show slightly smoother temporal
gradients than those observed. Simulated 03 concentrations also display a
trend similar to observed 03. The afternoon predictions actually track the
upper boundary of the concentration envelope better than they track the average
observed concentrations. The estimated 03 maximum is 0.223 ppm at 1500 CST,
while the average observed peak is 0.183 ppm at 1400 CST, a 21.9 percent
overprediction; nearly the same as that in Report I for Day 275.
Model performance on these 5 days indicates the PBM can simulate the
average urban air quality for selected pollutants in relatively stagnant
conditions. The weakest performance seems to be with the primary pollutant CO
because of the non-homogeneous character of its emission and dispersal. While
the areas of maximum observed 03 on the remaining 15 days listed in Table 1
did not occur within the PBM domain the model was exercised for these days as
well in order to provide a consistent statistical summary with the other models
discussed in this report. In the tables and figures that follow data from
model performance on all 20 days are included. :
18
-------
One significant aspect of photochemical transformation omitted from these
discussions on PBM simulation of an urban atmosphere is the presence of vola-
tile organic compounds, or hydrocarbons (HC), in the smog mixture. While the
chemical kinetic mechanism in the PBM does account for HC interactions with
other reactive species, a discussion of the details of the mechanism is beyond
the scope of this paper.
Table 2 summarizes the 03 maxima predicted and PBM domain-average observed
maximum concentrations and an elementary statistical analysis of the results
for all 20 modeled days. The "specific" model predictions correspond to the
same hour as the observed maximum, and the "independent" predictions represent
the peak at any hour of the simulation. The specific and independent predic-
tions may not coincide, indicating a phase lag between observed and predicted
03 peaks. This lag often appears when the maximum observed 03 within the
model domain occurs before noon. Statistics on the residual concentration, AC,
have been computed for both the specific and independent predictions. Both the
average signed residual and absolute residual, and standard deviation, are
presented for the analyses. The average residual is negative in both the
specific and independent cases, indicating an overprediction of 03, although
the specific value is one-half the independent value. The values of the
average absolute residuals are both different from the average signed resi-
duals. This implies that there were underpredictions as well as overpredic-
tions among the individual days. The magnitude of the standard deviation
(s.d.) is slightly greater than the average residual. The discrepancy on Day
251 accounts for a large portion of the s.d. The average value of the model's
overprediction for 03 over all 20 days is 23%, compared to the 31% overpre-
diction for the original 10 days cited in Report I. It should be noted that
the average value of the overprediction over the 5 days where the maximum
measured 03 occurred within the PBM domain is 7.5%, considerably less than
that for the full set of days.
Although the analysis in Table 2 summarizes one aspect of model perform-
ance, it is clearly not a complete picture of the accuracy or precision of the
PBM. To more fully appreciate PBM performance, the model can be viewed from a
broader statistical perspective. Figure 9 presents frequency histograms (RS)
19
-------
TABLE 2. A SUMMARY OF PREDICTED AND OBSERVED 03 MAXIMA FOR THE PBM
Julian
date*
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
Hour
(CST)
12
• 14
13
13
13
14
10
13
13
10
15
13
14
15
15
12
12
15
11
14
Observed
at 4-meters
(ppm)
0.137
0.159
0.094
0.115
0.114
0.140
0.108
0.132
0.072
0.084
0.084
0.125
0.162
0.151
0.094
0.127
0.111
0.144
0.115
0.183
PBM
Specific
(ppm)
0.125
0.162
0.128
0.070
0.182
0.144
0.087
0.125
0.096
0.059
0.205
0.203
0.166
0.132
0.084
0.114 '
0.088
0.171
0.132
0.216
Predicted
Independent
(ppm)
0.148
0.173
0.140
. 0.097
0.182
0.145
0.122
0.125
0.119
0.101
0.205
0.210
0.170
0.137
0.088
0.121
0.092.
0.171
0.156
0.223
aTwenty days selected from 1975 and 1976.
For the data displayed above:
AC = Obs - Specific
IT = -0.012
s.d.UC) = 0.039
TACT = 0.029
AC = Obs - Independent
1C = -0.024
s.d.UC) = 0.036
= 0.031
20
-------
of CO, N02 and 03 residual concentrations, the observed minus predicted
model results, over all test days. An ideal distribution would be narrow with
a peak at the 0.0 residual. The 3 histograms shown display a range of distri-
butions. The RS for CO is skewed toward the underpredictive side while
03 is skewed toward the overpredictive. The nitrogen dioxide RS is more
symmetric than the others but does contain a few outliers on the underpredic-
tive side. All 3 distributions show a peak at or very near 0.0.
A plot of the residuals versus observed concentrations (RO) should indic-
ate whether a bias in the predictions displays any trend with the magnitude of
the measured concentrations. Figure 10 shows RO plots for the' same species as
presented in Figure 9. For CO and N02 there appears to be a trend toward
greater underprediction with higher ambient concentrations. The trend is more
clear with CO. Ozone however, displays no particular trend with observed
concentration, but the PBM generally overpredicts ambient values above 0.05
ppm.
The PBM has been exercised on a variety of days taken from the RAPS data
base. Prevailing meteorological conditions occurring some of the days may not
have been compatible with the environment for optimum model results. Using the
PBM in a regulatory capacity could be restricted to worst case stagnation
conditions where the model seems to perform best. Time series for pollutant
species on test days other than those discussed here are displayed in Appendix
A. Also, tabulated model performance statistics for data on mean concentra-
tions, analysis of trends, and concentration maxima from the PBM are provided
in Appendix B.
CONCLUSIONS
A thorough evaluation of the performance of the PBM has been completed.
This analysis is a major portion of the documentation of the model's abilities
for air quality simulation. Evidence shows that the model is a useful tool in
assessing urban air quality for photochemically reactive pollutants, especially
in stagnation conditions. The PBM is relatively simple to use and its data
requirements are far less stringent than most other numerical air quality
simulation models. Areas of further study that may be pursued include:
21
-------
(1) the hysteresis problem during advection conditions, discussed in
Report I;
(2) the relationship between the average 03 concentration observed
within the model domain and the maximum 03 level observed at a
single station;
(3) the continued testing and refinement $f the chemical kinetic mechan-
ism within the PBM;
(4) model sensitivity to variations in selected parameters such as
initial and boundary concentrations, initial cell depth, emissions,
wind speed and solar radiation.
22
-------
RISING MIXED
HEIGHT
1 ENTRAPMENT OF
POLLUTANTS ALOFT
Figure 2. Schematic drawing of the PBM domain.
23
-------
10
SCALE, Km
Figure 3. Map of St. Louis area depicting extent of PBM modeling
domain (boxed area), city of St. Louis (shaded area),
RAPS monitoring stations (circles), and radiosonde
ascent site (triangle).
24
-------
PBM SIMULATION-750522
a.
Q_
o
o
I I I 1 I I 1 f I I I 1 I I I I I I I I I I I I I I I
7.5
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure 4a. PBM simulation results for Day 142 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
25
-------
PBM SIMULAT10N-750522
I J I I 1 I. J__ L I . I I I I I I I I t I I I I I I I
7.5 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure 4b. PBM simulation results for N02 - Day 142 of 1975.
Keys to figure described in 4a.
PBM SIMULATION-750522
• -*^'
^*5r9—**^ I i I 1 1 J_ 1 L t i | | ' t i | I i j 111 II
7 A 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
Figure 4c. PBM simulation results for 03 - Day 142 of
Keys to figure described in 4a.
26
20.0
1975.
-------
Q.
Q.
O
o
PBM SIMULATION-750627
i I i i i i ^—r—r"i i l'*7~n~T~
5.0
Figure 5a.
7.S 10.0 12.3 15.0
TIME, HOURS (CST)
17.5
20.0
PBM simulation results for Day 178 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
27
-------
PBM SIMULAT10N-750627
Q.
Q.
CM
O
I 1 l__..i_.. I ..!_ J i 1 Ll J 1.1 II I I if I I I J_ I 1 ._!_.!. J
0.00
M>
10.0 12.5 15.0 17.5 20.0
TIME, HOURS (CST)
Figure 5b. PBM simulation results for N02 - Day 178 of 1975.
Keys to figure described in 5a.
PBM SIMUIAT10N-750627
10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure 5c. PBM simulation results for 03 - Day 178 of 1975.
Keys to figure described in 5a.
28
-------
0.
CL
o
o
,»--\
PBM SIMUL4TION-750726
i i i i i i i i t i i i i i i i i i i i i i i i
7J 10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure 6a. PBM simulation results for Day 207 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
29
-------
PBM SIMULATION-750726
I i i i i j i T v"i i T i i i i r
i i i i * i i i i r i r**T~r*i t r~~7~t*~i~TT*'i i • i < i
0.00
7 A 10.0 12.5 15.0 17.5
TIME. HOURS (CST)
20.0
Figure 6b. PBM simulation results for N02 - Day 207 of 1975.
Keys to figure described in 6a.
PBM SIMULATION-750726
Q.
Q_
IlllfJlllllilli
0.00
10.0 12.5 -15.0 17.5
TIME, HOURS (CST)
20.0
Figure 6c. PBM simulation results for 03 - Day 207 of 1975.
Keys to figure described in 6a.
30
-------
Q.
CL
o
o
PBM SIMULATION-760730
A
-•Kr i iN—I—r-f-TT-l—
54
7.5 10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure 7a.
PBM simulation results for Day 212 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model -simulated (solid
line) hour-average concentrations of CO.
31
-------
.100
.075
PBM SIMULATION-760730
o.
CL
.090
O
.025
9.000
! I 1 I I I I I f * III j I I I I I I I I I
o \
^
o o
i i i i I i i i i Nf*~i-T~rT"t-r"r"r-r"hT—r-i—i
5.0 7.5 10.0 12.5 15.0 17.5 20.0
TIME, HOURS (CST)
Figure 7b. PBM simulation results for N02 - Day 212 of 1976.
Keys to figure described in 7a.
PBM SIMULATION-760730
QL
Q.
i i l i I i i ( i i i t r i 1 i i
0.00
7A 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure 7c. PBM simulation results for 03 - Day 212 of 1976.
Keys to figure described in 7a.
32
-------
PBM SIMULAT10N-761001
Q.
a.
o
o
1W.W
7.5
5.0
Z5
1 I I 1 I 1 1 1 1 I 1 1 1 1 1 1 1 1 I I 1 I 1 1 I
' ' , /
' v i
• / * ,'
I V '
• / \ '
i \ i
"/
'/
••'
i
* \ :
_ \ !
\ '
\ (
• \ t
o \ <
x 1
-oo
^. V ' O
s~~~^ o \ .'
-/\^ • . •-
^x^ ^
• s ,-T::==:^ °- ' w y u
Y ^-' ^^ .
i i i i 1 i i i i ! t i i i I i t i i 7~ i ~i i i I
i i i i
-
•
•
—
•
-
-
-
—
_
-
~
-
till
5.0 7.5 iaO 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure 8a. PBM simulation results for Day 275 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
33
-------
PBM SIMULAT10N-761001
0.
o.
1(1(11111111
0.00
7A 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure 8b. PBM simulation results for M02 - Day 275 of 1976.
Keys to figure described in 8a.
PBM SIMULAT10N-761001
Ililllllilflllll
7A 10.0 12.5 1.5.0 17.5
TIME, HOURS (CST)
20.0
Figure 8c. PBM simulation results for 03 - Day 275 of 1976.
Keys to. figure described in 8a.
34
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES
250-1-001
<
Q
CO
LJ
cr
200+001 •
150+001 U»
.100-1-001 ***
.5001-000
000
-.500-i-OOO
- 100+001
-.150 + 001
-.200+001
- 250-00'
••••***
10 15
40 45 50
20 25 30 35
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 260 DATA POSSIBLE = 260
Figure 9a. Residual histogram from the PBM simulation results
for all 20 test days for CO.
35
-------
RESIDUAL HISTOGRAM DETERMINED OVER AU_ T:MES
.750-001
2
Q.
Q.
g
on
U-l
cx
.500-001 (•
450-001
.300-001
.150-001
.000
- 150-001
- 300-001 •
-.450-001 -
-.600-001
-.750-00'
10 20 30 40 50 60 70 80 90 '00
FREQUENCY
NITROGEN DIOXIDE
DATA AVAILABLE = 260 DATA POSSIBLE = 260
Figure 9b. Residual histogram from the PBM simulation results
for all 20 test days for N02-
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES
Q.
0^
_l
r>
Q
on
LU
^^u^uuu
.200-t-OOO
.150+000
.100+000
]
j
1
500-001 ~
000
8BSSSSSS9S3SS8ES*
- 500-001 ~— .
- 100+000?
if
-.150+000
-.200+000
- 250+000
•
30 40 50 60
FREQUENCY
OZONE
DATA AVAILABLE = 260
30 90 '00
DATA POSSIBLE = 260
Figure 9c. Residual histogram from the PBM simulation results
for all 20 test days for 03.
36
-------
RESIDUAL VS OBSERVED CONCENTRATION
2
Q.
Q.
*t
RESIDU>
.200+001
150+001
.100+001
.500+000
.000
-.500+000
-.150+001
-.200+001
- 250+001
000
a a
0 an °cP 3 °
MWM J3D *3ri 0 Q O
jj^jBS^am1 . °°
^ BJ P .3} ^o i01 a a !|li
o
QQ O
a
1 ' 1 ) t 1 1 1 ! ^_
LCCENO
FREO SYM
< i, i> a
< 2. 2> 0
< 3. 3> A
< 4,99> «
100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 260 DATA POSSIBLE = 260
Figure 10a. Residual vs. observed plot from the PBM simulation
results for all 20 test days for CO.
37
-------
RESIDUAL VS OBSERVED CONCENTRATION
s
a.
Q
cn
UJ
or
./3U— UU 1
SOC-001
.450-001
.300-001
.150-001
000
- 150-001
-.300-001
-.450-001
-.600-001
-.750-001
000
-i
O
a
Batz»
jtf^yf ,,,,,'
PT ffl O
•
•
.500-001 .100+000 150+000 200+000 250-t
LEC£NQ
FREQ STM
< i. i> a
< 2. 2> 0
< 3-
Figure lOb.
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 260 DATA POSSIBLE « 260
Residual vs. observed plot from the PBM simulation
results for all 20 test days for N02-
RESIDUAL VS OBSERVED CONCENTRATION
a
a
~
c/5
LJ
a:
x^u^uwnj
200+000
150+000
.100+000
.500-001
°
-.500-001
-.100+000
- 150+000
- 200+000
- 250+000
•
.
-
0
^MI Hi ii i n ii i'' I i* "^ !»! "1 f"T*f'l m
i&wrrtmrnvr'F® ^"rjtr^lgr^i tLlSiLj jfl ! ' ' 1
tvi ^jn&- JjiP' ft r^T?Tn.''ii<'!rn^^3^^^Q MJLjsi* c* i
*^L 2_ ^ z-rr^.ifr _ _^^ n jr
- LJ PLPl /TO- ! BIIHI *•" "
UJ el CJ ' *3Ufvn n
Jr 'Jffi*f n
m ™
a OD 3
i r I ,
LtCCNO
FREO SYM
< 1. !> 3
< 2, 2> 3
< J. 3> A
< *,99> «
QQQ 500-001 .100+000 ,150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 260 DATA POSSIBLE - 260
Figure lOc. Residual vs. observed plot from the PBM simulation
results for all 20 test days for 03.
38
-------
SECTION 5
LAGRANGIAN PHOTOCHEMICAL MODEL
OVERVIEW
The Lagrangian Photochemical Model (LPM) was developed by Environmental
Research and Technology, Inc. (ERT) and adapted under contract with EPA for use
with the RAPS data base. (The LPM is essentially identical .to the general-
use model named ELSTAR). The LPM envisions a portion of the atmosphere as
an identifiable parcel which can be tracked from early morning to late after-
noon. As the parcel moves over the various emissions sources, pollutants are
assimilated, vertically mixed, and subjected to photochemical reactions in the
presence of solar radiation. The LPM is attractive relative to grid models in
that it is fairly simple to execute and uses a moderate amount of computer
time. On the other hand, the LPM calculates concentrations only within a
parcel and not over a complete spatial field.
Previously, LPM simulations were conducted for 10 days from RAPS and
documented in Report I. Findings from that study pointed'to the need for
lateral diffusion to be included for the parcel and indicated the importance of
initial conditions to the final ozone predictions. Based on those findings,
the model has been modified to include a Gaussian-type lateral spread of the
parcel. Furthermore, the manner in which the initial condition profiles of NO,
N02 and HC are loaded has been modified from the original recommended method.
A series of simulations over 20 days (including the original 10) has been
executed, and results are summarized in this section of the report. Meteoro-
logical characteristics of the test days are listed i.n Table 1.
METHOD
The LPM is executed using a series of program modules. They are METMOD,
EMMOD and KEMOD, sequentially performing calculations on meteorology and air
quality, emissions and photochemistry. The input and running procedures
39
-------
described by Lurmann et al. (1979) have been generally followed, with some
modifications as discussed in Report I. In addition, running procedures for
the expanding Lagrangian parcel were established and the method of setting
initial concentration profiles was altered. As done previously, the parcel
start position was set so it would arrive at the station observing the maximum
1-h ozone value for the day.
Following the initial 10 simulations, attention was focused on the
results for Day 251 where the ozone prediction was a factor of 2 higher than
the maximum observed concentration. This single case was important in the
summary results since the large residual (observed minus predicted) accounted
for a substantial portion of the standard deviation of the residuals over the
10 days. Based on early morning surface concentrations, the parcel on Day 251
was initially loaded with very high reactive hydrocarbon (RHC) concentrations,
and parametric analysis indicated that the initial loading was totally respon-
sible for the prediction exceeding the observation. Accepting the initial
condition formulation, the concept of a fixed parcel size within the LPM seemed
a likely contributor to the over-prediction. It seemed more reasonable that,
as a parcel leaves the central, high-emissions areas, it should expand so as to
simulate lateral diffusion. The net effect would be to decrease concentrations
of precursors and, ultimately, ozone. Therefore, under contract to EPA,
Lurmann (1980, 1981) implemented an expanding parcel scheme in the LPM. The
rate of expansion was based on Gaussian plume ay values. The meteorological
preprocessor, METHOD, now produces a parcel-width schedule in addition to other
outputs. The METMOD input allows the user to specify a proportionality con-
stant for the a y expansion and to choose a time when the expansion begins.
To maintain an objective evaluation the proportionality constant was taken as
1, and the parcel started expanding at 0800 CST for all simulations. The
expansion rate was controlled by an internally computed stability class, but
generally parcel widths of 7-10 km were achieved by midafternoon (from an
original 5 km). The effect on pollutant concentrations was noticeable but not
entirely satisfactory. For example, the Day 251 maximum of 0.53 ppm from
Report I (Table 4) dropped to 0.37 ppm, and the s.d. ( AC) over the 10 days
dropped from 0.11 ppm to 0.068 ppm. However, in the second set of 10 days, Day
237 of 1976 also had high initial concentrations with resulting over-predic-
40
-------
tion, so that the s.d. (AC) over 20 days was an unacceptably high 0.090 ppm.
It was clear that the expanding parcel alone would not solve the problem of
substantial overprediction in certain cases.
These results prompted a re-examination of the method used for initali zing
the pollutant profiles for HC and MOX (only surface measurements were avail-
able). As originally implemented by ERT, the assumed initial profile showed
concentration decreasing to 1/2 the surface observation at the morning mixing
height - typically 200 m - and remaining constant at that value up to the
parcel top (the afternoon mixing height, typically 1500 m). The flaw in
the original scheme is that the assumed total mass above the morning mixing
height may be very large if the surface concentration is large. High HC and
NOX initial concentrations are usually the result of accumulation of near-
surface emissions during very calm nights. Therefore, it is reasonable that
the true vertical extent of the high loading is only up to the morning mixed
layer and that more-or-less background levels should exist above. Such a
scenario was implemented in the LPM for the final 20 simulations, with the
surface, observed concentrations of MO, N0£ and RHC assumed constant to the
lower boundary of Level-3 in the parcel (up to 150-200 m) and assumed to be at
the lower default through Levels-4 and 5. The initial concentrations are given
in Table 3, along with the defaults assigned in the upper parcel. The method
of estimating the individual hydrocarbon class concentrations in ppm based on
the reactive hydrocarbon measurement (RHC) in ppmC was revised from the orig-
inal proposal by ERT (Lurmann et al., 1979). The new method, also used
in the first 10 simulations, is:
[Alkanes] = 0.0786 [RHC]
[Alkenes] = 0.0328 [RHC]
[Aromatics] = 0.0248 [RHC]
[Formaldehydes] = 0.0405 '[RHC]
[Other Aldehydes] = 0.0405 [RHC]
The initial profiles of 03 were derived by the same method as described
in Report I. Initial profiles for CO were derived in a manner similar to that
for RHC and NOX described above. However, the background value for Levels-4
41
-------
TABLE 3. INITIAL SURFACE CONCENTRATIONS ASSIGNED TO THE PARCEL FOR REACTIVE
HYDROCARBON (RHC), MO, CO, 03, AMD INITIAL CONCENTRATIONS ASSIGNED
FOR 03. ALOFT.
Day RHC NO CO 03 03 Aloft
(ppm) (ppm) (ppm) (ppm) (ppm)
- --
142
178
182
183
184
207
209
221
230
231
251
159
160
195
211
212
225
226
237
- 275
0.26
0.19
0.52
0.15
0.39
0.11
0.32
0.14
0.12
0.15
1.28
0.22
0.05a
0.05a
0.07
0.10
0.05a
0.41
1.57
0.70
0.03
0.01
0.04
0.01
0.03
0.01
0.02
0.01
0.02
0.01
0.08
0.01
0.01
0.00a
0.01
0.01
0.01
0.01
0.06
0.10
0.44
0.41
0.73
0.18
0.77
0.27
0.73
0.69
0.50
0.48
2.60
0.86
0.32
0.17
' 0.13
0.11
0.32
0.97
2.3
2.82
0.00a
0.00
0.01
0.01
0.01
0.01
0.01
0.02
0.00a
0.01
o.ooa
0.01
0.02
0.03
0.01
0.01
0.04
0.01
0.01
0.01
0.06
0.08
0.12
0.12
0.12
0.07
0.08
0.08
0.06
0.06
0.06
0.11
0.11
0.08
0.05
0.11
0.07
0.09
0.11
0.06
aDefau1ts: RHC=0.05, N0=0.0025, 03=0.00'25
42
-------
and 5 was inadvertently set to a rather high 0.5 ppm. The general effect of
this oversight would be to add several tenths of a ppm to the afternoon CO
predictions. This should be kept in mind when viewing simulation results, but,
since CO has negligible influence on 03 production, there was little reason
to repeat the simulations to adjust for this error.
Expansion of the parcel implies entrainment of surrounding air during a
simulation. LPM therefore requires an estimate o-f the concentrations for side
boundary conditions. For RHC, subject to the split given above, the assumed
value was 0.063 ppmC. NO and NOg were assumed at continental background
values. Ozone was set to the observed midday inflow from upwind stations on
that day by a procedure described by Shreffler and Evans (1982). This is the
same 03 value given to the upper levels of the parcel in the initial profile
and generally varies from 0.06 to 0.12 ppm.
RESULTS
As was done in Report I, the time series results for Days 231, 251, 226
and 275 are presented in this section. In this way the effect of the new
initial conditions and parcel expansion can be judged. From the additional
set of 10 selected days, Day 237 (1976) time series results are also presented,
since high initial loadings of NOX and HC were observed on that day. Summary
statistics are presented on the residual (observed minus predicted) concentra-
tions at the time and position of the observed ozone maxima for all 20 days.
Time series for the remaining 15 days and temporal correlation coefficients are
presented in Appendix B.
Figure lla presents a trajectory map for Day 231. This map covers a
square 100 km on a side, and only 9 of the stations are shown to avoid clut-
tering the figure. These stations are (compare Figure 1) 101 in downtown,
122 and 114 to the north, 109 and 123 to the east, 118 and 124 to the south
and 120 and 125 to the west. The location-of the parcel each hour along the
trajectory is indicated. The start position is "indicated by "S" and the start
time is given above the figure. On Day 231, the parcel starts in a rural
locale south of the city. It moves over the high emissions areas at 0800-1100
CST and continues to the northwest. The 03 maximum was recorded at T400 CST
by station 121.
43
-------
Figure lib gives the 63 concentrations for Day 231. The RAPS observed
values are given by a solid line. The model-predicted values at Level-1 (L-l,
understood to be the surface) are given by a dotted line, and the Level-3 (L-3,
about 300 m in most cases) predictions are given by a dashed line. (The reader
is reminded that the time series represent the concentrations within a moving
parcel and not at a single point in space).
The results for 03 do not differ greatly from those for the original
test simulation. There is a slight underprediction up to the time of the
observed maximum. Beyond that time the prediction greatly exceeds the ob-
servation, but, since the parcel is beyond the measurement network, this
comparison has little meaning. The predicted NO and N02 peaks are diminished
somewhat from the original results, reflecting the lower initial loading and
parcel expansion. The CO predictions indicate the influx of morning emissions
with the L-l values showing greater peaks. In the afternoon, the model parcel
becomes well-mixed, and the decreasing tail reflects the expansion of the
parcel (compare with the original test).
The trajectory for Day 251 is given in Figure 12a. The start point is
near downtown St. Louis, and initial loading of the parcel with HC and NO is
high. Furthermore, winds are very low during the first 4 hours so that the
parcel accumulates large quantities of pollutants from emissions. After 1000
CST the parcel moves more rapidly and arrives at station 122 at 1400 CST. The
03 time series in Figure 12b shows a substantial response to the new initial
condition scheme. Ozone shows an initial surge in the morning, but, as the
mixed layer rises, the cleaner air aloft is mixed into the system. The pre-
dicted peak at 1400 CST is only slightly above the observation. As seen in the
original simulation of this day, the L-3 NO, N02 and CO predictions do a
better job of tracking the observations. The effect of parcel expansion is
evident in the tail of the CO predictions.
