-/EPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-83-003e
July 1987
Air
Evaluation of
Rural Air Quality
Simulation Models
Addendum E:
Graphic Summary of
the Performance of
Rural Air Quality
Models
-------
EPA-450/4-83-003e
Evaluation of
Rural Air Quality Simulation Models
Addendum E: Graphic Summary of the
Performance of Rural Air Quality Models
By
William M. Cox
Gerald K. Moss
Joseph A. Tikvart
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park NC 27711
July 1987
-------
This report has been reviewed by the Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
EPA-450/4-83-003e
-------
PREFACE
EPA is conducting an extensive program to evaluate the performance of
air quality simulation models using statistical comparisons between measured
and predicted concentrations. The operational performance of several rural
models has been documented in tabular* and graphical form (Addendum B) for
the Clifty Creek power plant and in tabular form (Addendum A, C and D) for
three other large midwestern power plants. The purpose of this addendum is
to summarize the performance of the rural models in a graphical format us-
ing the accumulated information presented earlier for the four plants. The
composite behavior of each model is examined from both an operational and
scientific perspective with particular emphasis on comparison of differences
in performance among the models.
*Londergan, R. J., D. H. Minott, D. J. Wackter, T. Kincaid and D. Bonitata,
1982. Evaluation of Rural Air Quality Simulation Models. EPA Publication
No. EPA-450/4-83-003. U. S. Environmental Protection Agency, Research Triangle
Park, N. C. (NTIS No. PB 83-182758).
-------
iv
-------
TABLE OF CONTENTS
Page
PREFACE iii
FIGURES vi
1. INTRODUCTION 1
2. ANALYSIS OF GRAPHIC SUMMARIES 3
3. SUMMARY AND CONCLUSIONS 15
REFERENCES 17
APPENDICES
INTRODUCTION TO THE APPENDICES
A. APPENDIX AKINCAID POWER PLANT
B. APPENDIX B--PARADISE POWER PLANT
' C. APPENDIX C--MUSKINGUM RIVER POWER PLANT
-------
FIGURES
Number Page
1 Fractional Bias By Model Using High 25 Values For
Each Data Base 9
2 Composite Fractional Bias By Averaging Period Using
High 25 Values 10
3 Composite Net Fractional Bias Using High 25 Values 11
4 Composite Fractional Bias By Model Using High 25 Values
For Each Stability Class 12
5 Composite Fractional Bias By Model Using High 25 Values
For Each Wind Speed Class 13
VI
-------
Section 1
INTRODUCTION
EPA has an extensive ongoing program to evaluate the performance of
air quality simulation models used for regulatory purposes. The perform-
ance of several rural models has been documented through extensive tabula-
tions of performance statistics using field data collected around four rural
power plants (1,2,3,4). The purpose of this report is to present graphically
the composite performance of four rural models using the performance data
assembled previously. At each site, SOg is the pollutant used for evaluation
purposes.
The four rural models evaluted are (1) MPTER developed by EPA (2) MPSDM
developed by ERT, (3) PPSP developed by the Martin Marietta Company, and
(4) TEM-8A developed by the Texas Air Control Board. This analysis is
confined to these rural models, since they essentially span the range of
technology represented by the rural models as a group. The data base con-
sists of two years assembled for Clifty Creek (1975 and 1976), two years at
Muskingum River (1975 and 1976), one year at Paradise (1976), and essentially
one year of data collected at Kincaid (1980 and portions of 1981).
The body of the report focuses on the composite operational performance
of each model with emphasis on differences among the models. The Appendices
provide a more in-depth analysis of the diagnostic performance components
of the four models for three of the data bases. The graphical format used
in the Appendices is similar to that used previously in which the performance
of these models was presented graphically using the data assembled for the
fourth data base at Clifty Creek (5).
-------
-------
SECTION 2
ANALYSIS OF GRAPHIC SUMMARIES
The graphic summaries of model performance (Figures 1-5) were developed
from information presented in the Appendices and from an earlier report in
which a comprehensive set of graphics were presented for Clifty Creek (5).
When possible, the summary statistic is a composite average of the same
statistic computed for each data base. In some instances, (e.g. individual
stations) composite results among data bases are not meaningful and hence,
only general trends are noted.
The performance of the models is displayed using the fractional bias
statistic in which model bias is normalized. The general expression for the
fractional bias is given by:
FB = 2* (OB - PR)/(OB + PR)
The fractional bias of the average is computed using this equation where OB
and PR refer to the averages of the observed and predicted values. The same
expression is used to calculate the fractional bias of the standard deviation
where OB then refers to the standard deviation of the observed values and PR
refers to the standard deviation of the predicted values. The fractional bias
was selected as the basic measure of performance because it has two desire-
able features. First, the fractional bias is symmetrical and bounded. Values
for the fractional bias range between -2.0 (extreme overprediction) and +2.0
(extreme underprediction). Second, the fractional bias is a dimensionless
number which is convenient for comparing the results from studies involving
different concentration levels or even different pollutants.
-------
The first figure displays four plots corresponding to the four models
being evaluated. Only the outcomes for 1-hour averages are shown; however,
the results using 3-hour and 24-hour averages are reasonably similar. In
this figure, the fractional bias of the average is plotted vs the fractional
bias of the standard deviation using only the 25 highest observed and pre-
dicted concentrations, unpaired in space and/or time. For a given model,
six points are plotted, one point for each of the six data bases used in the
evaluation. The placement of the points relative to the center of the plot
(0,0) is a measure of the bias of a particular model. Variation among the
data bases is indicated by the degree of spread of the points on each plot.
For reference purposes, a "box" is included near the center of each panel
which corresponds to performance that is within a factor-of-two for both the
average and the standard deviation. Values of the fractional bias (average
or standard deviation) that are less than -0.67 overpredict by more than a
factor-of-two while values that exceed +0.67 underpredict by more than a
factor-of-two. Values that fall entirely within the box correspond with
performance that is within a factor-of-two for both performance measures.
For the MPTER model (upper left panel of Figure 1), the bias of both
the average and the standard deviation is within a factor-of-two for each
of the data bases except one. For Muskingum River (1976) the standard
deviation of predicted values is larger than the standard deviation of ob-
served values by over a factor of two (fractional bias less than -U.67).
For MPSDM, there is a greater tendency for overprediction than evidenced
for MPTER. For Clifty Creek (1976) both the bias of the average and standard
deviation are overpredicted by a factor greater than two. At Clifty Creek,
for 1975, only the average is overpredicted by a factor greater than two
-------
while at Muskingum River, for 1976, only the standard deviation is overpre-
dicted by a factor greater than two. For the other three data bases,
(Muskingum 1975, Paradise, and Kincaid), both measures of bias are within a
factor-of-two.
For PPSP, large overpredictions are evident for each of the six data
bases. In each case, both the average and standard deviation of the high
25 values are overpredicted by a factor that exceeds two.
Among the four models, TEM-8A appears to have the smallest average bias
as indicated by the closeness of the points to the vertical axis. This con-
trasts with the bias of the standard deviation, which TEM-8A overpredicts
significantly at Clifty Creek (1976) and Muskingum River (1976) and under-
predicts significantly (fractional bias greater than +0.67) at Kincaid.
Figure 2 presents the composite performance of the models in which the
mean fractional biases of the six data bases are displayed. Separate panels
are shown corresponding to 1-hour, 3-hour and 24-hour averages. For 1-hour
averages (first panel), TEM-8A exhibits the least average fractional bias
followed closely by MPTER which is also relatively unbiased. MPSDM shows a
tendency for modest overpredictions while PPSP overpredicts significantly
for both the average and the standard deviation. The value plotted for
each model in this panel is positioned at the centroid of the values shown
on Figure 1 which displayed the performance of each data set for 1-hour
averages. As the averaging period increases to 3-hours (top right panel)
and to 24-hours (lower panel), there is a noticeable tendency for the points
to shift upward and to the right indicating smaller overpredictions (PPSP
and MPSDM) and larger underpredictions (MPTER and TEM-8A). For each of the
-------
three averaging periods, PPSP seems to be clearly outperformed by the other
three models.
Figure 3 redisplays the information shown in Figure 2 in a different
format in which only the three better performing models (MPTER, MPSDM and
TEM-8A) are compared. A composite measure of model bias is shown in the
form of "buildings" where each building corresponds to one of the three
averaging periods across the front axis and one of the three models across
the receding axis. The measure of model performance used is the "net frac-
tional bias", which is computed as the square root of the sum of squares of
the two measures of bias, i.e., the bias of the average and the bias of the
standard deviation. The composite measure of bias, which is the height of
each building, is just the net fractional bias averaged across the six data
bases.
Two features appear worth noting from this plot. One, there is a notice-
able contrast in performance with averaging period. For MPSDM, the composite
bias decreases with increasing averaging period while the opposite trend is
apparent for TEM-8A. Secondly, MPTER appears to perform marginally better on
average than the other two models.
The last two figures display information similar to that shown in
Figure 1 except the biases for selected meteorological categories are con-
trasted.
The data consists of the 25 highest observed and 25 highest predicted
values within each of four selected stability categories (Figure 4) and with-
in each of three selected wind speed categories (Figure 5). The stability
categories were selected as: very unstable (class A or B); unstable (class C);
neutral (class D); and stable (class E, F, or G). The wind speed categories
-------
were selected as: low wind speeds (less than 2.5 m/sec); medium (2.5 to
5 m/sec); and high (greater than 5 m/sec).
In Figure 4, the composite performance for each of the four stability
categories is shown for each of the four models. The first panel shows a
tendency for MPTER to underpredict the more stable conditions and to over-
predict the more unstable conditions. For MPSDM, (top right panel), the
spread between stability categories is more apparent; stable conditions
are associated with very large underpredictions while neutral, unstable and
very unstable conditions are associated with fairly large overpredictions,
on average. For PPSP, the composite performance indicates large overpredic-
tions for each stability category. For TEM-8A, the unstable categories are
associated with very low bias. However, as the stability class moves from
unstable to neutral and stable, an increasing tendency for underprediction
becomes obvious.
Figure 5 shows the bias of the four models for three wind speed cate-
gories. For PPSP and MPSDM, there is a clear tendency for low winds to be
associated with large overpredictions and high winds to be associated with
small overpredictions or small underpredictions. For MPTER, no trend is ap-
parent with wind speed class; only the medium winds show any tendency for
bias where slight (average) overpredictions are evident. For TEM-8A, the
high winds are associated with underpredictions. While the two lower wind
speed classes exhibit no noticeable bias. Although composite results are
not meaningful for individual stations, there were distinct patterns in the
bias of the models as a function of station to source distance (Figures 11-16
in each of the three Appendices and the analogous figures in the Clifty Creek
-------
stations accompanied by relatively large over and underpredictions at the
stations nearest the source. For very unstable conditions, the models as a
group displayed a tendency for decreasing overprediction as distance between
source and receptor increases. For stable conditions, the pattern appeared
to reverse (except PPSP), i.e., the models tended to underpredict at most
of the stations but the degree of underpredictiqn decreases as the source
distance increases.
The fractional bias of the models as a function of hour of the day
revealed very distinct patterns that were persistent among the six data
bases (See Figures 17-20 in the Appendices and the corresponding figures
in the Clifty Creek report). The typical diurnal pattern consisted of
marked underpredictions during the early morning and late evening hours
and moderate overpredictions during the midday hours.
-------
FR/>
10
u
p
Q
0-0 A
0-14
-U
-10
-2
^CTIONAL BIAS HIGH 25 CONCENTRATIONS
MODEL=MPTER AVERAGING PERIOD" 1
4 -14 -14
FRA
10
§ .4
1 "
0-0.5
a -14
-IS
-10
-2
+ *
X
O
A
-0.5 04 OS
BIAS OF AVERAGE
DATA BASES
+- CC76
X- CC76
D- MK76
0 - MK76
*- PARADISE
14 1.5 10 A" KINCAID
CTIONAL BIAS HIGH 25 CONCENTRATIONS
MODEL-PPSP AVERAGING PERIOD" 1
A ^
X
4 -14 -14
-
-0.5 0.0 0.5
BIAS OF AVERAGE
DATA BASES
+ - CC75
X- CC76
D- UK7S
0 » UK76
* - PARADISE
14 1.5 10 A- KINCAID
FRA
10
14
I .4
§ "
a
I "
0-A5
5-14
-1.5
-24
-I
CTIONAL BIAS HIGH 25 CONCENTRATIONS
MODEL=MPSDM AVERAGING PERIOD"!
X
4 -1.5 -14
A
*a
o
-OS 04 0.5
BIAS OF AVERAGE
DATA BASES
+ - CC75
X- CC76
D- UK7B
0 - UK76
* - PARADISE
14 14 24 A- KINCAID
FRA
10
o "»
i «
i-
O-OJ
a -1.0
-\s
-24
-2
CTIONAL BIAS HIGH 25 CONCENTRATIONS
MODEL-TEM-8A AVERAGING PERIOD-!
4 -1.5 -14
A.
