OCR error (C:\Conversion\JobRoot\00000AXF\tiff\2000XGBK.tif): Unspecified error
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S. Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are.
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8 "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects, assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-80-008
January 1980
POWER PLANT STACK PLUMES IN COMPLEX TERRAIN
Data Analysis and Characterization of Plume Behavior
Kenneth E. Pickering
Robert H. Woodward
Robert C. Koch
GEOMET, Incorporated
15 Firstfield Road
Gaithersburg, Maryland 20760
Contract No. 68-02-2260
Project Officer
George C. Holzworth
Meteorology and Assessment Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Science Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
n
-------
ABSTRACT
Data collected during a 16-month field program conducted near the Clinch
River Steam Plant in western Virginia were analyzed to characterize the
behavior of a power plant plume in complex terrain. Onsite measurements and
synoptic scale meteorological data are summarized for 31 periods of significant
S02 measurements and similar meteorological conditions.
Statistical analyses of S02, NO and NOx concentrations at fixed
monitoring sites and in one of the two plant stacks revealed significant
diurnal and seasonal variations. Ridge sites showed maximum concentrations at
night but without a well defined diurnal pattern; valley sites showed a well
defined late morning maximum hypothesized to be caused by plume fumigation.
Cross-section pollutant and meteorological measurements made from a helicopter
during July were analyzed to show the dimension and height of the plume as a
function of distance from the plant. Plume heights calculated from wind and
temperature profiles and plume widths calculated from the standard deviation of
wind direction produced improvements over standard estimates. The influence of
terrain features on the shape and path of the plume is clearly shown when
assessing the helicopter data and selected case studies of ground monitoring
data. Uncertainties in hourly measurements of sulfate due to laboratory
assessment accuracy, random natural deposition, duration of ambient exposure,
and other unexplained factors are undesirably large; however, there is evidence
in the measurements of contributions from both the nearby plant and more
distant sources.
Parameters for a Gaussian plume model were modified on a stepwise basis of
considering the impact of site specific meteorological and terrain data. The
study showed that the standard flat terrain model and modifications frequently
used to represent complex terrain influences can be improved by using data
available for the Clinch River site. The Gaussian plume model can provide
useful estimates of maximum concentrations to be expected, but it can not
represent on an hour-to-hour basis the influences of the complicated flow
found in a complex terrain setting. The data obtained for tie Clinch R-jvpr
site provide a useful record of the physical behavior of a power plant plume
in complex terrain and will be useful in testing hypotheses of plume behavior
and in diagnosing plume behavior.
This report was submitted in fulfillment of Contract No. 6H--0?-??60 by
Geoiiict, IIK. under the '.porisorr.hip of the II.'.. I nviroiiiiieiil.-i I lY
-------
CONTENTS
Abstract iii
Figures vi
Tables xiii
1. Introduction 1
2. Summary and Conclusions 3
3. Description of the Clinch River Power Plant,
Terrain and Monitoring Network .... 6
4. Data Period Analysis 13
5. Statistical Analysis of Plant Emissions and
Measured Ambient Pollutant Concentrations 24
6. Analysis of Plume Structure 50
7. Development of Profiles of Wind and Temperature,
Plume Heights, and Plume Trajectories 82
8. Analysis of Plume Impact in Relation to
Meteorological Conditions 99
9. Sulfate Analysis 171
10. S02 Dispersion Analysis 187
11. Application of Dispersion Model at Another Location. . . . 218
References 223
Appendices
A. Frequency Distributions and Mean Concentrations of
Mobile Van Data 225
B. Frequency Distributions of Meteorological Data 241
C. Frequency Distributions of S0? and N0y Vs.
Observed Winds 286
-------
FIGURES
Number Page
1 Location of the Clinch River Steam Plant 7
2 Topographical features and fixed monitoring sites.
Elevations are in feet above sea level 8
3 Diurnal variation of mean S02 concentrations at Lambert 39
4 Diurnal variation of mean S0£ concentrations at Kents 39
5 Diurnal variation of mean S02 concentration at Nashs 40
6 Diurnal variation of mean S0£ concentrations at Hockey...... 40
7 Diurnal variation of mean S02 concentrations at Johnson 4]
8 Diurnal variation of mean S02 concentrations at Tower 41
9 Diurnal variation of mean S02 concentrations at Munsey 42
10 Diurnal variation of mean S02 concentrations at Castlewood.. 42
11 Representative helicopter flight path at fixed distance
from the power plant „ 51
12 Traverse paths for cross section #18 and event numbers....„. 55
13 Traverse paths for cross section #20 and event numbers 55
14 Traverse paths for cross section #24 and event numbers 56
15 Traverse paths for cross section #26 and event numbers..„... 57
16 Traverse paths for cross section #28 and event numbers 57
17 Two-minute Hockey 30 m wind directions and locations of
peak S02 concentrations during cross section #4.
Numbers indicate the chronological order of the wind
directions and the traverses 59
VI
-------
NumDer
18 Two-minute Hockey 30 m wind directions and locations of
peak S0£ concentrations during cross section #6.
Numbers indicate the chronological order of the wind
directions and the traverses ............................... 60
19 Plume configuration at 15:36 EST July 26, 1977 with
helicopter traverse path. Point A represents endpoint
of the 15:36 vector; Point S represents endpoint of the
15:00 vector. Tic marks on axes represent 0.5 km incre-
ments of distance from the plant .......................... 62
20 Plume configuration at 15:40 EST July 26, 1977 with
helicopter traverse path. Point A represents endpoint
of the 15:40 vector; Point T represents endpoint of the
15:02 vector. Tic marks on the axes represent 0.5 km
increments of distance from the plant ..................... 63
21 Oy vs. x for cross sections #l-#5 ........................... 73
22 a vs. x for cross sections #6-#8 ........................... 73
23 Oy vs. x for cross sections #15, #17 ........................ 74
24 oy vs. x for cross sections #20, #22 ........................ 74
25 o vs. x for cross sections #24, #25 ........................ 75
26 ay vs. x for cross sections #26, #28, #29 ................... 75
27 Example of temperature profile construction methodology ..... 88
28 Plume position at 1428 EST, July 23, 1977 computed by
trajectory analysis for 60 minutes ........................ 97
29 Munsey wind roses for July-September 1977, for 3-hourly
periods during 0200-1300 EST ............................. 100
30 Munsey wind roses for July-September 1977, for 3-hourly
periods during 1400-0100 EST .............................. 101
3'; SOg pollution roses ....... . ................................. 105
32 Ratio of NO to N02 concentration ............. . .............. 109
33 Configuration of plume at 2300 EST, December 21, 1976 ....... 113
34 Configuration of plume at 2400 EST, December 21, 1976 ....... 114
35 Configuration of plume at 0100 EST, December 22, 1976 ....... 115
VI 1
-------
Number page
36 Wind direction profile at 1017 EST February 11, 1977, at
the power plant 117
37 Wind speed (m/s) profile at 1017 EST, February II, 1977,
at the power plant 118
38 Temperature profile based on fixed ground station
temperatures - February 11, 1977 120
39 Location of fumigating plume (slash lines) on
April 19, 1977 122
40 Wind direction profile at 1228 EST, April 19, 1977, at
the power plant 123
41 Wind speed profile at 1228 EST, April 19, 1979, at the
power plant 124
42 Temperature (°C) profile at 1228 EST, April 19, 1977,
at the power pi ant 125
43 Calculated configuration of plume at 1000 EST, May 15, 1977..128
44 Calculated configuration of plume at 1010 EST, May 15, 1977..128
45 Calculated configuration of plume at 1020 EST, May 15, 1977..129
46 Calculated configuration of plume at 1030 EST, May 15, 1977..129
47 Calculated configuration of plume at 1040 EST, May 15, 1977..130
48 Calculated configuration of plume at 1050 EST, May 15, 1977.,130
49 Calculated configuration of plume at 1100 EST, May 15, 1977..131
50 Configuration of plume at 0340 EST, June 30, 1977 135
51 Configuration of plume at 0350 EST, June 30, 1977 135
52 Configuration of plume at 0400 EST, June 30, 1977 136
53 Configuration of plume at 0410 EST, June 30, 1977 136
54 Configuration of plume at 0420 EST, June 30, 1977 137
55 Configuration of plume at 0430 EST, June 30, 1977 137
56 Configuration of plume at 0440 EST, June 30, 1977 138
vm
-------
Mumper Page
57 Configuration of plume at 0450 EST, June 30, 1977 139
58 Configuration of plume at 0500 EST, June 30, 1977 140
59 Wind speed (m/s) profile at 1005 EST, July 6, 1977 at
the plant 144
60 Wind direction profile at 1005 EST, July 6, 1977 144
61 Temperature (°C) profile at 1005 EST, July 6, 1977 145
62 Wind speed (m/s) profile at 1202 EST, July 6, 1977 146
63 Wind direction profile at 1202 EST, July 6, 1977 146
64 Temperature (°C) profile at 1202 EST, July 6, 1977 147
65 Location of helicopter cross-section - July 24, 1977 -
0837-0856 EST 149
66 Helicopter cross-section - July 24, 1977 - 0837-0856 EST 150
67 Longitudinal view of helicopter cross-section - July 24,
1977, 0837-0856 EST 151
68 Theoretical volume of air sampled during helicopter
traverse 152
69 Wind direction profile at 0847 EST, July 24, 1977 at
the plant 154
70 Wind speed (w/s) profile at 0847 EST, July 24, 1977 154
71 Wind direction at 0943 EST, July 24, 1977 155
72 Wind speed (m/s) at 0943 EST, July 24, 1977 155
73 Longitudinal view of helicopter cross-sections - July 24 5
1977 - 0837-0856 EST and 0915-0947 EST 156
74 Location of helicopter cross sections - July 24, 1977 -
1202-1304, 1306-1353, 1607-1621 EST 157
75 Helicopter cross section - July 24, 1977 - 1202-1304 EST 158
76 Longitudinal view of helicopter cross section - July 24,
1977, 1202-1304 EST 159
-------
Numoer Page
77 Helicopter cross section - July 24, 1977, 1306-1353 EST 160
78 Longitudinal view of helicopter cross section - July 24,
1977, 1306-1353 EST 161
79 Helicopter cross-section - July 24, 1977, 1607-1621 EST 162
80 Longitudinal view of helicopter cross section - July 24,
1977, 1607-1621 EST 163
81 Location of helicopter cross-section (dotted line) - Day
207, 1510-1619 166
82 View of helicopter cross-section from plant - Day 207,
1510-1619 167
83 Longitudinal view of helicopter cross-section - Day 207,
1510-1619 168
84 Wind direction profile at 1536 EST, July 26, 1977 169
85 Wind speed (m/s) at 1536 EST, July 26, 1977 169
86 Average sulfate concentration for each wind direction,
measured at the Hockey 30-m 1 evel 172
87 Sulfate mass due to natural deposition on filters vs.
exposure time (solid circles) and standard deviation
(open circles) of natural fallout where multiple
samples were available 173
88 Average sulfate concentration for each wind direction,
measured at the Hockey 30-m level, for filters exposed
no more than 12 days 174
£ (dotted line) and SO* concentrations at Hockey
July 8, 1977 7
89 SO? (dotted line) and SO/i concentrations at Hockey for
177
90 SO? (dotted line) and $04 concentrations at Johnson for
JulyS, 1977 177
91 SO? (dotted line) and S04 concentrations at Nashs for
JulyS, 1977 7 178
92 SOo (dotted line) and SO* concentrations at Hockey for
July 27, 1977 178
-------
Number Page
S3 SOp (clotted line) and SO* concentrations at Kents for
July 27, 1977 179
94 SOp (dotted line) and S04 concentrations at Hockey for
August 12, 1977 179
95 SOp (dotted line) and S04 concentrations at Kents for
August 12, 1977 180
96 S04 concentration at Nashs for August 12, 1977 (SO? data
was missing) 180
97 S02 (dotted line) and SO, concentrations at Hockey for
August 22, 1977 181
98 SOp (dotted line) and 504 concentrations at Johnson for
August 22, 1977 181
99 SOg (dotted line) and S04 concentrations at Kents for
August 22, 1977 182
100 SO? (dotted line) and SO* concentrations at Nashs for
August 22, 1977 182
101 SOp (dotted line) and SO^ concentrations at Hockey for
August 23, 1977 183
102 S02 (dotted line) and $04 concentrations at Johnson for
August 23, 1977 183
103 SOp (dotted line) and $04 concentrations at Kents for
August 23, 1977 184
104 SOo (dotted line) and S04 concentrations at Nashs for
August 23, 1977 7 184
105 SOp (dotted line) and $04 concentrations at Hockey for
August 26, 1977 185
106 SOp (dotted line) and $04 concentrations at Johnson for
August 26, 1977 185
107 SOp (dotted line) and $04 concentrations at Kents for
August 26, 1977 186
108 SOp (dotted line) and S04 concentrations at Nashs for
August 26, 1977 186
-------
Number Page
109 Cumulative frequency distribution of S02 for Tower 208
110 Cumulative frequency distribution of S0£ for Nashs 209
111 Cumulative frequency distribution of S02 for Kents 210
112 Cumulative frequency distribution of S02 for Johnson 211
113 Cumulative frequency distribution of S02 for Lambert 212
114 Cumulative frequency distribution of S02 for Munsey 213
115 Cumulative frequency distribution of S02 for Hockey 214
116 Cumulative frequency distribution of S02 for Castlewood 215
117 Chestnut Ridge Monitoring Network 219
XII
-------
TABLES
Number Page
1 Monitoring Site Characteristics 10
2 Summary of Measurements at Fixed Monitoring Sites 12
3 Example Classification of Data Periods 16
4 Data Period Summary 17
5 Percentage of Stability Classes for Local and Synoptic
Periods (Based on the 31 Periods with Significant
Concentrations in Table 4) 22
6 Number of Local and Synoptic Periods for Predominant Surface
Wind Directions (Based on Periods in Table 4) 22
7 Most Frequent Ranges of Plant Parameters 25
8 Most Frequent Ranges of Generator Load by Time of Day 26
9 Mean Pollutant Concentrations over the Monitoring Period 27
10 Ten Highest Hourly Average S02 Concentrations Observed at
Each Moni tori ng Si te , 29
11 Ten Highest Hourly Average NOX Concentrations Observed at
Each Moni tori ng Si te 30
12 Frequency Distribution of S02 vs. Hour from 76286 to 77273,
Tower Site (Row & Column Headings are High Ends of
Intervals) 31
13 Frequency Distribution of SO? vs. Hour from 76286 to 77273,
Munsey Site (Row & Column Headings are High Ends of
Intervals) 32
14 Frequency Distribution of S02 vs. Hour from 76286 to 77273,
Castle Site (Row & Column Headings are High Ends of
Intervals) 33
15 Frequency Distribution of S02 vs. Hour from 76286 to 77273,
Nashs Site (Row & Column Headings are High Ends of
Intervals) 34
XT n
-------
Number Page
16 Frequency Distribution of SO? vs. Hour from 76286 to 77273,
Hockey Site (Row & Column Headings are High Ends of
Intervals) 35
17 Frequency Distribution of S02 vs. Hour from 76286 to 77273,
Lambert Site (Row & Column Headings are High Ends of
Interval s) 36
18 Frequency Distribution of S0£ vs. Hour from 76286 to 77273,
Johnson Site (Row & Column Headings are High Ends of
I nterval s) 37
19 Frequency Distribution of SC>2 vs. Hour from 76286 to 77273,
Kents Site (Row & Column Headings are High Ends of
Intervals) 38
20 Mean Pollutant Concentrations for Mobile Van Data 44
21 Frequency Distribution of S02 Concentrations Measured in Van
While in Stationary Mode 46
22 Frequency Distribution of NOX Concentrations Measured in Van
While in Stationary Mode 46
23 Frequency Distribution of N02 Concentrations Measured in Van
While in Stationary Mode 47
24 Frequency Distribution of NO Concentrations Measured in Van
While in Stationary Mode 47
25 Frequency Distribution of 03 Concentrations Measured in Van
While in Stationary Mode 48
26 Ozone Concentrations Versus Distance from Plant for Mobile
Van Data 49
27 Summary of Helicopter Cross Sections for July 1977 53
28 Locations of Center of Mass of Individual Traverses 58
29 Comparison of Computed S02 Mass Flux and Plant S02 Emission
Rates 64
30 Location of Center of Mass of Plume 66
31 Measured Plume Heights 68
32 Plume Dimensions 71
33 Turner Stability Class Corresponding to Measured ay Values 72
xiv
-------
Number Page
34 Vertical Distribution of S02 Mass Computed from Helicopter
Traverses 77
35 Significant Fixed-Station and Ground Mobile Measurements
During Helicopter Cross Sections 80
36 Average Difference and Standard Deviation of Differences of
Temperature and Wind Speed Between Fixed Stations and
Pi bal /T-Sondes 83
37 Average Difference and Standard Deviation of Differences of
Temperature and Wind Speed for Huntington and Greensboro
Rawinsondes vs. Pibal/T-Sonde 84
38 Power Law Exponents 89
39 Comparison of Plume Height Measurements and Estimates 93
40 Summary of Plume Heights in Relation to Vertical Temperature
Structure, October 12, 1976 - September 30, 1977 95
41 Wind Speeds and °/\ Ranges Associated with Highest S02
Concentrations 106
42 List of Case Studies 110
43 Case Study I Data Ill
44 Case Study II Data 116
45 Case Study III Data 121
46 Case Study IV Data 127
47 2-Minute S02 Concentrations at Johnson 132
48 Case Study V Data 133
49 2-Minute S02 Concentrations at Hockey 141
50 Case Study VI Data 142
51 Case Study VII Data 148
52 Estimated Plume Bearing and Bearing of Center of Mass 164
53 Case Study VIII Data 165
xv
-------
Number Page
54 Data on Estimated Bearing of the Plume, Based on 30 M
Hockey Winds (July 8, 1977) 175
55 Data on Estimated Bearing of the Plume, Based on 30 M
Hockey Winds (July 20, 1977) 175
56 Data on Estimated Bearing of the Plume, Based on 30 M
Hockey Winds (July 27, 1977) ..175
57 Data on Estimated Bearing of the Plume, Based on 30 M
Hockey Winds (August 12, 1977) 176
58 Data on Estimated Bearing of the Plume, Based on 30 M
Hockey Winds (August 22, 1977) 176
59 Data on Estimated Bearing of the Plume, Based on 30 M
Hockey Winds (August 23, 1977) 176
60 Data on Estimated Bearing of the Plume, Based on 30 M
Hockey Winds (August 26, 1977) 177
61 Distribution of the Joint Occurrences of Classes of Bulk
Richardson Number and Pasquill Stability .189
62 RMSE of Model to Measurement Comparisons for All Stabilities..193
63 Correlation of Model to Measurement Comparisons for all
Stabi 1 i ties 194
64 RMSE of Model to Measurement Comparisons for Unstable
Conditions 194
65 Correlation of Model to Measurement Comparisons for Unstable
Conditions 195
66 RMSE of Model to Measurement Comparisons for Neutral
Conditions 195
67 Correlations of Model to Measurement Comparisons for
Neutral Conditions 196
68 RMSE of Model to Measurement Comparison for Stable
Conditions 196
69 Correlation of Model to Measurement Comparisons for Stable
Conditions 197
xvi
-------
Number Page
70 Number of Model to Measurement Compari sons ................... 198
71 Comparison of Highest and Second Highest 1-Hour S02
Concentrations Between Modeling Techniques ................. 199
72 Joint Frequencies of Seven Classes of SSCTM Calculated and
Observed Concentrations of SC>2 at Tower and Castle Sites... 203
73 Joint Frequencies of Sevel Classes of SSCTM Calculated and
Observed Concentrations of S0£ at Lambert and Johnson Sites 204
74 Joint Frequencies of Seven Classes of SSCTM Calculated and
Observed Concentrations of S0£ at Hockey and Munsey Sites.. 205
75 Joint Frequencies of Seven Classes of SSCTM Calculated and
Observed Concentrations of S02 at Kents and Nashs Sites ____ 206
76 Comparison of Linear and Transport Curvilinear Models ........ 216
77 Plant Operating Characteristics .............................. 220
78 Model to Measurement Comparisons of 1-Hour Averages for
Chestnut Ridge - All Conditions During 1975 (Station
Locations are Shown in Figure 117) ......................... 222
xvn
-------
SECTION 1
INTRODUCTION
In this study we have analyzed aerometric data collected in the vicinity
of a coal-fired power plant located in the complex terrain of southwestern
Virginia. This report describes the data used in these analyses, the types of
analyses performed and relationships that have been identified between terrain
characteristics, meteorological conditions, plant emissions and the resulting
ground-level pollutant concentrations. This analytical work represents the third
phase of a three-phase study of the behavior of power plant plumes in complex
terrain. The first phase (Koch et al. 1977) consisted of a literature review
of the current state of knowledge of plume behavior in complex terrain with
respect to both the transport and diffusion processes and the pollutant trans-
formation and removal mechanisms. The second phase (Koch et al. 1979) consisted
of an aerometric field study to collect data related to plume behavior in
complex terrain in the vicinity of Clinch River Steam Plant near Carbo, Virginia.
Sixteen months of ground-level pollutant concentration data and meteorological
measurements were gathered in an eight-station monitoring network. The fixed
stations were supplemented by upper-air meteorological observations, stack
emission measurements, and pollutant monitoring from a mobile van and a helicopter.
The study as a whole was undertaken due to the generally recognized lack of
complete understanding of plume behavior in complex terrain. Phase II provided
a set of SOp, sulfate, NO, NO^, and meteorological data from which conclusions
can be ascertained concerning pollutant transport, diffusion and transformation
in mountainous terrain and from which improved dispersion modeling techniques can
be developed. Accurate prediction of air quality levels in complex terrain is
becoming more critically important than ever before as major power plants
and industrial sources plan to locate in these regions. For each new major
pollution source studies must be performed to predict the potential air quality
impact of the operation of these sources for the purposes of showing compliance
with Prevention of Significant Deterioration regulations. Coal-fired electric
power plants are major sources of SO- and NO. Sulfur dioxide and sulfate,
which is formed from SO^ in the atmosphere, are recognized to produce
adverse health effects. Also, NO is converted in the atmosphere to NO-,
which has been determined to be a health hazard.
The Phase III analyses include (1) the identification of data periods con-
taining significant plume impact; (2) statistical summaries of emissions, ground-
level pollutant concentrations, and meteorological data; (3) statistical summaries
of relationships between concentration and meteorological data; (4) the analysis
of plume cross sections observed by helicopter; (5) the development of vertical
profiles of wind speed, wind direction, and temperature; (6) use of the vertical
wind and temperature profiles for estimating plume heights; (7) an analysis of
1
-------
observed sulfate data in relation to SO- concentrations and meteorology; (8) an
analysis of dispersion model estimates of SCL concentrations in relation to
measured values; and (9) a detailed examination of the causes for individual
cases of high SO^ concentrations.
The data from the Phase II field study were reduced, edited and placed on a
magnetic tape for distribution. The tape contains 34 files of data which include
8 fixed-station files, 1 sulfate file, 1 upper-air meteorological data file,
7 ground-mobile data files, 16 helicopter data files and 1 plant operating data
file. A copy of this tape is available from the National Technical Information
Service. The data formats are described in the Phase II report (Koch et al.
1979). The Phase III analyses have been based on these data sets along with
additional information contained in field notes and logs, daily weather maps,
and raw data files.
-------
SECTION 2
SUMMARY AND CONCLUSIONS
Field observations collected in the vicinity of the Clinch River
Steam Plant in southwest Virginia for a period of about a year were
analyzed to determine relationships which characterize the behavior of
the plume in complex terrain. The data include measurements from eight
fixed monitoring stations, a mobile van, one of the two plant stacks, a
helicopter, and balloon soundings. The following conclusions have been
drawn from the data analysis:
0 There are 31 significant periods of 2 to 20 days' duration
in which the meteorological conditions are consistent and
during which high SOg and NOX concentrations indicated that
the plant plume was observed.
• The significant data periods include 16 periods with locally
dominated flow patterns and 15 periods with synoptically
dominated flow. All but one period had a majority of the
hours having stable conditions as defined by the bulk
Richardson number.
t Valley stations have a diurnal pattern of concentration
with a maximum in the morning daylight hours. This differs
from the diurnal patterns of ridge stations which do not have
well defined maxima. The valley maximum is probably due to
fumigation of the plant plume which is trapped in nocturnal
valley inversion layer.
• Ridge,.stations are exposed to higher maximum 1-hour con-
centrations than are valley stations. The maximum concen-
trations at ridge stations generally occur at night. This
differs from valley stations which have maximum concentra-
tions during the day.
t ~he mobile van was capable of recording plume impac'; with
greater frequency tnan did any of the fixed monitoring
stations. The van measured hourly average S02 concentrations
greater than 20 ppb approximately 50 percent of the time',
while the fixed station (Tower) with the greatest number
of significant readings had concentrations greater than 20
ppb only 16.5% of the time. Higher S02 concentrations were
recorded by the van than by any of the fixed stations. This
result indicates that the van was under the plume centerline
more often than were any of the fixed stations.
-------
The mobile van was also useful in providing data for the
study of the NO/NOg ratio as a function of distance from
the plant. This ratio had a mean value of 5.6 for measure-
ments within 0.5 km of the plant and was found to decrease
to approximately 0.8 at distances of 3 km to 15 km from
the plant.
The plume cross section for a period of an hour as observed
by helicopter agrees well with dimensions indicated by vari-
ations in wind direction and computed plume paths using
2-minute winds. These results indicate that wind observations
taken on a 30 m tower at a ground elevation near the height
of the plume are useful in defining the dimensions of the
plume (i.e., during daylight hours when the flow at the
observation height is well mixed to ground level).
The horizontal dispersion parameter o^, as observed by the
helicopter is well represented by Pasquill's suggestion of:
°y = °A x f(x).
The mean ratio (measured Oy/Pasquill
-------
• A major problem with hourly sulfate measurements is the fact
that there is a large amount of inadvertent deposition on
the filters when the sampling blower is not operating.
Future measurements must be directed toward better control
of this interference. The data collected in this study
suggests that the nearby power plant makes some, but probably
not the major, contribution to sulfate concentrations in the
vicinity of the plant.
t The simple Gaussian plume model, when applied using site-
specific meteorological data to estimate plume dimensions
(e.g., through standard deviation of wind azimuth observa-
tions) and local terrain heights to estimate plume height
above ground, can significantly improve SC^ concentrations
for estimates made with a flat terrain model and standard
Pasquill-Gifford dispersion parameters. The use of terrain
adjustment factors such as are found in the CRSTER and
VALLEY models were found to not improve the standard flat
terrain estimates.