As seen in Figure 13a, the trajectory for Day 226 moves from the west side
of the city to the east, reaching station 109 at 1300 CST. It then turns and
moves back into the city towards evening. The 03 predictions for this simula-
tion must be deemed excellent. They have decreased from those in Report I.
44
-------
The L-3 predictions for NO, MO2 and CO are also quite good. As discussed in
Report I, the L-l (surface) predictions of primary pollutants seem to respond
in an unrealistic fashion when a new source is abruptly encountered. Use of
the L-3 predictions is recommended since they are essentially buffered from the
surface sources.
As shown in Figure 14a, the parcel on Day 237 started about 10 km north-
west of downtown St. Louis and moved to the west-by late afternoon. This day
was not treated in Report I, but is highlighted here because it shows the
highest RHC and third highest MO initial values over all 20 days. The resul-
ting 03 predictions show a rapid initial build-up with a dip at time of
inversion dissipation. The 03 predictions are reminiscent of those for Day
251 in this respect. The ozone prediction greatly exceeds the maximum ob-
served. This case again emphasizes the importance of initial conditions. To
reiterate, the conditions assume that the average RHC recorded by nearby
stations exists through the lower 200 m and over an area 5x5 km (the parcel
size). If, for instance, the vertical extent of the RHC accumulation were only
to 100 m, the final ozone predictions likely would be much lower. The kind of
detailed definition of the initial state needed for precise calculation is
unattainable with RAPS data.
The trajectory in Figure 15a indicates that Day 275 was a very calm day.
The parcel starts out southwest of the city center and reaches station 102 at
1400 CST. It then moves south and nearly returns to its origin. Two points
should be made concerning such calm conditions. First, the trajectory is
subject to great uncertainty, and all that is really known is that the air mass
was nearly stationary. Second, the LPM point-source module is not designed to
function well when the wind speed is so low that the emissions grid (5x5 k.m) is
not traversed in a single hour. The precise impact of this problem is unknown
but may be minimal. Ozone predictions are somewhat lower than those in Report
I and in better accord with observations. MO, M02 and C02 predictions are
quite similar to those in Report I.
45
-------
Table 4 presents a summary of the predicted and observed 03 maxima and a
statistical evaluation of the results for the 20 days of the study. The
prediction refers to the model prediction at the time and position of the
observed 63 maximum. The residuals, AC, are calculated for both the L-l and
L-3 predictions. The average residual for L-l indicates slight underpredic-
tion, while the average residual for L-3 is essentially zero. There would be
an inclination to choose the L-3 prediction because of the aforementioned
tendency to track the observations more consistently. For both levels, the
s.d. (AC)'S are identical, and the average absolute deviations are nearly
so.
CONCLUSIONS
Compared to results in Report I, the LPM has shown considerable improve-
ment in its predictive capabilities as indicated by reduced s.d. (AC). This
improvement resulted from implementation of an expanding parcel and changes in
the method of loading initial conditions. Both adjustments reduced the gross
overpredictions which occurred when initial surface concentrations were high.
AC has remained encouragingly small throughout the preliminary and final simu-
lations. The variability on a daily basis may arise from input uncertainties
(involving winds or initial conditions) rather than underlying model weak-
nesses.
The LPM has shown promise as an effective tool to understand 03 produc-
tion in an urban region. The model is relatively easy to use, inexpensive to
execute, and seems immune to various execution errors which tend to arise
unexpectedly in complex computations of this sort.
46
-------
TABLE 4. A SUMMARY OF PREDICTED AND OBSERVED QS MAXIMA
FROM THE LPM FOR 20 DAYS SELECTED FROM 1975-1976.
Day
Hour
Station
Observed
4-meters
(ppm)
LPM Predicted
Level -1 Level -3
(ppm) (ppm)
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
12
14
12
15
13
14
11
15
13
14
14
15
16
15
15
12
13
13
11
14
101
112
125
124
118
113
118
121
121
121
122
122
115
114
120
108
117
109
120
102
0.20
0.20
0.16
0.21
0.18
0.18
0.21
0.17
0.19
0.23
0.26
0.20
0.22
0.22
- 0.16
0.17
0.17
0.23
0.18
0.24
0.07
0.15
0.25
0.18
0.21
0.11
0.14
0.16
0.18
0.19
0.29
0.18
0.21
0.15
-0.11
0.10
0.08
0.20
0.30
0.26
0.13
0.31
0.24
0.18
0.22
0.13
0.15
0.16
0.17
0.19
0.30
0.18
0.23
0.15
0.12
0.11
0.09 '
0.22
0.29
0.30
For the data displayed above:
AC = Obs minus L-l Pred
1C = 0.023
s.d.UC) = 0.058
TATT = 0.052
AC = Obs minus L-3 Pred
1C" = 0.005
s.d.UC) = 0.058
TACT - 0.050
47
-------
75231. RAMS 121 AT 1400CST. START 0600
iW) f^"^™T i i j ^ i i i i i TC^^"*^^^^^""™^^™"^^
79
SO
or o
s •
II i i i i i i I i i ._!_.. i J
50
KM
79 100
Figure lla. LPM parcel trajectory for Day 231 of 1975.
48
-------
75231, RAMS 121 AT 1400CST. START 0600
CL
CL
uT .2
o
fSJ
o
.1
00 >->•
I I I I I I I I I I 1 I I I I
08S
PRED L-1
PRED L-3
•rfS-f-r . I .... I .... I .... I ....
54 7A iaO 12J 15.0 17.5 2aO
HOUR, CST
Figure lib. LPM time series of observed and predicted 03 hourly
concentrations on Day 231 of 1975.
.100
.075
75231, RAMS 121 AT 1400CST. START 0600
.050
O
z
.025
0.000
i I i i i i r
iiii
OBS
PRED L-1
PRED L-i
7 A
10.0 12.5 15.0
HOUR, CST
2ao
Figure lie. LPM time series of observed and predicted NO hourly
concentrations on Day 231 of 1975.
49
-------
es
o
.100
.075
J550
.025
75231, RAMS 121 AT 1400CST. START 0600
OBS
PRED L-1
PRED L-i
0,000* * i i ' i i « ' ' t t ' i t i i i t t i i t i i"n i i t
56 7.5 iaO 12J 15.0 17.5 20.0
HOUR, CST
Figure lid. LPM time series of observed and predicted N02 hourly
concentrations on Day 231 of 1975.
75231, RAMS 121 AT 1400CST. START 0600
CL
0.
o"
u
I I I 1
I '
^ i i i [ i 11 j I i i i j_ . 1. -i
• i , r j • . •
""—- OBS
PRED L-V
PRED L-1
54 7.5 10.0 1X5 15.0
HOUR. CST
20.0
Figure lie. LPM time series of observed and predicted CO hourly
concentrations on Day 231 of 1975.
50
-------
75251. RAMS 122 AT 1400CST. START 0600
• Wr r""^-^-^^^^^-^^^M«
73
50
2S
25
90
KM
75 100
Figure 12a. LPM parcel trajectory for Day 251 of 1975,
51
-------
75251, RAMS 122 AT 1400CST. START 0600
I I I I I I I I I I t I I I
— OBS
— PRED L-1
. PRED L-3
i i till | j i I L I I I
10.0 12J 15.0
HOUR. CST
20.0
Figure 125. LPM time series of observed and predicted 03 hourly
concentrations on Day 251 of 1975.
75251, RAMS 122 AT 1400CST. START 0600
CL
Q.
100 12J
HOUR, CST
OBS
PRED L-V
PRED L-i
20.0
Figure 12c. LPM time series of observed and predicted MO hourly
concentrations on Day 251 of 1975.
52
-------
,75251. RA.V1S 122 AT 1400CST. START 0600
r.
03S
PRED L-1-
PRED L-3
7.5
10.0 12.5 15.0 17.5 20.0
HOUR, CST
Figure 12d. LPM time series of observed and predicted N02 hourly
concentrations on Day 251 of 1975.
75251, RAMS 122 AT 1400CST. START 0600
— oes
— PRED L-M
— PRED L-3
9U) 7& 10.0 12 J 15.0 17 J 20.0
HOUR, CST
Figure 12e. LPM time series of observed and predicted CO hourly
concentrations on Day 251 of 1975.
53
-------
76226. RAMS 109 AT 1300CST. START 0600
'wr^ i i f""1^^~"i"~ i T^™r""T^""|O"T™'T"~T^^^^^ri""lT"—
90
i ' -" ' I
80
KM
75 100
Figure 13a. LPM parcel trajectory for Day 226 of 1976.
54
-------
76226, RAMS 109 AT 13QOCST. START 0600
• • i
- oss
- PRED L-1
— PRED L-3
* t I t I 1 I t t I I I
10.0 12.3
HOUR, l
19.0
17.5
20.0
Figure 13b. LPM time series of observed and predicted 03 hourly
concentrations on Day 226 of 1976.
.19
76226, RAMS 109 AT 1300CST. START 0600
.10
Q.
0_
.05
0.00
f
i | i i I^T r i i r T i i r
"~~~ 08S
----- PRED L-V
-- PRED L-i
10.0 12.5 1S.O
HOUR, CST
17.5
2O.O
Figure 13c. LPM time series of observed and predicted NO hourly
concentrations on Day 226 of 1976.
55
-------
.18
76226, RAMS 109 AT 1300CST. START 0600
.10
0.
Q.
CN~
O
.05
0.00
I I I I I
OSS
PRED L-
; \PRED L-i
I I | .._|__. L I I I I t I I I I ..._!.... 1
54
iao iis 15.0
HOUR, CST
17.5
20.0
Figure 13d. LPM time series of observed and predicted N02 hourly
concentrations on Day 226 of 1976.
76226, RAMS 109 AT 1300CST. START 0600
o
o
i I I \
I ' ' ' '
OSS
PRED L-1.
PRED L-i
7J 10.0 1Z5 15.0-
HOUR, CST
17.5
20.0
Figure 13e.
LPM time series of observed and predicted CO hourly
concentrations on Day 226 of 1976.
56
-------
76237. RAMS 120 AT 1100CST. START 0600
i W* P"^^^^"P""T^"HT"1 i~ T J1^^""^1 n* i ^^^^i|"'^p—]*~T~
SO
25
o o
25
50
KM
75 100
Figure Ha. LPM parcel trajectory for Day 237 of 1976.
57
-------
76237, RAMS 120 AT 11OOCST. START 0600
o.
Q.
ul
O
NJ
O .2
—— 08S
PRED L-1
PRED L-3
00
54
7.5 10.0 12J 15.0
HOUR, CST
17.5
20.0
Figure 14b. LPM time series of observed and predicted 03 hourly
concentrations on Day 237 of 1976.
76237, RAMS 120 AT 11 OOCST. START 0600
O
i i i i i i
OSS
PRED L-1
PRED L-i
0.00
7.5
10.0 1Z5 15.0
HOUR, CST
17.5
20.0
Figure 14c. LPM time series of observed and predicted NO hourly
concentrations on Day 237 of 1976.
58
-------
76237, RAMS 120 AT 1100CST. START 0600
I I I I I I I
— OBS
•— PRED L-1
— PRED L-i
7A
10.0 1ZS 15.0
HOUR, CST
17.3
20.0
Figure 14d. LPM time series of observed and predicted N02 hourly
concentrations on Day 237 of 1976.
76237, RAMS 120 AT 1100CST. START 0600
Cu
a.
o"
o
I ' ' ' '
OSS
PRED L-1
PRED L-S
OL-L-L,
SJO
. i .... I .... r?-. .. I .... i. ...
7.3 10.0 1Z3 15.0 17.5 2aO
HOUR, CST
Figure 14e. LPM time series of observed and predicted CO hourly
concentrations on Day 237 of 1976.
59
-------
76275. RAMS 102 AT 1400CST. START 0700
1W* p**^^T~"'~ J * I i TT^T""Hr"""^^f^HT" I V ' ~ "(•••^^^^"•p*
SO
I
30
KM
75 100
Figure 15a. LPM parcel trajectory for Day 275 of 1976.
60
-------
.4
76275, RAMS 102 AT 1400CST. START 0700
a.
a.
ul .2
o
.1
0.0
OBS
PRED L-1
PRED L-3
/-A
&0 7A 10.0 12J 15.0
HOUR, CST
17.S
20.0
Figure 15b. LPM time series of observed and predicted 03 hourly
concentrations on Day 275 of 1976.
76275, RAMS 102 AT 1400CST. START 0700
Q_
O.
.1
oo
I I I I I I
I ' ' '
OBS
PREO L-1-
PRED L-4
SU)
10.0 12.5 15.0
HOUR, CST
17.5
2ao
Figure 15c. LPM time series of observed and predicted NO hourly
concentrations on Day 275 of 1976.
61
-------
76275, RAMS 102 AT 1400CST. START 0700
Q.
Q.
*
CM
O
0.0
iiiiiiri i
—— 08S
PRED L-1
PRED L-3
7.5 10.0 1ZS 19.0
HOUR, CST
17.5
20.0
Figure 15d. LPM time series of observed and predicted N02 hourly
concentrations on Day 275 of 1976.
76275, RAMS 102 AT 1400CST. START 0700
10
Q.
Q.
O*
(J
08S
PRED L-1-
PRED L-3J
0
flU)
i i i I i i i i I i i i i I i i i- t I i i i i I i i
10.0 1ZS 15.0.
HOUR, CST
17.5
20.0
Figure 15e. LPM time series of observed and predicted CO hourly
concentrations on Day 275 of 1976.
62
-------
SECTION 6
URBAN AIRSHED MODEL
INTRODUCTION
The Urban Airshed Model (UAM) is a three-dimensional (3-D), grid-type
photochemical air quality simulation model (PAQSM) developed by Systems
Applications, Inc. (SAI) of San Rafael, California. The model structure
consists of a latticework array of cells, whose total volume represents an
urban-scale domain and in which the physical and chemical processes responsible
for photochemical smog ar.e mathematically simulated. These include the advec-
tion of pollutant species through the modeling domain, the species' entrainment
from aloft by a growing mixed layer, diffusion of material from cell to cell,
the injection of primary source emissions into- the modeled volume, and the
chemical transformations of reactive species into intermediate and secondary
products. The horizontal dimensions of each cell are constant but cell heights
vary throughout a model simulation as mixed layer depth in the UAM" changes
accordingly.
The area modeled in the St. Louis application was 60 km wide and 80 km
long. Individual cells were 4 km on a horizontal side. Vertically, cells
totaled 4 layers; the bottom 2 layers simulated the mixed layer and the
top 2 represented the region immediately above the mixed layer. The domain
of the UAM was centered just west of downtown St. Louis and included the entire
metropolitan area. Figure 16 shows a horizontal cross-section of the domain as
applied to the RAPS monitoring network. All 25 monitoring station locations
are given in relation to the grid pattern. The outer ring contains the bound-
ary cells for the model. Although model simulation does not occur in the
boundary cells, they are an important part .of the advective transport of the
UAM. Four monitoring stations, 122-125, lie outside of the computational
domain and are available for determining boundary concentrations. Generally
boundary concentrations for all species except 03 were small, near hemi-
63
-------
spheric background levels. Ozone levels at the boundary however were sometimes
significant compared to those within the model domain. The remaining sites lie
within the domain, no 2 residing in the same cell. The 3-D grid model is a
sophisticated type of PAQSM and provides both spatially and temporally resolved
concentration predictions. Thus, the UAM attempts to estimate the hour-average
observed concentration of a pollutant species at each monitoring site within
the model domain.
Details of the method for performing a St. Louis simulation with the UAM
are described in Report I. There are 12 data preprocessors in the UAM package
that must be exercised before the simulation model itself. They access the
RAPS data base and retrieve the appropriate parameters required by the UAM.
Model simulations begin at 0500 CST and continue through 1700 CST.
The UAM and its preprocessors remain structurally unchanged from the model
simulations described in Report I, except for one aspect. An independent test
of the numerical advection algorithm was described there and showed that the
algorithm in the model at that time produced excess numerical diffusion, a
highly undesirable trait for a grid model. A study of the problem revealed
that the numerical advection routine in the UAM should be replaced with a
sounder method that has the qualities of (1) minimizing the amount of numerical
diffusion produced and (2) advecting areas with large concentration gradients
such that the gradients are not destroyed. Several advection algorithms were
identified and tested as potential candidates for replacing the existing
method. The one chosen was a fourth-order version of the multi-dimensional
flux-corrected transport algorithms described by Zalesak (1979). This method
produced far less numerical diffusion and preserved peak concentration areas
better during advection than the algorithm that had resided in the model. An
added advantage of the Zalesak algorithm is that the horizontal advection in
the model could be performed in one step while the original routine needed a
step for each of the x and y directions, another potential source of error.
The penalty encountered by changing the numerical advection routine is about a
12% increase in the execution time of the simulation model (on EPA's UNIVAC
1180 computer). The potential for increased accuracy within the model with the
new algorithm far outweighs the small increase in overall computational costs.
64
-------
Model simulations with the revised UAM were conducted for each of the 20
test days listed in Table 1. These test days include the original 10 presented
in Report I but are now reflecting the model update described above. Although
results for 03 predictions on all 20 test days are summarized, detailed model
results are described for only 4 of these days. Simulation results for the
pollutants CO, N02 and 03 on Julian days 142 and 207 of 1975, and 195 and
275 of 1976 are emphasized.
RESULTS
Among the 20 days from the RAPS data base selected for model simulation
with the UAM were 4 where the afternoon measured 03 maximum occurred at
one of the monitoring sites located outside the model domain. All other days
showed 03 maxima occurring at sites within the domain. The UAM simulates air
quality within each grid cell. Species concentration predictions at particular
RAPS monitoring site locations are interpolated from predictions at the closest
cells by a distance-weighted (1/r) formula. Detailed simulation results are
presented here for 4 of the 20 days modeled, and then summary statistics are
given for 03 alone over all the days. »
A UAM simulation provides a wealth of information including the spatial
and temporal concentration variations of all modeled species throughout the
grid. It is not practical here to display the time series of predicted and
observed concentrations for all RAPS sites, even for a few pollutants on a
single day of simulation. A more concise method of displaying simulation
results is used here. It is based on an analysis of residual concentrations,
observed minus predicted, over the 21 RAPS sites within the model domain
(Bencala and Seinfeld, 1979). The particular analysis techniques employed' here
include residual histograms, denoted as "RS", which show the general distribu-
tion of residuals for a complete simulation, and the residual vs. observed
concentration plot, referred to as the "RO" plot, which demonstrates any trend
toward over- or underprediction by the model as a function of observed concen-
tration levels. For a given species on one simulation day, there were 21 sites
for each of 13 hours of simulation to account for, or 273 possible data points.
65
-------
Since this report emphasizes the various models' abilities to estimate
03 concentrations an additional technique is used for that species to supple-
ment the statistical analysis of the paired observed and predicted data. That
technique is the contour plot of observed and predicted 03 concentrations
over the model domain at the hour the maximum 03 level is observed. These
plots reveal the spatial orientation of the peak concentration areas. Visual
comparisons can be made that are not dependent on specific paired data. This
will help determine if the DAM is predicting the maximum concentration levels
well but at incorrect areas of the model domain. It would not be possible to
determine this situation from the residual concentration data alone.
The first simulation day discussed is Day 142 of 1975. The synoptic
meteorological conditions on this day were described in Section 4. Background
03 concentrations for Day 142 were 0.06 ppm.
Figure 17a shows the RS for CO on this day. The ideal RS distribution
would have a peak at 0.0 ppm and symmetric tails on either side sharply falling
off to 0.0 frequency. The figure displays a shape like this except for a bulge
in the peak toward overprediction. The RO plot in Figure 17b shows a trend
toward underprediction, especially evident in the 3 points corresponding to the
highest observed concentrations. Report I discussed the CO underprediction
problem for high observed concentrations. The clustering of points in this
plot at slightly negative values for very low observed concentrations helps
explain the bulge in the RS on the overpredictive side. Both figures are
required to provide a total picture of the distribution of the residuals.
The RS for N02 is presented in Figure 18a. While the distribution peaks
at the 0.0 residual it is slightly skewed toward overpredictions. In the RO
plot in Figure 18b the majority of residuals fall to the overpredictive side,
but mostly for smaller observed concentrations. At-higher concentrations there
tend to be more underpredictions. In any case a definite trend would be
difficult to discern here for most of the N02 residuals fall quite close to
0.0.
66
-------
Figure 19a displays the RS for 03 on Day 142. The distribution of
residuals here is clearly skewed toward overprediction, with a peak at the 0.0
residual. The RO plot in Figure 19b shows most of the overpredicted concentra-
tions lie in the range of 0.050 to 0.125 ppm observed 03. At higher observed
concentrations no clear trend appears.
•
The peak hour-average 03 level observed was 0.195 ppm at site 101 at
noon. The model-simulated concentration for the same time and location was
0.116 ppm, a 40.5% underprediction. Figure 20 presents the contours of 03
concentration for the model-predicted and observed cases for the hour beginning
at noon CST. In the model-predicted case the contours have been objectively
drawn based on the 300 regularly spaced grid values for the lowest layer of
cells in the UAM. For the observed case the contours are based on the 25
irregularly spaced monitoring locations, 4 of which lie outside of the model
domain. Here the objective algorithm responsible for generating the contours
works best toward the center of the domain where the density of sites, repre-
sented in the figure by small crosses, is greatest. The credibility of the
contours near the edges is not as high, especially in terms of the extent of
the area circumscribed by a given contour. In the figure for Day 142 the
maximum concentration areas appear near the center of the domain with magni-
tudes close to each other. In the model-predicted case the highest valued
contour is displaced slightly northwest of the observed maximum area and a
relative "valley" has developed in the immediate area of the observed maximum,
allowing the large underprediction to occur.
On Day 207-75 a large high-pressure center was located northeast of St.
Louis with light winds both at the surface and aloft. Measured wind speeds
were generally less than 1.0 m s'1 near the surface and directions, were
variable. Sunny skies prevailed through the duration of the modeling period,
although haze persisted during the day. Background 03 concentrations were
approximately 0.07 ppm.
Figure 21a shows the RS for CO on Day 207. The distribution is nearly
symmetric about the 0.0 residual with a few outliers on the underpredictive
side. The RO plot in Figure 21b reveals these outliers as the residuals
67
-------
corresponding to the highest observed concentrations. Most CO concentrations
on this day were less than 1 ppm. Nitrogen dioxide presents a broader histo-
gram than CO in Figure 22a that is symmetric and centered slightly on the
overpredictive siae. This bias is probably caused by the great many overpre-
dictions at observed N02 concentrations less than 0.02 ppm seen in Figure
22b. Aside from this phenomenon there is no clear trend apparent in the RO
plot. The RS for ozone in Figure 23a presents a histogram with the rough
outline of a normal distribution but contains,periodic gaps throughout. The
majority of the predictions are less than the observed concentrations and the
RO plot in Figure 23b demonstrates that a definite bias exists toward underpre-
dicting 03 as observed concentrations increase.
The maximum hour-average 03 measurement for Day 207 occurred at 1400 CST
at site 113, northwest 'of downtown St. Louis, at 0.185 ppm. The corresponding
simulated concentration for this location was 0.141 ppm, a 23.8% underpredic-
tion. The contours of observed 03 concentrations at this hour in Figure 24b
reveal a steep gradient aligned from the northwest to the southeast through the
city of St. Louis. The predicted concentration field in Figure 24a shows a
similar gradient but displaced slightly north and west. The alignment and
magnitude of the contours are, in fact, very much alike in the 2 cases.
Possible errors in the model's treatment of the wind field, or the generation
of the wind field itself, may have contributed to the displacement between the
peak areas of predicted and observed concentrations and allowed the apparent
underpredictions by the model at the monitoring sites.
The synoptic meteorological situation for Day 195 of 1976 contained an
upper air ridge over the center of the country with a warm front northeast of
St. Louis. Winds near the surface were from the southeast during the day
averaging just over 2.0 m s~l. Skies were mostly clear with occasional high
level cloudiness. Background 03 concentrations were nearly 0.08 ppm.
CO concentrations were low on this day, mostly less than 1 ppm. The RS
for CO for Day 195 is presented in Figure 25a. "It shows a narrow distribution
of residuals peaking slightly on the overpredictive side. In the RO plot in
Figure 25b this peak has been translated to the many very low observed concen-
68
-------
trations often near the detectable limits of the monitoring instruments. At
higher concentrations the model tends to underpredict, especially at the 2
distinct data points representing CO hot spots. The N02 residual distribu-
tion shown in Figure 26a displays a normal symmetric shape about the 0.0
residual. Most of the overpredictions occur at low observed N02 concentra-
tions and a trend toward more severe underprediction with increasing observed
N02 is apparent from the RO plot in Figure 26b. The RS for ozone in Figure
27a is skewed toward underpredictions. Below an- observed 03 level of 0.1 ppm
the RO plot in Figure 27b shows little bias in the residuals. Above this level
there is a clear trend toward underpredicting at higher observed concentra-
tions. This explains the skewness in the histogram as resulting from the
residuals of observations mostly greater than 0.1 ppm.
The highest hour-average 03 measurement on Day 195 was at site 114 at
1500 CST, north of the city of St. Louis, at 0.223 ppm. The UAM estimated an
03 concentration of 0.160 ppm at the same time and place, an underprediction
of 28.3*. The contours of 03 predictions and observations for this hour are
shown in Figure 28. Agreement in the alignment and location of the contours
between the 2 cases is good, although the magnitude of the maximum predicted
contour is considerably less than that of the observed. The advection problem
discussed in Report I for Day 195 appears to have been corrected here with the
substitution of the new advection algorithm.