*
< o
-0.5 0.0 0.5
BIAS OF AVERAGE
DATA BASES
+ - CC7B
X- CC76
0. UK7S
0 - UK76
* - PARADISE
\0 1.5 2.0 A- KINCAID
Figure 1. Fractional Bias By Model Using High 25 Values For Each Data Base
-------
FR/a
2.0
1J
0 1.0
1 "
fe-0.5
1-1.0
-1.5
-2.0
-2
^CTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
AVERAGING PERIOD- 1
a
X
»
.0 -1.5 -1.0 -0.5 0.0 0.5
BIAS OF AVERAGE
MODELS
+ - MPTER
X- MPSOM
a - PPSP
0- TEU-BA
1.0 1.5 2.0
BIAS OF STD DEVIATION
&is*----K 5
btabaiiaaibblg ;Q
K» ......... t>
iCTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
AVERAGING PERIOD-3
a
X
0 -1.5 -1.0 -0.5 0.0 0.5
BIAS OF AVERAGE
MODELS
+ - UPTER
x- ypsou
a - PPSP
O- TEM-BA
1.0 1.5 2.0
FRA
2.0
1.5
0 \JO
| 0.5
Q
fe-0.5
-1.5
-2.0
-2
.CTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
AVERAGING PERIOD-24
a
>
+ *
0 -1.5 -1.0 -0.5 0.0 0.6
BIAS OF AVERAGE
MODELS
+ - MPTER
X- MPSDM
a - PPSP
0- TEM-8A
1.0 1.5 2.0
Figure 2. Composite Fractional Bias By Averaging Period Using High 25 Values
-------
NET FRACTIONAL BI AS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
0.898231 / 0.672(63 / 0.420691
1 3 24
AVERAGING PERIOD
Figure 3. Composite Net Fractional Bias Using High 25 Values
-------
ro
BIAS OF STD DEVIATION
^iegggsee 3
i >.
iCTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
MODEL-MPTER AVERAGING PERIOD =1
J -1.5 -1.0
U
V
N
S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
-03 0.0 03 1.0 13 10
BIAS OF AVERAGE
FRA
10
13
I "
§ 03
-10
-2
CTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
MODEL-PPSP AVERAGING PERIODS
Nu*
.0 -13 -1.0
STABIL 1 TY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
-0.5 0.0 03 1.0 1.5 2.0
BIAS OF AVERAOE
FRfi
10
13
O U
i-
fe-03
i-i-o
-10
-2
iCTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
MODEL-MPSDM AVERAGING PERIOD- 1
/
N
J> -13 -1.0 -0.5 0.0 03
BIAS OF AVERAGE
S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
IjO 13 10
BIAS OF STD DEVIATION
SSSSggSCg 3
CTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
MODEL««TEM-8A AVERAGING PERIOD- 1
U
H -13 -14 -03 0.0 03
BIAS OF AVERAOE
S
N
STABILITY
1 - VERY UNSTABLE
J - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 13 10
Figure 4. Composite Fractional Bias By Model Using High 25 Values For Each Stability Class
-------
5 DEVIATION
: g 5 s 8 3
>»
e uju
to
-IJ
-10
-1
iCTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
MODEL-MPTER AVERAGING PERIOD- 1
a -IJ -IJI
U
M
-0.5 0.0 OS
BIAS OF AVERAGE
WINDS
L - LOW
M » MEDIUM
H - HIGH
IJ) IJ 10
FRA
10
u
0 1.0
fe-0.5
1 -IJI
-IJ
-10
-2
CTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
MODEL-PPSP AVERAGING PERIOD- 1
M
L
.0 -1.5 -IJI
-flj 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
M - MEDIUM
H - HIGH
1.0 1.5 2.0
FRA
10
IJ
0 IJ)
i «
o-oJ
-IJ
-10
-I
CTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
MODEL-MPSDM AVERAGING PERIOD- 1
M
L
H
WINDS
L - LOW
U - MEDIUM
H - HIGH
0 -IJ -IJ) -OJ OJ) OJ IJ) U 10
BIAS OF AVERAGE
FRA
10
u
0 IJ)
£ o*
fe-0.5
-1.5
-10
-2
CTIONAL BIAS HIGH 25 CONCENTRATIONS
COMPOSITE MEANS OF RURAL DATA BASES
MODEL-TEM-8A AVERAGING PERIOD- 1
H
M
WINDS
L - LOW
M - MEDIUM
H - HIGH
0 -1.5 -IJ) -OJ 0.0 OJ IJI 1.5 10
BIAS OF AVERAGE
Figure 5. Composite Fractional Bias By Model Using High 25 Values For Each Wind Speed Class
-------
14
-------
SECTION 3
SUMMARY AND CONCLUSIONS
The performance of four rural air quality simulation models has been
presented in the form of comprehensive graphic displays using S02 data col-
lected at four power plant sites. Detailed graphic displays of model perfor-
mance were presented in an earlier report for Clifty Creek, and for three
additional data bases in the Appendices to this report. The results from
the six data bases (two years of data were available at Muskingum River
and Clifty Creek), were used to construct composite measures of model perfor-
mance and displayed in the form of fractional bias plots using the mean and
standard deviation of the highest 25 concentrations. In addition, variations
in performance among the data bases, among averaging periods, and among meteor-
ological conditions were examined.
From an operational performance standpoint (Figures 1-3), PPSP performed
poorly across the six data bases, overpredicting the mean and the spread of
the highest 25 values by well over a factor-of-two. MPSDM also tended to over-
predict the highest concentrations but not by the degree found for PPSP. Both
MPTER and TEM-8A appeared to have relatively little bias, especially for 1-hour
and 3-hour averages. TEM-8A demonstrated a noticeble tendency for underpre-
diction of the highest 24-hour averages.
Breakdown of performance by stability category (Figure 4) indicated a
strong tendency for underprediction of the highest concentrations for stable
conditions (except for PPSP) and overpredictions for the more unstable condi-
tions. It is particularly noteworthy for very unstable and unstable conditions
15
-------
that the composite bias for TEM-8A across the six inventories is virtually
nill. Variations in model bias among wind speed categories (Figure 5) was
relatively small showing a tendency for slight underpredictions for high wind
speeds and little bias for the medium and low wind speed categories. Excep-
tions to this pattern were obvious for PPSP, where overpredictions were
evident for each wind category, and, for MPTER, where the differences in
performance among wind categories were very small.
Although composite results are not meaningful for individual stations,
the trend in the fractional bias with distance between the stations and the
source were similar among the data sets. There appeared to be a consistent
tendency for decreasing overprediction with distance for unstable conditions
and, conversely, a tendency for decreasing underprediction with distance for
the more stable conditions.
Diurnal patterns among the data bases were also reasonably consistent.
In most cases, early morning and late evening hours are associated with large
underpredictions while the mid-day hours are associated with small to moderate
overpredictions.
It is generally concluded that the graphic formats used to assess the
operational and diagnostic components of model performance provide insight
into model behavior not readily obtained from tabular presentations of data.
Hopefully a framework has been established that can be used to develop a
standardized graphics approach for displaying and intercomparing the perfor-
mance of air quality models.
16
-------
REFERENCES
1. Londergan R. J., Minott D. H., Wackter D. J. and Bonitta D., Eval-
uation of Rural Air Quality Simulation Models, EPA-450/4-83-U03,
Environmental Protection Agency, Research Triangle Park, N. C.,
October 1982.
2. Cox, W. M., and Moss G. K., Evaluation of Rural Air Quality Simulation
Models, Addendum A: Muskingum River Data Base, EPA-450/83-003a,
Environmental Protection Agency, Research Triangle Park, N. C., June
1985.
3. Cox, W. M., Rorex H. W., and Moss G. K., Evaluation of Rural Air
Quality Simulation Models, Addendum C: Kincaid S02 Data Base, EPA-4bO/
83-003c, Environmental Protection Agency, Research Triangle Park, N. C.,
- March 1986.
4. Cox, W. M., Rorex H. W., Moss G. K., and Baldridge E. W., Evaluation
of Rural Air Quality Simulation Models, Addendum D: Paradise S02 Data
Base, EPA-450/83-003d, Environmental Protection Agency, Research
Triangle Park, N. C., January 1987.
5. Cox, W. M., G. K. Moss, J. A. Tikvart and K. W. Baldridge, Evaluation
of Rural Air Quality Simulation Models, Addendum B: Graphical Display
of Model Performance Using the Clifty Creek Data Base, EPA/483-0035,
Environmental Protection Agency, Research Triangle Park, N. C., August
1985.
17
-------
18
-------
APPENDICES
INTRODUCTION TO THE APPENDICES
APPENDIX AKINCAID POWER PLANT
APPENDIX BPARADISE POWER PLANT
APPENDIX CMUSKINGUM RIVER POWER PLANT
1-1
-------
1-2
-------
INTRODUCTION TO THE APPENDICES
The performance of four rural models (MPTER, MPSDM, PPSP, and TEM-8A)
is presented graphically in a format similar to that used in the earlier
report for Clifty Creek. Results for the three inventories are reported
separately as follows: Appendix AKincaid power plant, Appendix BParadise
power plant and Appendix C--Muskingum River power plant, 1975 and 1976. The
purpose of the Appendices is to provide complete documentation regarding how
each model performs both from an operational and diagnostic viewpoint. Also
they provide the basis for material presented in the body of the report that
analyzes composite performance of the models across the inventories (includ-
ing Clifty Creek) and for evaluating differences in model performance among
the inventories. Throughout both the operational and diagnostic phases of
the discussion in these appendices, graphics are displayed for two basic
data sets (1) the 25 highest observed and predicted concentrations unpaired
in space and time and (2) all observed and predicted concentrations completely
paired in space and time.
The figures used to display the performance of the models are presented
in the same sequence for each of the three inventories. Figure 1 in each
of the three appendicies presents a map that displays the location of the
monitoring stations relative to the stack(s) including information regarding
terrain. Figures 2-4 presents information oriented to the operational per-
formance of the models. Graphs are provided by averaging period and data set
using fractional bias plots, quantile-quantile plots and frequency distribu-
tions. Performance is given only for all stations and meteorological condi-
tions combined.
1-3
-------
A portion of the diagnostic information is provided in Figures 5-10 in
the form of fractional bias plots. These plots display the fractional bias
of the standard deviation vs the fractional bias of the average for indi-
vidual stations, and for individual wind speed and stability categories.
Further diagnostic information is provided in Figures 11-16 in the form of
bias of the average as a function of station to source distance by stability
and wind speed category. Finally, Figures 17-20 present comparisons of model
performance in the form of diurnal plots in which model bias of the average
is displayed for each hour of the day for all stations combined and as a
function of station groupings by distance from the source. For further
information regarding interpretation of the various types of graphs, the
reader is referred to the Clifty Creek graphics report(5).
1-4
-------
APPENDIX AKINCAID POWER PLANT
Figure A.I shows the location of each of the 30 S02 monitoring stations
located around the Kincaid plant. The stations are located at distances
ranging between approximately 2 km and 20 km. Since the area surrounding
the plant is essentially flat, no terrain features are shown.
Figure A.2 summarizes the direction and relative magnitude of the
fractional bias of the average and the fractional bias of the standard
deviation for each of the four models. The left hand column displays three
panels corresponding to the three averaging periods for the 25 highest
observed and predicted concentrations. PPSP obviously overpredicts for
each averaging period while the other three models are relatively unbiased.
There is a tendency for TEM-8A and MPTER to underpredict the scatter as
evidenced by the consistently positive values for fractional bias of the
standard deviation. Figure A.3, which displays Q-Q plots of the 2b highest
observed and predicted concentrations, supplements the information shown in
Figure A.2. The magnitude of overpredictions by PPSP is clear from the
distance between the points and the diagonal line of "perfect" fit. Also
evident is the somewhat nonlinear relationship between the rank ordered
observed and predicted values, especially for PPSP and MPSDM.
The right hand side of Figure A.2 presents fractional bias informa-
tion using all data pairs. There is a clear shift of points to the right
in each panel which indicates a tendency for smaller overpredictions (PPSP)
and larger underpredictions (MPTER). Figure A.4, which presents the same
information in the form of cumulative frequency distributions, provides
additional insight into the nature of the disparity between observed and
A-l
-------
predicted values. For example, the disagreement between observed and pre-
dicted values again appears to be a function of the magnitude of observed
values; the difference is especially evident for 1-hour and 3-hour aver-
ages.
Figure A.5 displays the bias at each of the 3D monitoring stations,
by model, using the 25 highest concentrations at each station. Consistent
with results shown in Figure A.2, PPSP overpredicts at the majority of
stations for both the average and the "scatter". Although there is a
rather large range in the magnitude of the fractional bias, a more clearly
defined pattern will be evident in subsequent displays involving distance
between source and monitoring stations. Figure A.6 displays the bias for
each of the four stability categories using the 25 highest concentrations
for each stability category. Except for PPSP, there is a clear separation
in performance among stability categories with large underpredictions for
stable conditions and relatively modest overpredictions for the more unstable
conditions. Figure A.7 displays similar results for wind speed categories.
Compared with the spread among the data points seen for stations and stability
conditions, the range in bias among wind speed categories is quite small. For
MPTER, MPSDM and TEM-8A, the biases are all within a factor of two. Figures
A.8-A.10 are analogous to Figures A.5-A.7 except that all data paired in
time and space are used to compute the fractional biases. MPSDM and TEM-8A
appear to exhibit the least overall bias as indicated by the clustering of
points inside the framed area that corresponds to a factor of two. While
PPSP continues to overpredict for most data categories, the degree of
overprediction is not as severe as with the 25 highest concentrations.
A-2
-------
Figure A.11 displays the fractional bias of the average as a function
of distance between the source and each monitoring station. For the more
distant stations, each model exhibits a relatively low bias. The biases at
the closest stations (less than 4 miles) are obviously larger, ranging from
overpredictions by PPSP (less than -1) to underpredictions by MPTER (greater
than +1). Figure A.12 and A.13 contain displays identical to Figure A.11
except they are subdivided by stability category (Figure A.12) and wind
speed category (Figure A.13). While the pattern for very unstable condi-
tions is similar to that shown in Figure A.11, the range in the bias as
a function of distance is somewhat larger for the unstable and neutral condi-
tions. For stable conditions each of the models severely underpredicts at
all distances except for PPSP where large overpredictions close to the source
become slight underpredictions at the most remote station. Figures A.14-A.16
are identical to the previous three figures except that all paired data are
used. Results are reasonably similar to those for the highest 25 values
except there is a greater tendency toward underpredictions.
Figure A.17 displays fractional bias of the average as a function of
hour of the day using the 25 highest values. For all models, except PPSP,
large underpredictions occur early in the morning and again later in the
day. Midday hours are associated with relatively little bias. PPSP dis-
plays large overpredictions for all hours. Figure A.18 presents the same
information except subdivided by stations grouped by distance from the
plant. The stations are grouped as follows: Group A -- less than 5 km;
Group B -- between 5 and 10 km; Group C -- between 1U and 15 km; Group D --
greater than 15 km. The diurnal patterns for the more distant stations
(Group D) contrast with those at the closer stations (Group A) in that the
A-3
-------
range in bias is smaller for the more distant stations, especially for
PPSP. Finally, Figures A.19 and A.20 display information in an identical
format to Figures A.17 and A.18'except all paired data are used. Patterns
are quite similar; again, the tendency toward greater underprediction is
apparent.