• The simple Gaussian model cannot be expected to give better
results on a case by case (specific hour basis) by improved
model parameter estimates than were found in this study,
because it cannot be managed with the complex flow pattern
which is present at least 50 percent of the time at the
Clinch River plant site. The data contains numerous ex-
amples which clearly indicate the presence of a complex
flow pattern. However, the Gaussian plume model provides
valuable guidance regarding what the maximum concentrations
over a long period of time can be expected to be.
• The degree of accuracy found by applying the Gaussian
model with site specific data to central Pennsylvania sites
was found to be better than the accuracy found at the Clinch
River site. Over the stations considered the average model/
observed ratio for the highest $03 values was 1.23 at Chestnut
Ridge Monitoring Network compared with 1.58 at the Clinch
River Network.
• The data collected from the Clinch River plant site provides
a valuable source of data for testing hypotheses regarding
the physical and chemical behavior of power plant plumes in
complex terrain. Among the unique data available are hourly
wind and temperature measurements from eight sites for a
period of a year, NO and NOX measurements at six sites for
a period of ayear, several hundred hourly sulfate measure-
ments, and nearly daily temperature and wind profiles from
balloon observations. In addition, there are a large number
of other observations which are similar to data from other
sites including a year of S0_2 measurements at eight monitoring
sites; 24 hourly cross sections of S02, NO, NOX and 03 from
helicopter measurements; observations from a mobile van; and
hourly stack emission measurements of NO and S02.
-------
SECTION 3
DESCRIPTION OF THE CLINCH RIVER POWER PLANT, TERRAIN AND
MONITORING NETWORK
The site chosen for the aerometric field study was the vicinity of the
Clinch River Steam Plant in Carbo, Virginia, located approximately 200 km
west-southwest of Roanoke. This site was chosen because:
• It is located in a nonurban, mountainous, area, isolated from other
major sources of SO-.
• Terrain in the vicinity of the plant exceeds 1-1/2 times the stack
height.
• An adequate network of roads existed to facilitate mobile sampling
and access to fixed monitoring sites.
• The power company was willing to cooperate.
PLANT DESCRIPTION
The Clinch River Steam Plant has a total generating capacity of 712 MW,
fired by low-sulfur coal in each of three boilers. Exhaust gases are emitted
through two stacks, 46 m apart, each 138 m high. Stack 1 with a diameter of
4.76 m serves boilers 1 and 2, while Stack 2, with a diameter of 3.81 m serves
boiler number 3. The boiler water is cooled through the use of five mechanical
draft cooling towers, each 18.6 m high.
TERRAIN
The location of the plant in southwestern Virginia is shown in Figure 1.
The details of the surrounding terrain, including the fixed monitoring site
locations are shown in Figure 2.
The terrain is generally characterized by a series of parallel ridges
and valleys which run southwest to northeast. However, numerous pronounced
short valleys run perpendicular or askew to the larger valleys, making a very
irregular and complex layout of terrain. Although many terrain features are
inaccessible, about half of the terrain is occupied by small farms and rural
residences.
The plant elevation is 461 m, which puts the stacktops at just under 600
m. There are ridges exceeding 680 m (2200 ft) within 3 to 5 km of the plant
in all quadrants of the compass. Elevations exceeding 870 m (2900 ft) occur
within 8 km to the northwest and 14 km to the southwest. Clinch Mountain,
-------
•1"
Of.
u
c
CJ
cc*
A
c
o
T!
',)
o
cu
S-
u.
-------
o
.C
oc
c
o
J3
o
3
b£
-------
located 14 km southeast of the plant, is the highest terrain in the area. It
runs northeast-southwest for approximately 25 km and reaches elevations in
excess of 1260 m (4200 ft,). Big A Mountain, located 20 km northeast of the
plant rises to 1140 m (3,800 ft.). Flat Top Ridge, situated 8 km northwest of
the plant, is oriented northeast-southwest and reaches elevations up to 915 m
(3,000 ft.). The major pattern of ridges and valleys trending northeast-
southwest is transected by many streams running northwest-southeast to form
a truly complex terrain.
in the immediate vicinity of the plant (within 1 km) ridges extend to
to 100 m above the top of the stacks. The highest elevations lie to the
east-northeast and to the northwest. A pronounced ridge also extending 100 m
above the stack is located approximately 2 km to the southeast. Flow over
these nearby terrain features on windy days causes a pronounced downwash of
the stack plumes. Further to the southeast is Copper Ridge, which exceeds
the stack top elevation by 200 m. Buffalo Mountain, also reaching 200 m above
the stack, is located 6 km to the northeast.
MONITORING NETWORK
The eight fixed monitoring station locations were selected to meet design
criteria for climatology, local topography, existing roadways and utility
services, and the ability to negotiate for the land on which to place the
instrument shelter. Locations were selected using guidance from dispersion
model calculations and representative climatological data. The character-
istics of the selected sites are summarized in Table 1. Initially, only six
stations were operated. The Lambert site was brought into operation on
October 27, 1976, and the Johnson site became operational on November 15,
1976. The Castlewood station washed away in the flood of April 5, 1977,
reducing the number of fixed monitoring sites to 7. The Lambert instru-
mentation was installed at Castlewood during July 1977.
At Site No. 1 (Tower) a 30 m tower was installed to measure winds at
two levels and temperature at three levels. This site was 3.4 km to the
northeast of the plant at an elevation of 585 m. The site was in open
terrain and air flows in the region were unobstructed by vegetation or
nearby terrain features. The nearest terrain feature was a ridge located
1 K..I to tne south. This was the closest site to the power plant. It had
an elevation of 124 m above the base of the plant.
Site No. 2 (Munsey) was located 4 km to the southeast of the power
plant. It was approximately 1/3 of the way up the hillside from the valley
floor to the Hockey site, located 200 m higher. The two sites are on about
the sa,.,e bearing from the plant and were used to resolve the characteris-
tics of tne plume perpendicular to the ridge on which these two sites were
located.
Site No. 3 (Nash's Ford) was located approximately 11 km to the east of
the power plant in the Clinch River Valley. The site was situated on a small
pleteau between the Clinch River and a small creek, and was representative of
upvcjley flews from the power plant due to thermodynamic effects and due to
channeling of the wind.
-------
CS
s
U
I
pa
IS
1 *»
: C S
fc t>
u
a
n *-*
1 if
™ .= 30
U <
£ IT
s
£ ?
U 4)
i O
^* tt
£ ***
C rt
5 1
w **
^ s
1 5
V
p
z
•»«
CO
H
3
2
S
91 V
"3. 3
lilt ! f 1 ! !
a. a. as ss B:
j
oo o\ \o o^ 1^ i/5 T^ '0^ vo
oooooooo
CO ^ Irt 00 ^ 1^
oo^o- oooo^o
^H MM >4
•a "2 -n*
fi 5-^11? | 1 *
1 1 1 1 5 1 i S 1
10
-------
Site No. 4 (Hockey) was located on a ridge (792 m msl) just above the
Munsey site and on a line with Munsey and the power plant. This site was
selected to observe the concentration characteristics of the power plant plume
in relation to the ridgetop.
Site No. 5 (Castlewood) was located approximately 8 km west-southwest of
the power plant. It was very close to the Clinch River; however, it had an
unobstructed upvalley view so that winds from that direction would be charac-
teristic of drainage flows if they were occurring. This site was about 10 m
below the plant elevation. The station washed away in the flood of April 5,
1977.
Site No. 6 (Kent's Ridge) was located 30 km east-northeast of the power
plant. This site was chosen for both its elevation and distance and was the
most distant site from the Power Plant.
Site No. 7 (Johnson) site which was located south-southwest of the power
plant on a ridge that runs from the west-southwest to the east-northeast.
This is the same ridge line that Hockey was located on.
Site No. 8 (Lambert) was located approximately one-third of the way up
the side of a valley. It was chosen to determine the plume characteristics
in that valley. Thf instrumentation was moved to Castlewood in July 1977.
A stack monitor was located in the duct work at the base of the Unit 3
stack. The monitor was between the electrostatic precipitators and the base
of the stack. Valid data was collected after December 1, 1976.
Table 2 summarizes the types of data that were available from each of the
fixed stations. A mobile van monitored S02, NOX, NO, and 03 while moving
under the plume and also at stationary locations within the footprint area of
the plume on many days througnout the field program. A helicopter flew
traverses through the plume during two 10-day periods also measuring S02, NOX,
NO, and Og. The helicopter monitoring periods occurred in November 1976 and
July 1977? The fixed-station meteorological data were supplemented by upper
air wind and temperature data obtained by the use of pilot balloons with
attached T-sondes. The National Weather Service surface and upper air analyses
were received at the plant by facsimile circuit.
11
-------
t
_ s
K K K K
VI
H
O
Z
5
g
llHT
1 J * • £
5 J ' = "
Hill
S I » 5 "
lll^7
• 6
iHll
5l-.l =
w
Z
w
w
K
11!
b
O
a
II!
CM
03
n
X X X X X X
12
-------
SECTION 4
DATA PERIOD ANALYSIS
In order to synthesize the data from the 16-month Clinch River field
study in a form which can be used to identify periods of interest for future
studies, the data were organized into cohesive periods based on observed
characteristics of the airflow.
DATA ORGANIZATION
The data were organized through the compilation of 11 parameters on a
da'ly basis over the entire monitoring period at the Clinch River Steam Plant.
The following parameters were included in the data listing:
1. Predominant Airflow Influence—To determine airflow influence each day
was divided into two~T?-hour periods; daylight (0700 to 1800 local standard
time) and nighttime (0100 to 0600 and 1900 to 2400 local standard time).
Then, all the hourly fixed station wind directions, with the exception of
Kents Ridge, were analyzed on a daily basis for both daylight and nighttime
by summing the number of wind directions in eight 45° sectors (i.e., 0°-45°,
45°-90°, etc.). If the sector with the highest frequency consisted of more
than 41 percent of the total cases, then the winds for that 12-hour period were
designated as synoptically influenced; if less than 41 percent, the winds for
that period were designated as locally influenced. The cutoff value of 41
percent was determined by noting that the mean percentage of observations in
the prevailing wind direction sector over a sample group of days was 41 percent.
Lacking any good guidelines, we decided that, if the frequency of the prevail-
ing wind octant exceeded this mean frequency, the wind direction was relatively
consistent in time arid space and primarily synoptical ly influenced. This
classification scheme resulted in four possible categories for a day:
1. Synoptic daylight and night (S)
2. Local daylight and night (L)
3. Synoptic daylight, local night (SL)
4. Local daylight, synoptic night (LS).
2. Fixed Stations Exhibiting Significant Concentrations — For each day
the stations which had significant concentrations of S0~ and NO were
listed. The cutoffs for significant concentrations at each station were
determined by assigning a significant concentration to the Tower site and using
a log-linear decrease with distance from the plant to assign values to other
13
-------
sites. The following relationships were used to determine significant concen-
trations of S09 and NO :
C- A
n^ T* x 60 ppb = Significant SO,, Concentration at Station X.
ln(Dx)
ln(DT)
x 100 ppb = Significant NO Concentration at Station X,
ln(Dx) X
where Dj is the distance from the plant to Tower, and
DX is the-distance from the plant to Station X
NOX is total of N02 and NO concentrations,,
This resulted in the following cutoffs of significant concentrations for each
fixed station:
Station SO^ ppb NO,, ppb
Tower
Castlewood
Hockey
Munsey
Lambert
Johnson
Kents
Nashs
Diurnal Variation in Vertical
60
35
41
50
36
41
22
31
Temperature
100
58
68
83
--
—
37
52
Difference
of vertical temperature differences measured over the interval 0.5 m to 30 m at
the Tower site were categorized as small (range less than 2.0°c), medium (range
from 2.0°C to 3.2°C) or large (range greater than 3.2CC). The categories were
established by examining the distribution of the daily ranges over a portion of
the data set. The categories were selected so that each contains about one-third
of the days
14
-------
4. Ambient Temperature Range - The minimum and maximum temperature
at Hockey observed for each day was listed. The Hockey site was chosen
because it is the closest fixed station to the majority of the plume heights.
5. Total Precipitation - The total amount of precipitation in inches
that fell during each day was provided.
6. Range of UV Eradiation Peaks - The daily maximum UV radiation
measurements in mW/cm was presented for each day.
7. Hind Speed and Direction - The 1200 GMT, 850 mb facsimile maps were
examined for wind speeds and directions for southwest Virginia. The speeds and
directions were tabulated for each day.
8. Percent of Hours: Unstable, Neutral, Stable - The number of hours
for each stability class was calculated for each day. The method utilized in
determining stability is discussed in Section 10.0. The method is based
on the use of the Bulk Richardson Number computed from Tower data.
9. Prevailing Circulation - Each day was assigned to one of seven
categories as follows: strong cyclonic, weak cyclonic, cyclonic, strong
anticyc Ionic, weak anticycIonic, anticycionic, or mixed, based on examination
of the 1200 GMT, 850 mb facsimile maps.
10. Prevailing Surface Wind Direction - The 45° sectors of wind directions
which occurred most frequently among the hourly fixed-station observations were
noted for each day.
11. Plume Height Distribution - The number of estimated hourly plume heights
falling into four height ranges was tabulated each day for each stack. The
plume height estimates were made using vertical profiles of temperature and wind
speed. The method is discussed in Section 7.0.
Groups of days were then assigned into data periods designated as being
predominantly synoptic or local based on the predominant airflow influence.
For example, if the situation in Table 3 had occurred, the first 5 days would
be classified as a synoptic period, and the last 3 days would be a local
period.
15
-------
TABLE 3. EXAMPLE CLASSIFICATION OF DATA PERIODS
Day Classification
March 20, 1977
21
22
23
24
25
26
27
SL
L
S
S
S
Snoptic
L )
LS > Local
SL j
Table 4 presents the 11 parameters for each of 31 data periods from November
22, 1976 through the end of the field study observed to have a relatively large
amount of significant SO^ and NO concentrations at the fixed monitoring
stations. This table was derivea from the complete compilation of data in which
the same parameters were tablulated on a daily basis by considering only those
periods in which significant SO^ or NO concentrations occurred at more than
one station for at least three-fourths of the days during the period or signifi-
cant SO/, or NO occurred at four or more stations on at least 2 days during
the period. Tfte table begins on November 22, 1976 since pollutant instrument
calibration problems existed prior to this date.
The number of hours in each data period which each station recorded sig-
nificant SCL concentrations are presented in Table 4. In the table each
fixed station is represented by the first letter appearing in its name (i.e.,
C-Castlewood, T-Tower, etc.) followed by the number of significant concentra-
tion hours. The two plume height distribution columns of Table 4 apply to
Stack 1 and Stack 2, respectively. The first number in both columns is the
percentage of estimated hourly plume heights less than 250 m above plant base,
the second number is the percentage between 250 m and 400 m, the third is
the percent between 400 m and 600 m, and the fourth is the percent greater
than 600 m. For the other parameters the values given in Table 4 are either
maxima and minima of the parameter or the prevailing condition during the
period.
RESULTS
By examining Table 4, several distinctions between synoptic and local
periods become apparent. For instance, local periods generally had slightly
lower plume heights; for Stack 1 the mean percent of plume heights lower than
250 m above plant base was 43.2 percent for synoptic periods compared to 50.5
percent for local periods. However, this difference is not statistically sig-
nificant at the 95 percent confidence level. Local and synoptic periods also
differed in the frequency distribution of the stability classes as evidenced
16
-------
• £ c
S ts
v i ? ?. i
i
*~* s fe- t3
oll5
Hiss
Pllf
O r 2 S" U X
5 o c
E "E S
17
-------
11 s;
3 C "~ !•
s -u
If
e
! S*i?
t = 14 ^
II-ss.
H I 8.SS
2 1 " t S
[_; ^ ^t v oi
111 £
5.^ E S
Z"T3
= ~ t-5tr
S* * T«l-
» "i
•*- C t- O O
c C **•»-
3 3 • *> •
1s.if:
ill**
| | c
->cc.o ^r1-0 .^~ = **'
o
:D z ^ ~ z
(MO^ ^™S^ — c: Oi ^
-------
- „
" " S
t -I S
n|i.
2 | III
_ ? •= S S.
!II15
^ S £ I -5
flsfls
; 5 U * f
o o o o
19
-------
1
8
.
i
£ 8 - *
t ^ * i f
s 7 * : _
£ = ~ -
tS,
6t c C
llSi
°-l "
u s
c c
» T!
U
= ii si
t - 1 ° f.
t -i 1 1 -
•e -5
• Ills
^1111
",sg
5 "M ** " — '
4J0.* JJ B
5 w»T3 S »
f|»°'"
§
_ « •" « t
* 1 5 c
H T T = ~
I ~
e i =
|i|f|£
jijij
Ji c\, *•
*. r -
_ 1 * = 5
mJ!
2 S "
c 4)
1 £ S
ill
£
_
K ^t-
o
o
IN
o1
S <
5. C
^ z
S £
l/l VI
O
_;
1/5
C
c
0^
5
£
QC C
h- U
J
i
3
l/>
C
^r o
5
.,_
^^s; oo - "* vi^00 ^vE ^u-- ^52^,
0 O O o o
o o c 0 c i/, c
Ci (SI i i C- i ^ i
0
£• sit 1 1 sit | | £
S =3S S=S P-S -S3 ?'- S~?
= Z ^ = Z », = Z » = Z - DZ. = Z .
? =? !l -B$ ?* '5 - ?
J* ^ ^ £?' £§ H3
^ rr; O1 in C S
_« rO rsj IN rJ .«
2 r^ or J~ ri ^
-' « 6 c
« 5f
1 c " C '^ N C 1
O O^ OC OO OO OO OC^"
«-> T- o LT ^r <-j vo 1/1 li o" C T c:
! il ! 1 ! I
,_
c^ ficcc ooof^o ^.C(N^; ~°fsi" Z 2 •- "" £S~~
u u
*n o" "? c" "^ ~
u c 0 c C S
Q >• Q S- 0 —
t g J It 5 E 1 ^ t t - '=
-• S V ? £ - -- 3 ^ S o ^- ^- o « ..- c. S
I 22 il 11 I 1 1 \ 11
20
-------
1 1 ? _ x I ,j
S. z
" z
O O 0 C
21
-------
by Table 5. The more frequent incidence of stable conditions for local periods
may have been a contributing factor to the lower plume heights calculated
during these periods. Table 5 also shows there was a substantial difference in
the frequency of neutral conditions between local and synoptic periods. This
difference is significant at the 95 percent confidence level. The greater
frequency of occurrence of neutral stability during synoptic periods may be
attributed to greater wind speeds and the resulting enhancement of vertical
mixing.
TABLE 5. PERCENTAGE OF STABILITY CLASSES FOR LOCAL AND SYNOPTIC PERIODS
(BASED ON THE 31 PERIODS WITH SIGNIFICANT CONCENTRATIONS IN TABLE 4).
Local
Synoptic
Unstable
9.5%
7.5%
Neutral
13.8%
26.1%
Stable
75.8%
63.3%
Table 6 shows the frequency distribution of prevailing surface wind direc-
tion sectors for synoptic and local periods. Synoptic periods are predominantly
associated with a wind direction in the 225°-270° sector. The most frequently
observed wind direction during local periods is more varied with both the
180°-225e and the 225°-270° sectors being common. There is a secondary peak in
the 45°-90° sector.
TABLE 6. NUMBER OF LOCAL AND SYNOPTIC PERIODS FOR PREDOMINANT
SURFACE WIND DIRECTIONS (BASED ON PERIODS IN TABLE 4)
Number of Cases
Wind Direction Sector Local Synoptic
0°
45°
90°
135°
180°
225°
270°
315°
- 45°
- 90°
- 135°
- 180°
- 225°
- 270°
- 315°
- 360°
0
5
0
0
8
8
2
2
0
0
0
0
0
12
4
1
22
-------
A striking difference between synoptic and local periods is illus-
trdted in Table 4 for periods between December 18, 1976 and January 2, 1977.
The two local periods during that time had lower plume heights and markedly
less incidence of neutral stability compared to the two synoptic periods.
A noteworthy period among the periods with the highest frequency of
significant concentrations was the 2-day synoptic period of February 9-10,
19/7. During this period, all the fixed stations with the exception of
Lambert, had at least 1 hour of a significant SO- concentration. An
examination of the calculated plume height distribution for this period
revaals that all the plume heights for Stack 1 were less than 400 m and all
;:ha plume heights for Stack 2 were less than 250 m above plant base. A
contributing factor to the lower plume heights and hence the large concen-
trations may have been the large incidence of stable conditions (79 percent
of the hours were stable) for this period. The predominant surface wind
direction for this period was 225°-270°. Six of the eight fixed stations
are downwind when winds are within this sector or within 20° of this sector.
CONCLUSIONS
It was possible to divide the monitoring data into data periods in which
the airflow was locally or synoptically dominated based on the observed char-
acteristics of the winds within the monitoring network. This delineation
of the data allowed the identification of periods of significant plume impact
on the fixed-station network. Differences in frequencies of wind direction
sectors, stability classes and plume heights were noted between the synoptic-
flow and local-flow periods.
23
-------
SECTION 5
STATISTICAL ANALYSIS OF PLANT EMISSIONS
AND MEASURED AMBIENT POLLUTANT CONCENTRATIONS
General trends and patterns in pollutant data usually are apparent in
joint frequency distributions of concentration and other physical parameters.
Such analyses have been performed with the hourly concentration data from
each of the eight fixed stations and with the plant emission data. In addi-
tion, the dates and times of the highest S02 and NO concentrations have
been produced for each station. Mean pollutant concentrations for the entire
monitoring period were also computed. A review of these statistics has guided
subsequent analysis of the data.
PLANT EMISSIONS DATA
For each generating unit the following joint frequency distributions were
produced:
• Generator load versus hour
• Exhaust gas temperature versus hour
• Exhaust flow rate versus hour
• Estimated SO^ emission rate versus hour
• Estimated NO emission rate versus hour.
In addition, for Unit #3, frequency distributions of measured SO- emission
rate versus hour and measured NO emission rate versus hour were generated.
To typify the Clinch River plant operating conditions, the most frequent
categories of generator load, exhaust gas temperature, exhaust flow rate,
estimated SOp emission rate, and estimated NO emission rate were identified.
These ranges appear in Table 7. The most frequently measured S02 emission
rates for Unit #3 (Stack 2) were found to be between 250 and 300 g/sec;* the
* The measured SO- and NO stack concentration distributions were compiled
in 50 g/sec increments and 25 g/sec increments, respectively.
24
-------
most frequently measured NO emission rate was between 200 ana 225 g/sec. The
estimated pollutant emission rates for all three units were computed in Phase
II of the study using linear regression formulas derived from the available
measured emission rates for Unit #3 and the generator load and sulfur content
data provided by the power company. The estimates are useful for periods when
the stack monitors were not available or not working. For other periods use-
ful emission estimates can be computed by using the ratio of measured
emission to load for Unit #3 to scale the emissions for the other units to
the measured loads.
TABLE 7. MOST FREQUENT RANGES OF PLANT PARAMETERS
Exhaust
Temperature
Unit # (°K)
1 280-390
2 390-400
3 380-390
Exhaust
Airf lovs
(n3/sec)
32C-340
360-380
320-340
Generator
Load
(MW)
200-225
225-250
225-250
Estimated
SGp Emission
(g/sec)
300-325
275-300
275-300
Estimated
NO Emission
(g/sec)
200-225
200-225
200-225
Table 8 shows the typical diurnal variation of generator loads for each
unit. The pattern indicates a rather abrupt increase in load for Unit #1
between 7 a.m. and 8 a.m. and for Unit #2 between 6 a.m. and 7 a.m. A more
gradual early morning increase is noted for Unit #3. Most frequently, the
generator load is 200-225 MW from 8 a.m. through 11 p.m. for Unit #1 arid
225-250 Mkf for Units #2 and #3. Another abrupt change in load frequently
occurs between midnight and 1 a.m. for Units #1 and #3, when the loads typi-
cally drop from 200-225 MW to 125-150 Mw. The same drop in load occurs from
11 p.m. to 1 a.m. for Unit #2.
Table 7 shows that the most frequent SOp emission rates are 300-325 g/sec
for Unit #1 and 275-300 q/sec for each of Units #2 and ?3. Emission rates for
SOp as high as 525-550 g/sec were estimated for Units #1 and $3, while three
cases of SO/, emissions in the 550-575 q/sec range were estimated for Unit
#2. The majority of the very high SOp emission rates occurred in the afternoon
when typically dispersion conditions are best. The highest estimated NO
emission rates were 300-325 g/sec for Unit #3 and 275-300 g/sec for Units #1
and #2. Similar to SOp, generally the highest NO emission rates occurred in
the afternoon.
Exhaust air flow rates and exhaust temperatures were generally well cor-
related with the trends in generator load conditions, as peak flow rates and
exhaust temperatures generally occurred in the afternoon. A few cases of flow
rates as high as 420 m /sec and temperatures as high as 450° i< were noted.
25
-------
TABLE 8. MOST FREQUENT RANGES OF GENERATOR LOAD BY TIME OF DAY
Hour
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Unit 1
Load (MW)
125-150
125-150
125-150
125-150
125-150
125-150
125-150
200-225
225-250
225-250
225-250
200-225
200-225
200-225
200-225
200-225
225-250
200-225
200-225
200-225
200-225
200-225
200-225
200-225
Unit 2
Load (MW)
125-150
150-175
150-175
125-150
150-175
-150-175
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
200-225
150-175
Unit 3
Load (MW)
125-150
125-150
125-150
125-150
125-150
125-150
150-175
200-225
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
225-250
200-225
225-250
200-225
200-225
200-225
200-225
26
-------
AMBIENT POLLUTION CONCENTRATIONS AT FIXED STATIONS
Ambient SO^ concentrations at the monitoring stations never exceeded
the secondary National Ambient Air Quality Standard of 500 ppo averaged over
3 hours, the primary standard of 140 ppb averaged over 24 hours, or the annual
standard of 30 ppb. The low sulfur coal (generally less than 1 percent sulfur)
burned by the plant was probably a major reason why higher values were not
observed.