Day 275 of 1976 is typical of an extreme stagnation episode in St. Louis.
The synoptic meteorological conditions on this day were described in Section 4.
Background concentrations of 03 were about 0.06 ppm.
The histogram of residuals for CO is shown in Figure 29a for Day 275. The
distribution peaks slightly on the overpredictive side of the 0.0 residual
although there is a pronounced skewness toward underpredictions, including
several extreme cases. The RO plot in Figure 29b displays the trend toward
underprediction of CO at higher observed concentrations. The extreme stability
of the atmosphere on the morning of Day 275 is reflected in the highest CO
concentrations, on the order of 10 ppm, over twice the level of the observed
peak on most of the other modeled days.
69
-------
The RS for M02 in Figure 30a is normally distributed about the 0.0
residual and the RO plot in Figure 30b shows little bias with increasing
concentrations for observed N02 levels of 0.125 ppm and less. Beyond this
point there is evidence of consistent underprediction. The 63 residual
analysis in Figures 31a and b indicates good model performance in the overall
predictions for the day. The residual histogram has a broad normal distribu-
tion about the 0.0 residual and the RO plot demonstrates no clear trends toward
over- or underprediction.
The peak 03 measurement on Day 275 was 0.244 ppm at site 102, near
downtown St. Louis, at 1400 CST. The model prediction for this time and
location was 0.220 ppm, a~ 9.8% underprediction. The contour plots of predicted
and observed 03 concentration for this hour are shown in Figures 32a and b.
In both cases the highest concentrations appear near the center of the model
domain and are of comparable magnitude. Several pockets of reduced 03 concen-
trations appear in the pattern of predicted contours that break up an otherwise
consistent pattern with the observations. These appear to be in the vicinity
of large NOX emissions sources that have an exaggerated scavenging effect on
nearby 03 predictions. The observations do not substantiate this effect.
The rather simplistic treatment of point source emissions by the UAM fs probab-
ly a major factor in this problem.
Table 5 summarizes the predicted and observed 03 maxima and an elemen-
tary statistical analysis of the UAM results for all 20 modeled days. The
"specific" model predictions correspond to the same location and time as the
observed maximum, and the "independent" predictions represent the maximum
hour-average 03 generated by the model in any cell of the lowest grid layer
at any time during the simulation. The independent maximum is the greatest
surface value the model can produce within its domain. The specific and
independent predictions, as seen in Table 5, typically do not coincide. This
indicates that the maximum 03 value produced by the model did not correspond
either in space or time, or both, to the maximum value observed in the atmo-
sphere. Statistics on the residual concentration, AC, as shown in Table 5,
were computed for both the specific and independent predictions. The average
residual and standard deviation are also presented for the analysis. Values
70
-------
TABLE 5. A SUMMARY OF PREDICTED AND OBSERVED 03 MAXIMA FOR THE UAM
Julian
date3
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
Hour
(CST)
12
14
11
10
13
14
11
15
13
14
12
14
16
15
15
12
13
13
11
14
Station
101
112
121b
119b
118
113
118
121
121
121
121b
114b
115
114
120
108
117
109
120
102
Observed
at 4-meters
(ppm)
0.195
0.202
0.142
0.171
0.184
0.185
0.209
0.166
0.193
0.233
0.179
0.172
0.221
0.223
0.155
0.170
0.166
0.225
0.176
0.244
UAM
Specific
(ppm)
0.116
0.156
0.083
0.154
0.132
0.141
0.128
0.118
0.095
0.133
0.146
0.142
0.121
0.160
0.143
0.128
- 0.050
0.077
0.119
0.220
Predicted
Independent
(ppm)
0.238
0.243
0.166
0.209
0.234
0.252
0.195
0.149
0.214
0.205
0.178
0.312
0.190
0.174
0.169
0.155
0.073
0.124
0.203
0.246
aTwenty days selected from 1975 and 1976.
bOveral1 maximum at 122,123,124, or 125; outside UAM domain.
For the data displayed above:
AC = Obs - Specific
IT = 0.062
s.d.UO = 0.035
"ITr - 0.062
Obs - Independent
-0.006
0.053
0.041
71
-------
for the average residual for the specific and independent cases, 0.062 and
-0.006 ppm respectively, differ noticeably. The independent predictions show
more promise toward maximum 03 simulation. The average absolute residual is
-------
TABLE 6. A SUMMARY OF LEAST SQUARES REGRESSION STATISTICS FOR
UAM PREDICTED AND OBSERVED 0a
Julian
dateb
•
142-75
178-75
182-75
183-75
184-75
207-75
209-75
221-75
230-75
231-75
251-75
159-76
160-76
195-76
211-76
212-76
225-76
226-76
237-76
275-76
URBAN
Correlation
coefficient
0.805
0.864
0.875
0.864
0.726
'0.927
0.759
0.795
0.819
0.814
0.821
0.859
0.937
0.947
0.833
0.912
0.753
0.796
0.687
0.844
AIRSHED MODEL:
Slope
0.887
0.867
0.610
0.614
0.470
0.654
0.570
0.516
0.729
0.645
0.618
0.862
0.676'
0.720
0.773
0.632
0.310
0.426
0.432
0.896
03
intercept
(ppm)
-0.0057
0.0183
0.0263
0.0284
0.0436
0.0373
0.0347
0.0433
0.0059
0.0176
0.0232
0.0021
0.0464
0.0296
0.0010
0.0244
0.0504
0.0574
0.0581
-0.0004
Sum of
squared error
( ppm2 )
. 0.186
0.186
0.142
0.193
0.317
0.173
0.192
0.268
0.095
0.152
0.150
0.158
0.247
0.112
0.104
0.189
0.247
0.445
0.383
0.340
Includes observations and predictions for all times and locations.
bTwenty days selected from 1975 and 1976.
73
-------
CONCLUSIONS
The potential use of a grid-type PAQSM, such as the UAM, clearly is great,
although model complexity often makes the solution of problems which arise more
difficult. Trouble-shooting and test application exercises have taken place
over the past 3 years at EPA after the adaptation of the UAM to St. Louis
by SAI. The constituent components of the model have already been investigated
and upgraded where necessary. The numerical ^advection routine was recently
replaced by one with less inherent pseudo-diffusion.
The model evaluation discussed here has pointed toward some of the
strengths and weaknesses of the UAM. In studying the results for the 3
pollutant species focused on in this report it is clear that while the model
typically produces a normal distribution of residuals in a simulation, those
estimates corresponding to the highest observed concentrations are often
underpredicted. This is especially true of primary species such as CO and
M02, but has also been indicated in the case of 03 as well. It is these
higher observed concentrations that are of greatest interest from a regulatory
perspective. For the case of those species emitted from and affected by
emissions from point sources, their simulated performance results may improve
with a more physically realistic treatment of point emission sources.
It is quite interesting to note that the average of independent residuals
for 03 maxima over the 20 test days indicates an almost exact match between
predictions and observations. Studying the contour maps on the test days
reveals that the model usually produces 03 levels as high or higher than
those observed and the contours of predicted and observed often have similar
shapes. However, the variance in these residuals of the 20 03 maxima is
quite large and this is caused, in part, by the different spatial orientations
of the predicted and observed contours. Indeed, •even the paired predictions
and observations typically show poor spatial correlation. This is particularly
true for 63. The spatial alignment of the concentration fields is principal-
ly governed through transport in the model. The' treatment of winds in the UAM
is comparable to most schemes available today for interpolating a wind field
across a grid (Schere, 1981). However, even in a relatively simple flow
74
-------
situation such as St. Louis, problems can arise. An error of only a few
degrees in wind direction can lead to gross errors in specific air quality
predictions. It will be a major responsibility of any UAM user to substantiate
that the wind processor is providing a credible simulation of wind flow in the
application area. For complex topographic situations or land-sea interactions
within a model domain, a wind field analysis unique to that application may be
required.
In summary, the UAM is a powerful tool available for use by the air
quality analyst already experienced in working with complex simulation models.
Model results should be carefully studied in each application and may not
necessarily be used in an absolute sense in any given simulation. An area of
future study that may be pursued is model sensitivity to variations in selected
parameters, such as initial and boundary conditions, emissions, wind fields
and solar radiation.
75
-------
122
121
114
125
121
l.i
112
102
07
06
111
10
105
08
05
04
10
109
10
117
125
I
-> x Jx = Jv = 4 km
118
12
Figure 16. Schematic drawing of horizontal section of UAM domain
showing locations of RAPS monitoring sites.
76
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL T'MES AND LOCATIONS
OAY 142-75
2
Q.
Q.
^~*
_l
CO
LJ
or
.3UVJ+UU 1
I
4CO+001 -
300+001 t
.200+001
.100+001
t
K
BH""^^
- 100+001
-200+001
-.300+001
-.400+001
- 500+001
Figure 17b.
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 1 74 DATA POSSIBLE = 273
Residual vs. observed plot for CO from the UAM
simulation results for Day 142-75.
77
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 1*2 -75
750-001 i 1 1 1 1 1 1 1 • 1
0.
0.
600-001
450-001
.300-001
.150-001 J45"
.000
|
Q
00
j£ -150-001
-.300-001
-450-001
-600-001
-.750-001
•*»*
•*»*»••*••
;««.
5 10 15 20 25 30 35
FREQUENCY
NITROGEN DIOXIDE
40 45 50
DATA AVAILABLE = 195 DATA POSSIBLE
273
Figure 18a. Residual histogram for N02 from the UAM simulation
results for Day 142-75.
2
Q.
a
3
9
LJ
or
.750-001
.600-001
.450-001
.300-001
.150-001
-.150-001
-300-001
- 450-001
-.600-001
-.750-001
RESIDUAL VS OBSERVED CONCENTRATION
DAY 142-75
< i. i> a
< 2. 2> 0
< 3. 3> *
100-001 .200-001 .300-001 400-001 500-001
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 195 DATA POSSIBLE = 273
Figure 18b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 142-75.
78
-------
RESIDUAL HIS'OGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 142-75
25OOOO
- 100 -i-OOO j*
- 150+OOOf-
-.200^0001-
-.250+000 I—
10 20 30
40 50 60 70
FREQUENCY
80 90 100
OZONE
DATA AVAILABLE = 21
DATA POSSIBLE = 273
Figure 19a. Residual histogram for 03 from the UAM simulation
results for Day 142-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 142-75
^v
0_
Q^
SIDUAL
LU
o;
200+000
150+000
.100+000
.500-001
.000
- 500-001
- 100+000
- 150+000
-.200+000
- 250+000
000
.
' . jfrA** . . f '
9j^ *&&^E?y9&&<$ rtftP ° * oo ' '
9 Q
LEGEND
FREO 5TM
< i, i> n
< 2. 2> 0
< 3. 3> *
< *,99> *
500-001 .100+000 .150+000 .200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 214 DATA POSSIBLE = 273
Figure 19b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 142-75.
79
-------
ST. LOUIS HOT 22.197S I DDT
^eoicTEQ co«et
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 207-75
Q.
a
~
SIDUAL
LU
or
~I_iU *-UU "
400"**00 I
300+001
.200+001
.100+001
000
-.-100+001
-.200+001
-.300+001
-.400+001
-.500+001
i ii j
'
I
5F'"" " i i i i i i
•
•
•
50 100 150 200 250 300 350 400 450 500
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 237 DATA POSSIBLE = 273
Figure 21a. Residual histogram for CO from the UAM simulation
results for Day 207-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 207-75
a
o_
*"
RESIDUAL
.400+001
300+001
.200+001
.100+001
.000
-.100+001
-.200+001
-.300+001
-.400+001
-.500+001
a
-l^ref3 ,,,,,'
pF^ .ii...
•
L£CENO
FREO STM
< i, i> a
< 2. 2> O
< 3. 3> *
.000 100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 237 DATA POSSIBLE = 273
Figure 21b. Residual vs. observed plot for CO from the UAM
simulation results for Day 207-75.
81
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 207-75
Q.
a
"— '
j
3
co
LLt
CC
3UU— UUI
400-001
3CO-001
.200-001
100-001
AfVt
uuu
- 100-001
- 200-001
-.300-001
-.400-001
-.50Q-CQ1
1
.
-»*
J»
*****
HiHfims** , ,
tit « it:: m::i:itti::tht:i.»
»*•»*' li»*»»*»*»*«»-»*-*»
»*•<•*' n» •»*£**•»»**»»»
"••*!*?' '*{' '•**»•»»*»
|!"
*
*
, , i
10 15 20 25 JO 35
FREQUENCY
NITROGEN DIOXIDE
50
DATA AVAILABLE = 234 DATA POSSIBLE = 273
Figure 22a. Residual histogram for N02 from the UAM simulation
results for Day 207-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 2G7-75
a.
Q.
n
on
UJ
CC
.400-001
.300-001
.200-001
.100-001
000
- 100-001
-.200-001
-.300-001
-.400-001
-500-001
.000
.
•
a a
_ a
a a o ° a
^gP^-jopa^a^
O
a
a
LEGEND
FREO STM
< i. i> a
< 2. 2> 0
< 2, 3> *
< +,99> <»
.150-001 .300-001 .450-001 .600-001 750-001
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 234 DATA POSSIBLE - 273
Figure 22b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 207-75.
82
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 207-75
Q.
CL
_i
•J
LU
or
1 UU^UWW f
.800-001 .
.600-001 .
.400-001 -
.200-001 -
QOfl
WWW
-200-001 .
-.400-001 •
-.600-001 -
-.BOOrOOl •
- 100+000 •
a
[
t
|m«*
fj jgsjHim*"
: ': i: ': \\ til'****
•
•
5 10 15 20 25 JO iS 40 45 5
FREQUENCY
OZONE
)ATA AVAILABLE = 210 DATA POSSIBLE =
Figure 23a. Residual histogram for 63 from the (JAM simulation
results for Day 207-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 207-75
RESIDUAL (PPM
.800-001
.600-001
.400-001
.200-001
000
-.200-001
-.400-001
-.600-001
-.300-001
- 100+000
Q
Q
^c&i^^yF °i °
l"!0**^ * ° .
a
LEGEND
FREQ SYM
< i, i> a
< 2. 2> O
< 3, 3> *
< *,99> »
000 500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 210 DATA POSSIBLE = 273
Figure 23b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 207-75.
83
-------
ST. LOUIS JULY 28.1975 I OUT 207)
'ReOICTEO C8NCENTIBTIQNS flf 03 UN PPfl I
HOUR 1400-1500
ST. LOUIS JULY 28.I97S I OUT 2071
39SEMVCO COMCENTMaTlONS Of 03 'IN
HOUR 1400-1500
(a)
(b)
Figure 24. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 207-75.
84
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 195-75
500-i-OOl
CL
CL
g
400+001 •
.300+001 •
•
.200+001 •
.100+001 L
000
-.100+001
-200+001
-.300+001
-.400+001
-.500+001
0 10 20 20 40 50 60 70 30 90 100
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 240 DATA POSSIBLE =
273
Figure 25a. Residual histogram for CO from the DAM simulation
results for Day 195-76.
Q.
CL
Q
CO
RESIDUAL VS OBSERVED CONCENTRATION
DAY 195-76
.3UU^VW 1
.400+001
300+001
.200+001
.100+001
000
-.100+001
-.200+001
-.300+001
-.400+001
-.500+001
000
a
a
.*£ ,,,,,,,'
^*?T i i , i i , ,
•
•
100+001 .200+001 .300+001 400+001 .500+
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 240 DATA POSSIBLE = 273
Figure 25b. Residual vs. observed plot for CO from the UAM
simulation results for Day 195-76.
85
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 195-76
3UU-UU I r
.400-001
^ .300-001
a
^ .200-00!
-! .100-001
Q 000
OH
£J -.ioo-oor
-.200-001
-.300-001
-.400-001
-500-001
»
.
|.
'• |jf*j£t««««***
It!: :l 1||ii||IS|lItl1}|«*«
If?: :: lltf ||*«4«M***MiM4M*iMMSMtWt* '
|j:;t;~.
l.»
-*••«•
10 15
40 45 50
20 25 20 35
FREQUENCY
NITROGEN DIOXIDE
DATA AVAILABLE = 253 DATA POSSIBLE
273
Figure 26a. Residual histogram for N02 from the UAM simulation
results for Day 195-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 195-75
2
a.
p^
D
9
Ld
a
.3V.XI— UUI
400-001
.300-001
.200-001
.100-001
-.100-001
- 200-001
-.300-001
-.400-001
-.500-001
000
a
a
i o °a B a a° ;
3 nor a a
Q a a mm
_ fl ""BW rt^^p^wnwn hr^r^
f;.-*,
L£CENO
FREQ SYM
< 1, t> O
< 2. 2> 0
.150-001 .300-001 .450-001 600-001 .750-001
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 253 DATA POSSIBLE = 273
Figure 26b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 195-76.
86
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 195-76
.750-001 , 1 1 1 1 1 1 1 1 ,
I
s
o_
g
Q
LU
Qi
.600-001
.450-001
300-001
.150-001
nnn
WIAJ
-.150-001
- 300-001
-.450-001
-.600-001
- 750-001
r
•i 4
i:
.
Stt
ia
*• ' '5f**2Tf * ****** I i ~ i i
• VvS
*••
•
-
•
10 15 20 25 30 35
FREQUENCY
OZONE
DATA AVAILABLE -
40 45 50
184 DATA POSSIBLE » 273
Figure 27a. Residual histogram for 03 from the UAM simulation
results for Day 195-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 195-76
Q
uo
UJ
Qi
- 750-001
000 .500-001 .100+000 ,150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 184 DATA POSSIBLE = 273
Figure 27b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 195-76.
87
-------
ST. LOUIS JULY 13.1 ait I OUT 1951
WeQICTEO CBNCENTHHTIONS Of 83 I IN CPB I
HOUR t500-1500
ST. LOUIS JUur 13.1975 I OUT 1951
ossemo CONCCNTRSTIONS or as it*
HOUR 1SOO-1600
(a)
(b)
Figure 28. Contours of (a) predicted and (b) observed fields
of 63 at hour of observed maximum on Day 195-76.
88
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 275-75
750+001 |—
10
JO
80 90 100
40 50 60 70
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 250 DATA POSSIBLE
273
Figure 29a. Residual histogram for CO from the UAM simulation
results for Day 275-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 275-76
J
a^
_,
3
Q
l/>
LJ
rw
. / ^\j^vu '
SOO+001
450+001
300+001
.150+001
000
- 150+001
- 300+001
- 450+001
- 600+001
- 750+001
Q |
a « I
.
a°a B
%-n*d m
H, 3 a H
^-AnS^pg^ a . |
IBKjtgpPaH . . . 1 i 1 — (
P^F^^
•
'
1 i ' ' ' i i i i
LEGEND
FREO SYM
< i. i> a
< 2. 2> O
< <99> »
000 200+001 .400+001 .600+001 300+001 100+002
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 250 DATA POSSIBLE = 273
Figure 29b. Residual vs. observed plot for CO from the UAM
simulation results for Day 275-76.
89
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 275-75
250-OOC i — 1
3
9
(S)
200+000 I-
150*000 r
.100+000 j-fM
.500-001 |-?s,
.000
-500-001
- lOC+OCOJ-'
-.150+000 I-
I
-200+0001-
0 5 10 15 20 25 • 30 35
FREQUENCY
NITROGEN DIOXIDE
0
< 2. 2> O
< 3, 3> »
< *,99> »
.500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
QATA AVAILABLE = 228 DATA POSSIBLE = 273
Figure 30b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 275-76.
90
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL ~'MES AND LOCATIONS
DAY 275-75
s*
0.
•— '
_J
CO
Ul
DC
^•\J "UU J
200+000
.150+000
.100+000
.500-001
000
-500-001
j^
JL
=-
s»
- 100+000 fU
-.150+000
-.200+000
-.250+000
f
.
^
'0 20
30
40 50 60 7Q
FREQUENCY
ao 90 iQO
OZONE
DATA AVAILABLE = 248
DATA POSSIBLE = 273
Figure 31a. Residual histogram for 03 from the UAM simulation
results for Day 275-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 275-76
/•^
2
Q.
a
*"*
g
Ul
a:
200+000
.150+000
.100+000
.500-001
f\f\f\
.UwU
-.500-001
- 100+000
- 150+000
-.200+000
- 250+000
,
-
a
n
DQO u a
0 °g g JJD CD ° .
B| Q Jp^|l |3 —»
ffi^^Sg^iff^ ^^nj^"..^. Q
^ CJBU uuf oO^n0 ^"y "I
iii ta o
tf
Q
a
•
UCCCNO
FREO STV
< i. i> a
< 2. 2> 0
< 3, 3> *
< 4,99> »
000 .500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 248 DATA POSSIBLE = 273
Figure 31b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 275-76.
91
-------
ST. LOUIS OCT. l.ir?8 I OUT 273)
PREDICTED CBNCENT»ST1BNS OF 03 UN Pffl I
HOUR t400-1500
ST. LOUIS OCT. 1.1976 IOBT 27SI
OBSERVED CONCENTRATIONS OF 03 ItN Pf8J
HOUR MOO-1 SCO
(a)
(b)
Figure 32. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 275-76.
92
-------
REFERENCES
Bencala, K. E. and J. H. Seinfeld, 1979: An air quality model performance
assessment package. Atmos. Environ., 13, 1181-1185.
Littman, F. E., 1979: Regional Air Pollution Study—Emission inventory summari-
zation. Report No. EPA-600/4-79-004, U. S. Environmental Protection
Agency, Research Triangle Park, NC 27711.
Lurmann, F., D. Godden, A. C. Lloyd, and R. A. Nordsieck, 1979: A Lagrangian
Photochemical Air Quality Simulation Model. Vol. I-Model Formulation, Vol.
II-Users Manual. EPA-600/8-79-015a,b (available NTIS).
Lurmann, F., 1980: Modification and Analysis of the Lagrangian Photochemical
Air Quality Simulation Model for St. Louis.. Environmental Research and
Technology, Inc. Document No. P-A095. Westlake Village, CA. 25pp.
Lurmann, F., 1981: Incorporation of Lateral Diffusion in the Lagrangian Photo-
chemical Air Quality Simulation Model. Environmental Research and Tech-
nology, Inc. Document No. P-A748. Westlake Village, CA. 32pp.
MacCracken, M. C., D. J. Wuebbles, J. J. Walton, W. H. Duewer, and K. E. Grant,
1978: The Livermore Regional Air Quality Model, I. Concept and develop-
ment. J. Appl. Met., 17, 254-272.
Schere, K. L. 1981: Air quality model response to objectively analyzed wind
fields. In Proceedings of the Fifth Symposium on Turbulence, Diffusion,
and Air Pollution, American Meteorological Society, Atlanta, GA, March
9-13, 1981.
Schiermeier, F. A., 1978: Air monitoring milestones: RAPS field measurements
are in. Environ. Sci. Techno!., 12, 644-651.
93
-------
Shreffler, J. H. and R. 8. Evans, 1982: The surface ozone record from the
Regional Air Pollution Study, 1975-1976. Atmos. Environ., 16, 1311-1321.
Shreffler, J. H. and K. L. Schere, 1982: Evaluation of Four Urban-Scale Photo-
chemical Air Quality Simulation Models. EPA Report, U.S. Environmental
Protection Agency, Research Triangle Park. NC (in press).
Zalesak, S. T., 1979: Fully multi-dimensional- flux-corrected transport algo-
rithms for fluids. J. Comp. Phys., 31, 335-362.
94
-------
APPENDIX A
PBM - TIME SERIES FOR REMAINING TEST DAYS
•
This appendix contains the time series plots for the pollutant species
CO, N02 and 03 from the PBM test results for those days that were not spe-
cifically discussed within the body of this report. The plots presented here
are from test days 182-75, 183-75, 184-75, 209-75, 221-75, 230-75, 231-75,
251-75, 159-76, 160-76, 195-76, 211-76, 225-76, 226-76 and 237-76 of the RAPS
data base.
95
-------
PBM SIMULATION-750701
a.
CL
o
o
r
7A 10.0 12.5 19.0
TIME, HOURS (CST)
17.3
20.0
Figure A-la.
PBM simulation results for Day 182 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
96
-------
.100
PBM SIMUIATION-750701
............
7.3
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-lb. PBM simulation results for N02 - Day 182 of 1975.
Keys to figure described in A-la.
PBM SIMULAT10N-750701
a.
a.
0.00
7'A 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-lc. PBM simulation results for 63 - Day 182 of 1975.
Keys to figure described in A-la.
97
-------
PBM SIMULATION-750702
2
CL
O.
o
O
5.0
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-2a.
PBM simulation results for Day 183 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
98
-------
PBM SIMULATION-750702
.TOO
.075
^^
^2*
Q.
Q.
.•OSO
CS
0
.025
5.
' \
' \
/ \
' V
' \
1 V
' >
— / \
\
\
\
\
\
• / \ \ '
fc- \. \ r '*<»«•"'
• o \ \ N /
: \ °
- ^v.° * °
••
0 7A 10.0 12.5 15.0 17.5
.
™
—
^
^
"
•
—
-
-
_
.
-
I
20,
TIME, HOURS (CST)
Figure A-2b. PBM simulation results for N02 - Day 183 of 1975.
Keys to figure described in A-2a.
PBM SIMULAT10N-750702
Q.
Q.
fO
o
0.00
10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-2c. PBM simulation results for 03 - Day 183 of 1975.
Keys to figure described in A-2a.
99
-------
PBM SIMULAT10N-750703
a.
o_
o
o
A
j
.a.
5.0
7-5 10.0 12.5 15.0
TIME. HOURS (CST)
17.5
20.0
Figure A-3a.
PBM simulation results for Day 184 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
100
-------
PBM SIMULATION-750703
a.