A-4
-------
EPRI STATION
A CECO STATION
Figure A.I Field Monitoring Network Near The Kincaid Power Plant
A-5
-------
K
ID
IS
I"'
£ u
a
gu
^
m-lJl
-13
-U
-1
NCA1D S02 HIGH 25 CONCENTRATIONS
YEAfl-1980 AVERAQNC PEWOD-1
a
i -13 -\a
V
A
t-
V U 19
MODELS
+- ypTE*
X- yPSDU
a - PPSP
0- TEU-BA
U U U
KINCAID S02 ALL CONCENTRATIONS
YEAR-1980 AVERAGING PEHOD-1
U
13
i u
i u
a
gu
fc-u
-U
-U
-U
-i
a
t HI -*i
>
43 U U
HAS OF AVERAGE
t-
MODELS
+ - MPTER
x- upsoy
a - PPSP
«- TEU-«A
Ul U U
K
u
13
| ID
B-U
-13
-U
-I
NCAID «502 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIODS
a
1 -13 -1JO
x
^ *
-u u u
BIAS OF AVERAGE
MODELS
+- ypTER
x. ypsoy
D - PPSP
0- TEH-RA
1J 13 U
u
13
1 "
i-1,
-13
-U
-2
KINCAID S02 ALL CONCENTRATIONS
YEAR-1980 AVERAGING PEWOD-3
, n
t -13 -U
}
-QJ ftfi U
BIAS OF AVERAGE
+
MODELS
+- MPTER
x> ypsOM
a . PPSP
«- TEH-«A
U 13 U
K
u
13
i «
P
a «
Q
|M
S-u
B-U
-13
-U
-a
INCAID S02 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PEWOD-24
a
JD -13 -1JD
X
Jj|
+
-U U U
BIAS OF AVERAGE
MODELS
+- «PT«
x- ypsoy
0 . PPSP
« TEH-GA
U 13 U
U
13
I-
a
a u
i"
-13
-U
-J
KINCAID S02 ALL CONCENTRATIONS
YEAR-19BO AVERAGING PER10D-24
a
j» -13 -U
c
-U U U
BIAS OF AVERAGE
t-
MODELS
+- yPTER
x- ypsoy
a - PPSP
>- TEV-SA
U 13 U
Figure A.2 Fractional Bias By Averaging Period Using High 25 Values And All Paired
Values
A-6
-------
KINCAID SO2--HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1
MOO
MOO
11600
a
100
200
200
400
MODELS
+ - UPTER
X- UPSDU
D - PPSP
O- TEU-8A
I IUO
OBSERVED
3200
(400
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR-1960 AVERAGING PERIOD-3
400
200
400
600 1800
OBSERVED
3200
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-24
400
5200
100
SO
c
MODELS
+ - UPTER
X- UPSDU
a - PPSP
O- TEU-BA
SO
100
200
OBSERVED
400
BOO
MODELS
+ - UPTER
X. UPSDU
a - PPSP
O- TEU-BA
Figure A.3 Predicted Quantiles Vs Observed Quantiles By Averaging Period Using High 25 Values
-------
00
KINCAID S02 ALL CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD=1
10000
n
2
O 1000
100
10
15.5
MODELS
OBSERVED
yPTER
MPSDU
PPSP
88.5 97.7 M.7
CUMULATIVE FREQUENCY
t»M
99.999
KINCAID SO2 ALL CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD<=3
10000
n
2
C5 1000
100
10
A?/ /
515 17.7 M.7
CUMULATIVE FREQUENCY
KINCAID S02 ALL CONCENTRATIONS
YEAR=19BO AVERAGING PERIOD=24
10000
1000
too
10
MODELS
OBSERVED
_yPTE_R_
UPSOU
PPSP
TEU-6A
65.9
U.5 87.7 99.7
CUMULATIVE FREQUENCY
91.98
99.999
MODELS
OBSERVED
IIPTER
PPSP
TEy-BA
M.9M
Figure A.4 Cumulative Frequency Distributions By Averaging Period Using All Paired Values
-------
UD
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD=1 MODEL-MPTER
10
13
z 1J(
P
OX,
O-03
i-u
-.3
-10
-10 -13 -IJ)
4
5
F
R
D "
1 * E
*c 0 L
'
-0.5 0.0 03
BIAS OF AVERAGE
' 3
*
STATIONS
1-9 - STNS 1-9
A-U - STNS 10-30
1X1 U 10
KINCAID S02
HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 MODEL-PPSP
10
13
1 w
P
§ «
a
0-03
3
0-1X)
-13
-10
1
a*
ft M<
B
A*
d" E T
a K
/. u -
3
{ N
'
-10 -1.5 -IX)
H
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-9 - STNS 1-9
A-U - STNS 10-30
1.0 1.5 10
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 MODEL°MPSDM
10
1.5
0 ">
P
g "
0-OJ
i-1X>
-u
-10
A
,/* 1
7 N °^ Q
-10 -\A -1X1
i
*T
STATIONS
1-9 - STNS 1-9
A-U - STNS 10-30
-03 OX) 03 IX) 13 10
BIAS OF AVERAGE
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 MODEL-TEM-8A
10
§ -
F
i «
Q
e-
it
OT
a -1X1
-'»
-10
-10 -1.5 -IX)
R
s
,,,tjr^
*. =.
7
J
STATIONS
19 STNS 19
A-U - STNS 10-30
-03 0.0 03 1X1 13 10
BIAS OF AVERAGE
Figure A.5 Fractional Bias By Model Using 25 High Values For Each Station
-------
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD=1 MODEL=MPTER
10
1.5
§ U
"
"
0-05
-U)
-10
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
-10 -1.5 -1J) -04 0.0 05
BIAS OF AVERAGE
1.0
1.5
10
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 MODEL-PPSP
10
-
b-o.5
1,.
m-1.0
-1.5
-10
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
-10
-1.5
-1.0 -OS 0.0 04
BIAS OF AVERAGE
1.0
1.5
10
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR=19BO AVERAGING PERIOD- 1 MODEL-MPSDM
10
IS
g U.
i «'
o
0-04
ffl-U
-14
-10
-I
t
a -is -1.0
V
N
-OS 0.0 04
BIAS OF AVERAGE
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
U> U 10
KINCAID S02 HIGH 25 CONCENTRATIONS
10
14
STD DEVIATION
g S £
O-04
0-1.0
-14
-W
-2
YEAR- 1980 AVERAGING PERIOD- 1 MODEL-TEM-8A
JO -IS -U)
U
V
-0.5 0.0 0.5
BIAS OF AVERAGE
S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
U 14 10
Figure A.6 Fractional Bias By Model Using 25 High Values For Each Stability Class
-------
3»
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 MODEL-MPTER
BIAS OF STD DEVIATION
fa 8 g 8 S £ S
-1.5
-10
-2
0 -1.5 -1.0
U
L
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - MEDIUM
H - HIGH
1.0 1.5 10
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 MODEU-PPSP
2.0
1.5
STD DEVIATION
So
b> <3
O-0.5
m-i.o
-1.5
-10
-2
L
M H
Jt -1.5 -1.0
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - MEDIUM
H - HIGH
1.0 1.5 10
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 MODEL-MPSDM
BIAS OF STD DEVIATION
fa & 8 8 o £ S
-1.5
-10
-2
J) -1.5 -1.0
L
M
M
-0.5 0.0 OS
BIAS OF AVERAGE
WINDS
L - LOW
U - MEDIUM
H - HIGH
14 1.5 10
KINCAID S02 HIGH 25 CONCENTRATIONS
10
1.5
I «
O
1-1.0
-1.6
-10
-2
YEAR-1980 AVERAGING PERIOD- 1 MODEL-TEM-8A
.0 -1.5 -1.0
K
L
H
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - MEDIUM
H - HIGH
1.0 \S 10
Figure A.7 Fractional Bias By Model Using 25 High Values For Each Wind Speed Class
-------
3>
ro
2.0
u
S"
0-0.5
0-1.0
-1.5
-2.0
-2
KINCAID S02 ALL CONCENTRATIONS
YEAR=19BO AVERAGING PERIOD=1 MODEL=MPTER
$ 3
0 0* ^
" B"* j ' C
F
STATIONS
1-0 - STNS 1-8
A-U - STNS 10-30
XI -IJ -IJ -0.5 0.0 0.5 1J> 1.5 2.0
BIAS OF AVERAGE
2.0
1.5
g IJ
1 «
Q
0-0.5
0-1.0
-2.0
-2
KINCAID S02 ALL CONCENTRATIONS
YEAR=19BO AVERAGING PERIOD=1 MODEL=PPSP
A D £
BA K L
_.i ^ p
. o *lnn
*3 C
1
*
STATIONS
1-0 - STNS 1-8
A-U - STNS 10-30
0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
BIAS OF AVERAGE
KINCAID S02 ALL CONCENTRATIONS
YEAR=1980 AVERAGING PERIODS MODEL=MPSDM
3 DEVIATION
: g s s 8
g "
o-w
ffl-1.0
-1J
-M
-I
1
Jl -U -ID
N °
-0.5 0.0 06
BIAS OF AVERAGE
T
IJ) 1.5 U
STATIONS
1-i - STNS 1-0
A-U - STNS 10-30
'
3 DEVIATION
; g s c s
b "
d.
a-tJ>
-2J)
-I
KINCAID S02 ALL CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD- 1 MODEL-TEM-8A
0 -1.5 -IJ)
7
W2 Li
C
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-0 - STNS 1-i
A-U - STNS 10-30
1.0 1.S 2.0
Figure A.8 Fractional Bias By Model Using All Paired Values For Each Station
-------
KINCAID S02 ALL CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD=1 MODEL-MPTER
10
\i
§ 141
1 »
Q
eu
0-0.5
0-1.0
-1.5
-2J)
-2
J> -1.5 -1.0
U
-O.S 0.0 OS
BIAS OF AVERAGE
N
V
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 2.0
KINCAID S02 ALL CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD=1 MODEL=PPSP
10
14
0 <<>
o
6-
0-0.5
ffl -1.0
-1.5
-2.0
-2
V
NU
S
0 -1.5 -1.0
-0.5 0.0 0.5
BIAS OF AVERAGE
STABIL ITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 2.0
KINCAID S02 ALL CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 MODEL-MPSDM
U
14
O >4
i-
O-0.5
ffl -1.0
-1.5
-10
-1
4 -14 -1.0
V
U
h
-04 00 04
BIAS OF AVERAGE
U 14 «
STABIL ITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
)
BIAS OF STD DEVIATION
i i i i
sssggssss
!> .
KINCAID S02 ALL CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD- 1 MODEL-TEM-8A
0 -1.5 -1.0
V
U
-0.5 0.0 0.5
BIAS OF AVERAGE
S
N
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 2.0
Figure A.9 Fractional Bias By Model Using All Paired Values For Each Stability Class
-------
-J>
I
KINCAID S02ALL CONCENTRATIONS
YEAR" 1980 AVERAGING PERIODS MODEL-MPTER
10
§ Ul
Su
ew
0-03
1-1.0
-13
-10
-10 -13 -1.0
-03 0.0 03
BIAS OF AVERAGE
1.0 13 10
WINDS
L - LOW
U - UEOIUU
H - HIGH
KINCAID S02 ALL CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 MODEL-MPSDM
10
13
BIAS OF STD DEVIATION
S £ C 8 5
-13
-10
JO -13 -U)
k
H
-03 0.0 03
BIAS OF AVERAGE
WINDS
L - LOW
y - MEDIUM
H - HIGH
14) 13 10
KINCAID S02 ALL CONCENTRATIONS
YEAR-19BO AVERAGING PERIOD- 1 MODEL-PPSP
10
13
STD DEVIATION
g S E
0-03
n-i.0
-13
-10
-I
H
LM
0 -13 -14)
-03 04) 03
BIAS OF AVERAGE
WINDS
L - LOW
y - MEDIUM
H - HIGH
1.0 13 10
KINCAID S02 ALL CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 MODEL-TEM-8A
10
13
0-03
i-U)
-13
-10
M
-10 -13 -1.0 -03 0.0 03 14) 13 10
BIAS OF AVERAGE
WINDS
L - LOW
y - MEDIUM
H - HIGH
Figure A.10 Fractional Bias By Model Using All Paired Values For Each Wind Speed Class
-------
KINCAID S02 -- HIGH 25 CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD=1 CATEGORY=ALL MET CONDITIONS
2.0
0.5
tn
o
-0.5 '
2 3 4 5 6 7 8
STATION DISTANCE DOWNWIND (Ml)
10 11 12 13
MODELS
+ =MPT.IF*
>i=MP_S_DM__
D =PPSP
O =TEM-8A
Figure A.11 Fractional Bias Of The Average Vs Station Distance Using High 25 Values
-------
en
U)
I «
i
5 **
j-u
-u
-u
u
u
*"
?*
-u
-u
KINCAID S02 HIGH 25 CONCENTRATIONS
YEAR- 1 980 AVERAGING PERIOD- 1 CATEGORY-VERY UNSTABLE
*
* » t
X^-- * ---** ' °* -gfr '. -*
** v * 't'"*"t " "o""""
^-r ^----^^^r^^^ "
a * __-~-"""'"**^
o
a a
STATION DISTANCE DOWNWIND (W)
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 CATEGORY-NEUTRAL
%x X « "* - J<"" ""^fc. « -
"'"JL-^-^r2^
^^m^^° ° °
°» ° ^*^ °
o
STATION DISTANCE DOWNWIND (U)
unnFi ^
-t- -UPTER
x -UPSOU
n -PPSP
0 -TEU-BA
MODELS
+ -UPTER
X -UP SOU
a -PPSP
0 -TEU-BA
u
u
iii "
1
p °
f-"
-u
-2.0
!"
1-
-1*
-u
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 CATEGORY-UNSTABLE
\
\
\» \ »
V *\ + -
\ xx * *i
x \ *"*» ».~* " o -y
^xv * * ft * -^.-£'~*i^c^~^^ "
-H-- ^--^.^.^.^.^^ff^ fi
* ^f^*l^
^^^ a "
STATION DISTANCE DOWNWIND (M)
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 CATEGORY-STABLE
9 * I* 5* --J.-1 __J
«
a ° s
/*
" " ^^^ °
" ^^^^ a a D
, -a^af ° a ° ". " "
a
i i i i i i . i i i
STATION DISTANCE DOWNWIND (U)
unnFi s
t -UPTER
X-UPSDU
a -PPSP
0 -TEU-1A
MODELS
-UPTER
X-UPSDU
a -PPSP
0 -TEU-BA
Figure A.12 Fractional Bias Of The Average Vs Station Distance By Stability Class Using High 25 Values
-------
KINCAID S02 -- HIGH 25 CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD -1 CATEGORY-LOW WINDS
U>
U
1.0
1"
[-OJ
-IJ>
-15
-JO
it!""**- "
MODELS
*_^yPTER__
x-ypsoy
o -PPSP
O -TEU-8A
0 I I 1 4 t t 1 0 I 10 II . 12 II
STATION DISTANCE DOWNWIND (Ul)
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-HIGH WINDS
U
1.0
W
-Id
-IJ
-10
o
MODELS
X;IJPSOy_
a -PPSP
o i i i 4 t i ; i to n n u
STATION DISTANCE DOWNWIND (Ul)
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD-1 CATEGORY-MEDIUM WINDS
MODELS
x-upsoy __
n -PPSP
O -TEU-BA
Figure A.13 Fractional Bias Of The Average Vs Station Distance By Wind Speed Class Using High 25 Values
-------
KINCAID S02 -- ALL CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD=1 CATEGORY=ALL MET CONDITIONS
2.0
I
00
-2.0 '
0123456789
STATION DISTANCE DOWNWIND (Ml)
10 11 12 13
MODELS
+j^MPJER__
*.»MP_SDM_
D °PPSP
0=TEM-8A
Figure A.14 Fractional Bias Of The Average Vs Station Distance Using All Paired Values
-------
U
1.0
2 as
a «
?-
-u
-u
KINCAID SO2 ALL CONCENTRATIONS
YEAR- 1080 AVERAGING PERIOD- 1 CATEGORY-VERY UNSTABLE
4
* + * *
"*""*"* '"~"^F + * * " -tiJI *«%.*" "r*^~i
".t/ « ^r.--^v^"j^"^^--3iiii!:--^?
o;Y---^i2^v^r- :
" Q
a
a °
STATION DISTANCE DOWNWIND (Ml)
MODELS
+ -UPTER
x-upsoy
a -PPSP
0 -TEU-8A
10 1
u
IJ
!»