Mean pollutant concentration measurements were computed over the
period for which there is properly calibrated data (i.e., after Julian Day 286
of 1976 for S0? and after Julian Day 324 of 1976 for NO , N02, and NO).
The means appear in Table 9. The closest station (Tower") recorded the highest
mecn S0? and NO concentrations while the most distant stations (:
-------
Much greater confidence can be placed in individual high pollutant
readings. Tables 10 and 11 contain a tabulation of the 10 highest hourly
average S0~ and NO concentrations at each of the eight fixed stations.
A particularly interesting feature of the high SO- concentrations is the
fact that at four stations (Tower, Nashs, Castle and Lambert) most of the
high values occurred in the late morning. For example, all of the 10 high-
est SO- concentrations at Castle occurred between 9 a.m. and noon. Munsey,
along with the ridge-top stations (Hockey, Johnson, and Kents) do not show
this phenomenon. It is hypothesized that fumigation of the plume, as early
morning inversions were being eroded from the surface, is the cause of these
results. Investigation of this hypothesis is fully described in Section 8.
This pattern is less evident with NO , although Nashs and Castle each
showed 7 cases of the 10 highest NO Concentrations between 10 a.m. and
noon. All 10 of the highest concentrations at Hockey occurred between 3 a.m.
and 11 a.m., and all 10 highest NO values at Munsey occurred between 6 a.m.
and 11 a.m.
Tables 12 through 19 present the frequency distributions of measured
hourly average SO- concentrations at the eight fixed stations stratified
by hour of the day. Figures 3 through 10 show the mean S02 concentrations
for each hour of the day for each of the eight stations. Tower, Castlewood,
Munsey, and Nashs have distinct peaks in the late morning; the maximum aver-
age concentration is reached at 1100 EST at these stations. Less pronounced
maxima are also noted for Lambert and Kents for 1100 EST. These results also
support the hypothesis of prolonged fumigation incidents in the late morning
hours as a major cause for elevated SO- levels in the vicinity of the Clinch
River Plant. With the exception of Kents, which is the most distant station,
all six stations mentioned above are non-ridgetop stations. The ridgetop
stations, Hockey and Johnson, show substantially different diurnal patterns than
the other six stations. Johnson's average concentration is the highest during
the hours 0100-0900 EST; a distinct minimum is reached by late afternoon. At
Hockey two distinct peaks are evident, one occurring at 0300-0400 EST and the
second occurring from 0900-1000 EST. The high nighttime concentrations at
these two ridgetop stations are probably caused by stable plumes from the
power plant passing near ridgetop level. The 0900-1000 maximum at Hockey
may be related to the fumigation phenomenon at the six lower-level stations.
Listed below are the percent frequencies of occurrence of SO- concen-
trations less than 10 ppb:
Tower 75.5%
Munsey 89.2%
Castle 84.2%
Nashs 93.9%
Hockey 85.0%
Lambert 92.0%
Johnson 74.1%
Kents 95.7%
28
-------
TABLE id. TEN HIGHEST HOURLY iVERAGE SO CON CENTRA TICNS OBSERVED AT EACH MONITORING SITE
Rank
1
2
3
4
5
6
7
3
9
10
SO,
Cone.
(?Pb)
344
317
308
274
263
261
260
259
252
251
Tower
Date
7/25/77
"/2S/77
11/25/76
7/18/77
7/25,77
12/25/76
3/14/77
7/6/77
8/7/77
3/1/77
Hour
12
13
10
10
11
2
11
3
10
12
SO
Cone.
(PPb)
318
219
215
190
ISO
152
150
148
146
138
Hockey
Date
6/30/77
12/14/76
7/4/77
10/23/76
12/21/76
10/12/76
12/22/76
7/5/77
2/2/77
11/4/76
Hour
5
9
3
3
24
9
2
3
3
9
SO,
Cone.
(PPb)
200
ISO
134
12-
35
30
~5
71
66
61
Kent
Date
2/22/77
10/23/76
2/22/77
12/10/76
12.' 10/ 76
12/19/76
12/19/76
12/10/76
12/19/75
12/25/76
Hour
13
3
12
9
0
6
7
-
5
6
50 ,
Cone.
(PP'o)
105
••7
76
75
71
70
67
66
66
63
Nash
Date
1/23/77
6/7/77
5 / 17/~*7
11/2/76
2/10/77
2/10/77
1/28/77
6/5/77
11/26/77
6/1/77
Hour
14
12
10
11
13
12
11
11
11
3
Rank
1
2
3
4
5
6
7
3
9
10
so.
Cone.
C?pb)
268
260
241
21S
197
172
149
141
137
135
Castle
Date
11/2/76
2/8/77
10/29/76
2/3/77
3/21/77
2/8/77
10/29/76
2/11/77
3/21/77
2/11/77
Hour
9
11
11
10
11
12
9
10
12
12
SO
Cone.
(PPb)
182
178
167
135
125
119
119
ill
99
99
Munsey
Date
7/5/77
7/5/77
2/2/77
3/15/77
2/18/77
12/22/76
2/18/77
5/16/77
2/16/77
2/16/77
Hour
3
9
3
15
16
2
15
10
i i
13
SO
Cone.
(PP°)
246
199
197
151
136
131
127
124
110
199
jonnson
Date
3/1/77
5/10/77
5/11/77
A/2/77
5/15/77
3/1/77
1/11/77
1/11/77
3/2/77
1, 11/77
Hour
24
23
3
;
11
23
6
5
3
7
sc
Cone.
(PPb)
77
71
70
63
67
57
43
i4
42
33
Lambert
Date
5/ 16/77
4/19/-7
2/11/7-
1/23/7-
2/16/77
1 ' 23/ 77
2/ 11/77
2/-1, 77
12/23/75
12/1S/76
Hour
1!
11
12
12
10
.2
11
1 3
17
19
29
-------
TEN HIGHEST HOURLY AVERAGE
AT EACH MONITORING SITE
,0x CONCENTRATIONS OBSERVED
Rar,<
1
7
3
4
D
0
/
3
9
10
Rank.
1
2
3
4
5
6
7
8
9
10
NO
Cone.
(opo,
619
549
432
467
457
443
430
420
414
397
NO
Cone.
(ppb)
334
601
553
262
235
229
223
226
211
207
Tower
Date
1/28/ 77
12/30/75
3/15/77
12/30/75
7/7/77
1/26,77
1/23/77
12/23/76
12/30/75
12/23/75
Nasns
Date
8/10/77
6/7/77
8/10/77
2/10/77
7/8/77
2/10/77
7/12/77
2/H/77
6/7/77
5/6/77
-our
1C
'n
] ~
zc
3
f.0
:3
1C
10
^
Hour
16
12
15
12
10
1 1
1C
12
13
10
NO
Cone*
jpC 1
563
399
329
255
219
205
i /U
164
146
130
NO
Cone.
(pob)
316
589
464
435
410
394
375
365
350
290
hockey
ja'.e
5/30/ 77
7; Si 71
8/15/77
6/30/77
7/5/77
7/2/77
4/20/77
3/21/77
5/14/77
7/20/77
Castle
Jat9
2/15/77
2/8/77
2/8/77
2/11/77
2/11/77
2/11/77
2/3/77
2/11/77
2/10/77
2/1 177
-our
0
3
9
d
3
5
d
"j i
i V
•5
J
Hour
15
11
10
9
10
i i
12
12
10
a
NO
X
Lone.
i DDD /
353
103
101
92
89
o3
30
79
79
73
NO
Cone'.
tPP°)
419
408
297
230
275
252
206
186
173
170
,
-------
TABLE 12. FREQUENCY DISTRIBUTION OF S0? VS. HOUR FROM 762R6 to 77773, TOWER SITE
(ROW 8, COLUMN HIADINGS ARE HIGH ENDS OF INTERVALS)
5.0
1.0 J11
? . fl 7(0
H '•" "'
0 4.!' 717
R "•'' ? »
1 H . (I 714
1 i . '1 ? in
711.0 771
71... M5
7,1.0 711
7 1 . 11 711
71.') 7 4 II
pnTM.s 40^0
?« KOJ VM,
1 0 . 1)
77
76
}(>
14
27
7.6
31
7»
71
1?
15
10
14
37
1?
15
16
17
7»
<6
71
7.
741
IH S 1,1
SOE (rg/
15.0 70.0
6 »
1 9
1 1
11 11
4 14
17 17
HI ^
I ?
15 10
If, 11
16 16
16 16
13 n
33 1 J
16 11
M 15
II IJ
17 11
0 1 6
17 11
14 17
^ 1 5
1 1 t
1 4 »
715 115
• R1 THAN 0 .
m')
7-i.O
4
11
4
5
o
17
0
6
17
6
1
15
17
7
1 1
10
1
1
7
6
4
1
175
.000
10.0 40.0 50.0
776
741
n 4 6
547
1 « 7
1 7 5
1 11 ft
11 11 16
10 11 «
it 30 (1
in 71 1
6 16 11
4 16 14
16 17 It
10 15 »
755
441
754
6 7 1
754
751
H 5 4
171 717 155
0 Rfl7 V
60. O
7
0
7
0
4
6
11
U
14
1 3
10
14
1
7
6
1
0
1
1
7
1
1 17
M.lirl
10.0
1
t
7
4
1
4
a
11
ID
10
16
10
tfl
•
4
1
'
1
1
1
7
147
1 MTU
100. n
3
1
1
1
0
1
0
1
7
n
14
H
10
»
10
4
1
1
1
0
7
7
1
n
»1
< TH»M
150.0 7.00.0
1 0
7 1
1 0
1 0
7 1
1 0
7 0
11 7
10 1
1 *
7 1
n 7
7 3
5 I
5 0
3 0
4 0
1 1
7 t
0 0
1 0
1 0
1 t
16 15
600.0
100.0 500.0
0 0
1 0
o 0
0 0
1 0
0 0
o o
0 0
4 1
r, n
i 1
t 1
1 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
1 7 1
TOIM.R
11 3
719
1 t 1
317
101
inn
107
3 Of)
305
30)1
305
310
110
306
317
317
316
316
116
316
114
114
314
31 3
31
-------
TABLE 13. FREQUENCY DISTRIBUTION OF S0? VS. HOUR FROM 76286 TO 77273, MUNSEY SITE
(ROW & COLUMN HEADINGS ARE HIGH ENDS OF INTERVALS)
S02 (ug/m3)
S.O 10.0 1S.O 70.0 75.n 10.0 40.0 50.n lln.O dO.O 100.0 ISO.O 700.0 100.0 500.0 TOT»I,B
1.0 79ft 17
7.0 7»B 71
^| J.n 7 B «; 71
0 4.0 ?BS 7"
u
R s-° "« '*
*.0 7B5 7S
7.0 7H7 70
»,0 771 77
1.0 ?Sf> 70
lll.O 711 71
1 t .0 7)1 31
17.0 7 <1 41
H. 0 717 SO
14.0 744 SO
15.0 7S1 40
1H.O 7S4 SO
17.0 7l.fi 41
1».0 771 10
11.0 7»f, 17
20.0 7») 11
71.1 JOfl 7S
77.0 ?«f, 74
71.1 371 10
71.1 >
-------
TABLE 14. FREQUENCY DISTRIBUTION OF S0? VS. HOUR FROM 76286 TO //?7.i, CASTI E SITE
(ROW & COLUMN HEADINGS ARE HIGH ENDS OF INTERVALS)
S . 0
1.1 145
7.1 143
H '•" ""
0 4.0 1 4ft
u
n S.O 111
ft . 0 140
7. ft 113
S.1 177
i . n to?
10.0 'J t
1 1 .0 04
17.0 fi R
13.1 'M
14.0 |r>4
1 S . r) 110
1ft. n iis
17.0 1 75
1 n.o 117
11.0 IIS
?0 . 0 |40
71.0 140
77.0 11"
73. H 117
74.0 145
lo.o
IS
1 7
27
17
tft
17
71
74
71
7»
11
IS
14
1 1
1 ft
11
IS
Ift
71
I*
IR
70
11
1 4
rtlTHI.S 31Hh 47ft
it nit? vault s i,
S02 (ng/m3)
15.0 ?0.0 75.0
1 1 4
1 3 4
777
^ s 1
17 1 7
It 4 3
1037
11 4 3
11 1 *
17 1 7
14 1? IS
IS 10 1
1? 10 IS
15 1 J 7
17 S 7
10 fl ft
1 4 4
441
ft 4 7
ft 7. 7
417
ft 3 4
ft 1 '
71? 1 )•> 1 in
FSS THAN It. OOO
30.0 40.0
1 7
1 t
0 t
0 1
1 0
1 0
1 0
t 0
t 1
ft 3
.1 1
10 1
ft 5
1 ?
S 1
1 4
1 t
0 1
0 0
0 0
1 0
o 1
\ 0
SS 4ft
0
50.0 60.0
0 0
1 0
0 0
0 0
0 0
0 0
i o
1 0
0 1
1 0
* s
S 0
3 5
2 S
1 0
6 0
0 0
0 0
0 0
0 1
1 0
0 0
1 0
75 14
00.0
0
0
0
0
0
n
0
J
1
1
t
0
t
7
0
0
0
0
0
0
0
0
0
0
n
100.0 150.0 700.0
ooo
000
000
ooo
000
000
o o.o
,o o o
030
5 J 0
0 ? 1
} 7 1
2 ' 0 0
600
ft 0 0
ooo
000
1 0 0
0 1 0
000
000
000
000
000
S It 7
300.0 500.0
o o
0 0
0 0
0 0
o o
0 0
0 0
0 0
1 0
1 0
5 0
0 0
0 0
0 0
0 0
0 I)
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0
4 0
TOT«I
17S
I7J
i n
177
in
171
171
17o
U7
mft
Ifi7
IKK
167
164
1 ft 4
1 rm
1(7
U9
IKK
Id
tfifl
161
161
IF.Q
1 71
33
-------
TABLE 15. FREQUENCY DISTRIBUTION OF S0? VS. HOUR FROM 762R6 TO 77273. NASHS SITF
(ROW & COLUMN HEADINGS ARE HIGH ENDS OF INTERVALS)
1 k ft
3.0
1.0
4.0
s.O
(,.0
7.0
fl.O
O.I)
10.0
1 I.O
17.0
1 I.O
14.0
IS. II
d . <1
1 I.O
1 fi . 0
10.1)
70.0
71.0
77.0
71. T
74.0
mTM.I
•>.o
764
7M
7*7
?M
760
7">S
3*1
7SS
741
371
III
1'IB
700
371
711
7)7
Tin
7S7
7SB
764
7**
7*5
7*1;
7<,s
SO IB
10.0
17
11
17
14
11
34
14
44
1*
43
47
4S
47
41
4n
44
10
10
in
11
i*
17
17
H
01 1
SO
11.0
7
1
4
1
7
•5
1
'
»
13
17
17
R
in
7
in
i
4
4
1
7
1
7
1
17"
2 \ v9/
70.0
7
1
7
1
4
1
0
1
3
7
t?
II
1!
•s
7
*
f
6
4
7
1
1
•
7
10S
'm )
71.'
7
1
7
n
0
7
1
1
1
f,
in
1
7
S
0
s
7
n
1
0
0
1
o
1
ss
n I 1 o o n o a a i>
looionoooo
H I.O 7*7 17 47700010 R6000
^ 4.8 ?M 14 1 | n I 1 0 I 0 fl 0 0 0 0104
u
p S.O 760 11 740001 10 00000 101
oi oioonnon 107
i 1 I I o o n o o o lol
oool I D oooo 107
i l 1 o o o n n o OS9<>
1 1 7 I 7 '1 (I 5 n 03 04
B»?2<10000 30}
^nujjlBood loo
7(117110000 107.
3 f, O'l 0 0 I 0 0 0 100
74170(101100 105
0410000000 lot
ol oooonooo los
7700000000 10$
ol oooonooo loft
01 10000000 10S
onioooonoo loft
1 1 1 0 0 0 0 0 0 0107
0710000000 10*
'> 1 1 n o o o n o oiofi
1* SS 11 IB |l 1 1 0 O 0
1 51)7 VAt.HF.S I,FSS THAN O.Onn 0 Sri7 VRI.IfP'; r,RRT,FR THUN ">00.0
34
-------
TABLE 16. FREQUENCY DISTRIBUTION or S02 VS. HOUR FROM 7G2K6 TO 7/2/3, HOCKFY SHE
(ROW & COLUMN HEADINGS ARE HIGH ENDS OF INTERVALS)
(ug/m3)
1,0 7.M Ml 1 » <,
? , i ? ; 1 M* ) R ft
i . '» ? 1 ; *"i 1 1 Q
i , t> y t i »• "i 11 r-
•j , (1 ? 1 -s M> 17 17
M. J 7 ! S 57 1 *> in
i ti. f> ;>M in 1 7 in
11.0. 717 11 1 <) 0
,,.r, „, ,0 ?l 7
,,S -* 70 M
ll.l) >'i I 17 Is- M
1 '• 1 'M 1 < '.
70.0 7 < P 1 ' 11 11
71.0 71" 41 77 1 *
77.1 ) \1 '1 17 ! 1
,,.„
71.3 > i * si ;i 7
1 f) ^ 1 1 •; n
A S 1 1 1 7 1
•. I ) 7 1 7 7 1
1 H 1 f. 7 1 0
1 7 7 1 3 1 1
7377071
oil ) 1 4 4 J
17 ^ 7 J 1 1 1
•; 1 1 t, 4 7 o 1
ft 4 P 1001
1 r, , 1 | f. .1
7 7 7 1 n 1 0
4 1 7 1 1 0 fl
i i i > o o n
f, i l 7 l 1 n
5111111
',111100
? T S 7 0 ft 0
..114701
''',-,1771
1 "
1 r'
1 1
1 o
1 o
o n
n o
1 n
1 ]
7 0
1 0
o n
0 0
0 0
n o
0 1
o r*
n f>
n 'i
n o
n o
n n
r» i
i i
t) o iin
n n Ml
1 fl 179
o n 1 70
n l 17ft
n n 177
n o 177
0 n 176
t 0 1 J 1
o n 175
o n 174
0 0 17fi
r> f> 174
n o 1 7 1
i 0 17*
0 n 1 7 *>
0 f> 1 ? R
n 0 1 J *»
U o 179
o n 121
n n l in
f) n fin
o 0 Mt
n n 111
35
-------
TARLE 17. FREQUENCY DISTRIBUTION OF S0? VS. HOUR FROM 7f,7Bf, TO //?/!, LAMBERT SITE
(ROW & COLUMN HEADINGS ARF HIGH ENDS Of INTFPVAIS)
1
2
H 3
n
U 4
u
R 5
-
r
«
9
10
It
12
1 J
14
IS
f-
1 1
l»
1 1
20
21
23
71
74
rnrA
s.o
.0 !°2
.0 1 "1
.0 !•>/
,n fi
.0 f»J
.0 |05
.0 1 «&
.0 !»>.
.0 l»7
.0 lf.9
.1) IS1
.0 1M
.0 165
.0 i ;s
.0 |R*
.0 117
.0 1 1 R
.0 ?l>1
.n 70 1)
.0 111
.0 1 tj
.0 112
.0 I OR
.0 rn
I/S 4SO|
0 Sit 7 YAI,
1 1).
1 1
R
12
in
12
1
1 t
1 1
IS
IS
1 7
11
in
10
t 3
S
1?
7
«
q
R
17
\f,
1 1
717
IH-,
so2
n ts.o
q
•>
•s
4
4
7
4
1
7
9
1 1
J
1 1
12
4
1 1
4
R
S
4
7
10
1 7
10
t«.
I.FSS IHUN
([iQ/ni )
?o .rt ?s.n in. 'i
* ? o
7 1 0
1 0 t
4 1 0
7 * ?
1 4 2
S 1 0
ft 7 1
4 ? 1
ri 7 1
911
in i i
742
T 1 7
7 3 1
f. 1 0
S 1 (1
* 7 n
1 7 1
« 1 7
a 7 7
04?
710
? ; i
tin MI ; t
n .onn
1 o . n s f i . n *o . n n n . n i o r) . o
n o o o i'.
o n n o o
n n o f> (i
n o n o o
n 0 0 It ')
\ n n o n
1 o n f) o
n n n n (i
M *> i 'i n
it --.'I f v ni Hi «; t.^pi-1- 11 rii«n
1 so , n ?')(». <»
') n
n o
n (i
0 1)
f> I>
O (1
0 0
0 0
.1 0
n o
ft 0
n r>
f) 0
n n
(t ii
n i)
fl 0
O ')
(1 0
(1 0
n f)
n o
n (i
„
(i n
M,».n
) 1 R
71 0
j ;o
7IS
71 7
7IS
71 1
717
717
70D
710
710
7117
717
770
771
331
77?
37?
770
771
3 ? '
36
-------
TABLE 18. FREQUENCY DISTRIBUTION OF SO-, VS. HOUR FROM 76?86 TO 77273, JOHNSON SITE
(ROW & COLUMN HEADINGS ARE HIGH ENDS OF INTERVALS)
I . 0
?.f> M 7 is ?t 1*1 ^ <> B ? i ?
?.n MS 7S 4-5 If. f. 1 ft 1 n ? 7 i i .1 . n 1 o . it t s . o ?n . n /s . o
11? is 71 11 i
Ml f? 11 1 7 9
M9 7 t 11 7" 7
1 17 7| «1 71 )
Ml 7ft 11 11 ||
|}« 7x 1; in ,1
177 is n 1 1 it
171 IP *H 70 7
t?S 41 M 70 H
1 16 41 77 17 9
1^ H i7 9 4
IS? S? 10 B 2
I 71 11 |H in 7
I77 11 7t 1 1
1 7B 11 3h f, 1
170 11 14 Q )
J n . 0 1 (1 .
(> P
«. 1
> 1
J <
^ f-
S
1 1
I 2
7 •>
7 9
< 4
7 t
0 7
< I
') 5
7 7
I
I 0 . 0 171 4 f> * H 7 (t 7 I 2 ? (I 4 1
11. f> l?S 4^ M 70 4 7 S 1 1 7
12. f> Mb 4H 7? 17 9 7 1 7 o 7 '1 0 0 0 n 7 S 7
M.o t^*> 14 i7 Q 4 i 4 4 n t n n n i n?si
11.D I ft 7 s ? i q n 2 7 i ? n i o n n n ti 7 S I
ts.n
7 i « <* o n>S1
1 ') t> n .1 n 7Sfl
( (i 0 '» n O ?SR
|9.0 IMl 5J 1Q t» 7 7 2 7 0 1 " 0 0 n o 7 5 »
?O.D ISf^ SI 73 M S ; 1 1 I 1 o f! i) M M 757
; 1 . f) 1S1 *" M M 9 7 1 * n 1 o tt n n n ;s*
J^.l) 1 1 h 11 II t« 1 » S 1 I 'I 1 I' 11 ' n?sft
?j.o MS ^^ 11 ;; 1 i 1 i i ; t t i o "?s"»
i ISM **17 ; 4 2 1*1 i 117 F>I "' s^ i? *'J ;i
0 •>!)? Vfll.t!* S |,S ^S THAW 0 . ntMi (I ','17 V Hl.'ll- ^ ?,«("(•« fltH 'J
37
-------
TABLE 19. FREQUENCY DISTRIBUTION OF S0? VS. HOUR FROM 70280 TO 77^73, KENTS SI1F
(ROW & COLUMN HEADINGS ARE HIGH ENDS OF INTERVALS)
1 .0
7.0
H '•"
O. n
" . "
u
ft s-"
ft.O
7.0
1.0
10.0
11.0
17.0
1 1.0
14.0
1 S . 0
1ft. 0
17.0
m.o
11.0
70.0
71 .0
77.0
71.0
74. 3
rOTM.'
s.n
IOS
1"1
ini
7 'If,
>o
771
?»1
7R»
711
107
101
104
101
10,
10B
10,
101
>os
7171
S02 (ug/m
IO.O IS.O 70.0
127
J 7 1
777
7 ft 1
•ill
10 IS 1
in i 4
17 1 ft
Ift 7 1
in 11 4
71 7 7
11 II 1
11 t \
10 7 1
10 6 }
17 4 7
11 4 4
951
441
0 4 S
7 -S 1
» •( 1
741 110 ft'l
3)
7S .0 10.0
1 0
7 1
7 1
7 0
1 0
1 7
1 1
7 7
7 0
4 1
4 1
7 1
0 1
1 0
7 0
1 0
0 1
1 1
7 1
0 0
7 0
1 7
1» 7(
10, ,
1
1
0
1
J
1
1
7
1
1
4
1
1
7
1
0
0
0
0
0
0
1
7
1
11
ici.o *osn so.o $0.n (tn.n 100.o t^o.o 700.0 ino.o soo.o TTTM.R
t n n o o o o 017*;
o ? n o o D o o*;n
10000(00 *?7
1 7000000 1??
^ 1 1 0 n o o OUT
n n ? i oooo 170
7 ft 7 0 0 0 0 0^(5
0 7 0 0 0 0 0 0130
I o o o t o n 011%
1 1 0 0 0 0 0 QMfi
7 « 0 0 0 0 n 0«Mfi
70001000 11*
t o n o o i o oii<*
n o f* n o o o 0^17
0 0 1 0 0 0 0 0^10
o o n i r> o o 0*71
oooooooo *?•;
i o o o o o o n <74
l o o n o o o n 174
i o o o o o o n i?^
I n n o o n o n i?s
OOOOOOOO 174
t n o o o n n 0174
70000000 13S
7714 S I ? ? 0 0
0 SH7 V M.UF ft t.HPG* R T"fl»l *0f).0
38
-------
20
IS
16
14
0~ 10
1 2 3 4 5 6 7 5 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour
Figure 3. Diurnal variation ot mean SO2 concentrations at Lambert.
" 12
f 10
1 2 345 6
39 10 11 12 13 14 15 16 17 IS 19 20
Hour
21 22 23 24
Figure 4. Diurnal variation ot mean SO^ coucentr^tions at K.enti.
39
-------
20
13
16
14
12
10
3
6
4
1 2 3 45 6 7 3 9 10 11 12 13 14 L5 16 17 IS 19 20 21
Hour
Figure 5. Diurnal variation of mean SO,, concentrations at Nashs.
22 23 24
20
18
16
— 14
j
S 12
(M
8 10
345673
10 11 12 13 14 15 16 17 18 19 20 21
Hour
22 23 24
Figure 6. Diurnal variation of mean SO., concentrations at Hockey.