Q.
CS
O
0.00
7.5
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-3b. PBM simulation results for N02 - Day 184 of 1975.
Keys to figure described in A-3a.
PBM SIMULATION-750703
QL
Q.
ro
O
i i T i r i TriTiiiiiiiir
0.00
5U) 7A iaO 12.5 15.0 17.5 20.0
TIME, HOURS (CST)
Figure A-3c. PBM simulation results for 03 - Day 184 of 1975
Keys to figure described in A-3a.
101
-------
0.
0.
o
o
2.0
1.3
1.0
PBM SlMUUVTlON-750728
QJQ
iiri|(i!iitiiiiiiiij
i i i 7">r'~ T
&0
73 10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-4a.
PBM simulation results for Day 209 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model -simulated (solid
line) hour-average concentrations of CO.
102
-------
PBM SIMULAT10N-750728
I i i i i I 1 i I 1 J I I 1 1 I i I I 1 I 1 I I
a.
CL
CM
o
0.00
i i i I t
!.....!_. 1 ..! I 1 I I I I J_ . t i I L II i_._.J_... _L I 1 L
5.0
7 A 10.0 12.3 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-4b. PBM simulation results for NOg - Day 209 of 1975,
Keys to figure described in A-4a.
PBM SiMULATlON-750728
t I ! I T L | | ill j I 1 1 I I
7.5 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-4c.
PBM simulation results for 63 - Day 209 of 1975.
Keys to figure described in A-4a.
103
-------
a.
a.
O
o
PBM SIMULATION-750809
i i r i I i i i i I i i i i 1 i i i"7'*l'"VTT"'i I I i i i
5.0
Figure A-5a.
7 A 10.0 12.3 13.0 17.5
TIME, HOURS (CST)
2aO
PBM simulation results for Day 221 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model -simulated (solid
line) hour-average concentrations of CO.
104
-------
Q.
Q.
CN
O
.08
.06
.04
.02
PBM SIMULATION-750809
0.00 L-1
I I I I I I I I I T i I I T j I 1 T I
-•*'** Y
I 1 I I I I I 1 1
._!...._L ..i Ll I i 1.1 I t t I I I I I t I 1 t I I
5.0 7A 10.0 12.5 15.0 17.5 20.0
TIME, HOURS (CST)
Figure A-5b. PBM simulation results for N02 - Day 221 of 1975.
Keys to figure described in A-5a.
PBM SIMUUTION-750809
*OJ ft "r - i i IT i i II i i [ l f i (T 1 III I 1 " ]~ I I I ~"i
7A 10.0 12.5 .15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-5c. PBM simulation results for 03 - Day 221 of 1975,
Keys to figure described in A-5a.
105
-------
Q.
CL
o
o
PBM SIMULXriON-750818
1 I I I
I
I
_ I
9.0
Figure A-6a.
7J 10.0 12.5 1S.O
TIME, HOURS (CST)
17.5
20.0
PBM simulation results for Day 230 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
106
-------
PBM SIMULATION-750818
o.
a.
CN
O
z
I **«»**.M«.^***B**«»«_ *»*»•«•_«..«..+ _.*.•»*-.*
titfltrifiitiiFiifiTitii
0.00
10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-6b. PBM simulation results for N02 - Day 230 of 1975.
Keys to figure described in A-6a.
PBM SIMULAT10N-750818
i i t i i i i i i i i i i i i < i i * i ' i * * •*
JA 10.0 ^2A 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-6c. PBM simulation results for 03 - Day 230 of 1975.
Keys to figure described in A-6a.
107
-------
PBM SIMULAT10N-750819
a.
a.
o
o
I i r i i ^i i i i j i i i i i i i i i i
9U)
Figure A-7a.
7A
10.0
TIME,
15.0
HOURS (CST)
17.3
20.O
PBM simulation results for Day 231 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
108
-------
PBM SIMULATION-750819
Q.
Q_
I I I I I I I ! I I I j I I I I I I I I I I I I I !
aoo L-1
&0 7.5 10.0 12.3 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-7b. PBM simulation results for N02 - Day 231 of 1975.
Keys to figure described in A-7a.
PBM SIMULATION-750819
.301 i i r >
i i i i ii i i i
i . i i i I i i i i I i i i i
10.0 1Z5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-7c. PBM simulation results for 03 - Day 231 of 1975.
Keys to figure described in A-7a.
109
-------
5
CL
Q_
O
o
PBM S1MULATION-750908
I '
\
\
I
\s
7.5 10.0 12.5 15.0
TIME. HOURS (CST)
17.5
20.0
Figure A-8a.
PBM simulation results for Day 251 of 1975. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
110
-------
.100
.075
PBM SIMULATION-750908
a.
CL
.090
s
.025
0.000
ir*| r r r r
r r j i i i r
—r'VTT'4~TTT~1 I
5JO 7.3 10.0 1Z5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-8b. PBM simulation results for N02 - Day 251 of 1975,
Keys to figure described in A-8a.
PBM SIMUUT10N-750908
-T , I . . , . 1 r . . . I . . . .
10,0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-8c.
PBM simulation results for 03 - Day 251 of 1975.
Keys to figure described in A-8a.
Ill
-------
PBM SIMULAT10N-760607
a.
Q.
O
O
T~ i i I I i r i r r T r * i i i r i r T i i i i i i i r
A
1A 10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-9a.
PBM simulation results for Day 159 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
112
-------
PBM SIMULAT10N-760607
CL
Q.
CS
O
0.00
&0
Figure A-9b.
10.0 12.3 15.0 17.5
TIME, HOURS (GST)
20.0
PBM simulation results for N02 - Day 159 of 1976,
Keys to figure described in A-9a.
PBM SIMULATION-760607
10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-9c.
PBM simulation results for 63 - Day 159 of 1976,
Keys to figure described in A-9a.
113
-------
PBM SIMUtATION-760608
a.
Q_
o
o
Figure A-lOa.
7.S
iao 12.5 15.0
TIME, HOURS (CST)
17.5
20,0
PBM simulation results for Day 160 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
114
-------
PBM SIMULAT10N-760608
7.5
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-lOb.
PBM simulation results for N02 - Day 160 of 1976.
Keys to figure described in A-lOa.
PBM S1MUIXT10N-760608
o.
0.
I I I I I I I I I I I i.^1^ I I I I I I I I I I I I
i i i i I i i i i I i i i i
0.00
7J 10.0 12^ 15.0
TIME, HOURS (CST)
17.5 2aO
Figure A-lOc. PBM simulation results for 03 - Day 160 of 1976.
Keys' to figure described in A-lOa. .
115
-------
1.00
.75
PBM SIMULATION-760713
a.
a.
.30
O
O
.25
aoo
10.0 12JS 19.0
TIME. HOURS (CST)
17.5
20.0
Figure A-lla.
PBM simulation results for Day 195 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model -simulated (solid
line) hour-average concentrations of CO.
116
-------
.06
PBM SIMULAT10N-760713
.04
Q.
Q.
cs
O
.02
0.00
, ,, ,
V
Till
I ill li I 1 I I I I 1 i I T I I
Figure A-llb.
7A 10.0 12.5 1S.O 17.5 20.0
TIME, HOURS (CST)
PBM simulation results for N02 - Day 195 of 1976.
Keys to figure described in A-lla.
PBM S1MUUVT10N-760713
a.
Q.
I 1 I I I I < 1 I I j I I f I I
I I t I 1 I t I 1 1 I I I I I
0.00
7A 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-llc. PBM simulation results for 03 - Day 195 of 1976.
Keys to figure described in A-lla.
117
-------
PBM SIMULAT10N-760729
2
0.
0.
o
o
I 1 L 1 I ( I I I I I 1 I I
Figure A-12a.
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
PBM simulation results for Day 211 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
118
-------
.I00r-r-r
Q_
Q_
CN
O
.075
.oao
.025
0.000
PBM SIMUUVT10N-760729
5U>
Figure A-12b.
A
\
\
i ( | i
7.5 10.0 12.5 15.0 17.5 20.0
TIME, HOURS (CST)
PBM simulation results for N02 - Day 211 of 1976.
Keys to figure described in A-12a.
PBM SIMULXnON-760729
a.
a.
r •^f'-"—
&f^^T~\* i i i i t I i i i i I i i i i I i i i i I i i i i
0.00
7A iaO 12.5 -15.0
TIME. HOURS (CST)
17.5 20.0
Figure A-12c. PBM simulation results for 03 - Day 211 of 1976,
Keys to figure described in A-12a.
119
-------
PBM SIMUUVTION-760812
Q.
0.
O
o
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-13a.
PBM simulation results for Day 225 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
120
-------
QL
Q.
.06
.06
.04
.02
PBM SIMULAT10N-760812
1 I f 1 1 1
1II111I1IJI1 J* 1 i I 1
; \
i \
' \
i i i i
0.00
XV
f i t i t I i i 1 i I i t i i T t i i i
5UJ 7A 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-13b. PBM simulation results for N02 - Day 225 of 1976.
Keys to figure described in A-13a.
PBM S1MULATION-760812
Q.
Q.
i i i i i i i i iiiiiii
0.00
10.0 12.5 15.0 17.5
TIME, HOURS (CST)
20.0
Figure A-13c. PBM simulation results for 03 - Day 225 of 1976,
Keys- to figure described in A-13a.
121
-------
PBM SIMULAT10N-760813
2
Q.
Q.
O
O
Figure A-14a.
7.5
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20,0
PBM simulation results for Day 226 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
122
-------
.100
.075 —
PBM SIMULATION-760813
Q.
Q.
.050 -
CM
O
Z
.025 -
0.000
7.5 10.0 12.5 15.0 17.5
TIME, HOURS (CST)
Figure A-14b. PBM simulation results for ^ - Day 226
Keys to figure described in A-14a.
20.0
of
PBM SIMULATION-760813
I I I I I I I I I I I I I 1 I ! I I I I I I I
7.5 10.0 12.5 -15.0
TIME, HOURS (CST)
17.5 20.0
Figure A-14c. PBM simulation results for 03 - Day 226 of 1976,
Keys to figure described in A-L4a.
123
-------
PBM SIMULATION-760824-
CL
a.
O
o
3
2
1,
I ,1 1 M | 1 1 1 1 | 1 1 1 1 | 1 1 1 1 | 1 1 I 1 | 1
I \
1 \
1 \
' 1 \
/ \
-r \
i \
•r \
l'
I O >
r \
• i
- °
i
i
\
0 V
,\ \ x' X
_ / \ N / N
" ' ^^^ \/
• / \ ^^^-^^.^ • o
L/ ^--^-a— ^ o * o
/ N "^
7 ,,,,!.,,, >VTT-i-4-i-^f-r-r ""?>>, 1 ,
™
—
™
.
—
~
-
^
_
-
—
•
.
""•
—
-
1 t
10.0 12.5 15.0
TIME, HOURS (CST)
17.5
20.0
Figure A-15a.
PBM simulation results for Day 237 of 1976. Time
series of average observed (circles), range of
observed (dashed lines), and model-simulated (solid
line) hour-average concentrations of CO.
124
-------
.15i i i i i
PBM SIMULATION-760824
12.5 15.0 17.5
TIME. HOURS (CST)
20.0
PBM simulation results for ^2 - Day 237 of 1976.
Keys to figure described in A-15a.
.201 i i i i
a.
a.
0.00
SUJ
PBM SIMUUM10N-760824
7A 10.0 124 15.0
TIME, HOURS (CST)
2Q.Q
Figure A-15c. PBM simulation results for 03 - Day 237 of 1976.
Keys- to figure described in A-15a.
125
-------
APPENDIX B
PBM - SUMMARY OF STATISTICAL RESULTS
This table summarizes the mean concentration data and the analysis of
trends data obtained from PBM simulations on all 20 days for the species
N02. NOX, 03, NMHC and CO. Note that the mode*! internally estimates a measure
of NMHC from its explicit structural classes. The table also includes summary
statistics on concentration maxima data for all 20 test days.for the species
N02, 03 and CO. The statistical parameters included in this table are
(Bencala and Seinfeld, 1979):
•
(a) mean observed and mean predicted concentration,
(b) mean residual (measures the average bias in the predictions and
indicates whether the model predominantly over- or underpredicts),
(c) root mean square (RMS) error (measures the average spread of the
residuals, but is insensitive to bias),
(d) mean residual/mean observed (measures the average percentage of
over- or underprediction), and
(e) error band (measures the percentage of predictions that fall within a
25% band around the observations).
The analysis of trends data indicates whether or not the predictions obey
the same fundamental relationships as the observations. The parameters con-
tained in this table include:
126
-------
(a) correlation coefficient (measures the overall degree to which the
magnitude of the predictions increases linearly with the magnitude
of the observations, but is insensitive to the extent of the in-
crease) ,,
(b) slope (the linear least-squares curve fit measure of the average
increase in the observations as the predictions increase),
(c) intercept (the linear least-squares curve fit measure of the bias
if the slope parameter is very nearly one; otherwise no independ-
ent meaning), and
(d) sum of the squared error (a relative measure of the scatter about
the linear least-squares fit to the data).
Concentration maxima data are presented for the species MOg, 03 and
CO for each test day. The parameters listed here are:
(a) peak observed concentration (and the time of occurrence),
(b) peak predicted concentration (and the time of occurrence),
(c) residual concentration of the peak, and
(d) residual/observed concentration (percentage over- or underpredic-
tion of the peak).
In all tables a positive residual indicates an underprediction by the
model and a negative residual indicates an overprediction.
127
-------
TABLE B-l. SUMMARY OF MODEL PERFORMANCE STATISTICS PROM THE PBMa
Mean Concentration Data
Species
N02
NOX
03
NMHC
CO
Mean observed
concentration
(ppm)
0.0285
0.0440
0.0725
0.402
0.369
Mean predicted
concentration
(ppm)
0.0281
0.0412
0.0879
0.350
0.676
Mean
residual
(ppm)
0.00039
0.00284
-0.0154
0.0520
0.193
-
RMS error
(ppm)
0.00931
0.0148
0.0289
0.188
0.356
Mean residual
Mean obs. cone.
(I)
1.4 -
6.5
-21.2
12.9
22.2
Error
band0
(?)
60.0
56.5
45.4
43.1
43.5
Species
Analysis of Trends Data
Correlation
coefficient
Slope
Intercept
(ppm)
Sum of
squared error
(ppm2)
N02
NOX
°3
NMHC
CO
,876
,923
,871
.837
0.848
0.753
0.832
1.09
0.666
0.599
0.00734
0.00976
-0.0230
0.169
0.464
0.0276
0.0685
0.269
12.1
50.6
128
-------
TABLE 8-1. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FROM THE PBM*
Concentration Maxima Data
Jul ian
date
142-75
178-75
182-75
183-75
184-75
207-75
209-75
Observed Peak
Species
N02
03
CO
N02
°3
CO
N02
°3
CO
NOg
03
CO
N02
03
CO
N02
°3
CO
N02
°3
CO
• Value
(ppm)
0.0367
0.137
1.31
0.0535
0.159
1.85
0.0408
0.0939
0.979
0.0526
0.115
1.57
0.0792
0.114
2.68
0.0416
0.140
1.24
0.0304
0.108
0.976
Concentration
Time
(CST)
0600 •
1200
0700
0700
1400
0600
0600
1300
0700
0800
1300
0700
0700
1300
0700
0600
1400
0500
0800
1000
0800
Predicted Peak
Value
(ppm)
0.0421
0.148
1.18
0.0456
0.173
1.07
0.0532
0.140
1.68
0.0594
0.0971
1.09
0.0753
0.182
2.11
0.0509
0.145
1.02
0.0406
0.122
1.06
Concentration
Time
(CST)
0700
1300
0600
0600
1600
0600
0600
1600
1500
0600
1700
0600
0600
1300
0500
0600
1500
0500
0700
1300
0600
Residual
(ppm)
-0.0054
-0.0110
0.130
0.0079
-0.0140
0.780
-0.0124
-0.0461
-0.7001
-0.0068
0.0179
0.480
0.0039
-0.0680
0.570
-0.00093
-0.0050
0.220
-0.0102
-0.0140
-0.0840
Residual
Obs. cone.
(*)
-14.7
-8.0
9.9
14.8
-8.8
42.2
-30.4
-49.1
-71.6
-12.9
-15.6
30.6
4.9
-59.6
21.3
-22.4
-3.6
17.7
-33.6
-13.0
-8.6
129
-------
TABLE &-1. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FROM THE PBM*
Concentration Maxima Data
Julian
date
221-75
230-75
231-75
251-75
159-76
160-76
195-76
Observed Peak.
Species
NO 2
03
CO
N02
03
CO
N02
03
CO
N02
°3
CO
N02
*3
CO
N02
°3
CO
N02
03
CO
Value
(ppm)
0.0453
0.132
1.94
0.0541
0.0717
2.08
0.0394-
0.0842
1.04
0.0667
0.0841
2.85
•0.0938
0.125
2.78
0.0652
0.162
1.56
0.0287
0.151
0.555
Concentration
Time
(CST)
0700
1300
0600
0700
1300
0600
0800
1000
0900
0900
1500
0600
0700
1300
0600
0700
1400
0600
1700
1500
0700
Predicted Peak
Value
(ppm)
0.0469
0.125
1.62
0.0499
0.119
1.72
0.0364
0.101
1.06
0.0461
0.205
1.72
0.0816
0.210
1.98
0.0693
0.170
1.50
0.0260
0.137
0.522
Concentration
Time
(CST)
0700
1300
0500
0700
1500
0600
0700
1300
0600
0700
1500
0600
0600
1200
0500
0600
1500
0500
0600
1600
0600
Residual
(ppm)
-o'.ooie
0.0070
0.320
0.0042
-0.0473
0.360
0.0030
-0.0168
-0.0200
0.0206
-0.121
1.13
0.0122
-0.0850
0.800
-0.0041
-0.0080
0.0600
0.0027
0.0140
0.0330
Residual
Obs. cone.
(%)
-3.5
5.3
15.5
7.8
-66.0
17.3
7.6
-20.0
-1.9
30.9
-143.9
39.6
13.0
-68.0
28.8
-6.3
-4.9
3.8
9.4
9.3
• 5.9
130
-------
TABLE 8-1. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FROM THE PBM*
Concentration Maxima Da.ta
Jul ian
date
211-76
212-76
225-76
226-76
237-76
275-76
Observed Peak Concentration
Species
N02
03
CO
N02
03
CO
N02
03
CO
N02
03
CO
N02
03
CO
M02
°3
CO
Value
(ppm)
0.0409
0.0939
0.869
*
0.0609
0.127
1.73
0.0421
0.111
1.08
0.0580
0.144
1.75
0.0832
0.115
2.45
0.108
0.183
4.17
Time
(CST)
1200
1500
0700
0700
1200
0600
1600
1200
0600
0700
1500
0700
0700
1100
0700 '
1700
1400
0700
Predicted Peak
Value
(ppm)
0.0338
0.0878
0.719
0.0528
0.121
1.42
0.0491
0.0919
1.26
0.0489
0.171
1.05
0.0521
0.156
1.05
0.0869
0.223
3.00
Concentration
Time
(CST)
0700
1600
0600
0600
1400
0500
0700
1100
0600
0700
1500
0600
0700
1400
0600
1000
1500
0700
Residual
(ppm)
0.0071
0.0061
0.150
0.0081
0.0060
0.310
-0.0070
0.0191
-0.180
0.0091
-0.0270
0.700
0.0311
-0.0410
1.40
0.0211
-0.0400
1.17
Residual
Obs. cone.
(*)
17.6
6.5
17.3
13.3
4.7
17.9
-16.6
17.2
-16.7
15.7
-18.8
40.0
37.4
-35.7
57.1
19.5
-21.9
28.1
Statistics for Mean Concentration and Analysis of Trends data are developed for all 20 test
days;
Statistics for Concentration Maxima are listed for each test day.
All concentration values represent an average over the model domain.
bPercent of predictions within 25% of observations.
131
-------
APPENDIX C
LPM TEMPORAL CORRELATION COEFFICIENTS
AND TIME SERIES FOR REMAINING TEST DAYS
This appendix presents temporal correlation coefficients for 03, NO, and
for all 20 days and time series from LPM simulations for the 15 days not
presented in the body of the report.
The temporal correlation coefficients were computed between observed and
predicted series from the start time to the lesser of 1700 CST or the end of
the run. These coefficients for Levels 1 and 3 appear in Table C-l. The
choice of a termination time can have significant influence on the correlation.
For instance, the correlation for 03 on Day 231 is below the average for the
20 days, but the simulation was fairly successful. A part of the problem with
the correlation is the discrepancy between prediction and observation after the
time of the observed maximum. As explained previously for Day 231, this
discrepancy is expected because the modeled parcel has moved away from the
stations contributing to the "observed" value. The correlations for NO and
N02 tend to be low, perhaps reflecting the low concentrations and the conse-
quent difficulty in obtaining accurate measurements.
132
-------
TABLE C-l. TEMPORAL CORRELATION COEFFICIENT BETWEEN OBSERVED AND LPM PREDICTED
POLLUTANT CONCENTRATIONS FOR THE PERIOD FROM
LATION TO 1700 CST OR EARLIER TERMINATION.
HE START OF THE SIMU-
Day
03
L-l L-3
NO
L-l " L-3
N02
L-l
L-3
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
0.53
0.52
0.99
0.95
0.89
0.73
0.64
0.82
0.85
0.64
0.67
0.97
0.94
0.94
0.90
0.80
0.11
0.92
0.89
0.90
7 = 0.78
°r 0.21
0.66
0.88
0.98
0.97
0.83
0.78
0.58
0.76
0.83
0.58
0.64
0.97
0.92
0.95
0.89
0.73
0.06
0.83
0.84
0.80
0.77
0.20
-0.03
0.54
0.95
0.90
0.90
0.18
0.53
0.82
0.30
0.79
0.91
0.94
-0.12
0.77
0.50
-0.01
0.13
0.78
0.73
0.45
0.49
0.42
0.34
0.51
0.70
0.39
0.48
0.41
0.17
0.56
0.20
0.15
0.87
0.88
-0.02
0.50
-0.28
0.07
0.34
0.48
0.70
0.88
0.42
0.30
-0.26
-0.39
0.94
0.65
0.77
0.32
0.93
0.77
0.43
0.62
0.84
0.89
-0.09
0.61
0.70-
0.30
-0.36
0.28
0.83
0.66
0.47
0.42
-0.03
-0.71
0.94
0.65
0.81
0.53
0.93
0.85
0.42
0.57
0.92
0.92
0.23
0.44
0.78
0.30
0.07
0.88
0.87
0.87
0.56
0.42
133
-------
75142. RAMS 101 AT 1200CST. START 0500
l(AJ •••p^T^"H^Z^™"^^T^
SO
25
0 25
50
KM
75 100
Figure C-la. LPM parcel trajectory for Day 142 of 1975.
134
-------
75142, RAMS 101 AT 1200CST. START 0500
iiiiiiiiiiiiii
OBS
. PRED L-1
. PRED L-3
i 1 I i i i I I r i t I i i i I I i i i i 1 t i i i
10.0 12.5 15.0
HOUR, CST
17 J
20.0
Figure C-lb. LPM time series of observed and predicted 03 hourly
concentrations on Day 142 of 1975.
.100
.075
.050
75142, RAMS 101 AT 1200CST. START 0500
0.000
I I I I
I I i I I I I I I I I I . I
I , I I
OBS
— PRED L-1
— PRED L-i
i i i i I I 1 t 1 I t 1 I I I t I
5.0
10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-lc. LPM time series of observed and predicted NO hourly
concentrations on Day 142 of 1975.
135
-------
0.
CM
O
.20
.15
.10
.05
75142, RAMS 101 AT 1200CST. START 0500
0.00
lIllIillIJtlliTI I I I | III
oes
PRED L-1
PRED L-4
A '\
/ \ •..
\v
t 11 1 1 i I I 1_ I 1 1 II
7.5 10.0 12.5 15.0
HOUR, CST
17,5
20.0
Figure C-ld. LPM time series of observed and predicted N02 hourly
concentrations on Day 142 of 1975.
75142, RAMS 101 AT 1200CST. START 0500
Q.
Q.
o"
O
I I I I ...
08S
— PRED L-t-
— PRED L-5.
* I I I I I
II 111 J i I • t I L _ t I I I
5J) 7A 10,0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-le. LPM time series of observed and predicted CO hourly
concentrations on Day 142 of 1975.
136
-------
75178. RAMS 112 AT 1400CST. START 0500
IW p™T"T~~T~ T ^H""rT^^^TT"^~™T'*"'TOFT™Tl"'"T^^P™1^™^^T1"™
SO
yK-^j o o
t II | I | | || 111 j 1 I I 1 1 I
0 29 90
KM
79 100
Figure C-2a. LPM parcel trajectory for Day 178 of 1975.
137
-------
75178, RAMS 112 AT 1400CST. START 0500
0.
O.
ul
O
rsj
O
OLO
OBS
PRED L-1
PRED L-3
/
/
/
/
10.0 12.5 15.0
HOUR, CST
17.5
20-0
Figure C-2b. LPM time series of observed and predicted 03 hourly
concentrations on Day 178 of 1975.
.4
75178, RAMS 112 AT 1400CST. START 0500
o
.1
T
T
I 1^ I I Til f
OBS
PRED L-1
PRED L-4
0.0«
54
7.5 iaO 12J 15.0
HOUR, CST
17.5
2ao
Figure C-2c. LPM time series of observed and predicted NO hourly
concentrations on Day 178 of 1975.
138
-------
75178, RAMS 112 AT 1400CST. START 0500
.4
Q.
Q.
*
CM
O
0.0
iiiiriiijiiijj
OBS
PRED L-1
PRED L-51
5U) 7.5 iaO 1ZS 13.0
HOUR, CST
zao
Figure C-2d. LPM time series of observed and predicted N02 hourly
concentrations on Day 178 of 19.75.