8 "^
|-4U
-!>.
-10
KINCAID SO2 ALL CONCENTRATIONS
YEAR- 1080 AVERAGING PERIOD- 1 CATEGORY-NEUTRAL
"V**^W-* **
--<.. ..-. r^u, . ,;
. ...... .-^
B «° Jl^'
n n -*- n
^*^- "
» ^ x a a
-CL" _>-*
o ° «
a
I 1 1 1 4 i 1 7 1 1 10 II It 1
STATION DISTANCE DOWNWIND (Ml)
MODELS
t-MPTER
X-UPSDU
a -PPSP
O -TEU-SA
)
KINCAID SO2 ALL CONCENTRATIONS
YEAR- 1080 AVERAGING PERIOD- 1 CATEGORY-UNSTABLE
5 = s 3 3 ? = = ?
SVB 1VNOU3VtU
*
V\ : - .
x *» \ °
x \ %. « « » ,
-»..* N * "> ~- -*t r *
; ^Tntrr^
^-< o^o *
+
+
-f"*-^x«f ---^*
_- s x
°o a
o" ^-^r* 8 0'
^ o °
a
a
STATION DISTANCE DOWNWIND (U)
MODELS
X -UP SDH
a -PPSP
0 -TEM-8A
t t 10 11 U 11
KINCAID S02 ALL CONCENTRATIONS
YEAR- 1080 AVERAGING PERIOD- 1 CATEGORY-STABLE
U
IjO
I--
2-ot
-1.0
-M
-u
""**""*» * *-
*
*
a
* ** /
a >^
a Sti
^S Q
a ^P^^ a a o
a
STARON DISTANCE DOWNWIND (Ml)
B
a
MODELS
+ -UPTER
X -UP SOU
a -PPSP
0 -TEU-8A
1
Figure A.15 Fractional Bias Of The Average Vs Station Distance By Stability Class Using All Paired Values
-------
ro
o
u
IJ
2 ^0
" " ^-.-"^ ^~~B °
~~B =g
a °
STATION DISTANCE DOWNWIND (U)
MODELS
+ -UPTER
X-UPSDM
D -PPSP
0 -TEU-8A
KINCAID SO2 ALL CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-HIGH WINDS
MODELS
X-IIPSDU__
o -PPSP
» -TEM-8A
711
STATION DISTANCE DOWNWIND (U)
ta ii 11 a
KINCAID SO2 -- ALL CONCENTRATIONS
YEAR- 1980 AVERAGING PERIOD- 1 CATEGORY-MEDIUM WINDS
MODELS
f_-UPT£R
X"UPSDU
a -PPSP
O -TEU-8A
« I I
10 II II 13
STATION DISTANCE DOWNWIND (Ul)
Figure A.16 Fractional Bias Of The Average Vs Station Distance By Wind Speed Class Using All Paired Values
-------
I
ro
-2.0 1
KINCAID S02 -- HIGH 25 CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD=1 CATEGORY=ALL STATIONS
MODELS
_M_PTER
MPSDM^
PPSP
TEM-BA
HOUR OF DAY
Figure A.17 Fractional Bias Of The Average Vs Hour Of The Day Using High 25 Values
-------
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-STN GROUP A
II
HOUR OF DAY
r>o
ro
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-STN GROUP C
14
MODELS
IJPTEJJ_
UPSDU
PPSP
TEU-BA
MODELS
JJPTER
UPSDU
PPSP
TEU-8A
HOUR OF DAY
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-STN GROUP B
HOUR OF DAY
KINCAID SO2 HIGH 25 CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-STN GROUP D
-It
n
HOUR OF DAY
UODELS
UPSDU_
PPSP
TEU-8A
UODELS
MPJER
upsnu_
PPSP
TEU-BA
Figure A.18 Fractional Bias Of The Average Vs Hour Of The Day By Station Grouping Using High 25 Values
-------
KINCAID S02 ALL CONCENTRATIONS
YEAR=1980 AVERAGING PERIOD=1 CATEGORY=ALL STATIONS
PO
00
2.0
1.5
1.0
O
fc
0.0
-0.5
-1.0 '
-1.5 '
-2.0 H
\ \
\
V
\ *
\x\
' /
12
HOUR OF DAY
\L
18
24
MODELS
_MPJER___
MPSDM^
PPSP
TEM-8A
Figure A.19 Fractional Bias Of The Average Vs Hour Of The Day Using All Paired Values
-------
KINCAID S02 ALL CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-STN GROUP A
M
HOUR OF DAY
ro
KINCAID SO2 ALL CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-STN GROUP C
MODELS
_IJPTER_
jUPSDU
PP8P
TEM-BA
MODELS
JIPTER
UPSQU
PP3P
TEU-8A
HOUR OF DAY
KINCAID S02 ALL CONCENTRATIONS
YEAR-1980 AVERAGING PERIOD-1 CATEGORY-STN GROUP B
MODELS
SKISL.
MPSDI}___
PP«P
TEU-6A
II
HOUR OF DAY
KINCAID S02 ALL CONCENTRATIONS
YEAR-1860 AVERAGING PERIOD-1 CATEGORY-STN GROUP D
MODELS
IIPTER_
ypsou
PPSP
TEU-8A
HOUR OF DAY
Figure A.20 Fractional Bias Of The Average Vs Hour Of The Day By Station Grouping Using All Paired Values
-------
APPENDIX BPARADISE POWER PLANT
Figure B.I shows the location of the 12 monitoring stations surrounding
the Paradise Power plant. The distance between the stations and source ranges
from approximately 3 km up to 17 km from the plant site. Although some terrain
features are shown on the figure, the elevation of the monitors relative to
plant grade are unimportant (less than 40 meters). When reviewing the graphics.
for this plant, the reader should be aware that stations 1, 2, 4 and 5 reported
slightly less than 40 percent of the possible hourly S02 concentrations.
This becomes particularly important for data sets that contain only a few
observations per data category, e.g., data subsets at individual stations by
meteorological conditions or hour of the day.
Figure B.2 presents the fractional bias of the average vs the fractional
bias of the standard deviation for each of the four models. The information
shown in the left hand side of this figure and from the companion Q-Q plot
(Figure B.3) indicate large overpredictions by PPSP for all three averaging
periods which contrast sharply with the relatively low bias for the other
three models. Also there is a general tendency for increasing underpredic-
tion as averaging period increases. Overpredictions of 1-hour averages by
MPSDM tend to vanish for 24-hour averages while a lack of bias for MPTER and
TEM-8A for 1-hour averages become underpredictions for 24-hour averages.
Similar trends are apparent from examination of the all data categories for frac-
tional bias (right hand side of Figure B.2) and the cumulative frequency
plots (Figure B.4).
Diagnostic information in the form of fractional bias plots for stations,
stability categories, and wind speed categories are shown in Figures B.5-B.7
B-l
-------
for the highest 25 values. The fractional bias of 1-hour averages indicates
that MPTER, MPSDM and TEM-8A are relatively unbiased as evidenced by the pre-
ponderance of the 12 stations found within the factor of two frame in the
center of each panel. Of these three models, MPSDM indicates a tendency for
overpredictions at stations 1, 8 and 12. Fractional bias for stability
categories (Figure B.6) shows a familiar pattern; stable conditions are
associated with large underpredictions while the more unstable conditions are
associated with low bias and/or overpredictions. PPSP overpredicts for each
of the stability categories including stable conditions. Fractional bias for
the wind speed categories (Figure B.7) shows little difference in the bias
among the three wind speed categories. MPTER and TEM-A both perform very
well as indicated by the close proximity of the three points to the center
of the fractional bias panel (0,0). As expected, the all concentration data
sets (Figures B.8-B.10) indicate a tendency for greater underprediction. For
wind speed categories (Figure B.10), a somewhat greater separation in perfor-
mance among the wind speed categories is evident compared with that for the
25 highest values (Figure B.7).
Fractional bias vs station distance (Figures B.11-B.16) show patterns
that may be affected somewhat by the missing data noted above for stations
1,2,4 and 5. The effect may be most pronounced for station 1 and station 2
since they are the closest and farthest stations, respectively. With this
caveat in mind, the patterns of over/under-predictions for the four models
appear to be a function of stability (Figure B.12) and wind speed (Figure
B.13) categories. For very unstable conditions, the models as a group shift
from overprediction toward slight underprediction as station distance in-
creases while the models take on different patterns as the stability increases,
B-2
-------
Neutral and stable conditions appear to be associated with decreasing under-
prediction with distance for each of the models except PPSP. Patterns for
wind speed categories are also obvious; large variations in bias near the
plant for low winds become somewhat smaller for medium winds.
Finally, the fractional bias for all stations (Figure B.17--high 2b and
Figure B.19all data) and for each of the 12 stations (Figure B.18 and
Figure B.20) is presented. The stations were not grouped by station distance
from the source to prevent possible masking of the effects of missing data.
Obviously the plots for stations having large amounts of missing data (sta-
tions 1, 2, 4 and 5) are much "noisier" than for the other stations. This
noisiness is probably a consequence of both the smaller quantity of data and
possible seasonal effects (monitoring did not begin at these stations until
mid-August). For most of the stations, familiar diurnal patterns do appear;
underpredictions occur for early morning and late evening hours accompanied
by relatively low biases near mid-day.