40
-------
20
18 |—
16 _
3 12
IN
'^ 10
~1
2 S
*i
6
4
J I
1 2 3 45 67 39 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
Figure 7. Diurnal variation of mean SO,, concentrations at Johnson.
3 4 5 6 7 S 9 10 11 12 13 14 15 16 17 18 19 20 2! 22 23 24
Figure S. Diurnal variation of mean SO-> concentrations at Tower.
-------
20
18
16
s 14
— 12
f 10
I 8
2
6
4
2
I i 1.1 J I i I 1 I 1 N 1 1 1 I I I 1 M I 1
2 3 4 5 6 7 3 9 10 11 12 13 14 IS 16 17 18 19 20 21 22 23 24
Hour
Figure 9. Diurnal variation of mean $®2 concentrations at Vtunsey.
20
18
16
14
S ,-5
-I
"oj 10
O
Wl
S 6
4
2
0
1 I 1 1 1 1 1 I i 1 1 i 1 1 i 1 1 J I i i S II
2345 6 7 8 9 10 11 12 13 14 IS 16 17 18 19 20 21 22 23 24
Hour
Figure 10. Diurnal variation of mean SO^ concentrations at Castlewood.
-------
These results Indicate that during the vast majority of the time the power
plr-nt plume was not impacting a particular station. Both Johnson, located
5.76 km from the plant, and the closest station to the plant (Tower) recorded
tht lowest frequency of extremely low S0~ concentrations (less than 10 ppb).
This indicates that either the power plant plume impacted these stations the
most frecuently or other smaller sources also affected the stations. The
Moss No. 3 Coal Preparation Plant which burns a small amount of low-sulfur
coal is located 0.84 km WNW of the Tower station and may have a small impact
on SC? concentrations at this site. However, no known sources of SO,-,
existed in the immediate vicinity of the Johnson static1- .
Frequency distributions of S00 concentrations versus various metec-
roloqical parameters were produced6for the entire monitoring period and also
on an individual season basis. The relationships between the concentration
data and the meteorological conditions are discussed in Section 8. Differences
between seasons of the year nave been no! ed in the distributions, however. At
five stations (Johnson, Kents, Lambert, :',unsey and Sashs) the season with the
greaipst frequency of m'nn S07 concentre.Lions ('.reator than 1GC ,>pb; was
wirr>.,r. The Cctstlev.'ood results showed the niohe^t fror.uerxv 1-1 sarir.n, while
Tower ar,d hockey had tneir hpuhest frequency of '.iqb "C , concentrations in
summer and fail.
AMBIENT POLLUTION CONCENTRATIONS - ^OL^! f VAf-'
Trip mcni!*7. oround samplinq was conducted u--ina o vari equipped to measure
SO,,, NO,, NO, N00, and 0^, as well as wind speed, direction, anc'
temperature. Dufina the"earlv portion or t^> field ^tudy the vind directic^n at the Hockey 30 r l.'.-vol. Position' voro
wine's shifted durino the day. A total r,f r:66 hours -T.C' '<•'':'• ."i
was performed by the v^n. Of this total, M'P Hours f',rc '> . 'I'in
fo"med in the stationery mod? and ?J7 1"f.iir<, fn^i 37 ifMlr", ^rr
the van in motion.
T.ib'f ?0 presents i;he ;,ean i:clliit^nf c' frv.
(fixed) samples, all the mov PQ sa^'plps, JPC! fr'" a1! -iSi
For each pciliitant chese ^eans are rppaff^ than any -f r
co 'cer.^rdt ions computed for the eioht fixed station?., "ir
var, was located under tfie plunf more often t: an were on>
The means frr the stationary samp lino arn qr^ater th^n
sampling for both SO,, and NO indicatir'n that the
the plume more often during ^he stationary samples
samples or that the ctat n, orsry same IPS w^^c -lore ri
center 1 int .
43
-------
—« "^ 3D ./I —t
Q
z;
a
H
O
u,
O
P
PS
2
w
u
§
h
|
-i
2
w
~i
CQ
-i
t J
2
44
-------
The number of significant S0? and NO measurements also illus-
trate 'che usefulness of the van in making plume measurements. Taoles 21
through 25 present frequency distributions for the five pollutants for the
stationary sampling events. Only 49.8% of the sarnnles had mean SO,, con-
centrations less than approximately 20 ppb, whereas the aercentaqe^for the
fixed station (Tower) with the least number of hourly average S00 concentra-
tions less than 20 ppb was 83.5%. Therefore, a power ulant plume can be
detected at the ground considerably more often by a van than ny a fixed scat ion.
The mobile van data were also analyzed through the i or struct i or f~or
eacr, ooliutant of two-way tanles strati fyinc the stationary sample data by the
following three methods: (1) distance from the uloni versus elevation; (2)
hour versus elevation; and (3) distance from the pl^nl; versus Direction fron
the plant. These tables, presented in Aopencrix /,, crive the- mean '.oncentrat ion
ano' number of samples for each combination of distance and elovalion, hour and
elevation, and distance and direction. The category with the hiqnesl mean
SO^ concentration was 0.0 to 0.5 km and 457-488 m MSL (1500-1600 feet MSL)
(i.e., very close to the plant). These concentrations may have resulted
from downwash of the plume near the plant caused by a nearby upwind ridge.
Seventy-five nercent of the fixed samples were taken from 3 km to 10 km
from the plant. The most frequent sampling location was the category 3 to
5 km from the plant and 518-579 m MSL (1700 to 1900 feet MSL). Most of the
sampling was performed between 0800 and 1800 local standard time. At dis-
tances greater than 5 km the direction classes 30° to 60° and 60° to 90°
contained the highest mean SOp concentrations. This result may be caused
by the prevailing southwesterly flow over the region.
The van data are also useful in examining the NO/NO ratio as a
function of distance from the plant. Ground-level NO concentrations are much
higher than N00 close to the plant, in the 0.0 to 0.5 km distance ranoe and
457-483 m MSL (1500 to 1600 feet MSL) category, the mean NO concentration was
410 ppb while the mean N02 concentration was 73 ppb, giving a mean ratio
of 5.6. In the 2 to 3 km range the mean NO to N02 concentrations give a ratio
of 0.8 in the 518-579 m MSL (1700 to 1900 feet MSL) elevation category, for
which there are 44 measurements. The NO to N02 ratio seems to be leveled off
at about the 0.8 level for the remainder of the travel distances, suggesting
that the reaction rate has greatly slowed down or reached some sort of equi-
librium. Table 26 presents the average ozone concentrations for each of 9
distance ranges and the number of cases in each range. There is the suggestion
of evidence of ozone depletion in the plume at the 1 to 2 km and 2 to 3 km
enhance ranges as these ranges show lower 03 concentrations tnan both shorter
ana longer distances from the plant.
I1.
-------
TABLE 21. FREQUENCY DISTRIUBTION OF SO CONCENTRATIONS MEASURED IN VAN
WHILE IN STATIONARY MODE
UNI is: pfn
C JM I'tUlKA I UHJ KftlJdK b
h u U k. N C »
<-!23b.
. 2 1
.bfc
. 4t
,0b
. b4
.24
'.42
.02
.bl
. 20
. au
.39
. yb
. bB
. . _r
1
1 /I
4i!b
i JW
I4b
93
I.1 . J
14.3
ib.b
7 .«
4.U
i .y
U.H
U.4
0. 3
U.b
U.b
0.3
O.U
U.U
1. /
TABLE 22. FREQUENCY DISTRIBUTION OF NOX CONCENTRATIONS MEASURED IN VAN
WHILE IN STATIONARY MODE
UNUS:
C J
393
430
4b«
. 1 3
. 19
.bl
.H3
. Ib
.4b
. 7H
. Ill
. 42
. /4
.Ue>
. 3H
. 10
.02
, 20
b /
, y4
132
, i fay
, 20b
, 244
, 2« I
, 11M
, !'>()
, 39 i
, 430
, 4 b H
, bOb
>bOb.
. i y
. 3 1
.H3
.Ib
.4b
. /B
. 1 0
. 42
. / 't
. Ob
. *H
. 10
. 02
.34
34
0
44y
3fa9
123
12
M
U
0
1 1
46
-------
;ABLE 23. FREQUENCY DISTRIBUTION OF NO? CONCENTRATIONS MEASURED IN VAN
WHILE IN STATIONARY MODE
> 1 H A I I UN H
UNJUS:
r K t, 'J u t, iv
PK.Ul. KH (
-3
3
10
i ;
^4
31
3b
4b
b^
by
bb
73
bO
8 7
94
101
. bH
. -43
.4b
. ^b
.47
.44
.so
.bl
.b/.
.b4
.bb
.bb
. bH
.by
.bO
.bl
<-3
3
10
i ;
/!4
31
3a
4b
b;
b^
bb
7 3
ao
a 7
y4
101
10H
>10H .
. b» J
. 43
.4b
. 4b
.47
. 4y
.bO
.bl
.b/
.b4
. bb
.bb
. bB
.b9
.bO
.bl
.b3
b3
1
1
1
faO
oy
01
73
3y
3b
i \
I'l
1 j
1 0
10
4
b
i
1 1
17.0
U .b
10. ;
7. 7
t. 1
*• • /
i. i
1.3
1 .4
1 . 1
1 . 1
0. 4
O.b
0.3
1 . i.
TABLE 24. FREQUENCY DISTRIBUTION OF NO CONCENTRATIONS MEASURED IN
WHILE IN STATIONARY MODE
o N1 1 •}: p t-> h
- J,
Jbb!«l j
^y» . y4 ,
33J .07,
3bb. 19,
3 y H . 3 i ,
4^
<~ 3 i . 3 4
o . 7 y
33. y^
fa 7 . Ob
100.17
133.30
lbb.43
199. bfa
***• bt)
^y« . 94
33^.07
3bb. 19
39«. 3
4 . t
I . H
0. 1
0. i
(I .':
o. u
i/..'
i) . V1
o. 1
47
-------
TABLE 25. FREQUENCY DISTRIBUTION OF CONCENTRATIONS MEASURED IN VAN
WHILE IN STATIONARY MODE
uw 11b: PPM
C JU JLU1 K» 1 iUN KH,,Lil-b i- He. JJK'i:. ( ''i-.H',. I- >•! I
-i.4 i 0 0.0
-j 13, /.4
iiy.bi, Jb.Do yfe B •'
Jb.OO, 40.49 '1 b- 4
40.49, 4b.9b 4H 4-4
4b.y«, 31.47 44 4.0
51.4/, bb.yb 3^ -i.y
bb.yb, 6
73.43, 7b.9^ 9 0.8
7b.9ii,S4.41 y u>;
B4.41, «y.yo y 'J-8
HW^Wll 12 1.1
48
-------
TAB:; 26. ozo^t . • iCLNrRAroN: v^sus
DISTANCE FROM PLANT FOR MOBILE vAi\ DAT
Oistace Ranoe
(km)
Averaae l'o
Concentratior (
sumrsr vi
Cases
0.0
0.5
1.0
2.0
3.0
5.0
7.0
10.0
- 0.5
-1.0
-2.0
- 3.0
- 5.0
-7.0
- 10.0
- 15.0
> 15.0
34
.?5
If:
n
'/','
') "7
!_/
i!8
:c
! o
CONCLUSIONS
Statistics produced from the pollutant monitoring data for the eight
fixed stations show generally low annual-average concentrations of 502
and NOX» The ten highest individual hourly values of S02 and NOX at the
non-ridgetop stations show a tendency for the highest concentrations to
occur in the late morning, suggesting that fumigation of the plume during
inversion breakup is an important process. The diurnal variation pattern
of the mean S02 concentration at these stations also supports this hypo-
thesiSo The mobile van was capable of recording plume impact with greater
frequency than did any of the fixed monitoring stations,, The van measure-
ments of NO and NO? are useful in studying the NO/N02 ratio as a function
of distance from trie plant. This ratio decreased through the first 2 to 3
from the plant and then leveled off at approximately a value of 0.8.
km
49
-------
SECTION 6
ANALYSIS OF PLUhE STRUCTURE
Except in the situation where two or more fixed stations are located
near enough to each other and at different elevations such that both are
affected by the power plant plume simultaneously, fixed stations at ground
level reveal little concerning the structure of the plume. In the Clincn
River monitoring network two fixed stations were situated such that some
plume structure information could be obtained. These were the Hockey and
Munsey stations, both located almost on a direct line from the plant. During
operation of the mobile van, useful data for plume definition were available
for three points. Results of the analysis of the Hockey/Munsey relationship
appear in Section 8, both from a statistical approach and from individual
case studies.
The most useful information concerning plume position and structure is
available from the plume measurements taken by helicopter. The airborne
observations were gathered during two intensive study periods - November 8
to 17, 1976, and July 20 to 28, 1977. The July 1977 data, however, are of
much greater value as the entire data set for July is composed of plume cross
sections flown between specific landmarks. However, these data require
extensive processing to determine the position and dimensions of the plume.
DATA PREPARATION
The helicopter data on the Clinch River Oata Tape* contains a great
degree of detail which must be processed and interpreted to determine
representative cloud dimension parameters such as the Gaussian plume dimen-
sions a and a . Based on notes taken during the flights a sot of plume
travers^ endpoints was compiled. Coordinates for each of these endpoirits were
obtained from U.S. Geological Survey maps. The helicopter dace were recorded
as scans of the instruments at a rate of between 4 and 5 scans per second.
The data from the Clinch River field study is available from the national
Technical Information Service on magnetic tape as LPA Report Number
EPA-600/7/79-OlOb.
50
-------
Eaci scan became a separate record on tape identified by a time, based on the
coordinate ana time information, x and y coordinates had been computed for
each scan. An elevation above sea level was also associated with each recora.
The first step in processing the scan data was to average the data over
periods comparable to the response time of the instrument and to adjust the
data in time and space to account for the instrument response ana lag times.
The lag time is the time required to pump the sample gas into the instru-
ments from the base of the helicopter. The data were separated into identi-
fiable cross sections. A cross section was defined as a group of traverses
flcwn at several different altitudes between a set of fixed r-ndpoints at a
particular distance downwind of the power plant (sec hiuure 11).
N
N
/
777777
Figure 11. Representative helicopter flight path at fixed distance from the power plant.
51
-------
Twenty-four plume cross sections were identified in the July 1977 hel • -
copter data. Table 27 summarizes the times, altitudes and number of traverses
for each cross section. The original scan data in each cross section were
averaged over each 9 successive data points to obtain a value approximately
every 2 seconds. The following instrument response time correction was
applied to these values:
dC.ft) «.(!
cr(t)-c1(t)ta-^+f._Jr
where C (t) is the real pollutant concentration at, time t, C,»t) is the
observed concentration at time t, and a and p are constants'relatec 10 the
time constant of the instrument. Based on information froi:t ttie manufacturers
the following a and p values were used:
S09 NOV, NO
C~ A
ot 4.9 2.0
B 2.56 G.O
No
instrument response adjustment was necessary for tne ozone data.
The corrected cross-section data were plotted and examined for any incon-
sistencies or problems. It was noted that in portions of some cross-sections
the peak SO,,, NO and NO concentrations as well as the minimum 0- concen-
trations for eacft traverse appeared to move back and forth along^the cross-
wind axis with each successive traverse. In addition, the S0? peak always
occurred after the NO peak. The lag of S0? peak after the No peak
averaged 4.5 seconds Tor all traverses except those labeled as events 749-783,
for which the time difference averaged 10.5 seconds. It was hypothesized that
this lag was caused by the delay in SO- gas reaching the instrument vJhich had
a smaller flow rate than the NO instrument. The S00 analyzer drew ?:83
cm /min, the NO -NO analyzer drew 1000 cm /min, and the ozone analyzer
drew 275 cm /min. Using these flow rates and the diameters ana lemjlhs of
the tubes, the times required for the gases to reach their respective instru-
ments were computed. For S0? this computation resulted in 5.91 oeccnds;
for NO and NO the lag time was 1.67 seconds, while for ij , re was only
0.71 seconds. The difference between the S(L lag ana the" NU( lag was
4.24 seconds which is very close to the 4.5 seconds difference between the
S09 and NO peaks observed in the data. The much larger difference
between the times of the S0? and NO peaks for events 749-7h.i was tun
C A
52
-------
1 —
en
>_
=
o
LJL.
oo
SZ
0
1— 1
h-
C__)
UJ
oo
o
rv
(__)
CC
UJ
1—
Q.
O
(__}
»— *
1
UJ
^
O
>_
a:
JE
:z>
CO
CM
UJ
— 1
CO
I—
4- oo
O O)
00
S- S_
fcz ^
^ i.
O)
T3 _J
3 OO
UJ 4J E
OJ
13 1
C 3 OO
•r- 4-J ^~
cn-i- — -
OJ 4J E
CO i —
fi3
__,
T3 C
C O)
UJ >
O)
C -I-J
•r- C
cn oj
CQ OJ
O)
"^ E
C '^
LU +->
•^ O)
cn E
O) ••-
-^ -(->
cu
_J )
.13
c
tn O s_
1/1 T- OJ
O -f-* O
i- U E
C-J) OJ Z!
oo 2:
OO S3- OO CO LO vO CO v"O 1-^ LO i— LO CO CO 'oD CTi O O 30 vi
o f
i— cn o f-- CM
CM ro 'ct '— cn
O
O '•>
Cxi CO CM
O -~O C*1) i
C\J CTi CTi
CNJOJCOLOLO.— OOrOCOOOOCMLD^jOOi— ODiO T vtl OO CT> LO
O ' — CM CO S3- O i —
i — si-
cricrii-DCMCNjrooocrii— co
LOOrOCMOOOOOOOO
'>o i— ro-^-Lor^.ooorO'd-
'JD CNJ X! oo r^. r^ LO r-. CM 1
LO OJ ro O C\l ro i — O O
CTi O O •* s3- co O~i O OJ
r*^ •— O r^O '.D i-O — X) '£> >. jr, ro =3-
O -=3- > — -3- • — O sr OJ > — oo i — i — o
^D O LO ^O cr. PO "O ^r r; r--. cr, CM ro
ro
I I
'— CM co LO
— CXI CM .XI 'XI CM Aj CM CXI
53
-------
result of the use of a new transducer in the S0? analyzer whicn was
initially very slow in response.
The cross section data were modified in the following manner to account
for the lag involved in the gas reaching the instruments. For all events
except 749-783 the SCL data was moved 6 seconds (3 records) back in time,
while the NO , NO and N0? data for all traverses were moved 2 seconds (1
record) back. No adjustment was made to the 0,, data. For events 749-783 the
S02 data was moved 12 seconds back (6 records) to account for both the
additional instrument response time and the regular 6 second lag time.
Nineteen of the 24 cross sections were observed between two end points;
however, Cross Sections #18, #20, #24, #26, and #28 are composed of tra-
verses along more than one line (different sets of checkpoints). Shifting
winds during the cross section necessitated changes in the flight path.
Cross Sections #18, #24, #26, and #28 were flown along three paths, while
Cross Section #20 utilized two traverse paths. Figures 12 through 16 depict
the multiple-path cross sections and also show the mean Hockey 30 m wind
direction for the period of the cross section.
After the final adjusted cross section data were plotted, considerable
movement of the position of the peak concentrations from traverse to travel se
was still evident in portions of some cross sections. The center of mass for
each traverse was computed, and the same phenomenon was observed in these
positions. Movements of the center of mass ranging from 200 to 1000 m along
the cross-wind axis were evident between successive traverses. Table 28
presents the variation in position of the center of mass from traverse to
traverse for three cross sections. It was hypothesized that the cause of this
phenomenon was primarily the meander of the wind direction during successive
traverses which make up the cross section,, Two approaches were used to
explore this idea. The first approach was as follows. The variation
in wind direction between successive Hockey 30 m wind data read outs (2
minutes apart) were reviewed for two of the cases (Cross Sections #4 and
#6). The range of the 2-minute wind directions recorded at the Hockey
station at the 30 m level (approximately the plume height) during the cross
section either equaled or exceeded the range of directions defined by the
range of the positions of the peak S02 concentrations of the individual
traverses (see Figures 17, 18). In these figures the arrows represent the
two-minute wind directions during the cross section while the tic marks on
the cross section line represent the locations of the S02 peaks of the
traverses.
The second approach for analysis of the plume shift phenomenon in-
volved determining the plume centerline position along the traverse path
as defined by the calculated plume configuration at successive 2-minute
intervals. The plume configuration is determined through the use of Hockey
30 m level 2 minute wind data. A computer program was written to advect
every two minutes each previous two-minute wind vector, representing
54
-------
833'
Plant
1 Km
Figure 12. Traverse paths for cross section *13 and
event numbers
Figure 13. Traverse paths for cross section *20 aid
event numbers
55
-------
W
84° Wind
2 Km
Figure 14. Traverse oaths for cross section ^24 and
event numbers
56
-------
Figure 15. Traverse paths for cross section ?26 ana
event numbers
1 Km
Figure 16. Traverse paths for cross section s*28 and
event numbers
57
-------
Cross
sect ion
number
Ait ituae
(m '-iSL)
Distance of
center of mass
from enapou.t (nr.j
oc3
762
792
823
853
883
914
833
914
914
944
975
1005
1036
58
-------
Plant
800 m
Figure 17. Two-minute Hockey 30 m wind directions and locations
of peak SOo concentrations during cross section #4. Numbers
indicate tne chronological order of the wind directions and the
traverses.
59
-------
N
A
800 m
Figure 18. Two-minute Hockey 30 m wind directions and locations
of peak SOo concentrations during cross section #6. Numbers
indicate the chronological order of the wind directions and the
traverses.
60
-------
successive two-minute sections of the plume. The latest vector represents
the latest section of the plume to leave the stack. Printouts were
obtained depicting the positions of these successive vectors using alpha-
betic characters. The procedure is described more fully in Section 7.
Figures 19 and 20 show plume configurations at times 2 minutes apart during
Cross Section #15. A 650 m shift in the plume along the traverse path
occurred during this interval, based on the Hockey winds. Cross Section
#15 was composed of 18 traverses along the same traverse path. The mean
shift of the calculated plume configuration between traverses over the
entire cross section was 290 m, while the mean shift of the peak $62 con-
centration from traverse to traverse was 321 m. The individual shifts in
plume configuration ranged from 40 m to 1080 m, while the shifts in peak
concentration ranged from 15 m to 850 m. However, no correlation was found
between the individual shifts determined from the helicopter data and those
determined from the calculated plume configurations, indicating that
although the turbulence intensity in the vicinity of the plume was similar
in magnitude to that at Hockey, the particular wind direction shifts
occurred at different times.
The usefulness of the helicopter data can be assessed also by compar-
ing the total mass flux of S0£ in the plume across the plane of the cross
section with the plant emission rate. The mass flux was computed by inte-
grating the concentration data across each individual traverse to obtain
a cross wind integrated concentration (CWIC). Each CWIC was then multi-
plied by a height interval extending from halfway between the traverse
elevation and the next lower traverse to halfway between the traverse
elevation and the next higher traverse. The result was a value of the
mass of S02 contained in each horizontal strip of the plume per unit meter
in the downwind direction. The mass for each traverse was then multiplied
by a wind speed appropriate for each altitude, obtained from the pibal
release closest in time to the cross section (sometimes as much as two
hours different). Table 29 presents the computed mass flux for each cross
section along with the SC>2 emission rates for the corresponding hours.
The majority of the mass fluxes are considerably less than the cor-
responding emission rates. The major reason for this result is the lack
of sampling by the helicopter of the entire vertical extent of tne plume
in many cases. Cross sections numbers 9, 10, 11, 18, 25 and 26 show
reasonably good comparison between the computed fluxes and the emission
rates; these cross sections were each composed of a "large numoer of tra-
verses, allowing sampling of the entire plume. Reasonable estimates of the
vertical plume dimension can only be made for these particular cross sec-
tions. The mass flux computation is also subject to error through the use
of wind speed data not observed at either the same time or ot tne same
location as the plume cross section.
PLUME POSITION
The helicopter data provide an excellent means of determining tru> actual
location of the plume for the purposes of comparison witn the location pre-
dicted by tne use of observed wind data. The nelicoptrr measurements are
not useful, however, for the ourposo of obtaining a<" ar.curaU: r
61
-------
s_
O
r- c o
CT> M-
-
CL o
OJ C
I- (0
r? •*-> ~o
c
H- -r- **-
CO O O
LU Q.
c ro
o •• E
C O OJ
O t/i
U 4-> (U
C t-
a) ••- a.
E o a>
r) ex s.
cn a. s-
•r- QJ fO
u- s- E
62
-------
s-
o
O C
CL ITJ
-O r-
C 0.
aj
a
r»- c
Ol 0) E
' — u-j O
IC O.
oj a;
t- -O
o
L.
O 4->
u CJ
OJ =
S- OJ
•«- o
4— 4—
CO-*-1
o c:
U i-J 4-
C I/}
-------
oO
UJ
i
_
o
»— *
00
CO
E"1^
2:
UJ
CM
O
I—
z:
•a:
j
Q.
O
X
• — i
— J
u_
CO
CO
2:
ex.
o
CO
o
LU
o_
«&.
o
0
u_
o
z:
o
CO
o:
CL.
as:
o
o
CTi
CM
UJ
v I
CO
I
OJ
^ j
O i- —
CO U
C 0)
4-> O (/)
rO l/l C">
1 — I/) — '
Q_ -r-
OJ
rd O
S 0)
yi
o a-.
LO ' — ^
03 72
-M ' —
0 4-
<4- U1
O HI
i- S-
^E ^0
^ ^
c:
o
'^3
u s-
O) O)
u> HI
r.
i/l 5
f~^i r_\j CM r*~«» ^— co r*— . co
t o (_^ L£"J r^^ f^-. CO CO co
If) 10 IO p^^ p^^ L£*j LO ^.^
i — CO' — LDOO' — CO
ro i — r^*- CM o^ co CTi ^d~
m ^- m co tn vo co ro
r—
.— >Xl 0 ^t ^"r ^ , 1 - 0
OO CM
i— i—I
- *XJ '^l" CJi i—- LO ..T: i— r>-) <^ i_o OxJ LI^ O
f -O '— ^ CM i— mo >."v' 'o ro «vt~ r^» co LO
i— r~-. r— ~ <^i- -^r oo CM •— — CM co CM CM r—
i — i — CM < — •
i — CM
i
o-i
O
o
(J
o
r?
c
o
u
QJ
o
0)
4->
64
-------
of the structure of the entire plume at a particular instant in time. Plume
dimensions averaged over the period of the cross section (10 minutes to 2
hours) can be derived; these computations are discussed in the subsection
entitled "Plume Dimensions."