10.0
75178, RAMS 112 AT 1400CST. START 0500
I
O*
O
2J>
QJQ
i I r i §
I i i i i
OBS
PRED L-1
PRED L-4
I | i . L....._!_ I I i 1 I I I f 1 I t 1 I I
&0 7.5 10.0 12.S 15.0
HOUR, CST
17.5
20,0
Figure C-2e. LPM time series of observed and predicted CO hourly
concentrations on Day 178 of 1975.
139
-------
75182. RAMS 125 AT 1200CST. START 0500
•w ^T-MT"^r"TMT™™^^t i~" v I k TCHT"^^"^^
90
o o
29
30
KM
100
Figure C-3a. LPM parcel trajectory for Day 182 of 1975,
140
-------
75182, RAMS 125 AT 1200CST. START 0500
. PRED L-3 A
i i i i I i i i t I I 1 i I I I I * I I I I I t
&0
10.0 12J 15.0 17.5 2aO
HOUR, CST
Figure C-3b. LPM time series of observed and predicted 03 hourly
concentrations on Day 182 of 1975.
751-82. RAMS 125 AT 1200CST. START 0500
t i i t i i i i j i i i r
- oss
...... PRED L-
PRED L-31
I
CL
O
0.00
7A 10.0 1Z3 19.0 17.5 2aO
Figure C-3c. LPM time series of observed and predicted NO hourly
concentrations on Day 182 of 1975.
141
-------
75182, RAMS 125 AT 1200CST. START 0500
OSS
PRED L-1"
PRED L-i
0.000
iao 12.3 15.0
HOUR, CST
17.5
20.0
Figure C-3d. LPM time series of observed and predicted N02 hourly
concentrations on Day 182 of 1975.
75182, RAMS 125 AT 120QCST. START 0500
Q.
0.
o"
O
OBS
PRED L-1-
PRED L-3
8.0 7A 10.0 12.5 15.0 17.5 20.0
HOUR, CST
Figure C-3e. LPM time series of observed and predicted 03 hourly
concentrations on Day 182 of 1975. :
142
-------
75183, RAMS 124 AT 1500CST. START OSOO
• W r™"T^"r ~T~~ i "*T~™J*"" | | j V ^ ^ | Lf 11 in ^ | | |
SO
I ' "" ' I
I i L | I 1 I i I I I II 1
25 90
KM
75 100
Figure C-4a. LPM parcel trajectory for Day 183 of 1975,
143
-------
75183, RAMS 124 AT 1500CST. START 0500
Q.
Q.
UJ
O
r>4
O .1
0.0
oas
PRED L-1
PRED L-3
I 1 I I I I I I I I I I I I I I I I I I I I I I I I I
7.9 10.0 12.9 19.0 17.9 20.0
HOUR, CST
Figure C-4b. LPM.time series of observed and predicted 03 hourly
concentrations on Day 183 of 1975.
75183, RAMS 124 AT 1500CST. START 0500
OBS
PRED L-1
— PRED L-1
10.0 12.9 19.0
HOUR, CST
17.9
20.0
Figure C-4c. LPM time series of observed and predicted NO hourly
concentrations on Day 183 of 1975.
144
-------
.06
.04
Q.
Q.
CM
o
.02 -f
aoo
75183, RAMS 124 AT 1500CST. START 0500
I I I r I F I T
1 I i r i i jT i I i
OBS
PRED L-1
PRED L-J1
7.5 iaO 1Z3 13.0 17.3 20.0
HOUR, CST
Figure C-4d. LPM time series of observed and predicted N02 hourly
concentrations on Day 183 of 1975.
75183, RAMS 124 AT 1500CST. START 0500
1.5
I
O
O
i r i i i r i i r r i i i i i i i i i
OBS
PRED L-1.
PRED L-i
. ... I .... I ..
10.0 1Z5 15.0
HOUR, CST
17.3
20.0
Figure C-4e. LPM time series of observed and predicted CO hourly
concentrations on Day 183 of 1975.
145
-------
75184. RAMS 118 AT 1300CST. START 0500
• *^ ^^T™"^^HHT""T™"^""TT r T TTO^"^TT^m
75
90
25 -
25
50
KM
75 100
Figure C-5a. LPM parcel trajectory for Day 184 of 1975.
146
-------
75184, RAMS 118 AT 1300CST. START 0500
i
OBS
PRED L-1
I PRED L-3
i i -**V liiiiliiifititiiiiiiliitt
10.0 1Z5 15.0
HOUR, CST
17 J
20.0
Figure C-5b. LPM time series of observed and predicted 03 hourly
concentrations on Day 184 of 1975.
.too
.073
75184, RAMS 118 AT 1300CST. START 0500
Q.
O*
.090
.025
0.000
i r i T j i i i i i i i i i
- OBS
------ FRED L-1
-- PRED L-i
fluO
10.0 12.3 15,0
HOUR, CST
17.3
2O.O
Figure C-5c. LPM time series of observed and predicted NO hourly
concentrations on Day 184 of 1975.
147
-------
.20
.15
Q.
0- .10
CN"
O
z
75184, RAMS 118 AT 1300CST. START 0500
OBS
PRED L-1
PRED L-3
0.00 *-
5.0
10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-5d. LPM time series of observed and predicted N02 hourly
concentrations on Da^y 184 of 1975.
75184, RAMS 118 AT 1300CST. START 0500
i ii r r i 7 i
i • • • • i ' • • •
OBS
PRED L-1
PRED L-4
oL_i_a
54 7.5 10.0 1Z5 15.0
HOUR, CST
17.5
2ao
Figure C-5e. LPM time series of observed and predicted CO hourly
concentrations on Day 184 of 1975.
148
-------
75297. RAMS 113 AT 1400CST. START 0600
1(W r"T*"T~^~T" t i T^T^T i V n* i T"TP"7"""1""^~^"~
75
90
I ' ' ' ' I
90
KM
79 100
Figure C-6a. LPM parcel trajectory for Day 207 of 1975.
149
-------
75207, RAMS 1 1 3 AT 1 400CST. START 0600
2 •*
CL
Q.
UJ
O
r^i
O .1
0.0
I I I I I I
OSS
PRED L-1
PRED L-3
I I 1 I I F 1 i
i r~ i i i r
iO 7J 10.0 12.5 15.0 17.5 20.0
HOUR, CST
Figure C-6b. LPM time series of observed and predicted 63 hourly
concentrations on Day 207 of 1975.
.100
.075
75207, RAMS 113 AT 1400CST. START 0600
O
J325
ill! 1 I T I Illlllflili
—~~" OSS
PRED L-1.
PRED L-i
0.000 L
5U>
J. I 1 L
iaO 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-6c. LPM time series of observed and predicted NO hourly
concentrations on Day 207 of 1975.
150
-------
.100
.073
Q.
Q.
O
.023
75207, RAMS 113 AT 1400CST. START 0600
~~~~ 08S
PRED L-1
PRED L-1
OJOOI ' ' ' ' I ' ' ' ' I ' ' ' ' I ' i ' ' I ' ' ' • I ' ' » '
5.0 7.5 iaO 12.3 13.0 17.5 2aO
HOUR, CST
Figure C-6d. LPM time series of observed and predicted N0£ hourly
concentrations on Day 207 of 1975.
75207, RAMS 113 AT 1400CST. START 0600
Q.
Q.
O
O
0
30)
I ...
OBS
PRED L-V
PRED L-1
i 111 [ 11 iiliiiiitiii I i i i IT I
i i i i
7.3 10,0 1Z3 13.0 17.5
HOUR, CST
20.0
Figure C-6e.
LPM time series of observed and predicted CO hourly
concentrations on Day 207 of 1975.
151
-------
752Q9. RAMS 118 AT 11OOCST. START 0600
ItX? p'T^TLTJlT™1illlTT--r' r 1 i i^nW"^" i i r"~T""^™T1™"
90
29
ao
KM
73 100
Figure C-7a. LPM parcel trajectory for Day 209 of 1975.
152
-------
0.
Q.
ul
z
O
M
O .1
0.0
75209, RAMS 1 18 AT 11OOCST. START 0600
i i i i i i i i i r ' i i ' i i ' i ' I ' ' ' ' i * » ' •
08S
PRED L-1
PRED L-3
"»l-< It I I ! f I f I 1 ! I I 1 I I I I 1 I ! _ I
5.0 7.5 10.0 12.5 15.0 17.5 20.0
HOUR, CST
Figure C-7b. LPM time series of observed and predicted 03 hourly
concentrations on Day 209 of 1975.
75209, RAMS 118 AT 110OCST. START 0600
Q.
Q_
O
OBS
PRED L-1
PRED L-31
0.00
10.0 1ZS 15.0
HOUR, CST
17.5
20.0
Figure C-7c. LPM time series of observed and predicted NO hourly
concentrations on Day 209 of 1975.
153
-------
.100
75209, RAMS 118 AT 110OCST. START 0600
oas
PRED L-1
PREO L-i|
L__..!__. 1 L_..J_.. .!_ I I I I I I i i L i
10.0 12.5 15.0 17.5 20.0
HOUR, CST
Figure C-7d. LPM time series of observed and predicted N02 hourly
concentrations on Day 209 of 1975.
75209, RAMS 118 AT 1100CST. START 0600
Q.
Q.
O*
O
~~~~" 08S
PRED L-1-
PRED L-4
1 t I J I I I
&0 7A 10.0 12.3 15.0
HOUR, CST
17.5 20.0
Figure C-7e. LPM time series of observed and predicted CO hourly
concentrations on Day 209 of 1975.
154
-------
75221. RAMS 121 AT 1500CST. START 0600
iiA/ ^*T"^^™™TOT"i™r^™r'""^^^j^T^»^^
75
2 »
j-r I«-T < r
o o
29
80
KM
73 100
Figure C-8a. LPM parcel trajectory for Day 221 of 1975.
155
-------
.20
.15
75221. RAMS 121 AT 1500CST. START 0600
0.001-1
i i i I i i i i
DBS
PRED L-1
PRED L-3
I
I
7.5 10.0 12.5 15.0
HOUR. CST
17.5
zao
Figure C-8b. LPM time series of observed and predicted 03 hourly
concentrations on Day 221 of 1975.
.020
75221. RAMS 121 AT 1500CST. START 0600
1 I ' '
OBS
PRED L-1
PRED L-i
0.000 u
54
10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-8c. LPM time series of observed and predicted NO hourly
concentrations on Day 221 of 1975.
156
-------
.06
.04
I
CM
O
z
.02
0.00
75221, RAMS 121 AT 1500CST. START 0600
"" 08S
PRED L-1
PRED L-J
54 7.5 iaO 12J 13.0
HOUR, CST
17 A
20.0
Figure C-8d. LPM time series of observed and predicted NOg hourly
concentrations on Day 221 of 1975.
2JQ
1.5
I
O
O
OO
75221. RAMS 121 AT 1500CST. START 0600
1 ) I I I 1 i I
08S
PRED L-1
PRED L-i
1_ i _J_ 1 ]_ I. 1. I ll I ! I t \ I I I I I t JIL
54 7.5 10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-8e. LPM time series of observed and predicted CO hourly
concentrations on Day 221 of 1975.
157
-------
75230. RAMS 121 AT 1300CST. START 0600
• W n^T^T^HT"^™ \L i T™™T^^ \ r™"^^^^^^^^T^~P™T™TM"
90
9 O
I I 1 I I I 1 1 I I I I I I
25
ao
KM
75 100
Figure C-9a. LPM parcel trajectory for Day 230 of 1975,
158
-------
75230. RAMS 121 AT 1300CST. START 0600
r r t 11 i * T IT T 1 i i I i i r i j r i i r j IT i
—— OBS
PRED L-1
. PRED L-3
O.
Q.
ul
O
N
O .1
OO
\
&o 7.5 10.0 1^5 15.0
HOUR, CST
17.5
zao
Figure C-9b. LPM time series of observed and predicted 63 hourly
concentrations on Day 230 of 1975.
75230, RAMS 121 AT 1300CST. START 0600
.15
•'«
.05 -
0.00
l I I l J I I T T
OBS
PRED L-1.
PRED L—3.
10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-9c. LPM time series of observed and predicted MO hourly
concentrations on Day 230 of 1975.
159
-------
.100
75230, RAMS 121 AT 1300CST. START 0600
1*11111111111
I ' ' ' '
OSS
PRED L-1
PRED L-il
iaO 12J 15.0 17.5
QjQOO
20.0
Figure C-9d. LPM time series of observed and predicted N0£ hourly
concentrations on Day 230 of 1975.
75230, RAMS 121 AT 1300CST. START 0600
o_
a.
o"
o
I I . I
OBS
PRED L-1-
— PRED L-i
0
6.0
t t ill | 1 < J | i I 111 1 | 1 I 1 1 1 I I I 1 I I i
7J 10.0 12.S 1S.O 17.5 20.0
HOUR, CST
Figure C-9e. LPM time series of observed and predicted CO hourly
concentrations on Day 230 of 1975.
160
-------
76159. RAMS 122 AT 15QOCST. START 0500
•WJ p*^^^^••»•^M^m
90
o o
, , I . , , , I ,
29 90
KM
79 100
Figure C-lOa. LPM parcel trajectory for Day 159 of 1976.
161
-------
76159, RAMS 122 AT 1500CST. START 0500
o.
o.
ul
z
o
M
O .1
oo
I
T i I i r r
OBS
— PRED L-1
— PRED L-3
II I I I I I lilt I I I I 1 It t
SuO 7A 10.0 1Z5 15.0 17.5 2aO
HOUR. CST
Figure C-lOb. LPM time series of observed and predicted 03 hourly
concentrations on Day 159 of 1976.
76159, RAMS 122 AT 1500CST. START 0500
7J
10.0 12.5 15.0
HOUR, CST
17.5 zao
Figure C-lOc.
LPM time series of observed and predicted NO hourly
concentrations on Day 159 of 1976.
162
-------
76159. RAMS 122 AT 1500CST. START 0500
» i i i i i i i
OBS
PRED L-1
PRED L-i
10.0 12.9 19.0
HOUR, CST
17.9
20.0
Figure C-lOd. LPM time series of observed and predicted N02 hourly
concentrations on Day 159 of 1976.
76159, RAMS 122 AT 1500CST. START 0500
— 08S
— PRED L-1
•— PRED L-i
i i -i i I i i i i I i i
7.5
10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-lOe. LPM time series of observed and predicted CO hourly
concentrations on Day 159 of 1976.
163
-------
76160. RAMS 115 AT 1600CST. START 0500
i w* r^T^™^^p^r^"™T^'T^^T™*T*™T^~i^^™T^™riTnnr™T"^^~
25
o 0
I .... I
o
. I
i til I i i i L
29
90
KM
75 100
Figure C-lla. LPM parcel trajectory for Day 160 of 1976.
164
-------
76160, RAMS 115 AT 1600CST. START 0500
I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' '
OBS
PRED L-1
. PRED L-3 ^.-
/J
O.
CL
ul
O
0.0
7.5 iaO 12.5 15.0
HOUR, CST
17.5
zao
Figure C-llb. LPM time series of observed and predicted 03 hourly
concentrations on Day 160 of 1976.
.04
.03
£ •«
.01
76160, RAMS 115 AT 1600CST. START 0500
0.00
r T^T | T . r
OBS
— PRED L-1.
— PRED L-1
y , , i 1 ,
5.0 7.5 10.0 12.5 15.0- 17J
HOUR, CST
20.0
Figure C-llc. LPM time series of observed and predicted NO hourly
concentrations on Day 160 of 1976.
165
-------
.100
.075
Q.
Q-.050
.025
0.000
76160, RAMS 115 AT 1600CST. START 0500
i I i I i i r r
I I ( I I I T
OBS
PRED L-1.
PRED L-i
jit i i j i i i I i i i i 1 i t ii lii i i 1 i i i i
10.0 12 J 15.0
HOUR, CST
17.5
20.0
Figure C-lld. LPM time series of observed and predicted N02 hourly
concentrations on Day 160 of 1976.
76160, RAMS 115 AT 1600CST. START 0500
10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-lle. LPM time series of observed and predicted CO hourly
concentrations on Day 160 of 1976.
166
-------
76,135. RAMS 114 AT 1500CST. START 0500
IQOi i i i i | i i i i | i vi I I I I I I I
79
SO
29
I 1 J I t 1
. s
29
SO
KM
79 100
Figure C-12a. LPM parcel trajectory for Day 195 of 1976.
167
-------
76195, RAMS 114 AT 1500CST. START 0500
i
OSS
PRED L-1
PRED L-3
I I I I I f I I I I t 1 I I I I ! I I I I 1
10.0 12J 15.0 17.5 20.0
HOUR, CST
Figure C-12b. LPM time series of observed and predicted 63 hourly
concentrations on Day 195 of 1976.
.01
.01
0.00
76195, RAMS 114 AT 1500CST. START 0500
o
1 1 1 I
i I 1 1
I ' ' '
OBS
PRED L-1.
PRED L-i
SU3 7.3 10.0 12.3 15.0 17.5 2aO
HOUR, CST
Figure C-12c. LPM time series of observed and predicted NO hourly
concentrations on Day 195 of 1976.
168
-------
.100
.079
76195, RAMS 114 AT 1500CST. START 0500
I
.080
CN
O
.025
0.000
i r r i | * i r i i r
I j I I I 1
OSS
PRED L-1
PRED L-4
I t I _j ._.J_ | | I I t I j I I i 1._._!_
54 7.5 iaO 12J 15.0
HOUR, CST
20.0
Figure C-12d. LPM time series of observed and predicted N02 hourly
concentrations on Day 195 of 1976.
76195, RAMS 114 AT 1500CST. START 0500
1
o"
O
0.0
r i i i i i i i
i I I I I i I I r
— OBS
PRED L-1.
PRED L-i
ill!
7.5 10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-12e. LPM time series of observed and predicted CO hourly
concentrations on Day 195 of 1976.
169
-------
76211. RAMS 120 AT 1500CST. START 0600
• W ^^^"T^f^j^ —... ( "•—j—.y—^-—••—j—^^^—j—j—»yi—^n^^-^-
90
o o
40
KM
79 100
Figure C-13a. LPM parcel trajectory for Day 211 of 1976,
170
-------
76211, RAMS 120 AT 1500CST. START 0600
Q.
Q_
ul
O
M
O .1
0.0
OBS
PRED L-1
PRED L-3
xX~"
-, i ,» i i
i i »j--L-J— »-^«-y* i I I I i I I I I I I
I 1 I I I I
5.0 7A 10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-13b. LPM time series of observed and predicted 03 hourly
concentrations on Day 211 of 1976.
.100
.075
76211, RAMS 120 AT 1500CST. START 0600
-030
o
z
.025
0400
—— OBS
PRED L-1
PRED L-i
< *-...»•
i i i
SJO
7J&
10.0 1££ 15.0
HOUR, CST
17.5
20.0
Figure C-13c. LPM time series of observed and predicted NO hourly
concentrations on Day 211 of 1976.
171
-------
O.
O
.100
.075
.090
.025
76211, RAMS 120 AT 1500CST. START 0600
0.000
T
OSS
PRED L-1
PRED L-4
i i I i i i ! I i i i i I i i i i I
i i
5UJ 7A 10.0 1Z5 19.0 17.9
HOUR, CST
20.0
Figure C-13d. LPM time series of observed and predicted NQ2 hourly
concentrations on Day 211 of 1976.
76211, RAMS 120 AT 1500CST. START 0600
CL
Q.
o"
O
™~~~ OSS
PRED L-V
PRED L-5.
6U} 7J 10.0 12.9 15.0
HOUR, CST
17.5
20.0
Figure C-13e. LPM time series of observed and predicted CO hourly
concentrations on Day 211 of 1976.
172
-------
76212. RAMS 108 AT 12QOCST. START 0600
lyQ .—•yni-ji ( ( j_. .._j_. j j_ • (- ^ .j T (—- | ( (
25
I ' ' ' ' I
. . . I
I I I I I I I I T
50
KM
75 100
Figure C-14a. LPM parcel trajectory for Day 212 of 1976,
173
-------
76212, RAMS 108 AT 1200CST. START 0600
CL
CL
m
LU
Z
o
M
O .1
0.0
I \ i t
I '
OBS
PRED L-1
PRED L-3
i I t i i i I i i i i I i i i i I i i i i
7A 10.0 1i5 15.0
HOUR, CST
17.5
20.0
Figure C-14b. LPM time series of observed and predicted 03 hourly
concentrations on Day 212 of 1976.
76212, RAMS 108 AT 1200CST. START 0600
.02
0.
Q.
.01
0.00
OBS
—~ PRED L—1-
PRED L-4
1
1
5U> 7.3 10.0 124 15.0.
HOUR, CST
17^
20.0
Figure C-14c. LPM time series of observed and predicted MO hourly
concentrations on Day 212 of 1976.
174
-------
.06
76212, RAMS 108 AT 1200CST. START 0600
.04
0.
a.
•
CN
O
.02
0.00
OBS
PRED L-1
PRED L-51
till
1 j j 1 L I I I 111 t 1 l_ .1 J I L L \ I i . I. .!_
7.5 10.0 12.5 15.0
HOUR, CST
2ao
Figure C-14d. LPM time series of observed and predicted N02 hourly
concentrations on Day 212 of 1976.
2JO
1.5
76212. RAMS 108 AT 1200CST. START 0600
I
o"
O
0.0
I ' ' ' ' I ' ' '
—— 08S
PRED L-1.
PRED L-i
&0
10.0 12.5 15.0
HOUR, CST
17.5
20.0
Figure C-14e. LPM time series of observed and predicted CO hourly
concentrations on Day 212 of 1976.
175
-------
76225. RAMS 117 AT 1300CST. START 0600
i^W "^^"f*"T*™T™'TOiri^^^"T^T^^ii^^
90
25
o*-4-
j. I I 5 I
29
30
KM
100
Figure C-15a. LPM parcel trajectory for Day 225 of 1976.
176
-------
.20
.15
O.
Q.
ul .10
O
M
O
.05
76225, RAMS 117 AT 1300CST. START 0600
I .... I
5U>
7A
12J 15.0
HOUR, CST
17 J
zao
Figure C-15b. LPM time series of observed and predicted 03 hourly
concentrations on Day 225 of 1976.
o_
O*
.020
.015
.010
76225, RAMS 117 AT 1300CST. START 0600
.005
0.000
OBS
PRED L-1
PRED L-i
x\\
\g£> I I i 1^^"^
tiriliiiiliiitiiiiilri
5U) 7 A 10.0 12.5 15.0
HOUR, CST
17.5 20.0
Figure C-15c. LPM time series of observed and predicted NO hourly
concentrations on Day 225 of 1976.
177
-------
.04
.03
O.
Q. .02
CM
O
.01
76225, RAMS 117 AT 1300CST. START 0600
0.001-1-1
I T I I I I i
i i i i I i i i i
OBS
PRED L-1
PRED L-4
i i i I i i i i I t i i i
&0 7J 10.0 1Z5 15.0
HOUR, CST
17.3
20.0
Figure C-15d. LPM time series of observed and predicted N02 hourly
concentrations on Day 225 of 1976.
2JO
76225, RAMS 117 AT 1300CST. START 0600
o
O
OSS
PRED L-1.
PRED L-i
&0
10.0 12.5 15.0 17.5 20.0
HOUR. CST
Figure C-15e. LPM time series of observed and predicted CO hourly
concentrations on Day 225 of 1976.
178
-------
APPENDIX D
UAM - RS AND RO PLOTS FOR REMAINING TEST DAYS
This appendix contains the residual histogram (RS) and residual vs.
observed plots (RO) for the species N02, 03 and CO from the UAM statistical
analysis that were not specifically discussed within the body of this report.
The plots presented here are from test days 178-75, 182-75, 183-75, 184-75,
209-75, 221-75, 230-75, 231-75, 251-75, 159-76, 160-76, 211-76, 212-76,
225-76, 226-76 and 237-76 of the RAPS data base.
179
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 178-75
500+001
Q.
Q.
.400+001
.300+001
.200+001
-j 100+001
<
9 ooo
c/>
{£ -.100+001
- 200+001
-.300+001
-.400+001
-.500+001
5 10 15 20 25 30 35
FREQUENCY
CARBON MONOXIDE
40 45 50
DATA AVAILABLE
163 DATA POSSIBLE
273
Figure D-la. Residual histogram for CO from the UAM simulation
results for Day 178-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 178-75
2
Q.
Q.
^"'
RESIDUAL
.3UU+UUI
400 +001
.300+001
.200 -i-OOl
.100-1-001
000
- IOOt-001
- 200+001
-.300+001
-.400+001
- 500+001
•
- o
0°
0 ° 008
jji|JiSp&f-M^^C^^^" ' I ' ' '
3
•
•
-
000
100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 163 DATA POSSIBLE = 273
Figure D-lb. Residual vs. observed plot for CO from the UAM
simulation results for Day 178-75.
180
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 173-75
750-001
3
Q
on
.500-001
450-001
.300-001 H
.150-001 j-g;
.000
- 150-001
-.300-001
-.450-001
-.600-001
- 750-001
5 10 15 20 25 30 35
FREQUENCY
NITROGEN DIOXIDE
40 45 50
DATA AVAILABLE = 224 DATA POSSIBLE
273
Figure D-2a. Residual histogram for N02 from the UAM simulation
results for Day 178-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 178-75
Q_
Q.
Q
LEGEND
FREQ STM
< 1, 1> 0
< 2. 2> O
< J, 3> *
.150-001 .300-001 .450-001 600-001 750-001
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 224 DATA POSSIBLE = 273
Figure D-2b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 178-75.