B-3
-------
Figure B.I Field Monitoring Network Near The Paradise Power Plant
B-4
-------
p;
u
u
I u
i u
£ u
fe-u
sU
-u
-M '
-t
kRADISE S02 HIGH 25 CONCENTRATIONS
YOR-197B AVERAflNGPERKJO-1
o
i -u -u
+
X
i»
i
IAS Of AVCRAtC
UODELS
*- NTTER
X- MPIOM
O- fMP
«- m-M
PARADISE S02 ALL CONCENTRATIONS
YE/W-1S76 AVER/WHO PEROO-1
U
u
I u
1 u
gu
fe-u
lu
-u
O
X
-u -u -u -u u u
SIM OF AVERAGE
f
yOOELS
+- MPTER
x- NPSOV
a- rrsp
«- TEM-*A
U U 10
p/
w
u
0 U
1 u
6M1
fe-u
S-u
-u
-u
-1
WVDISE S02 HIGH 25 CONCENTRATIONS
YEMt-1978 AVERAGING PERUD-3
a
«
f
* -u -u -u t* u
IAS OF AVtRME
UODELS
+- WTEM
X- HPSOH
a- PMP
«- TDt-tA
10 U U
PARADISE S02 ALL CONCENTRATIONS
YEAft-1876 AVERAONC PEFBOO-3
If
U
i u
fc-u
-u
-u
-t
O
<
X
a -u -u -u u u
HAS Or AVERME
f
u u u
MODELS
f- MME*
x. Mr-ton
0- rrSP
«- TEM-U
p;
If
U
§ u
I"
-u
-u
-t
iRADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVBUOMO PSBOO-24
O
«
» -o -o -u u u
IAS Or AVBUME
u u u
UODELS
*- MTTER
X- MPUM
D-PMP
«- TEM-«A
PARADISE S02 ALL CONCENTRATIONS
YBR-1876 AVERMMGPBBOO-24
U
U
§ u
i u
0
&-U
s^
-u
^A '
a
«
X
-U -U -U -U U U
tt At or AVOUK
+
UODELS
+- Mr-nil
x. MPUM
a- wsr
« TEH-tA
u u u
Figure B.2 Fractional Bias By Averaging Period Using High 25 Values and
All Paired Values
B-5
-------
CD
I
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1
8400
3200
1600
800
800
MODELS
+ - UPTER
X- UPSDU
O - PPSP
O- TEU-8A
1600 3200
OBSERVED
6400
PARADISE SO2 HIGH-25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD=3
3200
1600
800
400
400
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-24
800
400
200
100
800 1600
OBSERVED
MODELS
+ - UPTER
X- HPSDM
D - PPSP
O- TEU-8A
too
200 400
OBSERVED
800
MODELS
+ - UPTER
X- UPSDU
D - PPSP
O- TEM-8A
3200
Figure B.3 Predicted Quantiles Vs Observed Quantiljjs By Averaging Period Using High 25 Values
-------
CD
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1
10000
is 1000
too
to V-
su
MODELS
OBSERVED
MPTEH_
UPSDU
PPSP
TEU-BA
88.5 97.7 99.7 99.98
CUMULATIVE FREQUENCY
9B.999
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-3
toon
1000
too
to
KS 88.6 97.7 99.7
CUMULATIVE FREQUENCY
PARADISE S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-24
10000
o' 1000
3
too
to
UODELS
OBSERVED
MPTER
UPSDU
PPSP
TEU-6A
6S.S
88.5 97.7 99.7
CUMULATIVE FREQUENCY
99.98
99.999
Figure B.4 Cumulative Frequency Distributions By Averaging Period Using All Paired Values
UODELS
OBSERVED
yPTER
ypspu_
PPSP
TEH-6A
99.999
-------
ro
00
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL-MPTER
ia
O 1-0
P
I "
Q
0-0.5
0-10
-1.5
-10
-2
0 -1.5 -1.0
e
%
A
7 2
a
4
-0.5 0.0 0.5
BIAS OF AVERAGE
1
STATIONS
1-9 - STNS 1-9
A-C - STNS 10-12
1.0 1.5 10
PARADISE S02 HIGH 25 CONCENTRATIONS
10
tj
BIAS OF STD DEVIATION
S g g g S
-1.5
-10
-2
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-PPSP
3
s
1 8
0 -1.5 -1.0
2
7
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-9 - STNS 1-9
A-C - STNS 10-12
1.0 1.5 10
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD- 1 MODEL-MPSDM
o l*
a
p QJ)
K
O-0.5
-2
1
-------
PA
2.0
1.5
| 1.0
I °-S
Q
eoji
0-0.5
2
m-1.0
-1.5
-2.0
-2
vRADISE S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=MPTER
0 -1.5 -1.0
V
1
-0.5 0.0 0.5
BIAS OF AVERAGE
S
N
STABIL ITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 2.0
PA
2.0
1.5
| M>
1 °'5
g 0.0
V)
0-0.5
m-i.o
-1.5
-2.0
-2
^RADISE S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=PPSP
.VN S
u
0 -1.5 -1.0
-0.5 0.0 0.5
BIAS OF AVERAGE
STAB 1 L 1 TY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 2.0
PA
2.0
1.5
0 I-"
P
1 "
£ ao
O-0.5
m-i.o
-1.5
-2.0
-2
RADISE S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=MPSDM
0 -1.5 -1.0
V
fj
U
-0.5 0.0 0.5
BIAS OF AVERAGE
S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 ti 2.0
PA
2.0
1.5
i ..o
1 "
a
r°
0-0.5
2
m-i.o
-1.5
-2.0
-2
kRADISE S02 HIGH 25 CONCENTRATIONS
YEAR= 1976 AVERAGING PERIOD- 1 MODEL=TEM-8A
0 -1.5 -1.0
L
V
-0.5 0.0 0.5
BIAS OF AVERAGE
S
N
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 2.0
Figure B.6 Fractional Bias By Model Using 25 High Values For Each Stability Class
-------
CO
1
o
PARADISE 502--
-HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL=MPTER
2.0
1.5
STD DEVIATION
g g 5
O-fl.5
m-l.o
-1.5
-10
-2
a -\i -1.0
H
M
L
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - UEOIUU
H - HIGH
1.0 1.5 2J>
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 MODEL-PPSP
2.0
U
l»
B"
0-0.5
n-1.0
-1.5
-2.0
-2
H
M
L
0 -1.5 -1.0
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - UEOIUU
H - HIGH
IJJ 1.5 2.0
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 MODEL-MPSDM
2.0
1.5
STD DEVIATION
g g S
fc-U
n-1.0
-1.5
-10
-2
I
0 -\A -1.0
M
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - UEDIUU
H - HIGH
1.0 1.5 10
PARADISE S02 HIGH 25 CONCENTRATIONS
2.0
1.5
g 1.0
1 «'
O
0-0.5
m-l.o
-1.5
-2.0
-2
YEAR-1976 AVERAGING PERIOD- 1 MODEL-TEM-8A
0 -1.5 -1.0
H
M
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - UEOIUU
H - HIGH
1.0 1.5 2.0
re^ B.7 Fractional Bias By Model Using 25 High Values For Each Wind Speed Class
-------
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD =1 MODEL=MPTER
2.0
1.5
0 U>
0-0.5
m-i.0
-1.5
-2.0
-2
0 -1J -1.0
-0.5 0.0 0.5
BIAS OF AVERAGE
2 1
a* *
"
STATIONS
1-8 - STNS 1-8
A-C - STNS 10-12
1.0 1.5 2.0
PARADISE S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 MODEL-PPSP
2.0
BIAS OF STD DEVIATION
Z £ g 5 Z
-1.5
-2.0
-2
a
i
0 -1.5 -1.0
7
8 6
«
2
A
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-8 - STNS 1-9
A-C - STNS 10-12
1.0 1.5 2.0
PARADISE S02 ALL CONCENTRATIONS
2.0
1.5
0 1-0
I "'
IS"'
0-0.5
0-1.0
-1.5
-10
-1
YEAR=1976 AVERAGING PERIOD=1 MODEL=MPSDM
0 -U -1.0
2
t* ^
w:
i
STATIONS
1-8 - STNS 1-8
A-C - STNS 10-12
-OS 0.0 03 1.0 1J 10
BIAS OF AVERAGE
PARADISE S02 ALL CONCENTRATIONS
2.0
1.5
BIAS OF STD DEVIATION
i
fa « g 5 S
-1.5
-2.0
-2
YEAR=1976 AVERAGING PERIOD- 1 MODEL- TEM-8A
0 -1.5 -1.0
*'A
* * '
STATIONS
1-8 - STNS 1-8
A-C - STNS 10-12
-0.5 0.0 0.5 1.0 I.S 2.0
BIAS OF AVERAGE
Figure B.8 Fractional Bias By Model Using All Paired Values For Each Station
-------
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=MPTER
2.0
1.5
O 1-0
I 0.5
O-0.5
m-l.O
-1.5
-2.0
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
-10 -1.5 -1.0 -0.5 0.0 0.5 1.0
BIAS OF AVERAGE
1.5
2.0
PARADISE S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-PPSP
J.O
IS
0 tJJ
I °-5
Q
e°-D
0-0.5
ffl-1.0
-1.5
-2.0
-2
VU
0 -1.5 -1.0
N
-0.5 0.0 0.5
BIAS OF AVERAGE
S
STABI L 1 TY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 2.0
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD- 1 MODEL-MPSDM
2.0
1.5
STD DEVIATION
g g s
0-0.5
0-1.0
-1.5
-2.0
-1
0 -1.5 -1.0
j
'
U
-0.5 0.0 0.5
BIAS OF AVERAGE
S
M
STABILI TY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 10
2.0
1.5
2
0 '<>
Q
0-0.5
in -1.0
-1.5
-2.0
-2
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD- 1 MODEL-TEM-8A
0 -1.5 -1.0
y
-0.5 0.0 0.5
BIAS OF AVERAGE
N
STABIL 1 TY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 2.0
^Figure B.9 Fractional Bias By Model Using All Paired Values For Each Stability Class
-------
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=MPTER
2.0
1.5
STD DEVIATION
g s s
U.
1-1.0
-1.5
-2.0
-2
0 -1.5 -1.0
-0.5 0.0 0.5
BIAS OF AVERAGE
L
M H
WINDS
L - LOW
U - UEDIUU
H - HIGH
1.0 1.5 2.0
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD" 1 MODEL=MPSDM
1.0
1.5
| IJ-
1 °'s
a
e°-°
li.
§
ffl-1.0
-1.5
-2.0
-2
0 -1.5 -1.0
L
M
-0.5 0.0 0.5
BIAS OF AVERAGE
H
WINDS
L - LOW
U - UEDIUU
H - HIGH
1.0 1.5 10
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD- 1 MODEL=PPSP
2.0
1.5
0 '0
P
I °-5
a
£ 0.0
in
0-0.5
ffl-1.0
-1.5
-2.0
-2
L
0 -1.5 -1.0
U H
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - UEDIUU
H - HIGH
1.0 1.5 2.0
PARADISE S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 MODEL-TEM-BA
2.0
1.5
!-
a
e-
0-0.5
m-1.0
-1.5
-2.0
-2
.0 -1.5 -1.0
M L
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
U - UEDIUU
H - HIGH
1.0 1.5 2.0
Figure B.10 Fractional Bias By Model Using All Paired Values For Each Wind Speed Class
-------
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 CATEGORY=ALL MET CONDITIONS
2.0
1.5-
1.0 '
i
-p*
5
o
i
u.
-0.5
-1.0 '
-1.5 '
-2.0 '
-«*
3 4 5 6 7 8 9 10 11 12 13
STATION DISTANCE DOWNWIND (Ml)
MODELS
X-MPSDM
D =PPSP
O =TEM-8A
Figure B.ll Fractional Bias Of The Average Vs Station Distance Using High 25 Values
-------
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1876 AVERAGING PERIOD-1 CATEGORY-VERY UNSTABLE
It II II 11
STATION DSTANCE DOKNWM) (U)
MODELS
X-ypsOU__
O -PPSP
O -TEU-BA
14 ^
14
14
H»
C""
-14
-14
-U
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-NEUTRAL
*x -»-Vr ~~*--\
><-.:.. \
f : '«j^:---^ *
^> ' -~Tr ^w-*
0 ^
a ^ ~ 5^"^
^ *~P u
STATION DISTANCE DOWNWIND (Ifl)
MODELS
f-yPTCR
X -UP SOU
a -PPSP
« -TEU-BA
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1876 AVERAGING PERIOD-1 CATEGORY-UNSTABLE
MODELS
±JdH!!E6._
x=MPSoy__
o -PPSP
O -TEU-BA
0 I I
» II II II
STATION DBTANCE OOWWIND (U)
14
14
i»
S-oj
-14
-14
-U
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR- 1876 AVERAGING PERIOD- 1 CATEGORY-STABLE
^Zli^^i^F^:
x ~ K \ ""-
x \ *
a \
a x \
\ 'S ^
o a
a
STATION DSTANCE DOWNWIND (U)
MODELS
+ -UP TEH
x>yp§Dii
a -PPSP
0 -TEU-BA
Figure B.12 Fractional Bias Of The Average Vs Station Distance By Stability Class Using High 25 Values
-------
00
1
_J
u
14
14
1-
!-«
-14
-10
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 CATEGORY-LOW WINDS
"*" +
/T~\*
x ./
N _v;\^ /s,
*Si_i«» HI * s
** a f *~ ^S'
.rrf^^" a " """"
STATION DISTANCE DOWNWIND (Ul)
-
MODELS
+ -UPTER
X-UPSDU
a -PPSP
0 -TEU-BA
1
14
14
!"
!--
-14
-10
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 CATEGORY-HIGH WINDS
\ * '
V\ x x ^-* ^ ._
^^^j^^^ir^'
* T * * "s^ ^^
" X or
« }r
o_ a ^/p
a »
STATION DBTANCE DOWNWIND (U)
MODELS
+ -UPTCR
X-UPSDU
a -PPSP
0 -TEU-BA
1
14
14 '
I"'
5 04
-14
-14 '
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR- 1076 AVERAGING PERIOD- 1 CATEGORY-MEDIUM WINDS
/
* * -'S*s
\ s ^-^ * J/S J*
*;-*$~1£*C* *^^
2^^
a
' >'« i
STATION DISTANCE DOWNWIND (U)
MODELS
+ -UPTER
X-UPSDU
a -PPSP
0-TEU-SA
Figure B.13 Fractional Bias Of The Average Vs Station Distance By Wind Speed Class Using High 25 Values _
-------
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 CATEGORY=ALL MET CONDITIONS
2.0
CO
345678
STATION DISTANCE DOWNWIND (Ml)
9 10 11 12 13
MODELS
X=MPSDM
D -PPSP
O =TEM-8A
Figure B.14 Fractional Bias Of The Average Vs Station Distance Using All Paired Values
-------
00
I
00
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-VERY UNSTABLE
MODELS
x=ypsou__
a -PPSP
O -TCU-BA
7 I I
STATION DISTANCE DOWNWIND (U)
10 II 11 U
IJ
IJ
Si«
I"**
-U
-U
-U
PARADISE S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-NEUTRAL
**>- ' i " i1*" ~~-.
lr---_ f . ..« ^
^>
s
fr
^o " "^% a
STATION DISTANCE DOWNWIND (U)
MODELS
+ -UPTER
x-ypsDy
a -PPSP
0 -TEU-8A
PARADISE S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-UNSTABLE
MODELS
O -PP8P
* -TtU-SA
0 I
II tl 11 II
STATION DISTANCE DOWWO (U)
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STABLE
U
IJ
IJ
M
OJ
-U
-14
-14
-U
MODELS
x-ijp8oy__
o -PPSP
0 -TEy-8A
4 I I 7 I I 10 II 11 tl
STATION DISTANCE DOWNWIND (U)
Figure B.15 Fractional Bias Of The Average Vs Station Distance By Stability Class Using All Paired Values
-------
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-LOW WINDS
14
14 '
14
CO
I
V-. * *
"*-**.
MODELS
x-upsoy __
a -PPSP
O -TCU-BA
10 It II 11
STATION DISTANCE DOHNIND (M)
PARADISE S02 -- ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-HIGH WINDS
14
14
14
-14
-14
-14
\
fc
0
snnoN DISTANCE DOVNWIND (U)
3 M
i
G
c"**
-14
-10
PARADISE S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-MEDIUM WINDS
k -J.^ ^
^^^^^"~~~~~-^^^^'^ /**
" ^ x >V o « x ^/
** _--*"l-^'
^^^^ O
^-^"p °
^
\ 1 1 1 I 1 I 1 1
STATION DISTANCE DOWNWIND (Id)
MODELS
t-UPTER
x-upsoy
a -PPSP
0 -TEy-SA
MODELS
-» -uPTCft
a -PPSP
« -TEU-8A
II 11 II II
Figure B.16 Fractional Bias Of The Average Vs Station Distance By Wind Speed Class Using All Paired Values
-------
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 CATEGORY=ALL STATIONS
2.0
CO
I
ro
o
-2.0 H
12
HOUR OF DAY
MODELS
MPER
PPSP
TEM-8A
Figure B017 Fractional Bias Of The Average Vs Hour Of The Day Using High 25 Values
-------
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 1
CO
I
t\J
HOUR OF DAY
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATIONS
MODELS
UPTER
UPSDU
PPSP
TEU-BA
MODELS
_UPTE_R
UPSDU
PPSP
TEU-BA
HOUR OF DAY
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 2
li
HOUR OF DAY
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 4
10
1.6
\£
&s
OJO
t-M
-ID
-u
T /\v
12
HOUR OF DAY
MODELS
UPTER
UPSDU
PPSP
TEU-BA
MODELS
UPTER
UPSDU
PPSP
TEU-BA
Figure B.18 Fractional Bias Of The Average Vs Hour Of The Day By Station Using 25 High Values
-------
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATIONS
HOUR OF DAY
ro
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 7
HOUR OF DAY
14
MODELS
UPTEjjt
UPSDU
PPSP
TEU-BA
MODELS
MPJEH
UPSDU
PPSP
TEU-BA
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-19 76 AVERAGING PERIOD-1 CATEGORY-STATION 6
HOUR OF DAY
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATIONS
HOUR OF DAY
MODELS
IIPTER_
UPSDU_
PPSP
TEM-8A
MODELS
IIPJER
UPSDU
PPSP
TEU-BA
Figure B.18 (Continued)
-------
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 9
U
HOUR OF DAY
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 11
MODELS
MPJ£R_ __
UPSDU
PPSP
IEU-BA
MODELS
JIPTER
ypsou
PPSP
TEU-8A
HOUR OF DAY
PARADISE S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 10
HOUR OF DAY
PARADISE SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 12
-U
II
HOUR OF DAY
MODELS
IJPTER
UPSOU
PPSP
TEi: 3A
MODELS
MPSoy
PPSP
Figure B.18 (Continued)
-------
CO
2.0
-2.0 1
PARADISE S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 CATEGORY=ALL STATIONS
MODELS
_MPTER__
MPSDM^
PPSP
TEM-BA
HOUR OF DAY
Figure B.I9 Fractional Bias Of The Average Vs Hour Of The Day Using All Paired Values
-------
PARADISE S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 1
CO
I
ro
en
HOUR OF DAY
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 3
HOUR OF DAY
MODELS
JJPTER
UPSOU
PPSP
TEU-aA
MODELS
UPIE_R
UPSDU
PPSP
TEU-aA
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATIONS
II
HOUR OF DAY
PARADISE S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 4
-IB
HOUR OF DAY
MODELS
MPTER__
UPSOU
PPSP
TEU-8A
MODELS
UPTER___
UPSOU
PPSP
TEU-aA
Figure B.20 Fractional Bias Of The Average Vs Hour Of The Day By Station Using All Paired Values
-------
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1076 AVERAGING PERIOD-1 CATEGORY-STATIONS
HOUR OF DAY
00
I
ro
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1876 AVERAGING PERIOD-1 CATEGORY-STATION 7
HOUR OF DAY
14
MODELS
IIPTEH
PPSP
TEU-8A
MODELS
UP SOU
PPSP
TEU-BA
PARADISE S02 ALL CONCENTRATIONS
YEAR-1076 AVERAGING PERIOD-1 CATEGORY-STATION 6
I*
HOUR OF DAY
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 8
HOUR OF DAY
Figure B.20 (Continued)
MODELS
KISS-..-
UPSDU
PPSP
TEU-8A
MODELS
MPTEft
UPSOU
PPSP
TEU-BA
24
-------
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 9
-10
HOUR OF DAY
ro
PARADISE S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 11
HOUR OF DAY
UOOELS
UPTER
UPSDU
PPSP
IEU-BA
MODELS
JIPJER
UPSOU
PPSP
TEU-BA
PARADISE SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 10
-U
HOUR OF DAY
PARADISE 502 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 12
HOUR OF DAY
UOOELS
UPTCR
UPSDU
PPSP
TEU-6A
MODELS
UPTER
UPSDU
PPSP
TEU-8A
(Figure B.20 (Continued)
-------
APPENDIX C MUSKINGUM RIVER PLANT
A sketch of the Muskingum River monitoring network is shown in Figure
C.I. The four monitoring stations are located at distances ranging from
approximately 4 km to 20 km from the plant. Elevations of the four receptors,
relative to the base of the stack, range from approximately 75 to 130 meters.