Table 30 presents the locations of the center of mass of the plume
(computed from the S02 data) in both the horizontal and vertical. Distances
from the plant were computed perpendicular to the paths of the traverses.
There is considerable difference between the direction in which the center
of mass of the plume was found by the helicopter and the direction in which
an hourly average transport vector, based on the Hockey 30 m winds, would
have carried the plume. These differences may be attributed to either
terrain effects on the airflow in the vicinity of the plume or to differ-
ences in the general flow over the region between Hockey and the location
of the plume.
Cross sections #1 through #5 all show substantial deviation between
the location of the plume center of mass and the predicted location based
on the Hockey 30 m wind data. Cross sections #1 through #3 were performed
between 0956 and 1048 EST on July 23, 1977, when winds in the valley may
still have been showing the effects of early morning inversion conditions
with associated valley-flow regime. A pibal released at 1030 EST showed
winds at approximately 500 m above plant base to be blowing toward a bearing
of 8° which compares well with the plume being found at azimuths ranging
from 7° to 19°. Cross sections #4 and #5 which were performed from 1403
to 1510 EST show less deviation from the Hockey 30 m flow than was evident
earlier in the day.
Cross sections #11 and #12 indicated the plume center of mass to be
at 54°, while the Hockey 30 m transport vector indicated the plume would
have been at 74° and 93° between 1436 and 1621 EST. Pibal data from 1527
EST on July 24, 1977 show winds at plume height to be 252° to 254°, trans-
porting the plume toward bearings of 72° to 74° respectively in agreement
with the Hockey 30 m winds. It is hypothesized that in these cases the
ridge immediately to the east of the plant diverted the plume to the north-
east, causing it to be found at 54°, which is along the northeast slope of
this ridge that parallels Sinkhole Valley (see Figure 2).
On July 27, cross section #18 was performed over a period of approxi-
mately 2 hours between 0910 and 1109 EST. Winds were shifting in a clock-
wise direction during this time, as three separate traverse paths were used
to complete the cross section. For this reason, it is difficult to make
any comparison between the hourly average Hockey winds and the plume center
of mass. Cross section #20, performed in the afternoon of the same day,
showed substantial deviation between actual and predicted plume location
(325° vs. 277°). Pibal winds at plume height from a release at 1355 EST
were blowing toward 270° in approximate agreement with the Hockey winds.
The ridge immediately to the west of the plant probably diverted the
plume to the northwest where lower terrain elevations are located. Later
-------
TABLE 30. LOCATION OF CENTER OF MASS OF PLUME
Cross section
number
1
{
•j
n
5
(i
1
a
9
1U
n
12
13
15
17
18
, ?()
n
23
24
25
26
28
29
Distance of
cross section
from plant
(kin)
<.(!
3.0
3.0
1.8
1 . 8
3.1
6.2
18.2
1.8
1.8
8.2
1.8
2.3
1.6
6.7
1.7
3.6
2.5
7.6
7.6
2.5
1.8
5.6
?.o
Direction towards
which the Hockey 30 m
transport wind (deqrees)
would transport the
plume
148
148
148
100
101)
3i
42
52
V)
66
74
9!
184
165
176
296-315
277*
253*
253*
242*
19fi*
26
6*
46
Bearing' of x-section
center of mass
from power plant
7
14
19
49
5!>
34
v>
M
'i4
',/
54
54
??}
1/5
179
353
325
267
272
270
257
13
28
19
•
Altitude
of center
of mass (m) MSI
974
•",01
795
773
93 I
823
902
790
895
855
I055
,",46
--
995
KMtt
I212
1448
487
H46
H93
1044
1266
1096
1 188
Based on Hockey 10 m winds.
1,1,
-------
in che afternoon a transport vector based on pibal winds at plume height,
(approximately 700 m above plant base) agreed with the direction of plume
travel in cross section #24 (approximately 270°), while the Hockey winds
indicated transport toward 240°. The pibal at 1758 EST indicated trans-
port toward 275° at plume height, agreeing much closer to the location of
the center of mass (257°) during cross section #25 than the Hockey wind.
On July 28, the location of the center of mass in cross section #29
deviated by 27° from the Hockey 30 m transport vector. Pibal data from
1300 EST indicated transport of the plume toward 26° to 38°, directions
between the actual plume location and the Hockey transport direction.
The plume may have been almost immediately diverted by ridges close
to the plant in several cases. Following the initial diversion, the plume
traveled along Sinkhole Valley during cross sections #4, 5, 10, 11 and 12,
indicating some evidence of channelization of the flow in whicn tne plume
was being transported. The "lowest traverse (792 m) of cross section #7
was flown at tree-top level over Buffalo Mountain. A peak concentration
of 127 ppb was found dun'nc this traverse; therefore, substantial plume
impact was likely at ground level on the ridge top.
Cross sections #6 through #8 provide a series of plume traverses flown
at three downwind distances from the plant ranging from 3.1 km to 18.2 km.
Cross sections #6 and #7 were found at 34° and 35°, respectively, which
initially places the plume within the valley through which Dumps Creek
flows down to the Clinch River. Subsequently, the plume passed over Buffalo
Mountain. Cross section #8 shows the center of mass to be located at a
bearing of 54° at 18.2 km; the shift to the right may have been caused by
the plume being diverted around Big A Mountain.
Table 31 presents an analysis of the heights of the plume center of
mass as computed from the helicopter data. The plume heights above sea
levels above plant oase, above stack top, above the highest terrain along
the traverse, and above the terrain at the center of mass are shown. Cross
sections #7 and #8 had the lowest plume heights above the terrain at the plume
center of mass (170 m and 58 m, respectively). Therefore, even tnough
there was substantial impact of the plume on Buffalo Mountain in cross
section #7, the plume center!ine remained 170 m above the terrain. Between
the locations where cross sections #6 and #7 were performed, 3.1 and 6.2 km
dowr.wina of the plant, che terrain rose 213 m. Table 31 indicates that the
p1u:ne center! ine was 244 ;,i aoove the highest point along the traverse in
#6, wrrile in #7 the plume canterline was only 109 m above the highest
terrain. This indicates that the plume height above the terrain was reduced
by 63 percent of the terrain elevation difference as the plume passed over
Buf-Vio Mountain. This parameter, the change in plume height in relation
to the elevation change as a plume approaches a ridge, is important for use
in dispersion models. In another case in which the plume approached Buffalo
Mountain (Cross Sections #26 and #28) the plume height was reduced by 52
percent of the terrain elevation change., Cross Sections #15 and ^17 can be
used to analyze the approach of the plume toward Copper Ridge south of the
plant. A plume height reduction of 71 percent of the terrain change was
noted in tnis case.
67
-------
TABLE 31. MEASURED PLUME HEIGHTS
Cross
Section #
1
2
3
4
5
6
7
8
9
10
11
12
13
15
17
18
20
22
23
24
25
26
28
29
MSL
Altitude of
Center of Mass
(m)
974
804
795
773
931
823
902
790
895
855
1055
845
*
995
1018
1212
1448
987
1146
1193
1044
1266
1096
1188
Plume Height
Above
Plant Base
(m)
511
341
332
310
468
360
439
327
432
392
592
382
•M
532
555
749
985
524
683
730
581
803
633
725
Plume Height
Above
Stack Top
(m)
373
203
194
172
330
222
301
189
294
254
454
244
*
394
417
611
847
386
545
592
443
665
495
587
Plume Height
Above Highest
Terrain Along Traverse
(m)
242
133
124
175
333
244
109
58
297
257
457
247
Jt
324
237
553
716
310
469
S16
367
717
303
517
Plume Height
Above Terrain at
Center of Mast
(m)
352
25S
246
249
382
335
170
58
346
306
506
296
i-
385
250
724
728
377
585
656
495
784
425
639
* Insufficient data due to instrument malfunction
68
-------
Turner stability classes were determined from cloud information rec-
orded by field team members and Hockey 10 m wind speed data (Turner, 1964)
for each of the above three cases of plumes approaching ridges:
Cross Section # Turner Stability Class
6, 7 C
15, 17 B
26, 28 B
The Turner Class C for cross section #6 and #7 may not be appropriate,
since the plume dimensions (to be discussed in the subsection on "Plume
Dimensions," which follows) appeared to more closely resemble those of
stable plumes. The results of the analyses of the plume height change
most closely match the plume height adjustments made in the ERT-LAPPES
dispersion model (Slowik et a!., 1977) which reduces the plume height
above elevated terrain for all stability conditions.
PLUME DIMENSIONS
The method described by Whaley (1874) was used to compute the Gaussian
dispersion parameters a and a from the helicopter cross-section data.
First, the coordinates fy~, 1) of the center of mass of the plume are computed
for the crosswind and vertical directions in the following manner:
ZZ C.. Y..
" _ 31 13 U
Z Z c
0 i ij
z z c
- = 3 i ij
z z -
3 i ij
• th
where C - - is the SO- concentration at the i ' point along traverse J..Y-..
is the distance of the i point from a reference endpoint of the j 1J
traverse, and I. is the elevation above sea level of the j traverse.
After these coordinates have oeen determined, the followioq equations are
used to compute a and a.
az
i/
' zz
j i
r z z
j i
C,.j t
z z c
j i •
cu (i
z z c
3 i i
'ij - rf
ij
:j-D2
ij
69
-------
Horizontal dispersion parameters were computed for all cross sections while
CT values were computed for only those Cdi.es where good comparison was
obtained between the total S0? mass flux observed by the helicopter and
the S0? emission rate, indicating that the full vertical extent of the
plume was sampled. Table 32 presents the computed a and a values
along with a comparison with Pasquill-Turner values ^Turner* 1970) and a
values computed by the method of Pasquill (1976) as described by Irwiri (^979).
Pasquill's recommendations for predicting a use measured standard devia-
tion of wind direction fluctuations (a,) da^a (in radians) as follows:
ay = Vf(x)
where x is thendownwind1distance in meters and f(x) is defined-,as
[(1 + 0.0308 xU^b48)] for x _< 10 km and as 0.333 (10,000/x) IL for
x > 10 km.
Cloud data from field observers' notes, along with wind speeds from the
Hockey 10 m level were used to determine Turner stability classes appropriate
for each cross section. When ;>) tor very unstable
conditions. The curves arc all approx imal.c I y parallel, vJilh llic < m vr I or
measured values ind icat. in<| h iqher o values Hun Ihr others. I iqur e ',','
shows similar types of curves lor cr'oss sections //b through //;;. ihr
curve for the C stability cases shows o steeper slope than the sUnclar.i <.mves.
Figures 23 through 26 show similar curves for cases #lb, #!/, '.ases )*/'0, %'£'•>.
70
-------
o
t/3
§
1
3
UJ
s
»—j
Q-,
OJ
CO
3
H
£
J, C~
'3
cr
a.
^ ^
0) "
B 3
i 9 -i
D
o O b
fr-t ^*
O)
2 c
u 2 -P
"C E
1*0
s ,x ""
3 ^
S
to 13
id
^
'Eo S
tf
E
§•2
^ o
U
1
z
tn
OOOOO^oiOOrOlOOLnOOO^jOOOOOOOO
(M
in
oooooLooot^otoooLnorjooooooom
vo oj rg i£> 10
CJ csj rr> *-n
10
o1
l/)
, J [ j
CO O
•^f in
tn rg M -H
CNlroCJOrvJcomcMCMcoro
(MojcsjMcMtMrjro
71
-------
TABLE 33. TURNER STABILITY CLASS
CORRESPONDING TO MEASURED a VALUES
Cross Section Actual Turner Turner Stability Class
Stability Class Corresponding to Helicopter a
y
1
2
3
4
5
b
7
8
9
10
11
12
13
15
17
• 18
20
22
23
24
25
26
28
29
A
A
A
A
A
C
C
C
C
B-C
D
C
A
B
B
A-B
A
A
C
C
C
B
B
B
A-B
F
A
B
«.A
A-B
F means more stable than F
72
-------
io2-
ASME (very unstable)
Pasqulll - Gifford
Class A
10° X(Km)—* Id1
Figure 21. Jy vs. K for cross sections fr'l—''5.
(July 23, 1977)
10'
1 : , ,
! 1 i / S
\ '' ' ! / f
1 i ' 1 ' ' 1 | , / S /
1 ( i|l , / /
. ' • ' i i ! "//'
s /S
J
? / f
/ /
/ //
S ;'/
Ss'/
;' J /
/ ,' /
/ /Q
/ , ,
Clinch River
.4SME ^unstable)
Pasquill - Gifford
Class C
X(Km)i^-»
Figure 22. 3y vs. x for cross sections ''6-<'8.
(July 24, 1977)
73
-------
102 :
ASME (unstaole)
Pasquill - Gifford
Class B
Figure 23. a vs. x for cross sections #15, 'f!7.
(July 26. 1977)
10
3
1 ,
1 ' , . ' . 1
i ' ' ' !
i ' ' '
1 i : | i 1
• i • j ! ' . ;
: ; , f •>
/ f
Clinch River
ASME (very unstable
lass A
10
X(Km)
101
Fizure 24. r vs. x for cross sections ''20, '''22.
Quly 27, 1977)
74
-------
103
10-
/
Pasquill - Gifford
Class C
/L
Ilinch River
(unstable)
'10°
X(Kffl)-
Figura 25. "y vs . x for cross sections
(July 27, 1977)
?24, ••'25.
10J
x X
Clinch River
ASME (unstable)
Pasquill - Gifford
Class 3
1
X(Km)
10
Figure 26. Jy vs. x for cross sections •'/26, '/28, -'29
(July 28, 1977)
75
-------
cases rf24, #25, and cases #26, #28, and ?29. tach of inese groups o~
values form curves for classes A, B, and C with substantially different
slopes than the standard curves.
For the seven cases for which az was computed, the average value of
the ratio (measured az/Pasquill-Gifford oz) was 1.04, representing very
little enhancement of the vertical dispersion over that for flat terrain
for the mostly unstable conditions associated with these cases. However,
the ratio ranged from values of 0.56 to 1.74.
Table 34 presents values of S02 mass per unit meter in trie downwind
direction for each helicopter traverse in the 24 cross sections. These
values were obtained by integrating in the crosswind direction and multi-
plying by the height interval for which each traverse is representative.
Multiple peaks can be noted in many of these vertical profiles of SO?
mass. Well-defined double maxima occurred in cross sections #5, 9, 10,
17, 28 and 29; it is uncertain whether the double maxima result from the
separate plumes being emitted from the two stacks and having undergone
substantially different amounts of plume rise, or whether the phenomenon
results from reflection from the surface or from a stable layer. The
vertical distribution of mass for cross section #7 is unique in that it
is quite uniform; this is a case of the plume passing over Buffalo Mountain.
Cross section #6 shows a very compact well-defined plume with e top which
is much more distinct than the bottom. Relatively high values of mass were
noted on the highest elevation traverse in cross sections #10, #12, #17,
#20, and #25. It is uncertain whether this phenomenon indicates a distinct
plume top or the possibility that the helicopter did not go high enough to
sample the entire plume.
A comparison was made between the locations where the Helicopter found
the plume and the fixed monitoring stations that were affected by the plume
during this period. In addition, the ground mobile data were examined for
sampling periods corresponding to times when the Helicopter was in the air.
Table 35 presents the available fixed-station and mobile dat4° at lb.2 f'n, respectively. Trie
Tower station was located at 46" at 3.36 km and approxinately 200 n UP low ^"it-
lowest helicopter traverse of cross section #7, yet it received pea* :-;00
concentrations of 165-176 ppD. The mobile van was located at '-• kir. and at a
bearing of 25° during cross section #7, recording a peak of orly 28 ppr,. This
location was 100 m lower in elevation than Tower. Earlier, or, tr.e same cay
cress section #6 had the Tower site as one of its endpoints, Approximately 80
m below the lowest traverse. Tower received an hourly average: ot 31 ppn wifi
a peak of 85 ppb during this time period. During cross section ?;10 tne plume
center of mass was located at 57° at 1.8 km from the plant. The mobile var,
was located at the same bearing during this cross section but at 11.2' Km
distant; an hourly average of 37 ppo was recorded which was probably most"
background since the peak was only 4? ppb, indicating verv little, if -mv.
plume imoact at 579 m elevation above sea level.
76
i f
-------
TABLE 34. VERTICAL DISTRIBUTION OF SO;
COMPUTED FROM HELICOPTER TRAVERSE!
MASS
Mass of
Elevation SO?
(m) MSL (g/m)
Cross
945
975
1006
Cross
762
793
823
853
Cross
823
793
762
Cross
884
853
823
793
762
731
701
671
Cross
701
732
762
793
823
853
884
914
945
975
1006
1036
1066
1097
1127
Section
9
9
7
Section
12
11
17
6
Section
7
11
5
Section
6
5
5
5
5
11
8
3
Section
3
4
4
4
3
4
6
8
9
4
5
5
8
5
6
#1
.16
.47
.54
#2
.56
.92
.09
.85
#3
.29
.22
.28
#4
.31
.44
.54
.18
.30
.08
.75
.48
#5
.89
.49
.21
.11
.97
.58
.90
.97
.00
.96
.35
.51
.51
.69
.45
Mass of
Elevation S02
(m) MSL (.
Cross
762
793
823
853
883
914
Cross
793
823
853
884
914
945
975
1006
Cross
853
793
732
Cross
701
732
762
793
823
853
884
914
945
975
1036
1067
1097
1127
1158
1188
1219
Section
5
9
14
13
2
0
Section
6
5
2
5
4
6
7
5
Section
7
9
8
Section
9
7
8
14
1
1
4
1
3
2
7
3
4
11
6
7
7
g/m)
#6
.25
.06
.33
.25
.77
.05
#7
.61
.31
.80
.41
.75
.01
.37
.25
#8
.78
.08
.59
#9
.67
.33
.86
.13
.26
.47
.23
.42
.72
.35
.09
.39
.73
.41
.91
.42
.79
Elevation
(m) MSL
Cross Sect
1158
1097
1037
975
945
914
884
853
823
793
762
732
701
671
Cross Sect
701
732
762
793
823
853
884
914
975
1006
1036
1066
1097
1128
1158
1188
1219
1249
1280
1311
Mass of
so2
('
ion
13
9
8
7
6
9
7
6
14
6
8
5
13
8
ion
6
4
3
3
2
5
3
4
5
5
3
7
8
7
10
9
12
11
10
4
g/m)
#10
.08
.72
.30
.70
.70
.59
.19
.79
.18
.98
.35
.35
.95
.81
#11
.78
.28
.41
.24
.75
.51
.64
.74
.11
.74
.79
.40
.23
.13
.32
.58
.89
.28
.25
.75
continued,
77
-------
TABLE 34. (continued)
Elevat
Mass of
ion S02
(m) MSL (g/m)
Cross
914
853
793
732
671
Cross
701
732
762
793
823
853
884
914
945
975
1006
1036
1067
Cross
762
793
823
853
884
914
945
975
1006
1036
1067
1097
1127
1158
1188
Section #12
38.27
32.82
21.44
12.53
2.12
Section #13
2.23
1.76
4.10
4.24
8.58
9.99
9.92
9.02
17.69
11.97
23.35
10.70
12.06
Section #15
1.70
1.16
1.77
1.95
11.22
12.70
6.37
16.24
7.73
12.95
11.64
15.84
11.00
13.51
5.68
Elevat
Mass of
ion s°2
(m) MSL (g/m)
Cross
1097
1036
975
914
853
793
Cross
762
732
793
823
853
884
914
945
975
1006
1036
1067
1097
1128
1158
1188
1219
1250
1280
1311
1341
1372
1402
1432
1463
1493
1524
1555
1585
Section #17
21.67
14.06
2.05
21.34
5.67
1.94
Section #18
3.06
4.18
11.89
13.13
12.10
7.02
11.17
11.01
7.55
17.63
10.56
5.09
4.67
0.15
17.08
18.96
44.48
26.63
20.54
24.16
22.45
27.12
20.69
23.03
26.30
8.62
10.64
2.86
2.37
Mass of
Elevation S02
(m) MSL
Cross
1524
1463
1402
1341
1280
1219
Cross
1219
1158
1097
1036
975
914
853
793
732
823
Cross
1036
1097
1158
1219
Cross
793
853
914
975
1006
1036
1067
1097
1127
1158
1189
1219
1250
1280
1310
1341
(g/m)
Section #20
26.76
28.94
14.19
4.12
3.92
2.28
Section #22
8.63
3.66
27.87
16.15
12.19
11.55
19.08
2.71
2.34
1.47
Section #23
0.07
3.14
13.38
0.04
Section #24
0.10
0.00
0.00
0.46
1.23
4.94
1.78
6.44
8.28
4.84
8.18
1.24
2.30
3.95
12.98
4.00
(continued)
78
-------
TABLE 34. (continued)
Elevation
(m) MSL
Cross Sect
1250
1189
1127
1067
1006
945
884
823
Cross Sect
732
762
793
823
853
884
914
945
975
1006
1036
1067
Mass of
S02
ion #25
29.64
4.60
10.66
28.61
45.27
28.67
14.67
7.89
ion #26
2.00
5.50
3.86
0.00
13.34
15.02
24.69
18.05
17.24
23.98
29.15
17.66
Elevation
(m) MSL
Cross Sect
Mass of
S02
(g/m)
ion #26
(continued)
1097 7.87
1128 22.03
1158 3.05
1189 12.72
1219 6.25
1250 --12.10
1280 7.34
1311 6.34
1341 6.40
1372 5.55
1402 14.12
1432 14.28
1463 1.93
Cross Section #28
823
884
945
1006
1067
1127
12.53
6.87
6.90
9.37
9.14
5.12
Elevation
(m) MSL
Mass of
S02
(g/m)
Cross Section #28
(continued)
1188 7.28
1250 8.35
1311 0.93
1372 3.07
1433 0.28
Cross Section #29
1433
1372
1311
1250
1189
1128
1067
1006
975
945
914
4.06
10.03
3.77
0.78
1.34
5.99
0.35
8.96
1.85
0.46
1.01
79
-------
TABLE 35. SIGNIFICANT FIXED-STATION AND GROUND MOBILE
MEASUREMENTS DURING HELICOPTER CROSS SECTIONS
Cross
Section #
1
2
3
4
5
6
7
8
9
10
11
12
13
15
17
18
20
22
23
24
25
26
28
29
Dist.
From
Hourly Avg SO9/Hourly Peak SO2 „,
Tower Johnson (km)
_ _
9/54
9/54
15/34
15/34
31/85
38/175 - 4.0
73/165
15/34
11.2
68/286 ^
-
45/82 S.2
-
-
-
-
-
-
-
-
2.5
-
_
Mobile 1
Bearing
From MSL
Plant Avg/Peak SO2 Begin End Elevation
(deg) (ppb) Time Time (m)
25 15/28 0919 1011 488
57 37/42 1249 1359 579
224 25/71 0955 1102 555
340 46/77 0953 1047 482
80
-------
On July 26, 1977, Johnson recorded significant S0? concentrations during
cross section #13 despite being 30° offset from the center of mass as indicated
by the helicopter at 2.3 km from the plant. At 5.2 km distant and only 3° off-
set from the center of mass bearing, the mobile van recorded an hourly average
of 25 ppb and a peak of 71 ppb. In this case the van was located approximately
150 m below the lowest traverse, indicating that the plume extended to lower
elevations than those flown by the helicopter.
CONCLUSIONS
The cross sections of the plume made by helicopter are useful for deter-
mining both plume position and plume dimensions and structure. All of the
cross sections are representative of daytime neutral to unstable conditions.
In six cross sections in which the entire plume was sampled the mass flux
of S02 compared well with the plant S0£ emission rate. The amount of fluc-
tuation of the plume from traverse to traverse as observed by the helicopter
during the cross sections agreed well with the variations in wind direction
and computed plume configurations using the Hockey 30 m 2-minute data. Con-
siderable deviation of the actual plume travel direction as detected by the
helicopter from the transport direction indicated by the Hockey 30 m wind
and the pibal winds was noted in about half of the cases. Many of these
deviations could be explained by diversion of the plume by particular terrain
features such as ridges near the plant. The vertical distribution of S02
mass in the plume was computed and such features as multiple concentration
maxima in the vertical and distinct plume tops were noted.
The helicopter cross section results have implications regarding
dispersion modeling techniques. The analyses of cross sections observed
as the plume approached a ridge show reductions of plume height of from
52 to 71 percent of the terrain elevation change. The horizontal Gaussian
plume parameter tfy, as observed by the helicopter is well represented by
Pasquill's suggestion of:
°y = CTA * f(x).
However, the standard flat terrain Pasquill-Turner dispersion curves under-
estimate the observations.
81
-------
SECTION 7
DEVELOPMENT OF PROFILES OF WIND AND TEMPERATURE,
PLUME HEIGHTS, AND PLUME TRAJECTORIES
Two important characteristics of plume behavior are its height and
horizontal position. Since these were not directly measured, except during
helicopter observations, procedures were developed for using the available
data to make estimates of the plume height and position. An important step
in making these estimates was found to be the construction of vertical pro-
files of wind speed and temperature for each hour.
DEVELOPMENT OF PROFILES OF WIND AND TEMPERATURE
As part of the analysis of the Clinch River data, plume rise calcula-
tions were needed on an hourly basis for the duration of the recorded data
to determine the position of the plume in the vertical. Briggs' (1975)
equations, as applied by Holzworth (1978), were used to determine the final
plume height (above the stack) at which the plume buoyancy flux equals zero.
In order to use the Briggs/Holzworth method, a temperature and wind profile
must be available for every hour of Clinch River data. Since pibal/T-sonde
soundings were not performed hourly, a method was developed to estimate
wind and temperature profiles.