181
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 178-75
loo+ooo p—-
.aoo-oo i
600-001
.200-001
.000
Q.
Q.
g
9
C/l
g -200-001
-400-001
-.600-001
-.800-001
- 100+000
-t—
H*
0 5 10 15
OZONE
DATA AVAILABLE
20 25 JO 35
FREQUENCY
40 45 50
203 DATA POSSIBLE
273
Figure D-3a. Residual histogram for 03 from the UAM simulation
results for Day 178-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 178-75
5
Q.
C/)
LJ
o;
.500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 203 DATA POSSIBLE = 273
Figure D-3b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 178-75.
182
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 132-75
D.
Q.
~
_j
Q
W
Ul
or
400+001
300+001
.200 rOC 1
.100+001
000
-.100+001
- 200+001
-.300+001
-.400.+00'
- 500+001
.
•
•
L
•
•
•
•
50 100 150 200 250 300 350 400 *50 500
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 221 DATA POSSIBLE
273
Figure D-4a. Residual histogram for CO from the UAM simulation
results for Day 182-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 182-75
G.
Q_
^
_J
C/}
UJ
or
.400+001
300+001
.200+001
.100+001
000
-.100+001
— ^yu+uu
- 300+001
-.400+001
- 5QQ+QQ1
a
a
O 03
T a 8 °
__90MMT_ fl a , i ,
^rr"
•
•
i i i
LcCENO
fT»EQ SYM
< i. t> n
< 2. 2> »
< J. 3> *
.000 .100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 221 DATA POSSIBLE = 273
Figure D-4b. Residual vs. observed plot for CO from the UAM
simulation results for Day 182-75.
183
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 182-75
.750-001
600-001 U
Q.
Q.
9
en
*0 50 60 7Q
FREQUENCY
NITROGEN DIOXIDE
90 90 100
DATA AVAILABLE = 213 DATA POSSIBLE = 273
Figure D-5a. Residual histogram for NO^ from the UAM simulation
results for Day 182-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 182-75
v-s.
a.
0.
RESIDUAL (
.600-001
>50-001
.300-001
.150-001
000
- 150-001
- 300-001
-.450-001
-.600-001
- 750-001
000
a
' a a i? ™, a° °
[•ijffl^T M a a ° a ° q
.sp°" " '°°'°° '
:*
•
uecENp
FREO SYM
< i, i> a
< 2. 2> 0
< 3. 1> *
< *,99> 9
200-001 .400-001 .600-001 800-001 100+000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 213 DATA POSSIBLE = 273
Figure D-5b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 182-75.
184
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 182-75
2
Q.
Q.
Q
u5
ui
or
600-001
.450-001
.300-001
.150-001
000
-.150-001
-.300-001
-.450-001
-.600-001
- 750-001
*
8
•»
,
"1
tit
5*r
£»*«<
t
1
£
t*
1
Eji
i.4
:.
.
'S5JS***** "
: T - **•
: 1 1 : i ff?f r***7***v* i i i i
• *• - ??*?** 1 ! ! 1 1
lfffl!!t::::r
•»*
^
Figure D-6a.
10 15 20 25 30 35 40 45 50
FREQUENCY
OZONE
DATA AVAILABLE = 249 DATA POSSIBLE = 273
Residual histogram for 03 from the UAM simulation
results for Day 182-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 182-75
Q.
Q.
9
CO
LJ
o;
- 750-001
000 .500-001 .100-1-000 .150+000 200+000 250-t-OOO
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 249 DATA POSSIBLE = 273
Figure D-6b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 182-75.
185
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 183-75
s
CL
CL
SIDUAL
Ul
ce
3UU-I-UU 1
400+001
.300+001
.200+001
.100+001
000
-.100+001
-.200+001
-.300+001
-.400+001
- 500+00 1
.
L
fr ,
SBS"^ m*i~maL**L*J~LA" *
•
•
10 20 20 40 50 60 70
FREQUENCY
CARBON MONOXIDE
ao so too
DATA AVAILABLE = 213 DATA POSSIBLE
273
Figure D-7a. Residual histogram for CO from the UAM simulation
results for Day 183-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 183-75
2
CL
O._
_l
O
UJ
a:
.3UU-MJU 1
400+001
300+001
.200+001
.100+001
onn
wuu
- 100+001
- 200+001
-.300+001
- 400+001
- SDQ+QO!
, . i . i i i i i
•
O
o
a o en o °
,„<§& *o %0|f
liiMjfy^ Q ^^TJT Q
p"^*"8
•
1 1 t 1 i • 1 I 1
LEGEND
FREO STM
< l. i> O
< 2. 2> 0
< 3. J> *
< *,99> «
.000 .100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 213 DATA POSSIBLE = 273
Figure D-7b. Residual vs. observed plot for CO from the UAM
simulation results for Day 183-75.
186
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 183-75
.600-001
^ .450-001
a
"•" .300-001
^ .150-001
a
ig -.150-001
-.300-001
-.450-001
-.600-001
- 750-001
.
•*»
!$.
|| p^:::: r , ,
• '-u«- i **»
•*
4
*
10 15 20 25 30 35
FREQUENCY
NITROGEN DIOXIDE
40
45 50
DATA AVAILABLE = 221 DATA POSSIBLE = 273
Figure D-8a. Residual histogram for N02 from the UAM simulation
results for Day 183-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 133-75
^^v
2
a.
Q.
~
_i
z>
Q
CO
UJ
o:
. / JV— UU 1
600-001
.450-001
.300-001
.150-001
000
n o
rS 0°
a-,- %
~3EffiK~f£* QCCCr"^ •• J3
3nrjflpCj cj u Q ^r
fflSEzTy d ' ' '
|4£^*~^on^j ~ o DB
- 150-001
-.300-001
-.450-001
-.600-001
- 750-001
Pa
TJJ
a
p
n
I i . i , i i i
LEGEND
FREQ SYM
< i. i> n
< 2. 2> 0
< 3. i> *
< ».99> »
000 .500-001 .100+000 .150+000 .200+000 250+000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 221 DATA POSSIBLE = 273
Figure D-8b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 183-75.
187
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 183-75
0,
RESIDUAL
800-001 4
.600-001 •
.400-001 -
.200-001 -
000
-.200-001 -
-.+00-001 •
-.500-001 •
-800-001 •
- 100+000 -
t
t*
*if%i" i i i i ii
***i iili !*********
10 15
OZONE
DATA AVAILABLE
20 25 30 35
FREQUENCY
40 +5 50
200 DATA POSSIBLE
273
Figure D-9a. Residual histogram for 03 from the UAM simulation
results for Day 183-75.
Q
RESIDUAL VS OBSERVED CONCENTRATION
DAY 133-75
- 100+000
000 500-001 .100+000 .150+000 .200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 200 DATA POSSIBLE = 273
Figure D-9b. Residual vs. observed plot for 63 from the UAM
simulation results for Day 183-75.
188
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 184-75
Q.
Q.
<
9
CO
10 20 30
80 90 100
Figure D-lOa.
40 50 60 70
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 223 DATA POSSIBLE
273
Residual histogram for CO from the DAM simulation
results for Day 184-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 184-75
a
a
*~s
-j
9
Uj
a:
SOO+UUI
400+001
300+001
.200+001
.100+001
000
- 100+001
- 300+001
- 400+001
-500+001
000
a
300
^^j^jBcDD "o^ «. ^i Q D
JpP^ a1 °, aa
a
-
lECCNO
FBEO SYM
< i. i> a
< 2. 2> 0
.150+001 .300+001 .450+001 .600+001 750+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE: = 223 DATA POSSIBLE = 273
Figure D-lOb. Residual vs. observed plot for CO from the UAM
simulation results for Day 184-75.
189
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
QAY 184-75
^^\
3
0_
•— '
g
00
LJ
or
^JU-*" um-i
.200+000
.150+000
.100+000
500-001
000
-.500-001
- 100+000
-.150+000
-.200+000
- 250+000
•
A^
^
•
10 20 JO *0 50 60 70
FREQUENCY
NITROGEN DIOXIDE
ao 90 100
DATA AVAILABLE
239 DATA POSSIBLE = 273
Figure D-lla. Residual histogram for N02 from the UAM simulation
results for Day 184-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 184-75
^*v
Q.
OL
1 —
_i
3
00
UJ
or
^3U+UUU
•200+000
150+000
.100+000
.500-001
OCX"!
.
a a
a 30
- aj8M.g,§°baB *
iarfr?!i?j'3.
-.500-001
-.lOu+OOO
-.150+000
- 200+000
- JSQ-'-QQQ
*
•
,
L£C£NO
FREO STM
< 1. l> O
< 2. 2> O
< *:»> *
000 500-001 100+000 .150+000 .200+000 250+000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 239 DATA POSSIBLE = 273
Figure D-llb. Residual vs. observed plot for MOg from the UAM
simulation results for Day 184-75.
190
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 184-75
2j
0^
^
D
£
^uufuuu r
.800-001
.500-001
.400-001
.200-001
000
-.200-001
-.400-001
-.600-001
-.300-001
- 100+000
-»
• H
liili3!
:1|*|:«!
Hi i
tn:;:i:
|||n£:t
- •?***
i|in
< *
' -••»»•«
:*"
(
3 5
!*.««..*
: !?«!(•*
tiiS<<*
j*« '
10 15 20 25 20 35 40 45 50
FREQUENCY
OZONE
DATA AVAILABLE = 229 DATA POSSIBLE = 273
Figure
D-12a. Residual histogram for 03 from the DAM simulation
results for Day 184-75.
RESIDUAL VS OBSERVED CONCENTRATION
CL
^
g
Q
i7>
OS
100-t-OOO r
aoo-ooi [
500-001 •
.400-001
200-001
000 >
i |
1° °
- 200-001 5a°
fen °
— AOfl — OO 1 J *ftt» m m n*
*vU Jw 1 '
-.600-001
DAY 184-75
• i i i i i i i
•
o
B-jg&S1 9 M°
..^ESffi*1' M ^ °
'•^^^^FlTfi g
a ^__ jJTOgrr Sri
1 i '.i ' • •rt-^TTTrP*' ' i
_ Jj "rt *^j
n ^r ri"
g I^T3 CT
a^a0 a
cn a o
r ^
OBO^
. o a
-800-001 ^f
w
OC Q •
LECENO
FREO SYM
< i, i> n
< 2. 2> O
< 4,99> *
000 500-001 .100-1-000 .150-t-OOO 200+000 250-1-000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 229 DATA POSSIBLE = 273
Figure D-12b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 184-75.
191
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 209-75
^" X
3
Q.
— '
RESIDUAL
^3U+UU^
.200+002
.150+002
.lOO-t-002
.500+001
.000
-.500+001
-.100+002
-.150+002
-.200+002
-.250+002
.
»
-
•
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 230 DATA POSSIBLE = 273
Figure D-13a. Residual histogram for CO from the UAM simulation
results for Day 209-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 209-75
s
CL
0-
<
9
c/i
UJ
or
.43V -rvu*
.200+002
.150+002
.100+002
.500+001
000
-.500+001
-.100+002
-.150+002
-.200+002
- 250+002
.
a
LEGEND
FREO SYM
< i, i> a
< 2, 2> 0
< J, 3> *
< 4,99> »
.0
500+001 .100+002 .150+002 200+002 250+002
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 230 DATA POSSIBLE = 273
Figure D-13b. Residual vs. observed plot for CO from the UAM
simulation results for Day 209-75.
192
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 209-75
. JVAJ— VU i
.400-001
£L .300-001
a
"-" .200-001
^ .100-001
Q 000
(75
^ -.100-001
-.200-001
-.300-001
-.400-001
-.500-001
< j
•;
; ',
. ",
~ *
' <
*
.
.
t.
• !****
' .i : ' i tiiitHi • • « 1 1 • • i
" '< ' >• T9*9*49$*
' \ ", t *••••-*»»•»»
i < • >. k9
»»**
I
•
-
10 15 20 25 30 35
FREQUENCY
NITROGEN DIOXIDE
40 45 50
DATA AVAILABLE - 225 DATA POSSIBLE = 273
Figure D-14a. Residual histogram for NC>2 from the UAM simulation
results for Day 209-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 209-75
a
a
a
^ -.100-001
.000 .150-001 .300-001 .450-001 .600-001 .750-001
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 225 DATA POSSIBLE = 273
Figure D-14b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 209-75.
193
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 209-75
2
Q.
Q.
3
Q
(/)
Ld
or
^3w»-umj r
.200+000
.150+000
.100+000
.500-001
000
-,500-001
-.100+000
-.150+000
-.200+000
- 250+000
.
•
" 1 *
^Hm^jg.j™.
• t; ; • 'Jlil ijllij tjj**»*****tl***»i*»iiJ»*ti»»J* ' '
.
•
3 5 10 15 20 25 JO 35 40 45 5
FREQUENCY
OZONE
DATA AVAILABLE = 218 DATA POSSIBLE =
273
Figure D-15a. Residual histogram for 03 from the (JAM simulation
results for Day 209-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 209-75
•5.
Q.
Q.
>~'
i/5
UJ
or
.200+000
150+000
.100+000
.500-001
000
-.500-001
-.150+000
-.200+000
- 250+000
.
•
° jr>0 o
&"n_'a ° °
^ qt^^hcffifoiSi o o °
—^^— — — ^— iTO I .I3IDB .^H^BdD LIII
BI "TnoD^^^
LEGEND
FREX! STM
< i. i> a
< 2. 2> 0
< 3. J> *
< 4,39> «
.000 .500-001 .100+000 .150+000 .200+000 .250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 218 DATA POSSIBLE = 273
Figure D-15b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 209-75.
194
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 221-75
.750*001
2
Q.
Q.
10 20 30 40 50 60 70
FREQUENCY
CARBON MONOXIDE
ao
90 100
DATA AVAILABLE = 222 DATA POSSIBLE = 273
Figure D-16a. Residual histogram for CO from the (JAM simulation
results for Day 221-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 221-75
'"•X
s
Q.
Q.
~
RESIDUAL
.750+001
.500+001
.450+001
.300+001
.150+001
000
- 150+001
- 300+001
-.450+001
-.600+001
-.750+001
000
a
a
lflg°B 0
'wBJ$^°i 8 i °i , , , , '
Q
.150+001 .300+001 .450+001 600+001 .750^
LEGEND
FREQ SYM
< 1, l> HI
< 2. 2> 0
< 3. 3> A
< *,99> 0
^001
Figure D-16b.
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 222 DATA POSSIBLE = 273
Residual vs. observed plot for CO from the UAM
simulation results for Day 221-75.
195
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 221-75
3UU-UU 1
.400-001
./— X
5J .300-001
0.
^ .200-001
gj .100-001
Q .000
C/l
^ - 100-001
-.200-001
-.300-001
-.400-001
-.500-001
I
»»*
SI"
i**
ljjlfu:t ,",,.'
iffpll. UWUtI*"
••
»*
«*
10 15 20 25 30 35 40 45 50
FREQUENCY
NITROGEN DIOXIDE
DATA AVAILABLE » 222 DATA POSSIBLE = 273
Figure D-17a. Residual histogram for N02 from the UAM simulation
results for Day 221-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 221-75
S-*
0.
0.
*-s
_J
^
Q
C/l
LU
cr
.JUU— W 1
.400-001
300-001
.200-001
.100-001
.000
-.100-001
-.200-001
-.300-001
-.400-001
- 500-001
D fl.
B a a
ir°aaa *
• ^\g*i***\^Jr
| .jft'f'^ 'Jl'?"\ pi n ] I D ' 31
9pc nu^lj^ |Tr>^ a o ' ;u
-K|k a°o° a
n 3 jr
Q
a a
a
CD
II ill
LECENO
FREQ SYU
< i. i> a
< 2. 2> 0
< +!99> «
000 .150-001 .300-001 .450-001 500-001 750-001
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE - 222 DATA POSSIBLE = 273
Figure D-17b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 221-75.
196
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 221-75
.ASW^VVW
.200*000
^ .150*000
Q.
~ .100+000
J .500-001
Q 000
Co
£ -500-001
-.100*000
-.150*000 •
-.200+000 •
- 250*000 •
'*•**•*****• ..
.
.
r*
I1 f3»f3ir"
iljSf jtxffijjtjs**11***1'**"********
•
>
5 10 15 20 25 JO .55 *0 45 5
FREQUENCY
OZONE
DATA AVAILABLE = 239 DATA POSSIBLE =
273
Figure D-18a. Residual histogram for 03 from the UAM simulation
results for Day 221-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 221-75
^^
2
0.
a
**~*s
3
5
t/5
Ul
ac
.^3U*VAAJ
200*000
.150*000
.100*000
.500-001
.
•
B n<*>
crp o —
'•••••^ *? rfr'fiip'ri !'i (*r m
rV»*Wp-^m^fiL«J^J g Q
nnn Ln n^h awfiUWi*T V7F n I i I
SSjjfSff'^ llj
-.500-001
-.100+000
-.150+000
-.200+000
- 250+000
•
< i . . i , i i i i
LEGEND
FREQ SYM
< 1. !> 0
< 2, 2> 0
< 3. 3> »
< +,99> »
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 239 DATA POSSIBLE = 273
Figure D-18b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 221-75.
197
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 230-75
.250-1-001
0.
Q.
Q
00
10 20 30
30 90 100
*0 50 60 70
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 219 DATA POSSIBLE
273
Figure D-19a. Residual histogram for CO from the UAM simulation
results for Day 230-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 230-75
RESIDUAL (PPM)
.^au-*uu >
.200 1-001
,150-t-OOI
.100^-001
.500-t-OOO
000
-.500+000
- 100+001
-.150 -i-OOl
-.200+001
-.250+001
.000
Q
O
„ '.
"'^,1, 0=
iffr ° iii ii
3 an
9
LEGEND
FREO SYM
< i, i> a
< 2. 2> O
< 3, 3> *
< *,99> »
.100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 219 DATA POSSIBLE = 273
Figure D-19b.
Residual vs. observed plot for CO from the UAM
simulation results for Day 230-75. -
198
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
CAY 230-75
auu — vu i r
.400-001 •
^^\
5| .300-001 •
Q.
^ .200-001 •
J .100-001 -
Qnnn
'UUU •
(75
y - 100-001 -
-.200-001 •
-.300-001 •
-.400-001 •
-.500-001 L
^
.
•* J2^* ^
tesil*i^ii•»«Il»••l * "* '
IH*****************
• • *•*•••••
!«•»
*
•
10 15 20 25 30 35
FREQUENCY
NITROGEN DIOXIDE
40 45 50
DATA AVAILABLE = 231 DATA POSSIBLE = 273
Figure D-20a. Residual histogram for N02 from the UAM simulation
results for Day 230-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 230-75
o.
Q.
Q
CD
UJ
cc
- 500-001
.000 200-001 .400-001 .600-001 .300-001 100-^000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 231 DATA POSSIBLE = 273
Figure D-20b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 230-75.
199
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 230-75
. IUU+UUU
.800-001
-^
^ .600-00?
Q.
^ .400-00?
-j .200-001
QflfYl
00
UJ _ 200-001
- 400-001
-.600-001
-.800-001
-.100+000
**
.
^
*••
" f
: HI Utg«;*|:**
- " " * " '??*' '??^??*?T?T^S**
" " '2' ^.^
. ^ .. ,{)
** *
*
10 15 20 25 .30 35
FREQUENCY
OZONE
DATA AVAILABLE = 204
40
45 50
DATA POSSIBLE
273
Figure D-21a. Residual histogram for 03 from the UAM simulation
results for Day 230-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 230-75
^-s
0.
Q.
N— '
Q
on
UJ
o:
. lUU-fUUU
800-001
.600-001
.400-001
.200-001
000
-.200-001
-.400-001
-.600-001
-.800-001
_ lOO-"-000
.
aaa
•
a
a OB
a a g
— 3
- Jn a _
(3 SB ~ a
Sw'feS^0 °a
Q Q ^*] Q
a
•
l£C£NO
FREO STM
< t. i> a
< 2. 2> 0
< 3. 3> *
000 500-001 .100+000 .15a+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 204 DATA POSSIBLE = 273
Figure D-215. Residual vs. observed plot for 03 from the UAM
simulation results for Day 230-75.
200
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 23' -75
:s
o_
^
RESIDUAL
JUU+VU 1
.400+001
.300+001
.200+001
.100+001
.000
- 100+001
-200+001
-.300+001
-.400+001
-.500+001
(
t
I
^
F"
40 50 60 70
FREQUENCY
CARBON MONOXIDE
30
DATA AVAILABLE = 205 DATA POSSIBLE
273
Figure D-22a. Residual histogram for CO from the UAM simulation
results for Day 231-75.
Q.
a
(Sl
UJ
or
RESIDUAL VS OBSERVED CONCENTRATION
DAY 231-75
.400+001
300+001
.200+001
.100+001
000
-.100+001
- 200+001
- 300+001
-.400+001
-.500+001
300
a
a a
__rip*? ,",,,, , "
f^a .
•
•
100+001 .200+001 .300+001 *00+001 .500
LEGEND
FREQ SYM
< i. i> a
< 2. 2> 0
< 3. 3> *
t-001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 205 DATA POSSIBLE = 273
Figure D-22b. Residual vs. observed plot for CO from the UAM
simulation results for Day 231-75.
201
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
QAY 231-75
on
3UU— «U 1
.400-001
.300-001
.200-001
100-001
000
-.100-001
-.200-001
-.300-001
-.400-001
-.500-001
.
.
t
'. !$„.»
" 3S3»«*
i i i izi ", 3* : :t***
; . 'i >lj .. .« J^ 'J >«^f £4 J. J. ^ L , 1
: ii :?i :: i: : :i JKI J|I jn|****'**'** "
: : i SJ IsSl5J**'*w
Ht*tt*
cn
•
.,,,,,,,,
25 30 35
FREQUENCY
NITROGEN DIOXIDE
DATA AVAILABLE = 248
40 45 50
DATA POSSIBLE - 273
Figure D-23a. Residual histogram for N02 from the UAM simulation
results for Day 231-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 231-75
RESIDUAL (PPM)
400-001
.300-001
.200-001
.100-001
.000
-.100-001
-.200-001
-.300-001
-.400-001
-.500-001
000
«5>°°a"aJ =>
fl i1^1 LL L3 .!• »i taj ^ ^(-W ™ ,«. A [ •• i Q i
IR % 5 ^^r^ff1™ ° a
• (HBTin o I'B^^I n
v^D uD jn Q OQ
y Q OHM L^L Q Q
" oo a 0
UECCNO
FRCO STM
< i. i> a
< 2, 2> O
< 3, 3> A
< +,99> «
.150-001 .300-001 .450-001 600-001 750-001
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 2*8 DATA POSSIBLE = 273
Figure D-23b. Residual vs. observed plot for M02 from the UAM
simulation results for Day 231-75. :
202
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 231-75
lUUfUUW -
.aoo-ooi \-
^ 600-001 -
Q.
^ .400-001 -
J .200-001 -
Q .000
to
J£ -.200-001 -
- 400-001 -
-.600-001
-.800-001 •
- 1 00+000 1
»
.
»f
| tit t**
: II 1: : iii Hi if Ui.«..l ' '
: ! : |||*it£»t»*{S*$ii$ •
. ||j«*<-»>
•
•
10 15
OZONE
DATA AVAILABLE
20 25 30 35
FREQUENCY
40 *5 50
246 DATA POSSIBLE
273
Figure D-24a. Residual histogram for 63 from the UAM simulation
results for Day 231-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 231-75
r— X
Q.
Q.
^
RESIDUAL
. lUU^VUU
.800^-001
.600-001
.400-001
.200-001
.000
-.200-001
- 400-001
-.600-001
-.800-001
-.100 +000
.000
a
•'. 1.
=B00 °a a o ° B
" Q Q Q Q QQ^_ Q Q
^W^H Q
(3D M 0 m'l'l Q, _
fflfci]rT_~fff jfBL n I 1 ' l
H^^^^t^ 'i 'i T* ' *" rfffji^n 1*1 ' ' iii
^B1 ri ff Wfp *^p' —
3X.O a offi^ffl
o Wy 3
LECENO
FTO3 SYM
< l. i> O
< 2. 2> 0
< 3. J> *
< 4,99> «
500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 246 DATA POSSIBLE = 273
Figure D-24b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 231-75.
203
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 251-75
Q.
O.
s"-l'
SI DUAL
Ld
or
./3U+UUI
oCOrOOI
450+001
.300+001
.150+001
.000
-.150+001
-.300+001
- 450+001
-.600+001
-.750+001
-
}
|=~
SfiP88BBB^rr '
* .
•
•
10 20 30 40 50 60 70 80 90 100
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 204 DATA POSSIBLE = 273
Figure D-25a. Residual histogram for CO from the UAM simulation
results for Day 251-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 251-75
LJ
' SOD-DO ^ (- °
A50+OC1 -
300+00' - OS °
• c. -^ >•->•• M> ~3.^3Q
"5« *vO " «•«. JMH 3ST H
- 300*001 j-
- 450 -CO' -
-60C-DO' f
lECENO
FREO SYU
< i i> n
< 2. 2> C
< J. J> *
< *,99> »
000 200+001 400-rOQI 500-001 300+00! '00+002
OBSERVED CONCENTRATION
CARBON MONOXIDE
3ATA AVAILABLE = 204.
QATA POSSIBLE = 273
Figure D-25b.
Residual vs. observed plot for CO from the UAM
simulation results for Day 251-75.
204
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 251-75
250+000
S
CL
w -.500-001
20 .30 40 50 60 ~ 70
FREQUENCY
NITROGEN DIOXIDE
30 90 100
DATA AVAILABLE = 215 DATA POSSIBLE
273
Figure D-26a. Residual histogram for N02 from the UAM simulation
results for Day 251-75.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 251-75
5
Q.
Q.