The results for both 1975 and 1976 are presented for Muskingum River,
primarily to provide a sense of the variation in performance that can be
expected from year to year. The figures that follow are presented in the
same sequence as for Kincaid and Paradise except results for both years are
presented within each figure. Results for each year will be discussed only
when differences appear to be worth noting.
From Figure C.2, it appears that the models as a group tend toward
larger underpredictions and smaller overprediction in 1976 compared with
1975. Generally, the points plotted in 1976 are shifted to the right rela-
tive to their position in 1975. While this shift toward underprediction
occurs consistently for the bias of the average, the bias in the scatter
occasionally shifts in the direction of greater overprediction in 1976.
For example, MPSDM is within a factor of two in 1975 for the scatter of
both 1-hour and 3-hour averages of the highest 25 values. In 1976 this is
not the case, as the fractional bias of the scatter for MPSDM falls outside
and below the value (less than -0.67) corresponding to overprediction exceed-
ing a factor of two.
In terms of the operational performance of the four models (Figures
C.2-C.4), it appears that PPSP is the most seriously biased, overpredicting
for most of the data categories. MPTER and MPSDM appear to be relatively
C-l
-------
unbiased while TEM-8A tends toward underpredictions. Exceptions to this rule
are evident for some averaging periods, data set types and years. The general
trends noted previously still hold, i.e., the longer averaging periods and
all data categories are associated with the largest underpredictions and
smallest overpredictions.
Figures C.5-C.1G delve deeper into the performance across the four
stations, the four stability categories and three wind speed categories.
For the 25 highest values, performance among the stations varies somewhat
(Figure C.5) but the pattern is fairly consistent among the four models.
Station 4 is associated with the largest underpredictions (and smallest
overpredictions). The spread in results among the four stations appears
smaller in 1976 than in 1975. Again, 1976 and 1975 show differences in
performance as exemplified by TEM-8A in which reasonably accurate predic-
tions in 1975 contrast with noticeable underpredictions in 1976. Differ
ences in performance among stability categories are also obvious (Figure C.6)
Generally, the stable categories are associated with underpredictions while
unstable categories are associated with relatively low bias. Differences in
performance among wind speed categories (Figure C.7) are relatively small for
MPTER but noticeably larger for the other three models. For the all data
category (Figures C.8-C.10), there is no particular feature that warrants
discussion other than to repeat that there is a general tendency toward
greater underprediction compared with the data set comprised of the 25 highest
values.
The next series of plots (Figures C.11-C.16) display the fractional
bias of the average as a function of station to source distance. Since only
C-2
-------
four stations are involved, the trend lines should be viewed with some caution.
Examination of these plots broken down by stability category (Figure C.12--
high 25 and Figure C.15--all data) reveals differences in performance among
the models as a function of station to source distance for neutral and stable
conditions. For very unstable and unstable categories, the trend line is
relatively flat for each of the models. Examination of these plots broken
down by wind speed (Figures C.13 and C.16), reveals no significant patterns,
probably due to the relatively few number of stations that are available.
Finally, Figures C.17-C.20 display the familiar diurnal pattern seen for
the other data bases. The models as a group show a tendency for underpre-
dictions in the early morning and late evening hours and a tendency for
relatively low bias in the mid-day hours. The tendency for underprediction
by TEM-8A and overprediction by PPSP during the mid-day hour is particularly
evident for the 25 highest stations where all stations are combined (Figure
C.17).
C-3
-------
C-4
-------
MUSKINGUM NETWORK
N
A
MUSKINGUM
CALDWELL 4
EL. 332 m
RICH VALLEY 3
EL.305 m I
HACKNEY
EL.275 m
POWER
PLANT
STACKS HEIGHT
ELEV.
\ - 448 m J
2 - 454 m
Figure C.I Field Monitoring Network Near The Muskingum River Plant
C--5
-------
MU!
u
u
a
E"
E
-13
-U
-2
5K1NGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD- 1
D
J H5 HO
X
-45 U U
BIAS OF AVERAGE
U U SI
UODELS
+- UPTER
X- HPSOM
n- PPSP
«» TEH-BA
K,
U
u
i u
!°
n
H5
-U
-J
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1875 AVERAGING PERIOD- 1
a
t HI H»
4*
-45 41 45
BIAS OF AVERAGE
UODELS
+- UPTER
X- HPSDU
n - PPSP
0- TEH-eA
U U U
MUSKINGUM S02HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PEROO-3
MODELS
+- ypTE*
X- UPSDM
a- WSP
« TEH-BA
u
u
u
I-
I HI
H9
-u
-*
USKINGUM S02 ALL CONCENTRATIONS
YEAR-1975 AVERAGING PEJ80D-3
n
1 HI HI
O
t*
-41 41 45
BIAS OF AVERAGE
UODELS
+- NPTER
X. tIPSOU
a PPSP
0- TEU-BA
U U U
MU!
u
i«
g"
O-05
1-u
-li
-«
-z
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERJOD-24
a
» -U HO
-45 41 45
BIAS OF AVERAGE
UODELS
+- MPTER
X- HPSDM
a - PPSP
« TEH-BA
IJ U U
b
11
u
0 «*
i u
S-41
-15
-U
-t
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1975 AVERAGING PERIOO-24
a
J HI -U
<
*
-45 41 41
BIAS OF AVERAGE
UODELS
+- UPTER
X- HPSOU
0 - PPSP
«» TEM-BA
u u u
Figure C.2 Fractional Bias By Averaging Period Using High 25 Values And All
Paired Values
C-6
-------
MU!
u
u
i **'
S-1S
1-u
-u
-u
-1
5KINGUM 302 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAfitNG PERIOD- 1
a
3-13-10
X
-u u u
BIAS OF AVERAGE
U U U
MODELS
+- MPTER
X. UP9DM
O- PPSP
0- TEM-SA
MUSKINGUM S02ALL CONCENTRATIONS
YEAR-1976 AVERMMG PBVOO-1
MODELS
+ - HPTER
X- MPSOH
a - PPSP
«- TEII-GA
-u
-t> U U
BIAS OF AVERAGE
u
u
u
MU!
u
u
§ u
i u
o
e-
6-15
I-U
-U
-u
-1
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERACD4C PEBOO-3
1 -U -U
a
f
X
-u u u
BIAS OF AVERAGE
»
MODELS
+- MPTER
X- HPSOM
a - PPSP
0. TEH-iA
U U 11
U
u
13
i u
a
S"
1-U
-U
-u
-J
USKINGUM S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PHWJO-3
1 -13 -U
a
+ X
-«3 U U
BIAS OF AVERAGE
0
MODELS
+- IIPTER
X. UPSDM
a - PPSP
0. TEH-BA
IJ 13 11
MU!
u
13
1 U
i u
i-l
6-ts
a-UD
-13
-U
-J
3KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAQNC PERIOO-24
J -U . -U
a
+
X
-U U 13
BIAS OF AVERAGE
«
MODELS
+- MPTER
X. MPSDH
O- PPSP
0- TEH-SA
IJ 13 U
M
U
13
0 U
i u
a
°
6-u
!.,
-u
-u
-1
USKINGUM S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PBBOD-24
1 -13 -tf
a
+:
-u u u
BIAS OF AVERAGE
0
MODELS
+- MPTER
x- HPSDH
O PPSP
0- TEU-SA
IJ U U
Figure C.2 (Continued)
C-7
-------
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD-1
3200
1MO
an
400
200
200
O
I
00
400
MODELS
f- UPTER
X- UPSOU
a - PPSP
O- TEU-BA
too
OBSERVED
1100
3200
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD-3
3200
200
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD-24
400
MODELS
+ - UPTER
X- UPSDU
a - PPSP
O- TEM-BA
MODELS
+ - MPTER
X. UPSDU
a - PPSP
O- TEU-BA
3200
Figure C.3 Predicted Quantiles Vs Observed Quantlles By Averaging Period Using High
25 Values
-------
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-=1976 AVERAGING PERIOD-!
3200
O
l
10
MODELS
+ - MPTER
X = UP SOU
a - PPSP
O = TEU-BA
3200
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-3
3200
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-24
MODELS
*-- UPTER
X = UPSDU
P - PPSP
O- TEU-8A
MODELS
+ - UPTER
X- UPSDU
D - PPSP
O- TEU-BA
3200
Figure C.3 (Continued)
-------
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1
10000
2
1000
lfl°
10
o
_J
o
MODELS
OBSERVED
JJPTER
UPSDU
PPSP
TEU-BA
65.5 88.5 97.7 99.7
CUUULATIVE FREQUENCY
n.M
it.ni
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=3
10000
n
c» 1000
*^
o
100
10
68.5
MODELS
OBSERVED
HPTEH
UPSOU
PPSP
TEU-BA
88.5 B7.7 M.7 M.M
CUUULATIVE FREQUENCY
99.999
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD-24
10000
1000
too
10
MODELS
OBSERVED
UPT_E_R
JJPSDU
PPSP
TEU-BA
65.5
88.5 97.7 99.7
CUMULATIVE FREQUENCY
99.98
99.999
Figure C.4 Cumulative Frequency Distributions By Averaging Period Using All Paired
Values
-------
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1
10000
10
2
O
o
o
o
1000
too
10
6U
o
MODELS
OBSERVED
JIPTER
UfSOy
PPSP
TEM-8A
88J 97.7 99.7
CUMULATIVE FREQUENCY
99.98
n.(M
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD=3
10000
o" 1000
100
10 H
MODELS
OBSERVED
ypsou
PPSP
TEU-BA
S8.S 17.7 N.7
CUMULATIVE FREQUENCY
99.98
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD-24
10000
MODELS
OBSERVED
UPSDU
PPSP
TEU-8A
U.S 97.7 99.7 99.98
CUMULATIVE FREQUENCY
99.999
Figure C.4 (Continued)
-------
o
ro
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 MODEL=MPTER
2.0
1.5
!-
a
P 0.0
in
0-0.5
m-i.o
-1.5
-10
-i
2
0 -M -1.0
4
3
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-4 - STNS 1-4
1.0 1.5 10
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD- 1 MODEL=PPSP
10
1.5
BIAS OF STD DEVIATION
5 g § g £
-1.5
-to
-I
3
2 1
0 -1.5 -1.0
4
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-4 _ STNS 1-4
1.0 1.5 2.0
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 MODEL=MPSDM
10
1.5
§ IJ
i «
a
0-0.5
m-1.0
-1.5
-10
-2
0 -1.5 -\S>
3
1
i
-0.5 110 0.5
BIAS OF AVERAGE
STATIONS
1-4 - STNS 1-4
1.0 1.5 10
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 MODEL~TEM-BA
10
1.5
a
0-0.5
m-1.0
-1.5
-2.0
-2
.0 -\S -1.0
3
4
1
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-4 - STNS 1-4
1.0 1.5 2.0
Figure C.5 Fractional Bias By Model Using 25 High Values For Each Station
-------
o
CO
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=MPTER
2X1
1.5
g U
P
§ a*
a
£ "
0-0.5
2
0-1.0
-1.5
-10
-10 -IJ -1.0
A^
2 '
-0.5 0.0 OJ
BIAS OF AVERAGE
STATIONS
1-4 - STNS 1-4
1.0 IJ 10
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=»1 MODEL=PPSP
10
1.5
0 U
p
Q
P 0X1
in
0-0.5
q
ID -1.0
-1.5
-10
1
2
-10 -IJ -1.0
4
3
-0.5 0.0 OJ
BIAS OF AVERAGE
STATIONS
1-4 - STNS 1-4
1.0 1.5 10
-
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=MPSDM
10
U
1 IJ
P
! «
Q
to M
b-OJ
3
ffl-IXI
-IJ
-10
-10 -IJ -IJ
2 *
-OJ 0X1 0.5
BIAS OF AVERAGE
STATIONS
1-4 - STNS 1-4
1.0 tJ 10
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD- 1 MODEL=TEM-8A
10
g «
P
Q
$ °*
b-OJ
3
-ij
-10
-10 -1.5 -1.0
1
4
2
-0.5 0.0 OJ
BIAS OF AVERAGE
STATIONS
14- STNS 1 4
1.0 1.5 10
Figure C.5 (Continued)
-------
O
MUSKINGUM S02HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD-1 MODEL-MPTER
10
14
I"
iu
I"
fe-04
S-,4
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
-10
-15
-1.0
-05 1.0 04
BIAS Or AVERAGE
tJ>
14
U
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 MODEL-PPSP
10
14
0 1'
| 05
iu
&-05
g
B-14
-15
-U
-1
V
S
ja -15 -1.0
-04 04 05
BIAS Or AVERAGE
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
14 15 10
MU<
U
15
§ 14
i-
a
"
0-«5
IH.