Several alternatives were available for the development of hourly sound-
ings; four types of data were available for use: (1) regularly scheduled
12-hourly rawinsonde data from the four National Weather Service stations
closest to the Clinch River Plant (Greensboro, Nashville, Huntington, and
Washington, D.C.); (2) pibal/T-sonde data observed at the Clinch River plant;
(3) fixed station data from the Clinch River monitoring network; and (4) data
from the 30 m mast at the Tower site. Since rawinsonde observations are
scheduled twice daily at 0000 GMT and 1200 GMT (1900 and 0700 EST), use of
this data requires time interpolation to obtain hourly soundings. In fact*
the balloons are released 45 minutes before the scheduled observation. Also,
a decision had to be made concerning which rawinsonde station or stations are
the best approximation of the actual conditions at Clinch River (i.e.: the
radiosonde data require testing against the pibal/T-sonde data). The pibal/
T-sonde data are the best set of data available for representing actual con-
ditions at Clinch River. However, these data are limited by the relatively
small number of observations; their best use is as a test against other
methods of estimating the profiles. Use of the meteorological data from
82
-------
the fixed stations to form profiles assumes that the stations can be used to
represent actual levels in the atmosphere. The advantage of this method is
that the data are available continuously as hourly averages. The disadavan-
tages are the proximity of the instruments to the ground at the fixed stations
(both temperatures and winds are affected by the ground surface despite the
instruments being at the 10 m level) and the fact that the highest station
is only approximately 350 m above plant base. The accuracy of this method
was tested by using the pibal/T-sonde data. The Tower site had temperature
sensors at 0.5, 4 and 30 m and vector vanes at 10 m and 30 m. The parameters
obtained from these instruments give an estimate of atmospheric stability
and also can be used in conjunction with other fixed station meteorological
data to obtain a profile.
A combination of fixed-station meteorological data for the lower levels
(0 to 350 m above plant base) and time-interpolated radiosonde data above
350 m has the advantages of having hourly data available for the low levels
adhere most of the diurnal variation occurs and also of having data to as high
•an altitude as desired.
All pibctl/T-sonde data were used to test this method by interpolating
the pibal/T-sonde values to fixed station elevations (above plant base) and
comparing these values with the fixed station data. The average difference
and standard deviation of the difference were computed for wind speed and
temperature at ea^h station (see Table 36). The differences were computed
as station value minus pibal value. The higher elevation stations (Johnson,
Kents, Hockey and Hockey 30 m) exhibit a larger standard deviation for wind
speed difference than the lower stations (Nash's, Munsey, Lamberts, Tower,
and Tower 30 m). Also, the higher elevation stations with temperature sensors
(Kent's, Hockey) exhibit a larger standard deviation for temperature difference
than the lower stations (Nash's, Munsey, Tower).
TABLE 36. AVERAGE DIFFERENCE AND STANDARD DEVIATION OF DIFFERENCES
OF TEMPERATURE AND WIND SPEED BETWEEN FIXED STATIONS AND PIBAL/T-SONDES
Wind speed (m/sec)
Temperature ("C)
Station
Nash's
Munsey
Lambert
Tower
Tower 30 m
Johnson
Kent's
Hcckey
Hockey 30 m
Height above
plant base (m)
82
140
143
143
168
296
302
341
366
Average
error
-1.4
-1.9
-1.7
-0.5
0.0
-2.0
-1.0
-1.7
-0.7
Standard
deviation
1.5
2.2
2.6
1.5
1.8
5.3
7.7
4.8
3.7
Average
error
-2.3
C.4
-1.0
-
-
0.5
0.2
Standard
deviation
1.9
1.8
1.6
-
-
2.8
3-i
.0
83
-------
Data from the two closest rawinsonde stations (Huntington and Greensboro)
were linearly interpolated in the vertical to the pibal observation levels
and then linearly interpolated in time between the 0000 GMT and 1200 GMT
soundings to the pibal release times. The average differences and standard
deviation of the differences were determined for temperature, wind speed, and
wind direction for 50 m layers from plant base to 5000 m above plant base.
Table 37 summarizes the results for the first 500 m and supplies the average
difference for temperature and wind speed over the first 2000 m.
TABLE 37. AVERAGE DIFFERENCE AND STANDARD DEVIATION OF DIFFERENCES OF
TEMPERATURE AND WIND SPEED FOR HUNTINGTON AND
GREENSBORO RAWINSONDES VS. PIBAL/T-SONDE
Huntington
Greensboro
Elevation
above plant
(meters)
0-50
51-100
101-150
151-200
201-250
251-300
301-350
351-400
401-450
451-500
(Ave.)0-2000
^Average
Difference
WS
„
1.0
1.0
1.4
0.3
0.6
1.1
0.5
1.4
1.7
0.95
T
-4.4
-1.6
-2.5
-1.4
-1.2
-2.0
-0.4
-0.5
-1.3
-0.5
-0.81
Standard
Deviation
WS
—
2.3
2.6
2.0
8.5
12.7
6.6
4.8
3.5
3.3
T
3.5
3.7
2.7
3.2
3.9
3.1
3.4
2.8
6.0
2.7
*Average
Difference
WS
_
1.3
1.2
1.7
0.9
0.4
1.1
1.6
1.7
1.8
0.18
T
-3.0
-1.1
-0.5
0.0
-0.1
-0.9
1.1
0.9
-0.5
0.8
0.35
Standard
Dev i at i on
WS
.
2.1
2.1
2.7
8.4
1.7
6.7
5.0
3.6
3.0
T
4.7
3.6
3.4
4.3
3.2
3.2
3.6
3.4
6.1
3.2
Average Difference = Radiosonde value - pibal value.
WS = Wind Speed, m/sec.
T = Temperature, °C
Comparing Table 36 with Table 37 it is apparent that the average differ-
ences are about the same for the fixed station vs. pibal test as compared to
the radiosonde vs. pibal test. In the lower 2000 m the average wind speed
difference was positive for both stations but approximately five times larger
at Huntington; the average temperature difference was approximately twice as
large in magnitude at Huntington than at Greensboro and of opposite sign.
Within the interval from the surface to 2000 m above plant base both
radiosondes had stronger winds than were recorded (0.95 m/sec greater for
Huntington and 0.18 m/sec greater for Greensboro) by the pibal. Perhaps the
84
-------
boundary layer winds at Clinch River are decreased due to the complex terrain.
Both radiosondes had an approximately equal number of layers in which their
error was smaller. Huntington was on the average colder than Clinch River by
0. 81°C. However, the standard deviation of the individual errors was approxi-
mately equal for both radiosonde stations for both wind speed and temperature.
The radiosonde comparisons with the pibal were not as good as the fixed-station
pibal comparisons in the layer from stack top up to 350 m above plant base.
The time interpolated (linear) Huntington and Greensboro radiosonde
data were plotted and compared to the pibal/T-sonde data for selected times
exhibiting varying synoptic conditions. After examining the plots, it
appeared that the radiosonde temperatures were accurate above 300 m, but the
linear time interpolation did not suit the lower 300 m; the morning ground-
based inversions on the interpolated radiosonde data burned off too slowly
and their effects were shown throughout the day (i.e., during ocurrences of
afternoon superadiabatic lapse rates near the surface recorded by the pibal/
T-sondes, stable layers were interpolated using the radiosonde data). For
most layers, especially at Huntington, the interpolated radiosonde wind
speeds were greater than the pibal/T-sonde wind speeds. On occasion, the
interpolated Huntington and Greensboro radiosonde wind speeds exhibited large
differences with the pibal wind speeds, especially after frontal passages or
during periods of cyclonic flow.
After reviewing the test results of the fixed station vs. pibal/T-sonde
and interpolated radiosonde vs. pibal/T-sonde, a methodology was developed
to incorporate both radiosonde and fixed-station data for determining hourly
profiles. The methodology consisted of using the hourly average wind speed
and temperature values of the fixed stations from stack top to approximately
200 m above stack height. Above this layer either the Huntington or^ Greens-
boro radiosonde data £r a weighted average of the two were used, depending
on which of the three gives a value at the Hockey elevation which is closest
to the Hockey observations. The weighted average for temperature used the
average difference for the lower 2000 m calculated in the interpolated
radiosonde vs. pibal analysis (all temperatures in Celsius):
T - TH * '' TG
1 + 81/35
where T = weighted average temperature
T,, = temperature from Huntington sounding
Tg = temperature from Greensboro sounding
81/35 = ratio of average differences for the lower 2000 m,
85
-------
The weighted average for wind speed used an inverse square distance weighting
relationship minus a term which was the inverse square distance weighting of
the average difference for the lower 2000 m, calculated in the time-interpo-
lated radiosonde vs. pibal analysis (all speeds in m/sec):
(1/DH2) WSH + 0/DG2) WSQ (l/DH2) 0.95 + 0/Dg2) 0.18
ws = - - -- - - - -- - -
1/DH2 +
2 WS,
_ (VDH2) WSH +
1/D2
where WS = weighted average wind speed
WS^ = wind speed from Huntington sounding
WSg = wind speed from Greensboro sounding
DH = Huntington to Clinch River distance
Dg = Greensboro to Clinch River distance
0.95 and 0.18 = average differences between radiosonde and pibal wind
speed for the lower 2000 m for Huntington and Greensboro,
respectively.
Specifically, the average of the hourly temperature at Tower and Munsey
and the hourly wind speed at Upper Tower were used for values at stack top;
the hourly wind speed at Johnson and the hourly temperature at Kents were
used for a level 150 m above the stack; and the hourly wind speed at Upper
Hockey and the hourly temperature at Hockey were used for the level 200 m
above the stack (see Table 1 for station elevations). A determination was
made as to which time-interpretated radiosonde or weighted average radiosonde
data were closest to the fixed station temperature and wind speed values at
200 m above stack top. The entire temperature or wind speed sounding (above
stack top plus 200 m) of the best fitting radiosonde data was then shifted to
conform to the highest fixed station values, thus providing a smooth transi-
tion between the fixed station and radiosonde data.
86
-------
Figure 27 illustrates the profile methodology and a summary of the metho-
dology follows.
Vertical Profile Methodology
1. Perform linear time interpolation of Huntington and Greensboro
radiosonde data on an hourly basis.
2. Determine weighted average of the Huntington and Greensboro data.
3. Determine which of above-mentioned three soundings is closest to
the Hockey value for wind speed and temperature, separately.
4. Shift wind speed or temperature values for the entire selected
radiosonde sounding so that sounding value at the Hockey level equals
the measured value at Hockey.
5. Merge fixed station and radiosonde data:
Elevation above
stack (m) Temperature Wind speed
0 Average of Tower and Tower 30 m wind speed
Munsey temperature
150 Kent's temperature Johnson wind speed
200 Hockey temperature Hockey 30 m wind speed
Above 200 Adjusted radiosonde Adjusted radiosonde
temperature wind speed
When wind speed readings are missing below 350 m a power law is applied to
obtain these values. The power law exponents are based on values obtained from
wind observations at O'Neill, Nebraska, Kerang, Australia and several meteoro-
logical towers at nuclear power plant sites throughout the United States. An
exponent has been derived for each of several classes of vertical temperature
gradient corresponding to the six Pasquill Stability Classes (U.S. AEC, 1972).
The vertical temperature gradient for the layer up to 350 m is used to determine
a power law exponent for use in determining a wind speed value for the missing
level. Unfortunately, no reliable power law exponents for complex terrain
exist.
Um ' U350
U is the wind speed for the missing level, U^™ is the wind speed at
350 m, Z is the height of the missing level ana p is the exponent chosen
from Tabfe 38.
87
-------
500
400
" 300
a.
200
100
O Radiosonde
• Shifted Radiosonde
X Fixed Station
\ Finalized
•^Profile
12 13
Temperature (°C)
15
Figure 27. Example of temperature profile construction methodology.
88
-------
TABLE 38. POWER LAW EXPONENTS
AT/AZ(°C/100 m)
< -1.9
-1.9 to -1.7
-1.7 to -1.5
-1.5 to -0.5
-0.5 to + 1.5
0.10
0.10
0.10
(0.12
)0.21
0.35
day
night
> +1.5 0.49
Profiles of wind speed and temperature were produced on an hourly basis
for the lower atmosphere at the Clinch River site using the above methodology
for the period October 12, 1976 through September 30, 1977. Data collected
prior to October 12, 1976 is less complete and of less interest because of
uncertainties and incompleteness in the ground level concentration measure-
ments.
PLUME HEIGHT ESTIMATES
The Environmental Protection Agency has been using the formulations of
Briggs (1969, 1975) for estimating plume rise. The 1975 work presented
formulas for plume rise through an atmosphere with vertically varying
profiles of wind speed and temperature. Holzworth (1978) applied Briggs1
method in estimating effective stack heights using standard rawinsonde
observations. This same technique was applied to the hourly profiles of
wind speed and temperature which had been generated for the Clinch River
site.
Briggs1 method of calculating plume rise through variable temperature
and wind speed profiles consists of following the buoyancy flux of a plume
segment through successive layers where it is either depleted or enhanced
until the level is reached where the buoyancy flux is zero. The initial
buoyancy at the top of the stack (F ) is:
FQ = 3.7 x 10"10QH
where Qn is the heat emission rate in cal/sec, computed from
QH = 83.45QVP0 S T °
89
-------
where Q is the volume emission rate from the stack in m /sec, P is the
atmospheric pressure at the top of the stack in mb, T is the air temperature
at stack top, and T is the stack gas exit temperature, both in °K. Hourly
values of Qv and T swere available for each unit of the Clinch River Plant
from the plant emissions file. Atmospheric pressure was also recorded at the
plant, while the stack top air temperature was taken from the level of the
temperature profile most closely corresponding to the level of the stack
top.
Each sounding is divided into successive layers in which the change of
temperature with height [AT/AZ(°C m )] and wind speed [AU/AZ(s~ }] both are
constant and linear. The height above the stack top of the bottom of each
such layer is specified as Z , and the top as Z . Thus, the stack top is
the bottom of the lowest layer, where Z , = 0 and FQ becomes F ,. The
buoyancy flux at the top (F ) of each successively higher layer i£ calculated
until it becomes negative. For the layer immediately above the last level
where the buoyancy flux was positive, F is set equal to zero and the equation
is solved for the plume rise (Z ) above the physical stack height. The effec-
tive stack height equals the physical stack height plus the plume rise.
There are two sets of equations for calculating F and Z The first
set is for no-wind conditions, e.g., nearly vertically rising plumes, where
T (°C) is air temperature and Z (m) is height above the top of the stack.
0.265F
1/3
Fn = Fn-l
0
273 + 0.5
Zn - Zn-l
V
3.77
n-1.
: 1/3
(273
n-1
) x
Zn - Zn-l
VTn-l+0-01
3/8
90
-------
The second set of equations is for with-wind conditions, i.e., nearly
horizontally rising plumes,
Fn ' Fn-l
0.523
Tn - Vl
273+0.5(7^7,,.,)
0.01
X U.
-el
Vi
Vl
2n - Vl
Tn - Vl + °'0'(Z
1/3
1.91F.
[273 + 0.5(Tn_1
Zn " Vl
1/3
where U is wind speed (m s ). Z , is a preliminary estimate of the
plume rise above the stack top. This step is necessary because the equation
for Z can be sensitive to the wind speed (U O at the plume-rise height
.. p . _ i " •
(i.e.? at Zel).
91
-------
Specification of sufficiently slow winds for use of the no-wind equations
is somewhat arbitrary in some cases and is based on the personal recommendation
of Briggs. For layers with U = 0.0, there is no ambiguity and the no-wind
equations are used. For all other layers we interpolate to the level Z , at
which the critical wind speed (U, ) occurs, i.e.,
If U. occurs in the particular layer, the no-wind equations are used in the
sublayer with U £ U, and the with-wind equations are used in the sublayer
with U ^> U, . If, in the particular layer, all U > U. , the with-wind equa-
tions are used for the entire layer; if all U
-------
TABLE 39. COMPARISON OF PLUME HEIGHT MEASUREMENTS AND ESTIMATES
Plume Center of Mass
(Helicopter Measurement)
(m)MSL
974
804
795
773
931
823
902
790
895
855
1,055
845
995
1,018
1,212
1,448
987
1,146
1,193
1,044
1,266
1,096
1,188
Holzworth- Briggs
Plume Height
(in) MSL
792
800
800
999
999
769
809
750
910
929
927
921
908
925
761
958
816
816
934
934
. *
*
*
Briggs BEH072
Plume Height
(m) MSL
1,838
1,494
1,494
1,255
1,255
828
816
784
781
832
866
1,065
1,185
1,100
1,134
2,096
1,566
1,494
1,589
1,522
1,564
1,321
1,537
Wind speed and temperature profiles not available.
93
-------
emission for the two stacks. The mean absolute error for the Holzworth/Briggs
method was 143 m, while for Briggs BEH072 subroutine the mean absolute error
was 321 m. Subsequently, it was determined that if a plume rise enhancement
factor were included, the mean absolute error for the Holzworth/Briggs method
would be reduced by 15 m; however, the absolute error for the BEH072 method
would be increased.
The plume heights based on separate estimates for each stack were com-
puted for the entire set of wind speed and temperature profiles (October 12,
1976 through September 30, 1977). The statistics in Table 40 were generated
in comparing the plume heights from the two stacks with the vertical tempera-
ture structure. The percent of the time that the plume was contained in an
inversion layer reached a maximum at 0500 EST and a minimum by 1300-1400 EST,;
this reflects the diurnal cycle of stability in the boundary layer. Also, at
0500 EST the fraction of the time when the plume was below an elevated inver-
sion was at a minimum. By 1800-1900 EST a very large percent of the cases (77
percent) showed plumes below elevated inversions. No inversion existed in the
profile least often (1 percent) between 0600 and 0800 EST and most often at
1300 EST (11 percent).
DEVELOPMENT OF PLUME TRAJECTORIES
The soundings and plume height estimates aid in determining the position
of the plume in the vertical. A method is also needed for estimating the
plume position in the horizontal on a routine basis; therefore, the develop-
ment of a method of predicting the plume trajectory was needed.
The trajectory development was initiated by reviewing the July 1977
helicopter data. Methods of estimating a plume trajectory were tested
against the position of the plume as determined by the helicopter cross
sections. The Hockey 30 m wind data were found to be the most reliable in
estimating the plume position. Other methods such as spatial interpolation
of fixed-station wind data were rejected due to an inadequate number of
stations located at or near typical plume heights.
The following plume trajectory criteria were established for the vici-
nity of the Clinch River monitoring network whicn extends to a maximum of
30 km. from the plant:
For daylight hours
1. Use hourly average Hockey 30 m wind vector when wine speeds
are greater than 3.0 m/sec and the standard deviation of the
wimi direction fluctuations ( a.) is less than 25°.
2. Use the 2-minute raw data from the Hockey 30 m level to con-
struct a curvilinear trajectory over the period of an hour
for wind speeds between 1.0 m/sec and 3.0 m/sec and OA
between 25° and 40°. M
94
-------
TABLE 40. SUMMARY OF PLUME HEIGHTS IN RELATION TO VERTICAL TEMPERATURE STRUCTURE,
OCTOBER 12, 1976 - SEPTEMBER 30, 1977
1
Hour
1
2
3
4
5
6
7
3
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
% In Inversion
Suck 1
57
58
61
61
65
61
61
58
54
43
34
29
21
22
22
23
21
IS
19
29
38
47
54
57
Stack 2
55
56
57
59
60
57
59
57
53
41
32
30
22
20
25
25
20
15
19
29
39
46
53
54
% Below
Inversion
Stack 1
32
32
32
33
30
32
32
35
35
46
53
55
58
62
64
61
65
72
71
57 •
48
41
34
33
Stack 2
39
38
39
39
38
39
39
40
43
51
58
58
61
66
64
63
69
74
75
62
53
45
40
39
% Above
Inversion
Stack 1
7
6
5
4
3
5
6
5
8
8
7
9
10
6
6
8
7
6
5
10
11
8
8
6
Stack 2
2
2
2
0
1
2
0
1
1
4
4
4
5
5
4
4
3
4
2
S
5
4
4
3
% No
Inversion
4
4
3
2
2
1
1
1
3
3
6
7
11
9
8
8
7
7
5
4
4
4
3
4
95
-------
3. If the wind speed is less than 1.0 m/sec or the a., ii
greater than 40° the trajectory can be categorized as
either (i) indefinite or (ii) a stagnation situation.
For night
1. For cases when plume rise was low (plume did not penetrate
inversion), assume stagnation conditions.
2. For cases of high plume rise (i.e., plume above inversion),
use same criteria as daytime.
Indefinite plume trajectories occur when the plume is transported in myriad
directions due to highly variable winds. Stagnation cases occur when the
plume is trapped beneath a stable layer and there is insufficient wind to
transport the plume. In this case, the plume just disperses in all directions
in the layer below the inversion.
Eighteen percent of the hours in a test sample of 46 days and nights fit
the criteria for curvilinear trajectories. An additional 38 percent of the
hours in the 46 day test sample were determined to be indefinite or stagnation
situations. The remaining 44 percent of the hours met the criterion for
uniform linear flow. The development of a technique for displaying the plume
position or configuration at any point in time was considered necessary in
analyzing the reasons for particular stations showing impact of the effects
of the terrain on the plume. To satisfy both the need for curvilinear
trajectories and the need for "snapshot" representations of the plume, a
computer program was developed to produce a graphical display of the plume
configuration using the 2-minute wind observations from the Hockey 30 m
level. The program advects each 2-minute section of the plume arleady in the
atmosphere using the latest 2-minute wind vector. The latest vector also
represents the latest section of the plume to leave the stack. The program can
print a "snapshot" of the plume at any given 2-minute increment of time, using
the latest set of 30 2-minute observations covering an hour of time. During
the helicopter monitoring period in July 1977 the plume configuration was
produced for every 2 minutes during severa"1 of the cross section observations.
For each of the case studies (described in Section 8) one plume configuration
was computed and displayed for each hour on the hour for a 48-hour period
containing a case of high pollutant concentration.
Analysis of these printouts can assist in determining if the plume fol-
lowed the Hockey winds or was affected by the terrain features. Figure 28
is an example of the plume configuration displays produced for the period of
the helicopter cross sections. The figure shows the plume position at 1428 EST
on July 23, 1977 during the helicopter cross section #4. The vector from the
plant to the second point D at the end of the plume represents the section of
plume emitted an hour earlier and transported by successive 2-minute vectors to
its position at 1428 EST. The graphical displays of the plume configuration were
produced on the line printer at the same scale as the 7.5 minute U.S. Geological
96
-------
l/l
,
>
ifl J3
a
a. -o
QJ
cu +->
E =5
a a.
ol o
o
CO
CM
Cl)
3
CD
97
-------
Survey maps (1 in. = 24,000 in.). An overlay was produced from the USGS
maps that cover the region surrounding the Clinch River plant for use in
analyzing the plume position in relation to the terrain. Tic marks on the
axes of the graphical displays represent 0.5 km increments of distance from
the plant.
Description,of the usefulness of this graphical technique in explaining
high concentration cases is contained in the Case Studies part of Section 8.
SUMMARY
A method was developed for constructing hourly profiles of temperature
and wind speed using hourly ground station observations at five sites repre-
senting three different elevations, and twice daily rawinsonde observations.
The constructed profile compared favorably with 350 local pibal/T-sonde
observations. The root-mean-square-difference in temperature was less than
2°C up to 200 m above the plant base, about 3°C for the layer from 200 to
400 m above the plant base, and generally about 4°C (although as high as
6°C) above 400 m. The root-mean-square-difference in wind speeds was about
2 m/sec up to 200 m above the plant base and averaged about 5 m/sec above
200 m. The wind speed and temperature profiles were used to determine plume
height by a method developed by Holzworth and originally proposed by Briggs.
In comparing plume height calculations with helicopter measurements of 20- plume
cross-sections, we found a mean absolute difference of 130 m and a mean mea-
sured plume height of 978 m.
An objective method was developed for calculating plume trajectories based
on average 1-hour winds (linear 1-hour trajectory) or based on 2-minute winds
(curvilinear trajectory). Linear trajectories occurred 44 percent of the time
and curvilinear flow was found 18 percent of the time. The remaining 38 per-
cent of the cases were found to consist of irregular and poorly defined flow.
Most of the irregular flow situations occur during nighttime hours, i.e.,
about 60 percent of nighttime hours compared to about 16 percent of daytime
hours.
98
-------
SECTION 8
ANALYSIS OF PLUME IMPACT IN RELATION
TO METEOROLOGICAL CONDITIONS
Two approaches were utilized in analyzing the plume impact within the
fixed-station network. The first method was a statistical approach, which
involved producing joint frequency distributions of the pollutant concentra-
tions in association with a number of meteorological parameters. During this
stage of the.analysis statistics were also generated on the meteorological
data itself. The second approach was a case study analysis in which eight
specific high pollution incidents were studied in depth to determine the
causes for the plume impacts and to gain a better understanding of the plume
transport and diffusion processes that were occurring.
METEOROLOGICAL DATfl
Frequency distributions of wind direction, speed, and a. by hour of the
day for each of the eight fixed stations are presented in Appendix B.
The frequency distributions of wind directions recorded at each station
show a significant influence of the terrain in the immediate vicinity. Examples
can be cited to illustrate the relationship of prevailing wind directions to
local terrain features shown in Figure 31. The most frequent 10° wind direction
class at Johnson is 310°-320°. This class occurred most frequently from late
morning through late afternoon and is upslope flow through a hollow in Copper
Ridge. Lambert shows most frequent wind direction classes of 250°-260° and
260°-270°, indicating channeling of the airflow along Reed's Valley. Both
the upper and lower levels at Hockey have most frequent wind directions
that are upslope along a hollow on Copper Ridge. The upper level has a
peak in the distribution 220°-230° class, while the lower level exhibits a
maximum in the 200°-210° class. Munsey's most frequent wind direction is
120°-130° occurring mainly in the morning and evening. This flow represents
drainage through a hollow in Copper Ridge. Nash's shows the highest fre-
quencies in the adjacent 220°-230° classes which are oriented along nearby
Thompson Creek. The upper and lower Tower levels have highest frequencies
of wind directions in the ranges of 220°-260° and 240°-270°, respectively.
These ranges are slightly more frequent from mid-morning through late after-
noon than any other part of the day, suggesting some upslope flow toward
the Tower site from Dumps Creek. Castlewood's most frequent wind directions
are in the range 280°-310° which is up-river flow along the direction of the
Clinrh River Valley.
wind roses have been generated from the Clinch River wind data by
•axi"e1l H979). A particularly interestinq terrain influence on the air flow
99
-------
is demonstrated by the wind roses for Munsey. Figures 29 and 30 contain wind
roses for the months July-September 1977 by individual 3-hour periods of the
day. A very pronounced downslope flow (approximately 120°) develops at night
at Munsey as demonstrated by the wind roses for hours 0200-0400, 0500-0700, 1700-
1900, 2000-2200 and 2300-0100 EST. This phenonemon may be important in bringing
the elevated power plant plume down to the elevation of Munsey.