RESIDUAL
^3U+UUU
200+000
.150+000
.100+000
.500-001
000
-500-001
-.100+000
-.150+000
-.200+000
- 250+000
000
.
•
0
fffl^0 ,
;
LEGEND
FPEQ STM
< I, !> D
< 2. 2> 0
< 3. 3> *
< 4.99> »
.500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 215 DATA POSSIBLE = 273
Figure D-26b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 251-75.
205
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 251-75
a.
a.
9
CO
J
600-001 •„
.450-001 - :
.300-001 \-'
. i
.150-001 -;
000 - *
.
-.150-001 . :
-.300-001 4
-450-001 •
-.500-001 •*
»
"- 750-001 L
.
lift"
til!!!?
t»»
10 15
OZONE
DATA AVAILABLE
20 25 30 35
FREQUENCY
40 45 50
238 DATA POSSIBLE = 273
Figure D-27a. Residual histogram for 03 from the UAM simulation
results for Day 251-75.
RESiDUAL VS OBSERVED CONCENTRATION
DAY 251-75
9
00
1 3U -UU 1 | . . II
1 if
600 -00' - 3d
450-00' - ',, S^c -_3
___ _ £J ^^8ff Q
3CO-00 h _ JHu JJ ^m
5Q"0°'pc 3Q 3™ a
- JOO-OOl L ~lyp a
- 450-CC1 j-
-6oo-:c'- "
7=Q 00' i ' ' ' - , i
t£C£NO
P9£0 SYM
< '. ! > 2
< 2. 2> C
< 3. 3> *
000 500-00' 'X-t-000 '50+000 200-000 250+000
OZONE
DA"A AVAILABLE
238
OA^A POSSIBLE = 273
Figure D-27b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 251-75.
206
-------
RESIDUAL ^ISTOGRAM DETERMINED OVER ALL ~M£S AND LOCATIONS
DAY 159-75
0.
Q.
^~'
SIDUAL
_j
a:
4CO 1-002 j
300+002
.200+002
100*002
000
- 100+002
- 200+002
-.300+002
- 400+002
- 500+002
•
•
*» 1 1 ' L 1 1
•
•
50 100 150 200 250 300 350' 400 450 500
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 259 DATA POSSIBLE = 273
Figure D-28a. Residual histogram for CO from the UAM simulation
results for Day 159-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 159-76
CL
'-'
RESIDUAL
.400+002
300+002
.200+002
.100+002
.000
-.100+002
- 200+002
- 300+002
-.400+002
- 500+002
.000
0
•
LECENO
FREO STM
< i. i> a
< 2. 2> ffl
.100+002 .200+002 .300+002 400+002 .500+002
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE "
DATA AVAILABLE = 259 DATA POSSIBLE = 273
Figure D-28b.
Residual vs. observed plot for CO from the UAM
simulation results for Day 159-76.
207
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL PMES AND LOCAPQNS
DAY 159-75
/ 3U — UU ' f
SOO-001 •
^ .450-001 -
a
"" 300-00! -
^ 150-001 -
Q 000
00
•£ -.150-001 L
r
.
:!$•*
' :: ill Unfit
its:: Iff !:l|H:::i:::::: ni: .» .
»:«:;«::«*«"*"•' ^ '
3«5?i*
r:.»
:j*.
-300-001 L*
t
-.450-001 1-
-.60C-001 •
" _ 750-001 -
••
i i i i
0 5 10 15 20 25 30 35 40 45 50
FREQUENCY
NITROGEN DIOXIDE
DATA AVAILABLE = 243 DATA POSSIBLE = 273
Figure D-29a. Residual histogram for N02 from the UAM simulation
results for Day 159-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 159-76
' 600-001 |- -j
"5 a _ a
£ 450-001 t- o °
Q.
"•" .300-001 •
a
" n a
mm u u o 3
3C ffi o _ a 30_
J .150-001 • Jgffi " 0 0 u a ^
on WBT^a' a*"0 %i
w _ 150-001 §
- 300-001 •
- 450-001 •
-.600-001 •
cT °
a
o a
a
LECENO
"^ STM
< i. i> 0
< 2. 2> O
< J. 3> *
< 4,99> «
000 .500-001 .100+000 .150-1-000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 243 DATA POSSIBLE = 273
Figure D-29b.
Residual vs. observed plot for N02 from the UAM
simulation results for Day 159-76.
208
-------
RESIDUAL HISTOGRAM DETERMINED OVER AL:_ TiMES AND LOCATIONS
OAY '59-75
2
c.
CL
^
<
Q
on
La
a:
.200+000 j-
150+000
.100+000
.500-001
000
- 500-001
- 100+000
- 150+0001-
H!f!!!!!f««..,
*^
*
»
- 200+000 1-
10 15
OZONE
DATA AVAILABLE
20 25 20 35
FREQUENCY
40 *5 50
234 DATA POSSIBLE
273
Figure D-30a. Residual histogram for 63 from the UAM simulation
results for Day 159-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 159-76
CL
CL
v*~ '
|
£
.200+000
150+000
.
.100+000 1-
i
.500 -C0 1
000
- 500-00!
' _ °ff ^POQ m
XpCDi pi? .gfip^n ^' ,^ i5»i^.'iij)fi'>ifc'i^' n "n
a a 30 _°a
D ° ~ °
1 00+000 ^
-.150+000
- 200+000
3
•
FREO SYM
< i. i> a
< 2. 2> O
< 3. J> *
000 500-001 .100+000 .150+000 200+000 250'+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 234 DATA POSSIBLE = 273
Figure D-30b. Residual vs. observed plot for 63 from the UAM
simulation results for Day 159-76.
209
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 160-76
.500+001
Q.
a
a
.400+001 •
.300+001 J
.200+001
.100+001
.000
-.100+001
- 200+001
- 300+001
-.400+001
-.500+001
10 20 JO 40 50 60 70
FREQUENCY
CARBON MONOXIDE
30 90 100
DATA AVAILABLE = 242 DATA POSSIBLE
273
Figure D-31a. Residual histogram for CO from the UAM simulation
results for Day 160-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 160-76
2
Q.
Q.
N— '
RESIDUAL
3UU+UUI
.400+001
.300+001
.200+001
.100+001
- 100+001
-.200+001
-.300+001
-.400+001
-500+001
000
.
-
«a tf f
.__*aiMtHjraWi 4. , i , , ,
5™^° °
•
•
LECtHO
FREO STM
< t. :> n
< 2, 2> 0
100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 242 DATA POSSIBLE = 273
Figure D-31b. Residual vs. observed plot for CO from the UAM
simulation results for Day 160-76.
210
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 160-76
Q.
Q^
_i
*<
RESIDUE
. / 3U — uu > r
.600-001 .
.450-001 -
.300-001 -
.150-001 -
.000
- 150-001 -
-300-001 •
-.450-001 •
-.600-001
-.750-001 L
»
|
: * *«••
• •' if «tt»S*t"t*m
' ". '\'. ii|*iiliijiiit|iiii ^**^**
• 9*?
*
10 15 20 25 20 35
FREQUENCY
NITROGEN DIOXIDE
40 45 50
DATA AVAILABLE = 223 DATA POSSIBLE = 273
Figure D-32a. Residual histogram for N02 from the UAM simulation
results for Day 160-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 160-76
Q.
Q.
g
00
LLJ
cc
,/3U— UUI
.600-001
.450-001
.300-001
150-001
000
- 150-001
-.300-001
-.450-001
-.600-001
- 750-001
QOO
a
•
a
Q
a
fl
d^3 ° ..-.,.'
^^uag9' ' .
3
.
.500-001 .1004-000 .150+000 200+000 .250+
IECENO
FREO sry
< t. i> a
< 2. 2> O
< 4.99> »
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 223 DATA POSSIBLE = 273
Figure D-32b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 160-76.
211
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 160-76
S
Q.
Q.
~
_i
Q
y5
LU
o:
Z3U+VUU
200+000
.150+000
.100+000
.500-001
000
-.500-001
- 100+000
-.150+000
-.200+000
- 250+000
.
•1
tJ?
! • npi! ji!||im!niHpiP:*;"
1 ' " 'i< Iiiziil««xi«»» ^
>"
-
•
• i i i . > i . i
0 5 10 15 20 25 JO 35 40 45 5
FREQUENCY
OZONE
DATA AVAILABLE = 229 DATA POSSIBLE =
273
Figure D-33a. Residual histogram for 03 from the UAM simulation
results for Day 160-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 160-76
2
Q.
0-^
on
UJ
or
.200+000
.150+000
.100+000
.500-001
000
-.500-001
- 100+000
- 150+000
-.200+000
-.250+000
,000
0 °
m TWfe rV^rSiTOii'ifri^ff'
^ [g . mp il7TUf*''r^T^1 TBTj i ^ ai oT^ i i
•
LEGEND
FREO SYM
< 1. t> O
< 2. 2> 0
< 3. J> A
< »,99> »
500-001 .100+000 .150+000 200+000 .250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 229 DATA POSSIBLE = 273
Figure D-33b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 160-76.
212
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 211-75
.200+001
^ 150+001
Q.
~~* .100+001
J .500+000
Qnno
.UWU
(J]
£ - 500+000
- 100+001
-.150+001
-.200+001
-.250+001
.
t
-»
ms
• " T"[ i jit iii ijiiitijiiit^*^ ******
•
10 15
40 *5 50
20 25 30 35
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 233 DATA POSSIBLE
273
Figure D-34a. Residual histogram for CO from the UAM simulation
results for day 211-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 211 -76
y^
2
Q.
Q.
*— '
_l
<
D
Q
c/5
Ld
o;
^3U+UU 1
.200+001
.150+001
.100+001
.500+000
QOO
.
O
a
a
nnoa
.^3 n a
fllKHCLw^"
m^SXMIMHni Q
i iMIII'HirUB , ,,iii
jBBB9tt9? a
< 2. 2> O
< 3, 3> *
< *.99> «
.000 .100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE -
DATA AVAILABLE = 233 DATA POSSIBLE = 273
Figure D-34b. Residual vs. observed plot for CO from the UAM
simulation results for Day 211-76.
213
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 21 1-76
./3U-UUI
500-001
X-K
OJ 450-001
Q.
^ ,300-001
J .150-001
Q .000
00
{£ -150-001
-300-001
-.450-001
-.600-001
-.750-001
*
.
XT
• •+*•
10 15 20 25 30 35
FREQUENCY
NITROGEN DIOXIDE
40 *5 50
DATA AVAILABLE = 218 DATA POSSIBLE = 273
Figure D-35a. Residual histogram for N02 from the UAM simulation
results for Day 211-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 21 1-76
S
Q.
RESIDUAL (F
. /ou— uu i
..SOO-001
.450-001
.300-001
150-001
000
- 150-001
-.450-001
-.600-001
- 750-001
.000
. . i i ' i .o
a
a
o
|!f&'s-1 r ' '
.
•
lECENO
FREO SYM
< i. i> a
<. 2. 2> O
< *]99> »
200-001 .400-001 .500-001 800-001 100-1-000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 218 DATA POSSIBLE = 273
Figure D-35b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 211-76.
214
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 211-75
a
a.
RESIDUAL
2001-000
.150+000
.100*000
.500-001
.000
-.500-00?
-.100+000
-.150+000
-.200-t-OOO
-.250+000
^ • , , ,
pPUl1' -
i
10 20 X *0 50 60 70
FREQUENCY
OZONE
30 90 100
DATA AVAILABLE = 225 DATA POSSIBLE
273
Figure D-36a. Residual histogram for 03 from the UAM simulation
results for Day 211-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 211-76
0_
Q^
RESIDUAL
200+000
150+000
.100+000
.500-001
-500-001
- 100+000
-.150+000
-.200+000
- 250+000
000
.
: ;
' J ^SW18 " °°°° i • ,
Q
a
LECCNO
FREQ STM
< i, i> a
< 2. 2> 0
< 3. 3> *
.500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 225 DATA POSSIBLE = 273
Figure D-36b.
Residual vs. observed plot for 0-
simulation results for Day 211-7(
215
from the UAM
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 212-75
g
a
on
.3UU+UUI
400+001
.300+001
.200+001
.
•
Mfr
.100+001 ff
^ ^_^
MHHB|^lMHHBH^HIM^^^^^^^^^^^^^^_
-.100+001
-.200+001
-.300+001
-.400+001
-500+001
r
•
•
1 i i i i i i i
0 10 20 30 40 50 60 70 30 90 10
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 2*6 DATA POSSIBLE = 273
Figure D-37a. Residual histogram for CO from the UAM simulation
results for Day 212-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 212-76
^^
Q.
Q^
RESIDUAL
3UU»UUI
•400+001
300+001
.200+001
.100+001
000
- 100+001
-.200+001
-.300+001
-.400^001
- 500+001
000
.
a
-^AjflrtSpi *"*,,,'
Q
•
•
LEGEND
FREO STM
< i. i> a
< 2. 2> 0
< 3, 3> *
< *,99> »
.100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 246 DATA POSSIBLE = 273
Figure D-375. Residual vs. observed plot for CO from the UAM
simulation results for Day 212-76.
216
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 212-76
400-001
3J .300-001
Q.
^ .200-001
-J .100-001
3
Q 000
(75
}£ -.100-001
-200-001
-.300-001
-.400-001
-.500-001
*>
*
-*
' 4
':"•
;!
•
-
.
••
M ^ •
M 2 [••«•
itll?«*; ii....
:[U:iu....
JS5J.
*•
10 15 20 25 30 35
FREQUENCY
NITROGEN DIOXIDE
40 45 50
DATA AVAILABLE = 234
Figure D-38a. Residual histogram for
results for Day 212-76.
DATA POSSIBLE = 273
from the UAM simulation
RESIDUAL VS OBSERVED CONCENTRATION
DAY 212-75
•s
^C
Q.
Q.
^"^
_i
Q
00
LJ
.3UU— UU'
.400-001
.300-001
.200-001
.100-001
.000
- 100-001
- 200-001
-.300-001
-.+00-001
- 500—001
a
a
a
o a _
9 0 ° 0
o ™ Q a a
TT .DO n o
j?"Sn Jw *
tjjBJIfT^ iTJL **? ^^
jftlf^jffft^ QDu n n $
•ffigBg™ °
.r?
.a
i i
UCENO
FREO SYM
< i. i> a
< 2. 2> 0
< 3. 3> *
000 200-001 .400-001 .600-001 300-001 .100-1-000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 234 DATA POSSIBLE = 273
Figure D-38b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 212-76.
217
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 212-76
•C"
?
Q.
s"""'
_l
Z!
9
or
Figure
a
^
_j
a
ui
'-U
CE
. / 350-001
-600-001
- 750-001
fpy,
\
1 1 1 1 1 1 I 1
LEGEND
FREQ STM
< i. 1> 0
< 2.2> 0
< 3. 3> •
< 4,99> «
QOO .500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 216 DATA POSSIBLE = 273
Figure D-39b. Residual vs. observed plot for 63 from the UAM
simulation results for Day 212-76.
218
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALi_ TIMES AND LOCATIONS
DAY 225-76
.250-1-001
.200*001
2
a
a.
g
Q
in
-.200+001
-.250+001
10
20 20
40 50 60 70
FREQUENCY
CARBON MONOXIDE
80 90
DATA AVAILABLE = 253 DATA POSSIBLE
273
Figure D-40a. Residual histogram for CO from the UAM simulation
results for Day 225-76.
CL
Q.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 225-76
.200+001
.150+001
.100+001
.500+000
000
-.500+000
-.100+001
-.150+001
-.200+001
- 250+001
000
.
. °
o no
« o a a
a
o
a
.500+000 .100+001 .150+001 200+001 250H
Q
CO
a —-' — fa a *" ] LECENO
FREO SYM
< i. i> a
< 2. 2> O
< 3, 3> »
< 4,99> «
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 253 DATA POSSIBLE = 273
Figure D-40b. Residual vs. observed plot for CO from the UAM
simulation results for Day 225-76.
219
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 225-76
s*
Q.
N— '
_i
(75
UJ
or
.400-001 •**
300-001 3*
.200-001 -a]
.100-001 -;;
000 "
- 1 GO-DO 1 -\':
-.200-001 :::
-.300-001 4
-.400-001 •*
-.500-001 1 —
,
•
: i " ;; «•**•
• : ;; ; ^{*««>i»I>tft
J: : lIl^MiHtis**** * '
| * *•»
>i '
1,1,111,
10 15 20 25 30 35
FREQUENCY
NITROGEN DIOXIDE
40
50
DATA AVAILABLE = 236 DATA POSSIBLE
273
Figure 0-41a. Residual histogram for N02 from the UAM simulation
results for Day 225-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 225-75
•s
a
a
g
a
(/5
£ - '""-^'l-ff*;* m „%„ amtt, u 1 LEGEND
FREQ SYM
< t. i> a
< 2. 2> O
< 3, 3> »
- 500-001
.000 .200-001 .400-001 .600-001 .300-001 .iOO-t-000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 236 DATA POSSIBLE = 273
Figure D-415. Residual vs. observed plot for N02 from the UAM
simulation results for Day 225-76.
220
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 225-76
2
a
a^
<
Q
vi
UJ
or
.l^UTUUVf
.200+000
.150+000
.100+000
.500-001
000
-.500-001
-.100+000
-.150+000
-.200+000
- 250+000
.
*•
: •$*•
§|fe||«:*|.
*i*-Si I it*! •» • • i J • S**"^* JT t • ' 1
**»«»«»•»** ' " ' '
•
-
•
-
1 1 1 , 1 , , 1 ,
10 15 20 25 JO
FREQUENCY
OZONE
DATA AVAILABLE = 209
40 45
50
DATA POSSIBLE = 273
Figure D-42a. Residual histogram for 03 from the UAM simulation
results for Day 225-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 225-76
2
a.
0.
_j
RESIDU/
..43U+UUU
.200+000
.150+000
.100+000
.500-001
000
-.500-001
- 100+000
-.150+000
-.200+000
-.250+000
.000
-
0=
t " CJW!^^
aajjjF^iS&S*!™^
LECCNO
FREO Srtt
< t. t> 0
< 2. 2> 0
.500-001 .100+000 .150+000 .200+000 .250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 209 DATA POSSIBLE = 273
Figure D-42b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 225-76.
221
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALi_ TIMES AND LOCATIONS
DAY 226-76
a
a^
_i
Q
.500+001
.450+001
.300+00)
.150+OOt
000
-.150+001
-.300+001
-.450+001
-.600+001
- 750+001
1
»
"Pill! | ( | '
r".' r '" ' ' .
•
•
10 20 20 40 50 60 70
FREQUENCY
CARBON MONOXIDE
30 90 100
DATA AVAILABLE = 207 DATA POSSIBLE
273
Figure D-43a. Residual histogram for CO from the UAM simulation
results for Day 226-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 226-75
2
a.
0.
^~ '
RESIDUAL
. / ^U *VW 1
.500+001
.450+001
.300+001
.150+001
000
-.150+001
-300+001
-.450+001
-.600+001
-.750+001
.000
a
&n Q
pP551' i i i . i . i ^
•
•
LEGEND
FREQ 5YM
< 1 1 > Q
< 2. 2> 0
< J. 3> A
.150+001 .300+001 .450*001 600+001 .750+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 207 DATA POSSIBLE = 273
Figure D-43b. Residual vs. observed plot for CO from the UAM
simulation results for Day 226-76.
222
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 226-76
. /^J—UVM 1-
600-001 -
^ 450-001 •
Q.
^ .300-001 -
^ .150-001 -
Q .000
CO
^ - 150-001 -
-.300-001 -
-.450-001 •
-.600-001 •
-.750-001 .
t
*++
. >. U A
1^: ,« ,,,,,"
i ii I Jflflpilil ""*"**•"**'
: •{•*•*•
:
*
-
10 15 20 25 JO 35
FREQUENCY
NITROGEN DIOXIDE
DATA AVAILABLE
40 45 50
233 DATA POSSIBLE
273
Figure D-44a. Residual histogram for N02 from the UAM simulation
results fo.r Day 226-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 226-75
2
CL
^
_1
n
en
LJ
.600-001
450-001
.300-001
.150-001
.000
- 150-001
- 300-001
-.450-001
-.600-001
-.750-001
.000
0-
a
a aa
ffl fl l*T*H PI
Q 153 ji^ IJT n fj
Q
< 2. 2> 0
< 3, 3> *
< *,99> 9
.200-001 .400-001 600-001 300-001 .100+000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 233 DATA POSSIBLE = 273
Figure D-44b. Residual vs. observed plot for N02 from the UAM
simulation results for Day 226-76.
223
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 226-76
.200+000
^ .150*000
Q.
^ .100*000
J .500-001
Q .000
£ -.500-001
- 100*000
-.150*000
-.200+000
-.250+000
.
{*
;; ijjjmmtM
'.'. "• i ' - 1' $$SI$$
jjpfcS.41*"
•
0 5 10 15 20 25 20 35 *0 *5 5
FREQUENCY
OZONE
DATA AVAILABLE = 1 78 DATA POSSIBLE =
273
Figure D-45a. Residual histogram for 03 from the UAM simulation
results for Day 226-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 226-76
RESIDUAL (PPM)
.200*000
.150*000
.100*000
500-001
000
-.500-001
-.100*000
-.150+000
-.200+000
- 250+000
t
a
a ° °
«8™f^|jp^a
a a rp£,ase ° ^ .^ ^°^'"H
Pr '= u ' '
;
•
LEGEND
FREO STM
< i, t> a
< 2. 2> 0
< 3, 3> *
000 .500-001 .100+000 .150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 178 DATA POSSIBLE = 273
Figure D-45b. Residual vs. observed plot for 03 from the UAM
simulation results for Day 226-76.
224
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 237-76
.500-1-001
10 20 30 *0 50 60 70 SO 90 100
FREQUENCY
CARBON MONOXIDE
DATA AVAILABLE = 234 DATA POSSIBLE = 273
Figure D-46a. Residual histogram for CO from the UAM simulation
results for Day 237-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 237-76
Q.
0^
3
RESIDU/
.JWU^WW 1
.400+001
.300+001
.200+001
.100+001
000
- 100+001
-.200+001
-.300+001
-.400+001
- 500+001
000
.
a
a
"C
.•3 gj ° °
. — JSOIj 3Q _
,J_flFJ?T^F 0
rrMMBfiPS^ 11 'II
D
•
LEGEND
FREQ SYM
< i, i> a
< 2, 2> 0
< J. J> *
.100+001 .200+001 .300+001 400+001 500+001
OBSERVED CONCENTRATION (PPM)
CARBON MONOXIDE
DATA AVAILABLE = 23* DATA POSSIBLE = 273
Figure D-46b. Residual vs. observed plot for CO from the UAM
simulation results for Day 237-76.
225
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 237-75
.100+000
3
a.
CL
I
Q
10 20 JO 40 50 60 70
FREQUENCY
NITROGEN DIOXIDE
ao 90 100
DATA AVAILABLE
212 DATA POSSIBLE » 273
Figure D-47a. Residual histogram for N02 from the DAM simulation
results for Day 237-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 237-76
2
a
a.
~
j
9
LU
.1UV+UUU
.800-001
600-001
.400-001
.200-001
000
-.200-001
-.400-001
-.600-001
-.800-001
0
•
a
% ^ a
. Jga _ o1 °
jdB&nSjp % a
jB^^rnffftp"*" g O
«%f a°
o a
a
a
- 100 1 000 ' L- ' ' ' * ' ' -
LCCCNO
F»EO ST«4
< i, i> a
< 2. 2> 0
< 3. 3> *
000 500-001 .100+000 150+000 200+000 250+000
OBSERVED CONCENTRATION (PPM)
NITROGEN DIOXIDE
DATA AVAILABLE = 212 DATA POSSIBLE = 273
Figure D-47b. Residual vs. observed plot for N02 from the DAM
simulation results for Day 237-76.
226
-------
RESIDUAL HISTOGRAM DETERMINED OVER ALL TIMES AND LOCATIONS
DAY 237-75
200+000
OJ .150+000
a
^ .100+000
5? .500-001
QGIY)
.WWU
in
g -.500-001
- 100+000
-.150+000
-.200+000
-.250+000
.
.
if mijj. .„.
" : ' i :: :: || :i illlll| UftSSS $***••**•
:]: |; lUtin**"*********"
. I •*
-
•
Figure D-48a.
10 15 20 25 30 35 40 45 50
FREQUENCY
OZONE
DATA AVAILABLE = 230 DATA POSSIBLE = 273
Residual histogram for 03 from the UAM simulation
results for Day 237-76.
RESIDUAL VS OBSERVED CONCENTRATION
DAY 237-76
a
a.
<;
a
a
LJ
C£.
.^au+uuu
.200+000
.150+000
.100+000
.500-001
000
-500-001
- 100+000
-.150+000
-.200+000
-.250+000
.
iw.-.jy*
Ttc/iflvjfif '
_a_12p!M6'rnlsj _ a
' • Lj^ujj iro^yTJ ^ ^
^H I'l ^ f PL ffy ' |H O
jSWyBJ? T3 ' ' ' ' Q ' ' '
5^ a
l£CENO
FREO STM
< i, i> a
< 2. 2> 0
< 1. 3> »
.500-001 .100+000 .150+000 200+000 250*000
OBSERVED CONCENTRATION (PPM)
OZONE
DATA AVAILABLE = 230 DATA POSSIBLE = 273
Figure D-48b. Residual vs. observed plot for 03 from the DAM
simulation results for Day 237-76.
227
-------
APPENDIX E
UAM - CONTOUR PLOTS FOR REMAINING TEST DAYS
The contour plots of predicted and observed 03 concentrations for the
hour of the observed concentration maximum are presented here for all test days
not specifically discussed within the body of this report. The Julian day
numbers are listed in Appendix D.