-15
-U
-i
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR- 19 75 AVERAGING PERIOD- 1 MOOEL-MPSDM
a -15 -1.0
V
U
-u u as
IAS OF AVERAGE
S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 1.5 10
MUS
U
15
1 M
1 "
g U
O-45
m-M
-15
-U
-I
5KINGUM S02 HIGH 25 CONCENTRATIONS
(TEAR- 1975 AVERAGING PERIOD- 1 MODEL-TEM-8A
JO -\* -1.0
II
i
-05 OH 0.5
BIAS OF AVfRAOE
S
N
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 15 2.0
Figure C.6 Fractional Bias By Model Using 25 High Values For Each Stability Class
-------
MU!
10
1.5
1 "
0.0
0^5
-10
-I
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 MODEL-MPTER
/t -1.5 -U)
U
V
N
S
-45 M Hi
BIAS OF AVERAGE
1.0 U It
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
MU;
10
1.5
§ 1.0
I «'
o
ll.
o-u
-10
-I
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 MODEL-PPSP
N
0 -1.5 -U
"S
-as 0.0 o.s
BIAS OF AVERAGE
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
1.0 U 10
5 DEVIATION
: s s s e i
E "
O-OJ
0-1.0
-u
-10
-1
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 MODEL-MPSDM
a -1.5 -1.0
V
U
N
-041 OJ OJ
BIAS OF AVERAGE
S
1.0 1.5 10
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
MU;
10
\s
BIAS OF STD DEVIATION
S £ g S 5
-U
-10
-1
5KINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-TEM-BA
41 -1.5 -U
u
-0.5 0.0 0,5
BIAS OF AVERAGE
s
N
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
U U 10
Figure C.6 (Continued)
-------
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIODS MODEL=MPTER
10
§ U)
1 0.5
BIAS OF STD 01
4 B E
-10
-i
»
a -13 -u
L H
-05 00 05
BIAS OF AVERAGE
WINDS
L - LOW
y - UEDiuy
H - HI OH
1.0 13 U
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD- 1 MODEL-PPSP
10
13
STD DEVIATION
g S 5
0-05
ffl-IJ)
-13
-10
-1
-
M H
L
a -1.5 -u)
-0.5 0.0 0.5
BIAS OF AVERAGE
WINDS
L - LOW
y - UEDIUU
H - HI OH
IJ) 15 10
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR~1975 AVERAGING PERIOD^ 1 MODEL=MPSDM
U
1"
§ "
a
6"
0-05
I-U)
-U
-J
J) -13 -\J>
M
L
H
-03 00 U
BIAS OF AVERAGE
WINDS
L - LOW
y - UEOIUU
H - HI OH
IJ) 13 U
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD- 1 MODEL-TEM-BA
U
13
STD DEVIATION
g g 5
0-05
m-1.0
-10
-2
J -13 -IJ)
..
M
H
L
-03 0.0 03
BIAS OF AVERAGE
WINDS
L - LOW
y - yEDiuy
H - HI OH
IJ) 13 10
Figure C.7 Fractional Bias By Model Using 25 High Values For Each Wind Speed Class
-------
MU:
10
13
F
§ "
Q
eoo
ffl-13
-13
-10
-2
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR" 1976 AVERAGING PERIOD- 1 MODEL-MPTER
3 -13 -13
MU!
10
13
0 1-0
P
i u
Q
&"
O-03
a -13
-10
-i
1
1
-03 03 OS
BIAS OF AVERAGE
WINDS
L - LOW
M - MEDIUM
H - HIGH
13 13 10
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-PPSP
M
L
0 -13 -1.0
H
-03 03 03
BIAS OF AVERAGE
WINDS
L - LOW
M - MEDIUM
H - HIGH
13 13 10
5 DEVIATION
» O M ^
. 5 h. ta b c
C IUI
n
0-03
0-13
-13
-10
-1
5KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 MODEL-MPSDM
L
a -13 -13
MU!
10
13
§ 13
0-03
n-13
-13
-23
-2
M
-03 00 03
BIAS OF AVERAGE
WINDS
L - LOW
M - MEDIUM
H - HIGH
13 13 10
3KINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 MODEL-TEM-8A
a -13 -13
i
M
-03 0.0 OS
BIAS OF AVERAGE
H
WINDS
L - LOW
M - MEDIUM
H - HIGH
13 13 10
Figure C.7 (Continued)
-------
<->
00
fv
2.0
1.5
0 '<>
a
P 0.0
in
0-0-5
-1.5
-2.0
-2
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 MODEL=MPTER
0 -1.5 -1.0
4
2
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-4 - STNS 1-4
1.0 1.5 10
MUSKINGUM S02ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 MODEL=MPSDM
s
g
P 0.0
0-0.5
0-1.0
-1.5
-2.0
STATIONS
1-4 - STNS 1-4
-10 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 10
BIAS QF AVERAGE
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 MODEL=PPSP
10
1.5
STD DEVIATION
g g 5
0-0.5
in -1.0
-1.5
-2.0
-2
3
2
0 -1.5 -1.0
4
1
-0.5 0.0 0.5
BIAS OF AVERAGE
STATIONS
1-4 - STNS 1-4
1.0 1.5 2.0
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 MODEL=TEM-8A
10
1.5
STD DEVIATION
§ g 5
0-0.5
-1.5
-2.0
-2
0 -1.5 -1.0
4
3
-0.5 0.0 0.5
BIAS OF AVERAGE
1
STATIONS
1-4 - STNS 1-4
1.0 1.5 10
Figure C.8 Fractional Bias By Model Using All Paired Values For Each Station
-------
o
I
10
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=MPTER
10
15
3 >.o
s
if. M
Q
0-0.5
* J
-1.5
-10
-10 -1.5 -1.0
1 1
4 2
STATIONS
14** STNS 1-4
-05 0.0 as 1.0 15 10
BIAS OF AVERAGE
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=PPSP
z ,.
o '»
p
3
£i °'5
a
£ "
ll.
* IU
in -1.0
2
-2.0 -15 -1.0
3
1
4
STATIONS
1-4 - STNS 1-4
-0.5 0.0 0.5 1.0 1.5 2.0
BIAS OF AVERAGE
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR= 1976 AVERAGING PERIOD=1 MODEL<=MPSDM
15
I
P
<
§ OA
O
ll
0-05
* .J
ta-w>
-.5
10
-10 -15 -1.0
r
STATIONS
1-4 *> STNS 1-4
-05 0.0 05 HI 15 10
BIAS OF AVERAGE
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 MODEL=TEM-8A
2
o u
p
i .
% 05
[j
a
r-u
u.
0-0.5
2! 1B
m-1.0
-10 -1.5 -1.0
2'*
STATIONS
1-4 - STNS 1-4
-05 0.0 05 1.0 15 10
BIAS OF AVERAGE
Figure C.8 (Continued)
-------
0
1
ro
0
H
u
14
| 14
| U
g 04
-14
-14
-1
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD-1 MODEL-MPTER
\
4 -14 -IJI
-O.S 04 04
BIAS Or AVERAGE
N
S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
t - STABLE
14 14 10
^ 32233322S1
T T i i
NouviAaa OLE jo cvie
USKINGUM S02 ALL CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD- 1 MODEL-PPSP
V
14 -14 -14
-0.5 04 04
BIAS OT AVERAOE
S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N NEUTRAL
S - STABLE
1.0 14 10
M
14
14
§ 14
i u
b
0-«5
0-1.0
-14
-14
-1
USKINGUM S02 ALL CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD- 1 MODEL=MPSDM
0 -14 -14
,, U
V
-04 04 04
BIAS Or AVERAOE
S
N
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
14 14 10
3223§3253
Till
NOUYI/N3O OiS JO SVI8
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD- 1 UODEL-TEM-8A
.0 -14 -14
V U
-04 04 OJ
BIAS Or AVERAOE
S
N
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
14 14 10
Figure C.9 Fractional Bias B> Model Using All Paired Values For Each Stability Class
-------
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-MPTER
10
13
BIAS OF STD DEVMTION
s fc e e s
-23
-2
3 -13 -13
V
U
-03 00 05
BIAS OF AVERAGE
N S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
13 13 10
BIAS OF STD DEVIATION
s £ g e s s e
-10
-I
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-PPSP
N
0 -13 -13
vu
-03 03 03
BIAS OF AVERAGE
S
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
13 13 10
MUSKINGUM S02ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 MODEL-MPSDM
to
1.5
-
O-O5
-UI
-1.5
-U
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
-10 -13 -13 -03 00 03
BIAS OF AVERAGE
U)
13
U
MUSKINGUM S02 ALL CONCENTRATIONS
u
1.5
STD DEVIATION
g e s
0-05
0-1.0
-U
-10
-2
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-TEM-8A
0 -13 -1J)
V
U
HU OJ) OS
BIAS OF AVERAOE
S
N
STABILITY
V - VERY UNSTABLE
U - UNSTABLE
N - NEUTRAL
S - STABLE
U> 13 10
Figure C.9 (Continued)
-------
o
1
ro
ro
U
10
13
0 1-0
P
0-03
n-tJ)
-13
-2J)
-2
USKINGUM S02 ALL CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 MODEL-MPTER
.0 -13 -1.0
M
H
-03 OJ) 03
BIAS OF AVERAGE
WINDS
L - LOW
U - MEDIUU
H - HIGH
IJ) 13 10
Iv
10
13
BIAS OF STD DEVIATION
I I i I
g * s e g B s
U
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 MODEL-PPSP
M L
J> -1.5 -1.0
hi
-0.5 0.0 03
BIAS OF AVERAGE
WINDS
L - LOW
M - MEDIUU
H - HIGH
IJ) 13 2.0
K/
10
13
1"
I "
0-45
1-1J,
-10
-2
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 MODEL-MPSDM
J> -13 -ID
1
| L
-03 OO 03
BIAS OF AVERAGE
H
WINDS
L - LOW
U - UEDIUU
H - HIGH
IJ) 13 10
K/
10
13
Q
£ *"
0-03
ffl-U>
-1.6
-10
]
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 MODEL-TEM-8A
J) -1.5 -IJ)
M
-as 0.0 as
BIAS OF AVERAGE
H
WINDS
L - LOW
U - UEDIUU
H - HIGH
IJ) 13 10
Figure C.10 Fractional Bias By Model Using All Paired Values For Each Wind Speed Class
-------
K/
U
1.5
BIAS OF STD DEVIATION
* & g g 5
-U
-10
-1
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-MPTER
a -u -1.0
H
M
-OS &0 U
BIAS OF AVERAGE
L
WINDS
L - LOW
U - UEDIUU
H - HI OH
1.0 M 10
K/
U
U
1"
§ "
I"
0-OJ
L.
-u
-10
-I
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-MPSDM
0 -U -14
M
L
-05 flj) 05
BIAS OF AVERAGE
H
WINDS
L - LOW
U - MEDIUM
H - HI OH
14 U 10
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 MODEL-PPSP
10
1.5
0 '*
I "
Q
0-05
ffl-IJ)
-10
-1
M
a -1.5 -u)
H
L
-0.5 00 O5
BIAS OF AVERAGE
WINDS
L - LOW
U - UEDIUU
H - HI OH
U) 13 10
3 DEVIATION
! 5 5 5 E
is ""
0-0.5
S-u
-us
-M
-1
IUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 UODEL-TEM-8A
A -1.5 -|jO
M
-05 0.0 05
BIAS OF AVERAGE
H
L
WINDS
L - LOW
y - yEoiuu
H - HI OH
1.0 U U
Figure C.10 (Continued)
-------
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 CATEGORY=ALL MET CONDITIONS
2.0
o
I
-2.0 H
0 1 2 3 4 5 6 7 8 9 10 11 12 13
STATION DISTANCE DOWNWIND (Ml)
MODELS
X=MPSDM
D «PPSP
O =TEM-8A
Figure C.ll Fractional Bias Of The Average Vs Station Distance Using High 25 Values
-------
o
i
ro
tn
MUSKINGUM S02rHIGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 CATEGORY=ALL MET CONDITIONS
2.0
1.5
1.0
o
o
-0.5 '
-1.0 '
-1.5 '
-2.0 1
1 2 3 4 5 6 7 8 9 10 11 12 13
STATION DISTANCE DOWNWIND (Ml)
MODELS
D -PPSP
O =TEM-8A
Figure C.ll (Continued)
-------
O
I
ro
14
14
!«
3 **
2-oj
-14
-14
-14
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 CATEGORY-VERY UNSTABLE
,- -^
*
STATION INSTANCE DONMMNO (U)
unnn 9
+ -yPTER
X-UPSDM
a -PPSP
« -TEM-8A
u
I-
o **
-14
-to
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 CATEGORY-NEUTRAL
XT"""-
x^*^^ -
; -..__-.-.-
^
o^r--^'
lltflllllfll
STATION DISTANCE DOWNWIND (III)
MODELS
+ -UPTER
x-ypsoy
a -PPSP
« -TEU-SA
14
14
l»
8 *
g-M
-14 '
-li
-U
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD- 1 CATEGORY-UNSTABLE
M
i^sf^... ...--"
a
STATION DISTANCE KNIMMND (U)
MODELS
+ -UPTCH
x-ypsou
a -PPSP
O -TEy-BA
1
14
14
3I«
5 lui
-14
-24
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD- 1 CATEGORY-STABLE
""^^^v..^ -.
^-_____.