Figure 29. Munsey wind roses for July-September 1977, for
3-hourly periods during 0200-1300 EST.
100
-------
vtUNSEY 77 182 271 20-22
MUNSEY 77132 273 23-1
C 3-1 1-3 5-5 5-10 OVW '.0
Figure 30. Munsey wind roses for July-September 1977, for
3-hourly periods during 1400-0100 EST.
101
-------
The most frequent wind speed interval at most of the stations is either
0-1 m/sec or 1-2 m/sec. The exception is the upper level of Hockey which has
a maximum in the 3-4 m/sec interval. All stations except Wash's, Castlewood
and Munsey have 10°-20° as the most frequent standard deviation of wind direc-
tion (ci ) over periods of 1-hour in length. The most frequent afl classes
are 30-40° at Nash's, 50°-60° at Castlewood, and 60°-70° at Munsey. The
large a values at these sites reflect the disruption to the general flow
caused By the local terrain.
Frequency distributions were also analyzed for the additional meteoro-
logical data recorded at Tower, including vertical temperature differences
and wind elevation angle. The strongest inversion from 0.5 to 30 m was an
increase of greater than 3.3°C, while the strongest unstable condition was a
decrease of more than 2.4°C. Strong lapse conditions between 0.5 and 30 m
generally occur simultaneously with strong lapse conditions between 0.5 and 4
m. However only a few cases existed showing strong inversions occurring
simultaneously in both layers. It was difficult to draw conclusions regard-
ing the temperature gradient from 0.5 to 4 m because the 4 m level was apparently
reading too high after the sensor was replaced in October of 1976. The most
frequent standard deviation of wind elevation angle (a-) values were in the
range 6°-9° at the lower Tower level. These values occurred most frequently in
the morning and evening while the maximum range recorded (36°-39°) occurred at
midday. The upper
-------
Copper Ridqe. For Hockey winds of 180°-210° the preferred Munsey directions are
centered around 240° and 130°. These directions represent up-valley flow and
drainage flow, respectively. The Munsey direction matches the Hockey direction
best for the Hockey ranqe 260°-360°. The Munsey winds are scattered for the
Hockey ranqe 210°-250°.
The Nash's site is located on a plateau with a steep upslope about
0.3 km to the SE. The plateau and ridge are aligned approximately 55° to 235°.
There is a definite preference for Nash's wind directions to be aligned along
the axis of the plateau and ridge. For Hockey wind directions, ranging 0° -
180° the Nash's most frequent directions occur in the 20°-70° range. This
direction corresponds to the axis of the plateau. A maximum frequency also
occurs for Nash's winds at 0-20° with Hockey winds at 0-20°. For Hockey
directions of 180°-310° a maximum exists for Nash's winds between 190°-260° with
a secondary maximum between 0°-60°. This again corresponds to the axis of the
plateau. Nash's winds correspond to Hockey's for directions 310° - 360°. This
may be due to either the fact that generally northwest winds are stronger or
that the nearest obstruction to the flow from the northwest is more than 0.5
km away.
The Lambert site is located at the base of Copper Ridge (to the southeast
of the site) that is aligned 55° to 235°. There are no major obstructions to
the air flow within several kilometers to the N or NE. There is a preference
for the wind directions at Lambert to be aligned along the axis of the valley.
Lambert matches Hockey best for directions of 270°-36b° at Hockey. This may be
due to the stronger wind speeds encountered with this direction.
The Tower site is located on a plateau (aligned 55°-235°) with a higher
ridge approximately (same alignment) 0.4 km to the southeast. A steep downs lope
is located about 0.2 km to the northwest. There is a slight preference for Tower
30 m winds to be aligned along the plateau; however, Tower 30 m winds generally
correspond well with the Hockey 30 m winds. On the average Tower is backed
about 10° with respect to Hockey for most wind directions. Tower 10 m winds
exhibit a slight preference for directions to be aligned along the plateau
ax i s.
The Johnson site is located near the top of a wide ridge with extremely
complex terrain in all directions within a kilometer of the station. There
appears to be a preference for wind directions at Johnson to be aligned
along Sexton Hollow (which has an axis of 290°-110°), especially for Hockey
directions of 270°-360° and as a secondary maximum for Hockey directions of
120°-230°.
There is some evidence of wind channeling caused by the Clinch River
vaney at the Castlewood station.
103
-------
POLLUTANT CONCENTRATIONS IN RELATION TO OBSERVED METEOROLOGICAL CONDITIONS
Frequency distributions of the measured SO- and NO data versus
the simultaneous Hockey 30 m level wind directions werexgenerated. These
distributions are found in Appendix C. Mean S0~ concentrations were also
computed for each of 12 30° wind direction sectors for each station, using
the Hockey 30 m transport winds. Pollution roses were drawn for each station
(see Figure 31). All stations show significant impact of the plume for direc-
tions coming from the plant. In addition all stations also show elevated
concentrations with southwest winds whether the plant is located in this
direction or not, leading to the suggestion that higher background levels
are associated with this direction than with others.
From an examination of the frequency distribution tables for NO in
Appendix C, the Hockey 30 m wind direction ranges associated with thl highest
recorded NO readings at each of the stations could be easily obtained.
These ranges are listed below along with the direction of the plant from
each station:
Hockey 30 m
Wind Direction Range Associated Direction of Plant
Station with Highest NOx Readings from Station
Hockey 260°-320° 287°
Castlewood 200°-240° 58C
Nash's 210°-240°; 260°-300° 244°
Tower 200°-270° 226°
Munsey 190°-230°; 300°-340° 291°
Kent's 210°-250° 248°
Hockey, Nash's, Tower, and Kent's appeared to record high NO concentrations
with wind directions generally blowing from the direction of the plant. How-
ever, Munsey and Castlewood experienced their highest NO concentrations with
Hockey's 30 m wind directions much different than the direction straight from
the plant. These results indicate that terrain-induced flow regimes may be
important for transporting the power plant plume to these two stations. Drain-
age flows occur in the Clinch River Valley, affecting Castlewood, and downslope
flow at Munsey may bring the plume down to the Munsey elevation.
The relationships between high SO., concentrations at the eight stations
and wind speeds and o as recorded at the Hockey 30 m level were investigated.
The wind speed ranges and cr. ranges associated with the high SO- values are
listed in Table 41. All frequency distributions generated for the pollutant
data used the ranges listed below:
0-5 ppb
5-10
10-15 "
-------
921
\
o
a
CM
O
t/3
105
-------
TABLE 41. WIND SPEEDS AND a RANGES
ASSOCIATED WITH HIGHEST S02 CONCENTRATIONS
Station Wind speed_ (m/sec) °A (degrees)
Munsey 4-7 0-20
Kent's 4-6 0-10
Castlewood 3-5 0-20
Tower 1-8 0-20
Nash's 2-4 10-30
Lambert 2-4 0-20
Johnson 1-4 0-20
Hockey 1-3 0-70
106
-------
15-20
20-25
25-30
30-40
40-50
60-80
80-100
100-150
150-200
200-300
300-500
"High" Concentrations
For the purpose of analyzing the meteorological conditions associated with
high concentrations, the occurrences in the four highest concentration cate-
gories for each station were considered. Munsey, Kent's, and Castlewood
showed highest SO- readings with moderate wind speeds, while Nash's and
Lambert had their highest readings with slightly lighter winds (2-4 m/sec).
The ridgetop stations (Johnson and Hockey) had the lightest wind associated
with their high S02 concentrations. This result supports the finding that
high concentrations at the ridgetop stations are associated with nighttime
stable conditions as demonstrated in Section 4. The Tower station showed
no distinct propensity for any particular wind speed for high concentrations
as high values occurred with speeds from 1 to 8 m/sec. The high values with
light winds at Tower may be associated with fumigation incidents after
nearly calm stable nights, while the higher speeds were most likely asso-
ciated with the plume passing over the Tower at relatively low heights
occurring with neutral conditions.
With the exception of Hockey, the highest S02 concentrations occurred
with relatively low values of o.. At Hockey high values occurred with
a. as low as the 0°-10° range and as high as the 60°-70° range, indicating
the possibility that high values occur with both nearly direct hits by narrow
plumes and also in stagnation situations with rather ill-defined plumes. The
Kent's results show that the highest S02 readings occurred with very small
OQ values (< 10°), indicating that for Kent's to receive a significantly
high S02 concentration, a very narrow plume was required, which is reasonable
since Kent's was the most distant station.
RATIOS OF NO TO N02
An important consideration regarding NO concentrations in a power plant
plume is the rate of conversion of NO to NO-. Since the oxidation of NO- by
ozone is a very rapid reaction, but the available ozone supply in any volume of
air is low, the formation of N02 is highly dependent on the turbulent mixing
process which replenishes the supply of ozone in the plume from the ambient air.
If there is more turbulence over complex terrain than over flat terrain, it
may be evident in the rate of conversion of NO to NO-. We have not attempted
107
-------
to investigate this question in detail. However, we have computed the ratios of
mean NO to mean NCL concentrations which were measured at the various fixed
monitoring sites for each hour of the day. These ratios are shown in Figure 32
for the Tower and Kent's sites.
A pattern of diurnal variation in the NO to NO,, ratio is most clearly
seen at the Tower site, while hour-to-hour variations are least pronounced for
the Kent's site. These two sites are closest and furthest from the plant,
respectively. They clearly demonstrate the decrease in the ratio with dis-
tance, and thus travel time, from the plant. The mean of 24-hourly ratios is
1.19 at the Tower site and is 0.38 at the Kent's site. If the rate of conver-
sion from NO to N0? were proceeding at a constant first-order rate, the
ratios at the two sites would be represented by a single exponential decay
constant. However, a constant exponential decay for the ratio at the Tower
site is over four times the decay constant indicated by the ratio at the Kent's
site. It is clear that turbulent mixing rather than the rate of chemical reac-
tion is controlling the rate of conversion by the time the plume reaches the
Kent's site.
There is an interesting bulge in NO to NO- ratio at the Tower site
for the hours of 0800 to 1600. Since wind speeds are normally stronger during
daytime hours than during nighttime hours, this would bring the plume to the
Tower site faster and with less time to convert NO to NO^. The bulge begins
earlier in the day than the normal increase in wind speea which is most
commonly apparent about noon each day. A more detailed examination of the
relationship of the ratio to wind speed is warranted to better explain this
phenomenon.
CASE STUDIES
Periods of high hourly average concentrations at the fixed monitoring
stations were examined as possible case study periods. From this list, eight
case study periods were chosen for detailed plume behavior analysis based on
their apparent applicability to other complex terrain sites. Table 42 pro-
vides a listing of the case study periods along with a brief description of
plume behavior for each.
Each case study is described in detail with special emphasis on the
interaction of the plume with terrain features and the meteorology that influ-
enced the plume behavior. Tables containing pertinent parameters for a period
of 5 hours before to 5 hours after the pollution episode are provided for each
case study. Table 1 should be referenced for locations and elevations of the
fixed stations.
Case Study I (2300 EST, December 21, 1976 - 0200 EST, December 22, 1976)
High S0? concentrations were recorded at both Hockey and Munsey tor (he
same hour period during this case study (see Table 43). Both stations have
approximately the same bearing from the plant with Hockey on the crest ol
Copper Ridge and Munsey on the side of the ridge facing t.he plant,,
108
-------
CM
s
a
Q
o
fl
o
u
g
o
2
m
o
o
O
o
(M
LO
O
109
-------
TABLE 42. LIST OF CASE STUDIES
Case study
number
Date
Description
I 2300 EST, Dec. 21, 1976 -
0200 EST, Dec. 22, 1976
II 0800 EST, Feb. 11, 1977 -
1400 EST, Feb. 11, 1977
III 1000 EST, April 19, 1977
1500 EST, April 19, 1977
IV 1000 EST, May 15, 1977 -
1200 EST, May 15, 1977
V 0400 EST, June 30, 1977 -
0500 EST, June 30, 1977
VI 0700 EST, July 6, 1977 -
1000 EST, July 6, 1977
VII 0900 EST, July 24, 1977 -
1700 EST, July 24, 1977
VIII 1100 EST, July 26, 1977 •
1700 EST, July 26, 1977
Plume rose to cross Copper Ridge
Down-valley flow during morning, stable
conditions followed by plume fumigation
Prolonged fumigation at lower elevation
stations
High concentrations on ridge during
unstable conditions
Extremely high concentrations on ridge
during stable conditions
Prolonged fumigation
Helicopter cross sections, plume rose
to cross ridge
Helicopter cross section, lower portion
of plume channelled by terrain
110
-------
Ill
I I I I I I S
s - s • a » •
t s s ? ^ a s
.s
Hi
i £ i
« •? ' * '
iKJSssxS
c
"S
111
-------
This case study occurred within an 11-hour period in which Upper Hockey
wind directions shifted from 321° at 1800 EST or, December 21 to 204° at 0400 EST
on December 22. The weather was being influenced on the synoptic scale by a
high pressure system centered over South Carolina. During the case study
Hockey 30 m wind speeds ranged from 2.1 m/s to 3.9 m/s and wind directions
from 279° to 315°. These wind directions would have transported the plume in
the general direction of both Hockey and Munsey. However, as illustrated in
Figures* 33 to 35, the plume position based on the 30 m 2-minute wind vectors
at Hockey are southwest of both the Hockey and Munsey sites. The calculated
plume heights for the case study period indicate that the plume center line was
approximately 90 m to 170 m below the 30 m Hockey level. Wind directions at
the 10 m Hockey level compared to the 30 m level show that the winds were
veering with height. Therefore, during this case study, the 30 m Hockey wind
data did not accurately estimate the position of the plume; the estimated
plume position was veered with respect to the actual.
During this case study the DT3 (30 m temperature - 0.5 m temperature)
values at Tower were strongly positive resulting in the stability classifi-
cation to be very stable. The calculated plume heights were relatively low,
ranging from 195 m to 272 m above plant base. These plume heights would have
resulted in the plume centerline being located approximately halfway between
Hockey and Munsey in the vertical. However, the SCL concentrations at Hockey
averaged about twice as high as the concentrations at Munsey. Therefore, the
plume centerline probably rose to cross Copper Ridge. Wind directions of
approximately 120° recorded at Munsey indicate the existence of downslope flow
on Copper Ridge. This flow may have entrained the lower portion of tne plume
resulting in the elevated concentrations observed at Munsey.
Case Study II (Q800-1400 EST, February 11, 1977)
During this case study (see Table 44) elevated S0? hourly average con-
centrations were reported simultaneously at Castlewood, Lambert, Munsey, Tower,
and Nash's. A high pressure system off the South Carolina Coast was influencing
the weather pattern over soutwest Virginia. A pibal without T-sonde launched
at 1017 EST showed the wind flow at 100 m above plant base to be light and
from a direction of 86°; at approximately 320 m the wind speed was 4 m/s and the
direction had veered sharply to 223° (Figure 36 and 37). Winds at the 30 m
level at Hockey were relatively consistent during this period. Wind speeds
ranged from 3.2 m/sec to 5.1 m/s, wind directions ranged from 213° to 232°,
and sigma azimuths ranged from 4.7° to 10.9°. Wind speeds at Castlewood were
very light and therefore the wind directions were probably not very indicative
*NOTE: Figures 33 to 35 and 43 to 58 show snapshots of the plume at the indicated time. Each lettered point
represents an increased travel increment of 2 minutes. Point A is the location of material which left the stack
2 minutes earlier; Point B is the location of material which left the stack 4 minutes earlier, el<_.
11?
-------
Plant
Miinsev
Hockey
Figure 33. Configuration of plume at2300EST, December 21, 1976
113
-------
Plant
Munsey
Hockev
\
Flgnr* 34. ConfljuMtton of plnme at 2400 EST, December 21, 1976
11/1
-------
Plant
•
Munsey
*
Hockey
Figure 35. Configuration of plume it 0100 EST, December 22, 1976
115
-------
g
M
S
S
a
a
£
III
'1
E
t E
igi
1
c o o c o
-------
a.
-------
ra
I
ra
eo
°«j
1200—|
1100—
1000—
900—
800—
700—
600—
500—I
40OH
300—1
I I I
246
Wind speed, m/sec
Figure VJ. Wind speed (m/s) profile at 1017 EST, February 11, 1977, at the power plant.
118
-------
of the plume flow. The calculated plume heights, with the exception of
hour 1400 EST, were relatively low, ranging from 197 m to 268 m. The stability
was computed to be very stable on hours 0800 to 1100 EST, slightly stable
on hours 1200 to 1300 EST, and neutral on hour 1400 EST.
From hours 0800 to 1000 EST the elevated S0? concentrations were con-
fined to Castlewood. This may be an indication that the plume was entrained
in a downvalley flow along the Clinch River Valley to Castlewood. From hours
1100 through 1400 EST, S0? concentrations at Castlewood decreased while
significant S0? concentrations were reported at Tower, Nash's, Lambert, and
Munsey. This may be the result of a prolonged fumigation incident that
occurred in the valleys surrounding the plant. In this case, a very intense
inversion existed between Nash's elevation and Hockey at 0900 EST and between
Castlewood and Hockey for 1000 and 1100 EST (see Figure 38).
Case Study III (1000 EST - 1500 EST, April 19, 1977)
This case study is a typical example of an event that occurred frequently
during the late morning hours in the vicinity of the Clinch River plant. As
evidenced by Tables 12 to 19 in Section 4, the lower elevation stations (i.e.,
Tower, Castlewood, Munsey, Lamberts and Nashs) show a preference for high SO,
concentrations between the hours of 0900 EST and 1400 EST. This phenomenon 1s
less pronounced at the higher elevation stations (i.e., Kents, Hockey, and
Johnson). Often a number of fixed monitoring stations at different bearings
simultaneously recorded significant concentrations during the late morning
hours.
In this case study, elevated concentrations of S0? (see Table 45)
were recorded between 1100 EST and 1300 EST at Tower, cambert, Munsey and
Nashs. The weather at Clinch River was being synoptically influenced by
a high pressure system off the North Carolina Coast. The OT3 values at
Tower for 0600 EST to 0900 EST indicate the existence of a ground-based
inversion. This inversion probably extended to a point above the lower
elevation stations and inhibited vertical dispersion and plume rise. This
is supported by the calculated plume heights for 0600 EST to 0900 EST which
ranged from 225 m to 232 m. As a result of the local flows within the inver-
sion, the plume was advected through the valleys and mountain gaps surrounding
the plant (Figure 39). The DT3 values for the period of 1100 EST to 1300
EST indicate the inversion was dissipating from the ground up. This eventually
resulted in an elevated inversion above plume height, therefore producing a
fumigating plume. By 1300 EST the elevated inversion was dissipating, bring-
ing a conclusion to the fumigation episode. This is supported by a pibal/
T-sonae launched at 1228 EST which showed a nearly adiabatic "lapse rate
(see Figures, 40, 41 and 42).
To more fully study this prolonged fumigation phenomenon at complex
terrain sites, an extensive pibal/T-sonde network would be needed (i.e.,
launch sites in a number of valleys and on ridge tops) with launchings
119
-------
u
o
s
.£>
0)
fe
I
-------
°- 8 &
CM ™ -~
W
t/3
3;-
t ^ ^
i b«
.2 ^ "C
- s ~
I£
I!
„« — OOOO — OOOOO
10 rJ LT; — i fJ
— r*-r^-er\D
COO '.c
— -OC
o o p
000
c^po^ p^^p^o
ooo coooo
LD 1/1 LT, L" LT1 in LO
bo
«
CX
X
•*
I « o
S *
J3 o
1 : . -'
' U PQ
I * $
121
-------
fv
K
Ol
s
o
1
0)
I
"5.
bo
fl
122
-------
O
,£)
1200
1100 —
1000 —
900 —
800 —
700
600 —
500 —
400 —
300
200 —
100
220° 240° 260° 280°
Wind direction
Figure 40 Wind direction profile at 1228 EST, April 19, 1977, at the power pla
123
-------
a.
a)
o
43
1200 —
1100 —
1000-
900 -
800 —
700 —
600 —
500 —
400 —
300 -
200 -
100 —
1 i i r~ir rnrni
234
Wind speed, m/sec
Figure 41. Wi nd speed profile at 1228 EST, April 19, 1979, at the power plant.
124
-------
I
,£
4J
tJ
00
N
8
P.
(J
o
CM
Tf
W)
E
O
O
CO
o
o
!
o
[ I
I I
o
o
Ol
8
8
88
8
88
125
-------
every 1/2 hour commencing at daybreak. In this way, an accurate account of
the behavior of the inversion and the local wind flow could be documented.
However, one of the problems encountered during the Clinch River Field Project
was the propensity of fog to form in the river valley overnight. Often the
fog would not dissipate until late morning. Obviously, this restricted the
number of pibal launchings during these hours. An acoustic sounder might have
been one method of sidestepping the fog problem in obtaining the structure of
the temperature profile. Perhaps wind data could have been obtained from a
meteorological tower of at least stack height located near the plant.
There are many questions as to how such a complex phenomenon as the one
described in this case study could be modeled, as the location of the fumi-
gation episodes is a function of the terrain configuration and the local flows
transporting the plume.
Case Study IV (1000 EST - 1200 EST, May 15, 1977)
This case study included several hours of high S0? concentrations at
Johnson (see Table 46), a ridge top station; no other fixed monitoring sta-
tion reported elevated concentrations during this period. The weather over
southwest Virginia was being influenced by a weak cold front moving down from
the north. Hockey 30 m wind speeds ranged from 1.4 m/s to 2.2 m/s, wind
directions from 359° to 15", and sigma azimuths from 35° to 45°. These
wind directions would have carried the plume in the general direction of
Johnson.
Figures 43 to 49 show the configuration of the plume based on 2-minute wind
data from Hockey at specific times between 1000 EST and 1100 EST. Table 47 pro-
vides the 2-minute S0? concentrations for the same time period. It appears
the plume swept by the Johnson site during the course of the hour.
For this case study period the calculated plume heights ranged from
367 m to 443 m above plant base or approximately 150 m above the Johnson site.
The stability classification was slightly unstable for 1000 EST. Due to the
unstable conditions, the high concentrations observed at Johnson could be
explained by a large vertical dispersion or looping of the plume.
Case Study V (0400-0500 EST, June 30, 1977)
This case study involves the highest hourly average SO- concentra-
tion measured at Hockey (see Table 48); no other fixed monitoring station
reported an elevated S0? concentration.
The weather over southwest Virginia was influenced by a high pressure
ridge over the Middle Atlantic states and a stationary front extending from
Iowa to North Carolina. The 30 m wind speeds were 1.2 m/s and wind directions
were 278° and 233° with sigma azimuths of 24° and 37°. Hockey 10m wind
speeds were very light and hence directions were probably not indicative
of the plume transport direction. Plume heights were computed to be approxi-
mately 300 m and the stability class was determined to be very stable.
The relatively small wind elevation sigmas (o,-) observed at the Tower 30 m
level may indicate that the vertical extent or the plume was narrow, i.e.,
vertical dispersion was small.
126
-------
H
t/3
w
CO
£ i
I - -^
I s =
S £
iii
^ c
= s
I i
^ *F <£
~ O O C
w
C
bo
•»-(
10
0)
•a
0)
bo
D
J3
ft -P
O
O
5
^
s
c
o
•o
o
re
ffl
*
*
127
-------
West
Plant
Johnson
Hjrure 43. Calculated configuration of
plume at 1000 EST, May 15, 1977.
West
North
Plant
Johnson
Figure 44. Calculated configuration of
plume at 1010 EST, May IS, 1977.
128
-------
West
North
i
Plant
Johnson
Figure 45. Calculated configuration of
plume at 1020 EST, May IS, 1977.
West
Johnson
North
Plant
c
a.
Figure 46. Calculated configuration of
plume at 1030 EST, May 15, 1977.
129
-------
West
Johnson
*
North
Plant
Figure 47. Calculated configuration of
plume at 1040 EST, May 15, 1977.
West
North
Plant
Johnson
t
Figure 48. Calculated configuration of
plume at 1050 EST, May 15, 1977.
130
-------
West
North
I
Plant
**
cr -. »
Johnson
Figure 49. Calculated configuration of plume at
1100 EST, May 15, 1977.
131
-------
TABLE 47. 2-MINUTE S02 CONCENTRATIONS AT JOHNSON
Day Time
May, 15, 1977 1000
1004
1006
1008
1010
1012
1014
1016
1018
1020
1022
1024
1026
1028
1030
1032
1034
1036
1038
1040
1042
1044
1046
1048
1050
1052
1054
1056
1058
1100
Johnson
S02 (ppb)
100
130
140
no
160
150
240
280
290
240
200
180
190
170
130
100
90
100
90
80
70
80
80
90
90
90
80
80
80
90
132
-------
H
D
Q
-------
Figures 50 to 58 display the conf iguation and direction of the plume
based on 2-minute scans of Hockey 30 m wind speeds and directions. When
the plume configurations in those figures are compared to the corresponding
2-minute scans of SO- concentrations in Table 49, there appears to be a good
correlation between plume position and S0? concentrations at Hockey, i.e.,
generally as the plume approached Hockey, S02 concentrations increased and
as the plume retreated from Hockey, S0? concentrations decreased. It can be
concluded from this analysis that the vector addition of 2-minute winds at
Upper Hockey predicted the position of the plume reasonably well for this case
study.
Using the Gaussian relationship,
* -
where x = center! ine concentration at ground
a = horizontal dispersion coefficient
a = vertical dispersion coefficient
H = effective stack height
Q = emission rate
for the calculated plume height of 314 m at the plant and for very stable
conditions, the ground level concentration (x) at Hockey (only 17 m from the
plume centerline in the vertical) was computed to be approximately 117 x 10^ ppb
using Pasquill-Gifford dispersion parameters and assuming the plume center-
line was directly over Hockey. The peak concentration observed at
Hockey was 1160 ppb. If this incident represented a direct hit by the
plume at the Hockey site, then ay and oz, as determined by Pasquill-
Gifford, are too small, thus resulting in an overestimation of the plume
concentration. Other explanations for the discrepancy are that the
horizontal position of the plume did not directly intersect the Hockey
site, or that the plume rose to cross the ridge.