228
-------
ST. uauia JUNE 27.1975 IOBT 1791
^COICTCO CONCCNTtRTIQNS OF 03 I IN
HOUR 1400-1500
ST. L3UIS JUNE Z7.1975 1OflT 1781
OBSERVED CUNCENTHBTIONS or 03 KN
HOUR 1400-1SCO
(a)
(b)
Figure E-l. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 178-75.
140
ST. IOU1! JULT 1.1975 1OBT 1821
'•WEOICTED COMCENTKATIONS ar
-------
\ \
ST. LOUIS JULT 2,1975 I CRT
FUEOICTEO CONCENT*ST 1QN3 OF 03 I IN
HOUR 1000-1100
ST. U9UIS JULT 2.1975 I OPT 1831
OBSERVED CONCENTRflTlflNS OF 93 I IN
HOUR 1000-1100
la)
(b)
Figure E-3. Contours of (a) predicted and (b) observed fields
of 63 at hour of observed maximum on Day 183-75.
ST. LOUIS JULT 3.1975 IDBT 1841
PREQJCTEO CONCENTHBTtONS OF 03 I IN PPB I
HOUR 1300-1400
ia)
ST. LOUIS JULT 3.197S I OUT 1841
OBSERVED CONCENTRATIONS. OF 03 I IN I*f81
HOUR 1300-1400
(b)
Figure E-4. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 184-75.
230
-------
ST. LOUIS JULT 2S.197S IOST 2091
MEOICTEfl CONCCNTKRTIONS 3F S3 I IN '
HOUR 1100-1200
ST. LOUIS JULT 28.1975 I OUT 2091
OBSERVED CONCENTHBTIONS OF 83 UN
HOUR 1100-1200
(a)
(b)
Figure E-5. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 209-75.
ST. LBUIS BUG. 9.1975 I OPT 2311
ST. LOUIS RUG. 9.I97S I OUT 2211
OBSERVED CONCCNTRRTKJNS OF 03 I IN
HQUR 1 SCO-1600
(a)
(b)
Figure E-6. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 221-75.
231
-------
SI- LOUIS BUD. 18.1975 I OPT 2JO I
PREDICTED C ONCE NT*BT IONS 8F 03 UN PP9 I
(300-1400
(a)
ST. LOUIS AUG. 18.13-75 I OAT 2301
OBSERVED CONCENTRSTION5 OF 03 I IN
HOUR 1300-MOO
(b)
Figure E-7. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 230-75.
ST. LOUIS flUC. 19.1975 I CRT 2311
PREDICTED C ONCE NT «f»T IONS If 03 I IN PP8 I
HOUR 1400-1500
(a)
ST. LOUIS RUG. 19.191S I OflT 2311
OBSERVED CONCENTRPTIONS Of 93 I IN Pf 81
HOUR MOO-1SOO
(b)
Figure E-8. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 231-75.
232
-------
SI*. LOUIS SEPT. J.I 9^5 I OUT 2511
PREDICTED COHCENTtBTlOM Of 03 I IN PP8 I
HOUR 1200-1300
ST. LOUIS SEPT. 8.1375 I0«T 2511
OBSERVED CONCEHTMTIOM OF as UN PPSJ
HOUR 1200-1300
(a)
(b)
Figure E-9. Contours of (a) predicted and (b) observed fields
of 63 at hour of observed maximum on Day 251-75.
ST. LOUIS JUNE 7.1976 I DOT 1591
PREDICTED CONCENTHflTlONS OF 03 I IN PPflI
HOUR UOO-1SOQ
(a)
ST. LOUIS JUN6 7.1976 I OAT 1S9I
OBSERVED CONCENTHPT10NS, OC 33 UN PT8)
HOUR 1400-1500
(b)
Figure E-10. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 159-76.
233
-------
ST. LOUIS JUNE 9.1978 I OUT ISO I
PREDICTED CONCENTRATIONS OF 33 MM P
HOUR 1600-1700
ST. LflUtS JUNE 8.1975 (001 1801
OBSERVED CONCENtRSTlONS OF 03 I IN
HOUR isoo-noo
(a)
(b)
Figure E-ll.
Contours of (a) predicted and (b) observed fields
of 63 at hour of observed maximum on Day 160-76.
V
ST. LOUIS JULT 29.1978 I DOT 2111
PREDICTED CONCENT(ATIQMS OF 03 I IN PP8)
HOUR 1500-1800
ST. LOUIS JULT 29.1978 I OPT 2111
OBSERVED CONCCMTKHTIONS' OF 03 I IN
HOUR 1500-1600
(b)
Figure E-12. Contours of (a) predicted and (b) observed fields
of 63 at hour of observed maximum on Day 211-76.
234
-------
ST. LOUIS JULT 30.1978 I OPT 2121
PREDICTED CONCENTRATIONS If 03 I IN f
HOUR 1200-1300
(a)
ST. LOUIS JULT 30.1976 I OUT 212)
OBSERVED CONCENTRATIONS Of 93 I IN
HOUR 1200-1300
(b)
Figure E-13. Contours of (a) predicted and (b) observed fields
of 63 at hour of observed maximum on Day 212-76.
o
—«0.
ST. LOUIS SUO. 12.1976 I OAT 2251
PREDICTED CONCENT HAT IONS OF 03 UN PPS1
HOUR 1300-1400
ST. LOUIS BUG. 12.1975 IQAY 2251
OBSERVED CONCENTRATIONS OF 03 1 IN P
HOUR 1300-MOO
(a)
(b)
Figure E-14. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 225-76.
235
-------
ST. LOUIS BUO. U.197S I OUT 2281
PREDICTED CBNC£NT««T!ONS Of 33 I IN F
HOUR 1300-1400
ST. LOUIS CUC. 13.1976 I BUT 2261
385EHVEO CONCeNTRSTlONS Of 03 I IM
HOUR 1300-1400
(a)
(b)
Figure E-15. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 226-76.
ST. LOUIS BUO. 24.1975 IQPT 2371
PREDICTED CONC£NT»BT10NS OF 33 I IN PP8 I
HOUR 1100-1200
ST. LOUIS BUG- 24.197B IOPT 2371
OBSEKVEO C3NCENTHHTIONS af 33 I [N
HOUR 1100-1200
(a)
(b)
Figure £-16. Contours of (a) predicted and (b) observed fields
of 03 at hour of observed maximum on Day 237-76.
236
-------
APPENDIX F
UAM - SUMMARY OF STATISTICAL RESULTS
Three tables are presented in this appendix. Table F-l summarizes the
statistics for the mean quantities on each of the 20 days for the species
NO, N02, 03, OLE, NMHC and CO. Note that the observed OLE (olefin) value was
obtained by a temporally and spatially constant splitting factor from NMHC.
This factor was based on an analysis of the composite of volatile organic
emissions over the entire St. Louis area.
*
Table F-2 summarizes the analysis of trends data for the species N02,
03 and CO, and Table F-3 presents concentration maxima data for the same
species. The statistical parameters included in the 3 tables are the same as
those defined in Appendix B, except for the following:
(1) in addition to an overall correlation coefficient there are-also the
following:
(a) average spatial correlation (average over correlation coeffici-
ents for each hour of simulation calculated separately), and
(b) average temporal correlation (average over correlation coeffici-
ents for each location calculated separately);
(2) for concentration maxima data the following descriptions apply:
(a) peak observed concentration (along with the location and time of
occurrence), and
(b) peak predicted concentration (at same location as peak observed
concentration, along with the time of occurrence).
237
-------
TABLE F-l. SUMMARY OF MODEL PERFORMANCE STATISTICS FOR MEAN CONCENTRATIONS FROM THE UAM.
Jul ian
date
142-75
178-75
182-75
183-75
Mean observed
Species concentration
NO
NO?
03
OLE
NMHC
CO
NO
N02
03
OLE
NMHC
CO
NO
N02
03
OLE
NMHC
CO
NO
N02
03
OLE
NMHC
CO
(pom)
0.0086
0.0134
0.0635
0.0080
0.183
0.481
0.0088
0.0212
0.0920
0.0147
0.334
0.809
0.008?
0.0172
0.0737
0.0118
0.269
0.517
0.0105
0.0227
0.0809
0.0126
0.285
0.655
Mean predicted Mean
concentration
(ppm)
0.0064
0.0181
0.0779
0.0044
0.258
0.479
0.0070
0.0259
0.0850
0.0062
0.384
0.723
0.0029
0.0128
0.0777
0.0041
0.202
0.307
0.0055
0.0224
0.0855
0.0059
0.298
0.466
residual
(ppm)
0.0022
-0.0047
-0.0145
0.0036
-0.0754
0.0016
0.0018
-0.0047
0.0070
0.0085
-0.0494
0.0860
' 0.0052
0.0045
-0.0040
0.0077
0.0669
0.210
0.0051
0.0004
-0.0046
0.0067
-0.0126
0.189
Mean residual Error
RMS error Mean obs. cone. banda
(ppm)
0.0094
0.0111
0.0317
0.0093
0.202
0.416
0.0116
0.0168
0.0314
0.0113
0.265
0.537
0.0105
0.0113
0.0186
0.0100
0.208
0.521
0.0144
0.0162
0.0248
0.0112
0.239
0.559
!a\
\*>l
25.6
-35.1
-22.8
45.0
-41.2
0.3
20.5
-22.2
7.6
57.8
-14.5
10.6
63.4
26.2
-5.4
65.3
24.9
40.6
48.6
1.8
0.3
53.1
-4.4
28.9
(%)
19.7
30.8
42.1
17.0
25.0
33.9
15.0
28.6
46.3
13.2
22.5
31.9
11.0
28.6
' 69.5
87.0
22.5
27.1
17.5
32.1
54.0
7.8
28.7
24.4
238
-------
TABLE F-l. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FOR MEAN CONCENTRATIONS FROM THE (JAM.
Julian
date
184-75
207-75
209-75
221-75
Species
NO
N02
03
OLE
NMHC
CO
NO
NO?
°3
OLE
NMHC
CO
NO
NO?
03
OLE
NMHC
CO
NO
N02
03
OLE
NMHC
CO
Mean observed
concentration
(ppm)
0.0118
0.0254
0.0814
0.0253
0.575
0.859
0.0070
0.0134
0.0802
0.0113
0.258
0.378
0.0077
0.0149
0.0698
0.0095
0.216
0.486
0.0097
0.0191
0.0703
0.0138
0.303
0.755
Mean predicted
concentration
(ppm)
0.0044
0.0212
0.0806
0.0131
0.470
0.612
0.0060
0.0182
0.0656
0.0063
0.233
0.377
0.0070
0.0190
0.0616
0.0056
0.289
0.442
0.0073
0.0185
0.0523
0.0063
0.221
0.476
Mean
residual
(ppm)
0.0074
0.0042
0.0008
0.0122
0.106
0.247
0.0010
-0.0048
0.0146
0.0050-
0.0246
0.0002
0.0006
-0.0042
0.0082
0.0040
-0.0736
0.0440
0.0024
0.0006
0.0180
0.0075
0.0919
0.279
RM.S error
(ppm)
0.0179
0.0220
0.0319
0.0208
0.412
0.671
0.0077
0.0079
0.0209
0.0149
0.341
0.331
0.0104
0.0095
0.0265
0.0090
0.242
0.986
0.0153
0.0127
0.0274
0.0128
0.294
0.831
Mean residual
Mean obs. cone.
(*)
62.7
16.5
1.0
48.2
18.4
28.8
14.3
-35.8
18.2
44.2
9.5
0.1
7.8
-28.2
11.7
42.1
-34.1
9.1
24.7
3.1
25.6
54.3
30.3
37.0
Error
band*
(*)
16.5
28.9
48.9
13.6
24.5
25.6
24.2
29.9
51.9
17.9
24.4
35.4
14.7
31.6
50.9
21.4
17.6
39.6
15.9
30.6
38.5
10.5
21.6
30.2
239
-------
TABLE F-l. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FOR MEAN CONCENTRATIONS FROM THE UAM.
Jul ian
date
230-75
231-75
251-75
159-76
Species
NO
NO?
03
OLE
NMHC
CO
NO
N02
°3
OLE
NMHC
CO
NO
NO?
03
OLE
NMHC
CO
NO
NO?
03
OLE
NMHC
CO
Mean observed
concentration
(ppm)
0.0130
0.0188
0.0449
0.0146
0.332
0.550
0.0084
0.0165
0.0518
0.0125
0.284
0.437
0.0198
0.0253
0.0507
0.0193
0.438
0.806
0.0163
0.0275
0.0880
0.0182
0.414
0.968
Mean predicted
concentration
(ppm)
0.0100
0.0215
0.0531
0.0065
0.317
0.468
0.0089
0.0194
0.0530
0.0061
0.337
0.456
0.0141
0.0279
0.0447
0.0087
0.449
0.648
0.0084
0.0230
0.0996
0.0095
0.401
0.693
Mean
residual
(ppm)
0.0030
-0.0026
-0.0086
0.0081
0.0148
0.0821
-0.0004
-0.0029
-0.0012
0.0064
-0.0535
-0.0190
0.0058
-0.0026
0.0061
0.0106
-0.0111
0.158
0.0079
0.0045
-0.0116
0.0088
0.0128
0.275
RMS error
(ppm)
0,0165
0.0091
0.0209
0.0117
0.243
0.420
0.0095
0.0087
0.0221
0.0101
0.209
0.355
0.0307
0.0190
0.0214
0.0166
0.349
0.777
0.0215
0.0130
0.0270
0.0149
0.303
2.46
Mean residual
Mean obs. cone.
(%)
23.1
-13.8
-19.2
55.5
4.5
14.9
-4.8
-17.6
-2.3
51.2
-18.8
-4.3
29.3
-10.3
12.0
54.9
-2.5
19.6
48.5
16.4
-13.2
48.4
3.1
28.4
Error
band3
(%)
18.3
38.5
29.9
17.5
25.9
32.4
16.0
37.1
45.5
14.0
15.4
33.7
19.0
39.1
38.2
13.3
22.2
27.9
17.5
35.8
53.0
15.2
31.5
35.1
240
-------
TABLE F-l. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FOR MEAN CONCENTRATIONS FROM THE UAM.
Jul fan
date
160-76
195-76
211-76
212-76
Species
NO
NO?
03
OLE
NMHC
CO
NO
N02
°3
OLE
NMHC
CO
NO
N02
°3
OLE
NMHC
CO
NO
NO?
03
OLE
NMHC
CO
Mean observed
concentration
(ppm)
0.0091
0.0198
- 0.107
0.0130
0.297
0.606
0.00*2
0.0159
0.0827
0.0093
0.212
0.360
0.0074
0.0208
0.0462
0.0127
0.288
0.472
0.0102
0.0181
0.0664
0.0155
0.352
0.572
Mean predicted
concentration
(ppm)
0.0063
0.0207
0.0890
0.0065
0.325
0.504
0.0038
0.0154
0.0738
0.0033
0.237
0.396
0.0034
0.0166
0.0584
0.0061
0.313
0.473
0.0045
0.0186
0.0665
0.0062
0.295
0.429
Mean
residual
(ppm)
0.0028
-0.0009
0.0176
0.0066
-0.0286
0.102
0.0005
0.0005
0.0089
0.0061
-0.0251
-0.0357
0.0040
0.0043
-0.0122
0.0066
-0.0252
-0.0019
0.0057
-0.0005
-0.0001
0.0093
0.0573
0.143
RMS error
(ppm)
0.0147
0.0132
0.0237
0.0126
0.276
0.440
0.0041
0.0105
0.0179
0.0086
0.179
0.400
0.0090
0.0112
0.0214
0.0118
0.245
0.342
0.0137
0.0109
0.0219
0.0129
0.283
0.397
Mean residual
Mean obs. cone.
(*)
30.8
-4.5
16.4
50.8
-9.6
16.8
11.9
3.1
10.8
65.6
-11.8
-9.9
54.1
20.7
-26.4
52.0
-8.8
-0.4
55.9
-2.8
-0.2
60.0
16.3
25.0
Error
band*
(*)
15.3
28.7
59.0
14.7
22.6
30.2
18.0
41.1
75.0
13.7
31.1
27.1
11.6
38.1
38.7
14.3
30.2
24.5
17.6
31.2
56.9
13.1
30.4
27.6
241
-------
TABLE F-l. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FOR MEAN CONCENTRATIONS FROM THE UAM.
Jul ian
date
225-76
226-76
237-76
275-76
Species
NO
NO?
03
OLE
NMHC
CO
NO
NOa
03
OLE
NMHC
CO
NO
NOe
03
OLE
NMHC
CO
NO
NO?
03
OLE
NMHC
CO
Mean observed
concentration
(ppm) '
0.0065
0.0208
0.0625
0.0116
0.264
* 0.479
0.0088
0.0236
0.0806
0.0147
0.324
0.693
0.0115
0.0192
0.0813
0.0225
0.511
0.649
0.0364
0.0507
0.0782
0.0304
0.691
1.74
Mean predicted
concentration
(ppm)
0.0044
0.0221
0.0391
0.0058
0.267
0.427
0.0056
0.0227
0.0545
0.0051
0.272
0.450
0.0052
0.0189
0.0538
0.0068
0.284
0.414
0.0264
0.0519
0.0878
0.0152
0.892
1.44
Mejm
residual
(ppm)
0.0021
-0.0013
0.0235
0.0058
-0.0030
0.0521
0.0031
0.0010
0.0261
0.0096
0.0527
0.244
0.0064
0.0003
0.0275
0.0157
0.226
0.236
0.0100
-0.0012
-0.0096
0.0152
-0.201
0.295
RMS error
(ppm)
0.0076
0.0124
0.0280
0.0110
0.257
0.343
0.0144
0.0155
0.0397
0.0142
0.285
0.570
0.0170
0.0140
0.0355
0.0212
0.448
0.562
0.0432
0.0319
0.0392
0.0197
0.394
1.35
Mean residual
Mean obs. cone.
(*)
32.3
-6.3
37.6
50.0
-1.1
10.9
35.2
4.2
32.4
65.3
16.3
35.2
55.7
1.6
33.8
69.3
44.2
36.4
27.5
-2.4
-12.2
50.0
-29.1
17.0
Error
band3
U)
22.4
35.2
26.8
18.9
23.7
32.4
18.5
27.9
20.2
15.4
33.5
25.1
14.6
30.2
20.4
7.8
24.0
32.1
13.5
35.1
36.7
11.8
25.1
27.2
aPercent of predictions within 25% of observations.
242
-------
TABLE F-2. SUMMARY OF MODEL PERFORMANCE STATISTICS FOR ANALYSIS OF TRENDS DATA FROM THE UAM
Julian
date
142-75
178-75
182-75
183-75
184-75
207-75
209-75
221-75
Species
N02
03
CO
N02
°3
CO
N02
°3
CO
N02
°3
CO
NOa
03
CO
N02
°3
CO
N02
03
CO
N02
°3
CO
Correlation
coefficient
•0.621
0.805
0.488
0.478
0.864
0.581
0.771
0.875
0.559
0.685
0.864
0.685 -
0.683
0.726
0.770
0.791
0.927
0.669
0.741
0.759
0.185
0.575
0.795
0.415
Average
spatial
correlation
0.689
0.174
0.633
0.598
0.357
0.735
0.518
0.571
0.395
0.466
0.435
0.527
0.549
0.444
0.583
0.734
0.461
0.533
0.536
0.411
0.524
0.612
0.395
0.321
Average
temporal
correlation
0.489
0.850
0.374
0.388
0.913
0.299
0.779
0.897
0.432
0.559
0.879
0-.479
0.521
0.696
0.435
0.682
0.959
0.528
0.682
0.670
0.650
0.615
0.818
0.308
Slope
0.665
0.887
0.246
0.447
0.867
0.275
0.523
0.610
0.172
0.574
0.614
0.326-
0.397
0.470
0.403
0.751
0.654
0.340
1.01
0.570
0.063
0.493
0.516
0.145
Intercept
(ppm)
0.0014
-0.0057
0.119
0.0096
0.0183
0.610
0.0105
0.0263
0.464
0.0099
0.0284
0.503
0.0170
0.0436
0.612
-0.0002
0.0373
0.249
-0.0044
0.0347
0.458
0.0100
0.0433
0.686
Sum of
squared error
(ppm2)
0.0183
0.186
5.330
0.0498
0.186
60.4
0.0380
0.142
75.4
0.0601
0.193
94.1
0.157
0.317
154.
0.0139
0.173
35.8
0.0209
0.192
224.
0.0419 '
0.268
177.
243
-------
TABLE F-2. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FOR ANALYSIS OF TRENDS DATA FROM THE UAM
Julian
date
230-75
231-75
251-75
159-76
160-76
195-76
211-76
212-76
Species
N02
°3
CO
N02
°3
CO
N02
°3
CO
N02
03
CO
N02
03
CO
N02
°3
CO
N02
°3
CO
N02
03
CO
Correlation
coefficient
0.824
0.819
0.764
0.772
0.814
0.685
0.600
0.821
0.630
0.887
0.859
0.294
' 0.796
0.937
0.628
0.660
0.947
0.286
0.621
0.833
0.511
0.798
0.912
0.759
Average
spatial
correlation
0.804
0.361
0.694
0.706
0.436
0.694
0.410
0.451
0.616
0.676
0.392
0.556
0.747
0.592
0.368
0.654
0.495
0.485
0.685
0.421
0.519
0.642
0.614
0.639
Average
temporal
correlation
0.606
0.885
0.698
0.552
0.804
0.719
0.576
0.864
0.571
0.732
0.929
0.652
0.637
0.961
0.529
0.321
0.964
0.282
0.373
0.715
0.304
0.538
0.742 -
0.546
Slope
0.774
0.729
0.419
0.790
0.645
0.483
0.585
0.618
0.359
0.740
0.862
0.064
0.535
0.676
0.340
0.457
0.720
0.113
0.415
0.773
0.274
' 0.503
0.632
0.439
Intercept
(ppm)
0.0022
0.0059
0.354
0.0012
0.0176
0.217
0.0090
0.0232
0.574
0.0105
0.0021
0.923
0.0088
0.0464
0.435
0.0089
0.0296
0.315
0.0140
0.0010
0.342"
0.0088
0;0244
0.383
Sum of
squared error
( ppm2 )
0.0188
0.0950
58.6
0.0165
0.152
31.5
0.0635
0.150
159.
0.0546
o;i58
1700.
0.0565
0.247
61.5
0.0335
0.112
40.4
0.0322
0.104
32:i
0.0424
0.189
57.8
244
-------
TABLE F-2. (Continued)
SUMMARY OF MODEL PERFORMANCE STATISTICS FOR ANALYSIS OF TRENDS DATA FROM THE UAM
Jul ian
date
225-76
226-76
237-76
275-76
Species
N02
03
CO
N02
CO
NOg
03
CO
N02
CO
Correlation
coefficient
0.704
0.753
0.540
0.561
0.796
0.607
0.799
0.687
0.734
0.597 -
0.844
0.533
Average
spatial
correlation
0.657
0.377
0.573
0.688
0.110
0.707
0.793
0.354
0.561
0.761
0.239
0.622
Average
temporal
correlation
0.578
0.780
0.588
0.368
0.605
0.516
0.648
0.759
0.640
0.329
0.916
0.410
Slope
0.657
0.310
0.396
0.331
0.426
0.170
0.669
0.432
0.279
0.604
0.896
0.219
Intercept
(ppm)
Q.0063
0.0504
0.310
0.0161
0.0574
0.617
0.0065
0.0581
0.534
0.0193
-0.0004
1.42
Sum of
squared error
0.0333
0.247
30.8
0.0670
0.445
86.0
0.0476
0.383
108.
0.183
0.340
572~.
245
-------
TABLE F-3. SUMMARY OF MODEL PERFORMANCE STATISTICS FOR CONCENTRATION MAXIMA FROM THE UAM
Julian
date Specie:
142-75
178-75
182-75
183-75
184-75
207-75
209-75
221-75
N02
°3
CO
NQ2
°3
CO
N02
03
CO
NOa
°3
CO
M02
°3
CO
N02
03
CO
M02
03
CO
N02
°3
CO
Observed
Peak Concentration Predicted Peak
2 Value Location
(ppm) (RAPS site)
0.0458
0.195
4.03
0.0701
0.202
4.04
0.0867
0.142
4.19
0.106
0.171
3.81
0.180
0.184
5.26.
0.0621
0.185
4.26
0.0501
0.209
14.4
0.0726
0.166
5.55
102
101
121
103
112
107
106
121
115
106
119
112
104
118
104
106
113
104
110
118
117
102
121
107
Time
(CST)
1300
1200
1700
0700
1400
0600
0800
1100
1400
0700
1000
0700
0700
1300
0700
0600
1400
0500
0800
1100
1700
0800
1500
0700
Value
(ppm)
0.0654
0.144
0.667
0.0517
0.162
1.22
0.0659
0.103
0.196
0.0955
. 0.154
1.39
0.0795
0.133
2.05
0.0530
0.141
1.63
0.0589
0.128
1.23
0.0478
0.144
1.44
Concentration*
Time
(CST)
1300
1400
1500
0900
1500
1400
0600
1400
1700
0800
1000 -
0700
0600
1400
0500
0600
1400
0500
0900
1100
0800
0500
1400
•0500
Residual
(ppm)
-0.0196
0.0460
3.36
0.0184
0.0400
2.82
0.0208
0.0390
3.99
0.0105
0.0170
2.42
0.101
0.0510
3.21
0.0091
0.0440
2.63
-0.0088
0.0810
13.2
0.0248
0.0220
4.11
Residual
Obs. cone.
m
-42.8
23.6
83.4
26.2
19.8
69.8
24.0
27.5
95.2
9.9
9.9
53.5
56.1
27.7
61.0
14.7
23.8
61.7
-17.6
38.8
91.7
34.2
13.3
74.1
246
------- |