V
_^^
S 5~ sv
/ *
a
9 1 " 1 1 4 6 7 1 1 10 11 11 1
STATION DISTANCE DOWNWIND (U)
MODELS
f-yPTER
x-ypsoy
a -PPSP
0 -TEU-BA
1
Figure C.12 Fractional Bias Of The Average Vs Station Distance By Stability Class Using High 25 Values
-------
O
ro
IJ
U
!«
|-u
-U
-u
-ro
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-VERY UNSTABLE
J^^f, ~:^^^
*
STATION DISTANCE DOWNWIND (11)
MODELS
+ -UPTER
X-MPSDM
D -PPSP
0 -TEU-6A
IJ
IJ
g»
5 M
-ij)
-u
-10
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-NEUTRAL
r^v~::
X ""-^_ *^«
^ -::::::^
~*S^^
STATION DISTANCE DOWNWIND (Ul)
MODELS
X -UPSOM
a -PPSP
O -TEU-8A
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-UNSTABLE
MODELS
X-MPSDII __
o -PPSP
« -TCU-SA
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOO-1 CATEGORY-STABLE
U
M
t-M
-1.0
-IJ
-U
MODELS
X-IIPSDII_
a -PPSP
0 -TEM-8A
10 II 11 13
STATION DISTANCE DOWNWIND (ID)
Figure C.12 (Continued)
-------
o
00
u
IJ
IJ
3
a u
I M
K
|-OJ
-u
-u
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR- 1075 AVERAGING PERIOD- 1 CATEGORY-LOW WINDS
-,_,^.r-=-r-^^. .
-.-' _. J ' ^
*- V ___ __ -
,_ * " ____J
if"*"" -J>- ^
"
. ^^~*
s ~^_ %>*-~> _ -ZZZZZ-*
r
o
STATION DISTANCE DWMHND (tl)
MODELS
* -UP TEH
X-UPSOU
a -PPSP
O -TEU-BA
IJ
IJ
1"
|U
-u
-10
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR- 1075 AVERAGING PERIOD- 1 CATEGORY-MEDIUM WINDS
* V»,_ ___--'"" ^
* - "^^ . '",: ^ag^8^^
0
i i i . i i i i
STATION DSTANCE DOTMAMND (U)
MODELS
-f-UPTER
X-UPSOU
a -PPSP
0 -TEU-BA
Figure C.I3 Fractional Bias Of The Average Vs Station Distance By Wind Speed
Class Using High 25 Values
-------
o
1
ro
14
14
l»
0 "
5-oj
-14
-u
-24
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-LOW WINDS
* *-.
*
<^^ ^___^~
a
STATION DISTANCE DOWNWIND (U)
MODELS
-I--UPTER
X -UP SOU
a -PPSP
O -TEU-8A
1
14
"'
1"
§ M
|-04
-14
-14
-24
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-HIGH WINDS
» *"
^"^^^^^^^^
*ii ^^
STATION DBTANCE DOWNWIND (U)
MODELS
+ -UPTER
a -PPSP
0 -TEU-BA '
1
U
14
I"
^
g *4
|HJJ
-14
-14
-24
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-MEDIUM WINDS
* ..--*
« > -._.___ ""
M ^
^^^..^^^^
a
STATION DISTANCE DOWNWIND (ID)
MODELS
+ -UPTER
X;MPSDM
a -PPSP
0 -TEU-8A
Figure C.I3 (Continued)
-------
o
I
00
o
MUSKINGUM S02ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 CATEGORY=ALL MET CONDITIONS
2.0
1.5
1.0
m
0.5
g -
u
£-0.5 1
u.
-1.0
-1.5 '
-2.0 i
^
5 .« "" - -
*""-- "~ ~~ '
1 2 3 4 5 6 7 8 9 10 11 12 13
STATION DISTANCE DOWNWIND (Ml)
MODELS
>=MPSDM
D "PPSP
O =TEM-BA
Figure C.14 Fractional Bias Of The Average Vs Station Distance Using All Paired Values
-------
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 CATEGORY=ALL MET CONDITIONS
2.0
o
CO
-2.0 1
345678
STATION DISTANCE DOWNWIND (Ml)
10 11 12 13
MODELS
>=MPSDM
D »PPSP
0-TEM-8A
Figure C.14 (Continued)
-------
o
I
CO
ro
14
14
i-
Q *
|-u
-IJ
-14
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1875 AVERAGING PERIOD- 1 CATEGORY-VERY UNSTABLE
__^
r'L _... ~^r=zz
^f
a
STADON DISTANCE DOWNWIND (U)
unnn Q
f-MPTER
X-yPSDy
D -PPSP
« -TEU-8A
14
14
l»
|-4J
-IJ
"10
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 CATEGORY-NEUTRAL
X"
* _-->-.,..
» ^'"^
" - ^^*
_- -"^
^
a
i i i i i i
STATION DISTANCE DOWNWIND (U)
MODELS
+ -HPTER
X -UPSOM
a -PPSP
0 -TEM-BA
14
14
I"'
a "J>
g-4t
-14
-14
-U
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 CATEGORY-UNSTABLE
$
--*-»
:--T-^ ^
s^ ^****iSS. ""*» ~~~~'~i^^%^^
a ""*
STATION D6TMKE DOWNWIND (UQ
MODELS
+ -yPTER
" ' '
X-UPSDU
a -PPSP
0 -TEM-8A
14
14
1"
g-04
-14
-14
-14
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-197S AVERAGING PERIOD- 1 CATEGORY-STABLE
^_^-~
°/^~a ^\
/ \
a
STATION DISTANCE DOWNWIND (Id)
MODELS
+ -yPTER
x-ypsoy
a -PPSP
0 -TEM-BA
Figure C.15 Fractional Bias Of The Average Vs Station Distance By Stability Class Using All Paired Values
-------
O
I
U)
oo
u
u
I"
5 **
JS-M
-14
-14
-u
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-VERY UNSTABLE
..'"Jl
»T ^~'-v. ""^^^-*"*
^ ^~*^1fSS^rr -^ZZ^s^
STATION DISTANCE DOVMHND (U)
MODELS
+ -MPTER
x-ypsou
a -PPSP
O -TEH-8A
IJ
14
i»
8 *
|-U
-IJ
-10
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD- 1 CATEGORY-NEUTRAL
4^\- -.
* ^*\_ "*"*"*--
^- -^ "
\^ "^
~^^"**~^l
^~ "
STATION DISTANCE DOWNIMD (U)
MODELS
+ -UPTER
X-UPSOU
a -PPSP
0 -TEM-BA
!
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-UNSTABLE
-to
114
STATION DISTANCE DOWNWIND (II)
10 II u 11
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STABLE
U -
U
-14
-u
-ta
o i
it 11 it u
STATION DISTANCE DOWNWIND (U)
MODELS
x-yp8j>y__
a -PPSP
« -TCU-1A
MODELS
+_-ttPT£H__
X-IIPSOM__
O -PPSP
O -TEM-8A
Figure C.I5 (Continued)
-------
MUSKINGUM SO2 -- ALL CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 CATEGORY-LOW WINDS
u
1.6
-U
MODELS
t.^PT£R-_
X-UPSDV
o -PPSP
« -TEy-SA
i i
O
CO
-F*
4 I 0 7 I 0 10 II » U
STATION USTANCE DOWNWIND (U)
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-1075 AVERAGING PERIOD-1 CATEGORY-HIGH WINDS
U
IJ
1.0
-u
STATION DISTANCE OOWNWD (U)
U
14
!«
2 -04
-u
-u
10
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR- 1975 AVERAGING PERIOD- 1 CATEGORY-MEDIUM WINDS
+
'>$==5&i.- ...-""^
s ^^^^^i-j.:^.'----:^"--^^
a
''''''
STATION DISTANCE DOWNWIND (U)
MODELS
+ -HPTER
X -UP SDH
D -PPSP
« -TEU-8A
MODELS
x>tipsoy __
a -PP»P
O -TCM-SA
10 It II II
Figure C.16 Fractional Bias Of The Average Vs Station Distance By Wind Speed Class Using All Paired Values
-------
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-LOW WINDS
-
-14
-14
MODELS
-hj-MPTER__
x-ypsoy__
O -PPSP
«-TEU-6A
I
it 11 a n
o
en
STATION DISTANCE DOWNWIND (U)
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-HIGH WINDS
U>
U
|JO
M
U
-U
-ro
- , . i at ^^^ ~""***^»
"~^
MODELS
x=upsoy__
P -PP8P
« -TEU-SA
W II 11 M
STADON DISTANCE DOWNWIND (W)
IJ
1.0
K»
s M
2-u
-14 <
-14
-It
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR- 1976 AVERAGING PERIOD- 1 CATEGORY-MEDIUM WINDS
t »
-^nr±5i.Vs«._.^.^IU,3_. ... ii^ll^ *
^-^ ^j.
" "
/ u
a
STA1KW DISTANCE DOWNIMD (Ul)
MODELS
+ -UPTER
x.upsoy
a -PPSP
<> -TEU-8A
Figure C.16 (Continued)
-------
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 CATEGORY=ALL STATIONS
-2.0 1
12
HOUR OF DAY
18
24
MODELS
MPTER
MPSDM^
PPSP
TEM-8A
Figure C.17 Fractional Bias Of The Average Vs Hour Of The Day Using High 25 Values
-------
r>
-2.0 1
MUSKINGUM S02 HJGH 25 CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 CATEGORY=ALL STATIONS
MODELS
MPTER
MPSDM^
PPSP
TEM-8A
HOUR OF DAY
Figure C.17 (Continued)
-------
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD-1 CATEGORY-STATION 1
O
I
00
HOUR OF DAY
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1073 AVERAGING PERIOD-1 CATEGORY-STATION 3
HOUR OF DAY
MODELS
_UPT_ER_
UPSDU
PPSP
TEU-BA
MODELS
UPTEJS
UPSDU
PPSP
TEU-BA
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD-1 CATEGORY-STATION 2
HOUR OF DAY
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-107S AVERAGING PERIOD-1 CATEGORY-STATION 4
HOUR OF DAY
MODELS
MPTM
UPSDU
PPSP
TEU-IA
MODELS
JJPJER
UPSDU
PPSP
TEU-BA
Figure C.18 Fractional Bias Of The Average Vs Hour Of The Day By Station Grouping Using High 25 Values
-------
MUSKINGUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 1
O
I
CO
UD
HOUR OF DAY
MUS GUM S02 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 3
MODELS
JIPTER
UPSDU
PPSP
TEU-8A
MODELS
UPTER
UPSDU_
PPSP
TEU-BA
HOUR OF DAY
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 2
HOUR OF DAY
MUSKINGUM SO2 HIGH 25 CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 4
MODELS
_UPT_ER____
UPSOU
PPSP
TEU-BA
MODELS
UPTE_R
UPSOU
PPSP
TEU-BA
HOUR OF DAY
Figure C.18 (Continued)
-------
o
o
2.0
-2.0 1
MUSKINGUM S02ALL CONCENTRATIONS
YEAR=1975 AVERAGING PERIOD=1 CATEGORY=ALL STATIONS
MODELS
MPTER
PPSP
TEM-8A
HOUR OF DAY
Figure C.I9 Fractional Bias Of The Average Vs Hour Of The Day Using All Paired Values
-------
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR=1976 AVERAGING PERIOD=1 CATEGORY=ALL STATIONS
o
m
-0.5 '
-1.0
-1.5
-2.0 H
12
HOUR OF DAY
18
24
MODELS
_M_PTER
MPSD)y[
PPSP
TEM-8A
Figure C.I9 (Continued)
-------
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-1970 AVERAGING PERIOD-1 CATEGORY-STATION 1
HOUR OF DAY
ro
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 3
HOUR OF DAY
MODELS
_UPTER_
UPSDU
PPSP
TEU-IA
MODELS
.UPTER
UPSDU
PPSP
TEH-aA
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-1975 AVERAGING PERIOD-1 CATEGORY-STATION 2
II
HOUR OF DAY
14
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 4
-U
HOUR OF DAY
MODELS
JJPTEJL
UPSDU
PPSP
TCU-6A
MODELS
JIPUR
UPSDU
PPSP
TEU-SA
Figure C.20 Fractional Bias Of The Average Vs Hour Of The Day By Station Grouping Using All Paired Values
-------
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 1
-U
HOUR OF DAY
MUSKINGUM S02 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 3
HOUR OF DAY
MODELS
IIPTJER_
upsoy
PPSP
TEM-BA
MODELS
JJPTEJS
UPSDU
PPSP
TEU-6A
MUSKINGUM :SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 2
HOUR OF DAY
MUSKINGUM SO2 ALL CONCENTRATIONS
YEAR-1976 AVERAGING PERIOD-1 CATEGORY-STATION 4
-U
HOUR OF DAY
Figure C.20 (Continued)
MODELS
UPSDU
PP8P
TCU-tA
MODELS
JJPTE_R_
UPSDM
PPSP
TEU-6A
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing]
1. REPORT NO. ' 2.
EPA-450/4-83-003e
4. TITLE AND SUBTITLE
Evaluation of Rural Air Quality Simulation Models
Addendum E: Gr.aphic Summary of the Performance of
Rural Air Quality Models
7. AUTHOH(S)
William M. Cox Joseph A. Tikvart
Gerald K. Moss
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Source Receptor Analysis Branch
Monitoring and Data Analysis Division
U. S. Environmental Protection Agency
12. SPONSORING AGENCY NAME AND ADDRESS
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
July 1987
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO. .
11. CONTRACT/GRANT NO.
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA-450/4-83-003e
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This addendum uses a variety of graphic formats to display and compare the composite
performance of four rural models using data bases assembled at Clifty Creek, Muskingum
River, Paradise and Kincaid. The four models included MPTER (EPA), PPSP (Martin -
Marietta Corp.) MPSDM (ERT) and TEM-8A (Texas Air Control Board). Three Appendices to
this addendum present the model performance separately for Kincaid (Appendix A),
Paradise (Appendix B) and Muskingum River (Appendix C) using a graphic format similar
to that used previously for Clifty Creek (Addendum B). Composite measures of perfor-
mance are constructed using the information from each data base and presented as frac-
tional bias plots where the composite fractional bias of the average vs fractional bias
of the standard deviation is shown by averaging period, and selected meteorological
categories. A net fractional bias is displayed for each model and averaging period
which combines the fractional bias of the average and standard deviation, averaged over
the data bases. The results indicate, inter alia, a consistent and significant bias
(overprediction) for PPSP, a decreasing bias with increasing averaging period for MPSDM,
and an increasing bias with averaging period for TEM-8A. Overall, MPTER indicated the
most consistent performance and had the smallest composite bias among the four models.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution
Mathematical Modeling
Meteorology
Sulfur Dioxide
Statistical Measure
Performance Evaluation
18. DISTRIBUTION STATEMENT
b. IDENTIFIERS/OPEN ENDED TERMS
Air Quality Impact
Assessment
19. SECURITY CLASS (This Report I
Unclassified
20. SECURITY CLASS /This page)
Unclassified
c. COSATl Field/Group
21. NO. OF PAGES
22. PRICE
EPA Form 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
------- |