Case Study VI (0700-1000 EST, July 6, 1977)
During this case study elevated $03 hourly average concentrations
were observed at Hockey and Munsey and slightly elevated concentrations
were observed at Tower (see Table 50). The weather over southwest Virginia
was synoptically influenced by a high pressure ridge extending from the
Gulf of Mexico to western Pennsylvania. The stability classifications
ranged from very stable to unstable as the vertical temperature difference
became more negative through the morning. The available Hockey 30 m wind
data showed the wind speeds to be light, the hourly average directions
highly variable and the sigma azimuths relatively large. A pibal/T-sonde
launched at 1005 EST showed the wind speeds in the lowest 500 m to be
134
-------
^
N?
O
tt ' I
a
W
i
i
i
0
H
7
*• .
t
X
M
ft 0 ,1 S '
H
0
<4
U
i M
In
T
f
I
|
H
;,
.H
I
C
t
t-i
o
7 U
A 4J
4* * " "" *
IX
Cj
o"
CT)
(d
O
8
•M
1)
£
,__!
v<
C
,2
•M
3
_W)
O
U
d
o"
«j
c
3
H"
(/5
W
o
m
ro
O
4^
m
E
^
Oc
"o
q
O
nl
(H
_M
"c
O
u
in
3
00
E
135
-------
; A
1!
X
?
a
i
d
~; o
N
W
• T
X
r
; i
H
3
; j
:•*
G
_
i v
vj
«
•M
a
«,
Ou,
[^
o\
1—*
0
r^
0»
C
3
f_T
a
8
S
TO
«
t>
3
M
*§
o
U
(M
to
!1)
^
ce
u.
i
1
!
I
J
:
M
A
; n
i
s
H
I
0
d
i
0
N
W
I
n
»
r
; i
H
a
; j
i
ii
i
-
M
«
•M
1 *
"o •
a
«
2
A
X
t^
2
0
ro
C
H"
W
2
o
4J
«D
OJ
|
"S.
«*-(
o
c
.2
4-1
«
(-
be
C
o
o
ro
in
(U
FH
3
.EP
u^
136
-------
!
>-
a)
y
O •
£
* kM
I
e"
V
i
2
X
X
• M
* ~
A
n
j
' X
s
a
1 o
d
0
! N
^
T
• NJ
1 *
r
t
I
H
'J
A
i 'i
(.1
^
I
H
y
•y
2 t
rt *
E
t^
t-^
o\
"""*
o"
CO
OJ
c
3
„ •—"%
H"
13
o
N
S
4-»
CO
.
E
^
J
137
-------
K
i—I
CU
4)
C
3
H
w
O
O
O
a
138
-------
u
fi
o,
H
8
S
3
DC
o
u
3
oc
139
-------
r~ •"
-------
TABLE 49. 2-MINUTE SO,, CONCENTRATIONS AT HOCKEY
Day Time
581 0338
0340
0342
0344
0346
0348
0350
0352
0354
0356
0358
0400
0406
0408
0410
0412
0414
0416
0418
0420
0422
0424
0426
0428
0430
0432
0434
0436
0438
0440
0442
0444
0446
0448
0450
0452
0454
0456
0458
0500
Hockey
S02 (ppb)
10
10
10
10
170
180
50
20
10
1090
1160
960
870
870
770
730
380
290
180
420
570
370
220
270
210
190
280
290
390
220
120
70
50
50
40
30
20
20
10
10
141
-------
I I I 3 5
s,
' ?
" = c o c' e c c c - r* ;' 3
jli ' s' K 5 S « «' a s =' R- i
^|SS5SS?5S
O O C O O O* 6 O
_£
^ ,J
o
S Hi
o
g -tl
W •; - o o - - o o o c
S
«i I;-
u if| a s ; ; s s s » s
«{ — --•--
E ^ S ^ S * S
I §§§§§§§1111111
m«
-------
light and directions backing from 95° at TOO m to 353° at 500 m (Figures
5S and 60). As shown in Figure 61, the temperature profile is adiabatic
to 500 m. Since the plume height is calculated to be 327 m at 1000 EST,
the fumigation should be complete as 1s confirmed by the S02 measurements
at all stations.
THis case study is another example of prolonged plume fumigation.
In this case the height of the inversion above plant base had exceeded
the height of all stations. A possible remnant of the inversion between
340 and 400 m in the 1005 EST T-sonde suggests the inversion extended to
at least 400 m. Nash's did not observe elevated concentrations although
this site was below both Hockey and Kent's. In order to understand the
low concentrations at Nash's, the plume transport wind regime must be
determined for the case study period. However, the wind regime is not
entirely clear since light wind speeds and variable wind directions were
reported at all stations. A more detailed analysis has not been per-
formed.
A pibal/T-sonde released at 1202 EST (Figures 62, 63 and 64) indicated
the atmosphere had a lapse rate that was approximately adiabatic to a
height of at least 900 m above plant base.
Case Study VII (0800-1700 EST. July 24, 1977)
This case study included elevated S0£ concentrations at Tower with
simultaneous helicopter cross sections (see Table 51). The synoptic
weather pattern was under the influence of a high pressure system off
the Virginia coast. The first helicopter cross-section was performed
between 0837 and 0856 at a distance of 3.1 km from the plant (Figures
65 and 66).
Figure 65 shows the major terrain features of the area surrounding
the plant. The location of the cross-section is illustrated with a dotted
line. Figure 66 presents the cross-section as it would have appeared
if viewing it from the plant,, Each point represents an S02 concentration
(ppo) observed during a helicopter traverse. Only those points greater
than 50 ppb and the adjacent points were plotted. Figure 67 presents the
cross-section as viewed frora a radial plane eminating from the plant,,
Each slash in the vertical represents a helicopter traverse elevation.
The associated numbar is the total mass (g) of SO^ in a theoretical homo-
genous volume in which the dimensions were determined by the length of
the traverses, the vertical spacing of the traverses, and a width of one
mevar (Figure 68).
The center of mass of the plume was calcualted to be at a bearing
of 34° and an elevation of 363 m above plant base. It appears that the
plume was dispersing into a stable layer thus inhibiting the vertical
spread of the plume. The Tower site, which was located in the same
plane as the cross-section and approximately 600 m laterally from the
center of mass, had an hourly average S02 concentration of 31 ppb and a
peak of 85 ppb for the hour ending 0900 EST. A pibal without T-sonde
released at 0847 showed the wind directions to be backing in the lowest
143
-------
88
8
£M
ui '
ui 'sspq juejd
o
. o
00
o
•8
u
-------
888888
ui 'sseq jue[d
145
-------
O) 00
8 8
CM rt
o
CM
CO
o
00
CM
O
IO
CM
O
CM
O
CM
CM
8
CM
O
00
*-H
o
*H
Q,
1 1
-
*>-
i — l
P
H"
a
c
1 s
JJ ^
.rri CO
10 ^
C ^H
.i-f o
^ B.
a
0
TJ
o
[_<
T3
8
O
00
3
be
ui 'aseq
V)
w
CM
O
CM
s
ex
s, s
•o
-------
ro
ro
(M
ro
o
ro
K
a\
H
C/5
w
s
— fM
U -H
O w
* M
CM 2 £
2 2
^ Pi
S u
a; o
oo u
N &
t-.
CM
8888888
CJi 00 K ID ID ^i CO
147
-------
a
t~t
i—-i
Q
W
t/5
U
m
c c. —
JS £ u
*• & £
o E M
!*
o>
X
£ '
Q 2-
~ £
Sr|7
TT ^ S?
*-
ft) C
g £.
s
rt
-------
w
ID
in
oo
O
tv
en
oo
O
CJ
r£~
"3
O
•{3
o
-------
0)
i*
H
'•ft
•s«s
*°»9
• 8
O r-<
•sr
8
Ol
0
traverse
8
00
to
w
kO
LO
00
O
en
oo
O
Tf
CM
0
OJ
3^
O
O
1)
I
8
150
-------
O IV CJ CO
O CJ rO •*
2 e
«5J
E-
oo
w
oo
O
'a
B
M
E
1
8
01
O
O
151
-------
4>
S
&
&.
a
3
ff
T3
(U
O
-------
500 m and veering aloft; wind directions at plume height were observed
to be 215° (Figures 69 and 70).
Another helicopter cross-section was completed at 0947 at a distance
of 6.2 km from the plant. The center of mass for this cross-section was
calculated to be at a bearing of 34° and an elevation of 442 m above
plant base. A pibal without T-sonde released at 0943 showed wind directions
to be backing in the lowest 400 m and veering above; wind direction at
plume height was observed to be 224° (Figure 71). Figure 73 shows a
longitudinal view of the two cross-sections ending at 0856 and 0947
and their relationship to the terrain. It appears from this figure that
the plume center!ine rose as it crossed Buffalo Mountain. This is sub-
stantiated by the higher elevation of the center of mass computed for
the latter cross-section.
Figures 74 to 80 pertain to three separate helicopter cross-sections
done at the same location for different times during this case study.
These three cross-sections show the plume to have greater spread in the
horizontal than the cross-section completed at 0856. Apparently, later
in the day, the plume shifted positions frequently between helicopter
traverses. This assumption is supported by the increases in sigma azimuths
observed at the 30 m Hockey level over the duration of this case study
period. With the exception of 1600 and 1700 EST, Hockey 30 m wind direct-
ions appear to have correlated well with the horizontal position of the
plume center of mass as determined by the helicopter cross-sections.
Perhaps the reason for poor correlation between estimated plume bearing
and center of mass bearing during 1600 and 1700 EST involves the nature
of the cross-sections (i.e., the plume shifted positions between traverses,
the full vertical extent of the plume was not sampled, etc.). These con-
clusions are in disagreement with the analysis of the helicopter cross-
sections conducted by Thompson (1979) in the EPA wind tunnel study at
Research Triangle Park. In that paper Thompson used several helicopter
traverses to support the wind tunnel findings that with wind directions of
238° the plume was being deflected to the north by a ridge immediately
to the northeast of the plant. However, Thompson based his conclusions
on the helicopter data that had not been corrected for instrument response
time and lag time due to the plumbing of the ambient air intake system.
Also, Thompson examined only several traverses and not the whole cross-
section. As illustrated in Figure 75 the location of the plume varied
significantly between traverses.
Table 52 provides the observed bearing of the center of mass and the
corresponding estimate of the bearing of the plume as determined by the
30 m wind directions at Hockey.
153
-------
I I
888
oo N \e>
I I I I
8888
H
c/5
tL)
O •*
_2> o
S «
_
-------
H
00
W
•a t3
(U (1)
5) «)
a. a
M M
^
OJ
IN
"F
•00
¥
on
'o .«
T3 "3
II
fe
3 >,
ui '
155
-------
f-
£3
tv
•*
S
fl «H IX
« O CO
hco S -Si
e > ^
* •s «^
112-
.C ^*
a CM
8
8
8 8
8
O
O
CO
1
156
-------
H
IS
*-4
tM
-------
• 8
• S
M
• «
.8
» S
. R
• S
• S'
•2.8
.0
«§
,,
,8
l/i
I
M
£
8
H
1/3
w
s
i
CO
O
CO
rf
(M
0
P.
o
z
-------
CT) CTl
CO O
CO tx
en t^ \o cr> *O en tx
CJ -H oo co
-------
•5.
,8
.8
•K?
•8 ~
•J5«| »
• 8
« a-!»s »=.o
•— — —
in »S
.? .£ "
.8.8 »I% s .s
*" *~ »i
* ^
• ~
• —
s
r8
VD
o
CO
i 5 J^
I Tf
^8* >
00 a1 "3
o
l-o
I
o
-------
rO
N
in
CO
i-i CTI o^ cioo oo in in m,-i
\o d Is* ^o-^* vo oo tn cooo
fi
•B
o
8
o
o
161
o
S
-------
r »
• 2
,8
» s
» 5
8
(M
»S • £
(M *•
• I »s
•I * 2
• C
• S
^r
•I
• °
• S
° o
»-i vo
i u
i S3 £
! (U J>
I > ?\
\ a rq
2
o
p.
o
o
C1
0)
rt *^J
S
8
8888
o
162
-------
o
<£>
CM
00
-------
TABLE 52. ESTIMATED PLUME BEARING AND
BEARING OF CENTER OF MASS
Day
July 24
Time
(EST)
0900
1000
1100
1300
1400
1500
1600
1700
Estimated
Plume Bearing
from Hockey Winds
34°
42°
52°
50°
66°
58°
74°
93°
Actual
Center of
Mass Bear
34°
35°
54°
54°
58°
54°
54°
54°
ing
Case Study VIII (1100 - 1700 EST. July 26. 1977)
This case study contains moderately elevated concentrations at Johnson
with simultaneous helicopter cross-sections of the plume (see Table 53).
Over the case study period hourly averages showed wind speeds ranging from
3,2 to 5.0 m/s, wind directions of 345° to 4°, and sigma azimuths of 11°
to 34° at the Hockey 30 m level. The weather was synoptically influenced
by a large high pressure system centered over the northern plains.
A helicopter cross-section was performed between 1510 and 1619 EST at
a distance of 1.6 km from the plant (Figure 81, 82 and 83). The plume
center of mass was calculated to be at a bearing of 175° and an elevation
of 534 m above plant base. The calculated plume heights which ranged
from 407 m to 504 m for the hours of 1500, 1600 and 1700 were a slight
underestimation of the actual plume height as determined by the center
of mass of the helicopter cross-section. The calculated mean of 463 m
is 13 percent less than observed mean of 534 m.
During this cross-section Johnson recorded only moderately elevated
concentrations of S02, probably due in part to the high plume heights and
plume bearing. Johnson was located at a
center of mass was at 175°,
bearing of 191°; the plume
A pibal without T-sonde released at 1536 EST showed wind directions
to be veering from 316° to 21° in the lowest 350 m above plant base
(Figures 84 and 85). The wind direction at plume height was observed to
be 3°. Since Figure 82 presents the plume as it would have appeared to
an observer viewing it from the plant, the plume center!ine was backing
with height. This is in disagreement with the pibal observations made
during the cross-section. It is likely that the section of the plume
below 1000 m was being channeled by a terrain feature, a gap in the first
ridge south of the plant.
164
-------
i
i
p
g
=
£
p
VI
W
CO
3
IT
a
c
s-
1
s j I
- """ J;
"o '^3-
—* t^-
c — -
S £ i
~*
a c .
™ " **
*0 *^ c
EC—*
— — C
c* R *
N*= ^.
8 1 '
•** £ "~*
fc O ti
!?<£
«* m ti
^ D 2.
"" &
X S ~
IS fc
- § 1
„
*w
V C
1 ~
u
J
z
X
C1
M
5
i
~
c
f -
c.
— §
^ J
w.
rt
1 t
t J
D ""
_O
"^rM
S *
1
0.
a
ifl^oovia
*-, -.-.rjrg^rj(N(N]-.m-<
«
CM c\j m ^* PO u^LOtn'V^
rt vo^^^^cnc.o^^r.O
* rtm«^rtfOr*i^rom«
^ S^^g^CWOX^oS
O CCCOOOOOCOC
o cooooooocco
"-.. •^.''s.'^"^,"^.'^,"^.'1^"^"^.^.
•V •v-.y-^^cjrJh-Tj'Os^C-,
O OOCOOOOOOOO
c oooooooooco
0 OO^D^^O-C^VO^^
o oocooooocoo
rg (NrrOCOCCOOOO
*-»-S = a5S2S£:
^c>
(M
•^ Is- r t tn v
\ii •— * \o w t^-
G C C O C
o o c o o
N f- 1** t*» O
0 0 C O C
* S Si1 S «
o c o c o
o o c o o
vO VD UP * 0
O O O •"* ^
o c1 c" o o"
T-H ^ ?J (M CNJ
165
-------
a
*-<
in
rt
o
166
-------
•" 00,,
• c •§ »K (
» O1.
; »0 •£,
:.g-s
£*-
o
»*s
s»?
• I/- Om ^ O
01 _ CO *2 *?•-•? *~
• 5 „ _ "•» -
.o«§^.
•I-?1
'a*s.s ^•I.S
•S
?»s
2
o
I
I
"o
•° 3 -
a s
o
0
ro
O
O
(M
O C
2 ?
D O
3 0
3 0.
oo
/
O
O
t^
/
O
o
1C
'1 1
8
in
167
-------
. CM
CM
o
T-4
in
t--.
o
00 -r-1 Q -t •
<£> in o oo
oo O CM tn I
CO t^ CM CT> I
O
K
CTl
CM
Q
I
o
.s
1
00
a)
-------
Ol 00
1 i 1 I I
o o o o o
LO rf m N «
t--.
N.
Oi
JX
3
•£ H
5 3
ex «
in TO
•a -^
c »
(U
-------
SUMMARY
Terrain influences are evident in the meteorological observations,
the ground level measurements of pollutants, and in helicopter measure-
ments of the power plant plume. Frequencies of occurrence of wind direction
tabulated at each monitoring site show that each site, located in a valley
or a pass where terrain influences are important, experiences a major
component of air flow that is aligned with terrain features. Sites in
valleys have lowered wind speeds than sites on ridges, and higher
turbulence levels, as indicated by the standard deviation of wind direction
for a 1-hour period.
Detailed studies of 8 selected cases of measured high SO? incidents
included 4 situations with high values measured on ridges with flow directed
across the ridges, 3 situations with high values measured in valleys due
to prolonged fumigation of the plume into the valleys, and 1 situation
of terrain channeling of the flow as shown by helicopter cross-sections.
These case studies illustrate the nature of the plume behavior problem
from a descriptive point of view. The behavior of nitrogen oxide gases
in the plume were also studied along with the effects of terrain influences.
The NO/NOX data show evidence of rapid initial transformation of NO to N0£
within the first 3 km of travel followed by a much slower, almost stagnant
rate of transformation beyond this distance out to 15 km.
170
-------
Section 9
SULFATE ANALYSIS
Sulfate concentrations were determined from hourly samples of particu-
late collected using hi-vol type monitors at four fixed stations. The
sulfate data were obtained in order to characterize sulfate concentrations
in the plume and to compare plume concentrations with general background
levels. A significant preliminary step in interpreting the sulfate measure-
ments was to separate interfering effects in the measurement process from
the influences of ambient air concentrations.
The sulfate measurements include components that were not deposited on
the filter during active sampling. One component is the sulfate found in
the filter before deployment to the field. This filter background level was
determined by analyzing the sulfate level for one out of every 50 filters.
Another component is the fallout of sulfate on the filters after exposure
to ambient air. This value was determined by measuring the sulfate level
on filters that were not exposed to a blower-forced sample of air. At least
one such "natural deposition" filter was used with each set of actively
sampled filters at each fixed monitoring site.
Figure 86 presents sulfate concentration-wind roses for the fixed
monitoring sites. The wind directions at the 30 m Hockey level were used
for these analyses. There appears to be no clear trend at any fixed station
for elevated sulfate levels with wind directions that would have transported
the plume to the station from the plant. However, each station had either a
primary or secondary peak of sulfate concentration with wind directions out of
the west-southwest, the prevailing wind direction. Perhaps this is an indication
of long-range sulfate transport.
It can be concluded from Figure 87 that the amount of sulfate that
accumulated on natural deposition filters was a function of time. We have
concluded, based on the information shown from this figure, that the variance
of the the amount of sulfate fallout on natural deposition filters became toe
large for useful analysis after approximately 12 days. The standard error of
estimate of the data about the linear regression line was 391 yg when con-
sidering all the plotted data. However, the standard error of estimate
decreased to 281 ug when considering only those cases with exposure times of
less than 13 days.
Also shown as open circle data points on Figure 87 are the standard devia-
tions of the fallout for those cases when there were more than one natural
deposition filter at a site. In general the standard deviations increased
with exposure time. There were 11 cases of multiple natural deposition exposure
171
-------
172
-------
o
o
o
o
o
173
-------
of which 6 were less than 13 days and 5 were 13 days or more. For all 11 cases
the standard deviation is 209 yg. For the six cases with less than 13 days
exposure the standard deviation is 100 yg. The variations in deposition data
at a single site shows that this accounts for a significant part of the variance
in the regression analysis of measurements at different sites and times
Figure 88 presents pollution roses based only on those cases in which the
filters were exposed to ambient air for 12 days or less. It may be noted
that there is a preference for higher sulfate concentrations with wind
directions from the west-southwest.
Tables 54 to 60 and Figures 89 to 108 pertain to seven cases in which two
or more fixed monitoring stations were simultaneously sampling sulfates, and
elevated SO. concentrations were observed at one or more of the stations.
As illustrated in these figures, there appears to be little correlation between
SO* and SO,, concentrations. -,It should be noted that in these figures SO- con-
centrations less than 30 yg/m are close to the measuring threshold of the
instrument. Tables 54 to 60 indicate that for these cases wind directions (30 m
level at Hockey) were generally from the southwest. There is no clear evidence
of stations downwind from the plant receiving higher S04 concentrations. This
supports the hypothesis that some of the observed sulfate concentrations in the
area around the plant came from a distant source, perhaps to the southwest.
Sulfati us/*'
M - Direction with no samples available
Figure 88. Average sulfate concentration for each wind direction, measured at the
Hockey 30-m level, for filters exposed no more than 12 diys.
174
-------
TABLE 54. DATA ON ESTIMATED BEARING OF THE PLUME, BASED ON 30 M HOCKEY WINDS (JULY b, 1977)
Date
July 8, 1977
Tine
(EST)
0900
1000
1100
1200
1300
1400
1500
1600
1700
Wind Speed
(m/s)
2.3
1.6
2.2
2.0
2.1
2.4
2.2
2.4
2.4
Direction
(deg.)
256
296
283
298
328
327
299
309
242
Sigma Azimuth
(deg.)
32
38
31
29
43
32
25
16
28
Estimated Plume
Bearing (deg.)
76
116
103
118
148
147
119
129
62
TABLE 55. DATA ON ESTIMATED BEARING OF THE PLUME, BASED ON 30 M HOCKEY WINDS (JULY 20, 1977)
Date
July 20, 1977
Time
(EST)
1000
1100
1200
1300
1400
1500
1600
1700
Wind Speed
(m/s)
1.4
2.6
1.2
1.5
2.3
2.8
2.9
2.8
Direction
{deg.)
297
311
308
325
327
285
296
305
Sigma Azimuth
(deg.)
34
16
62
56
28
30
22
14
Estimated Plume
Bearing (deg.)
117
131
128
145
147
105
116
125
TABLE 56. DATA ON ESTIMATED BEARING OF THE PLUME, BASED ON 30 M HOCKEY WINDS (JULY 27, 1977)
Date
July 27, 1977
|
Time
(EST)
0900
1000
1100
1200
1300
1400
Wind Speed
(m/s)
4.6
4.3
1.7
1.7
0.6
2.1
Direction
(deg.)
112
116
135
353
192
--
Sigma Azimuth
(deg.)
14
15
63
--
91
--
Estimated Plume
Bearing (deg.)
292
296
315
173
12
--
175
-------
TABLE 57. DATA ON ESTIMATED BEARING OF THE PLUME. BASED ON 30 M HOCKEY WINDS (AUGUST 12, 1977)
Date
August 12. 1977
Time
(EST)
0800
0900
1000
1100
1200
1300
1400
1500
1600
Wind Speed
{m/s}
3.6
3.6
4.6
2.7
3.7
4.1
2.7
2.2
5.2
Direction
(deg.)
224
222
232
284
238
219
304
245
302
Sigma Azimuth
(deg.)
11
10
16
31
20
15
44
44
39
Estimated Plume
Bearing (deg.)
44
42
52
104
58
39
124
65
122
TABLE 58. DATA ON ESTIMATED BEARING OF THE PLUME. BASED ON 30 M HOCKEY WINDS (AUGUST 22. 1977)
Date
August 22. 1977
Time
(EST)
1100
1200
1300
1400
1500
1600
1700
1800
Wind Speed
(m/s)
4.9
3.7
2.1
3.0
1.8
2.3
2.3
3.0
Direction
(deg.)
295
307
275
225
246
321
312
318
Sigma Azimuth
(deg.)
12
31
34
26
35
18
18
17
Estimated Plume
Bearing (deg.)
115
127
95
45
66
141
132
138
TABLE 59. DATA ON ESTIMATED BEARING OF THE PLUME, BASED ON 30 M HOCKEY WINDS (AUGUST 23, 1977)
Date
August 23. 1977
Time
(EST)
0900
1000
1100
1200
1300
1400
1500
1600
1700
Wind Speed
(m/s)
2.8
3.5
4.4
b.5
b.l
b.4
5.1
3.9
2.8
Direction
(deg.)
232
232
228
224
21 b
21U
224
245
227
Sigma Azimuth
(deg.)
22
37
16
Estimated Plume
Bearing (deg.)
52
52
48
IJ 44
11
1 1
12
16
17
Jb
.ill
44
65
47
176
-------
TABLE 60. DATA ON ESTIMATED BEARING OF THE PLUME. BASED ON 30 M HOCKEY WINDS (AUGUST 26, 1977}
Date
August 26, 1977
Time
(EST)
0800
0900
1000
1100
1200
1300
1400
1500
1600
1700
1800
Mind Speed
(m/s)
1.1
0.4
0.8
3.4
4.4
4.3
4.4
4.1
4.3
4.7
4.3
Direction
{deg.)
104
257
239
216
219
190
199
190
200
183
182
Sigma Azimuth
(deg.)
43
96
56
28
9
19
20
20
18
15
17
Estimated Plume
Bearing (deg.)
284
77
59
36
39
10
19
10
20
3
2
100
50
3
1
)900 1000
—-——-_
1
1100
—^ — ' —
1200
1300
1400
' " J
1500 1600
1700
100
Ending time of observation (EST)
Figure 89. SO (dotted line) and SO concentrations at Hockey for July 8, 1977.
(Bearing to plant is 287°)
00 1
000 1
100 1200 1300 1400 1500
1600
' •
1
1700
Ending time of observation (EST)
Figure 90. SO (dotted line) and SO concentrations at Johnson for July 8, 1977.
(Bearing to plant is 11°)
177
-------
150
100
B
•fc
a.
g 50
-------
a
o
tJ
«
V
O
-------
DC
ci
o
— §
•"*"* r*.
2
•g
2 P
in
w
rt
T)
O
180
-------
2
CM
CM
CM
tt
1
/
1
1
t
1
1
1
1
1
1 »
I 1
1
1
i
' 1
1
I
1
-g 1
8 1"
"-"5 "3
0
E
4->
s
o o
"""* H 1
c2 PJ
S. »)
_. o
5 c
O p
° tO o
^^•in cti ^
" ^ o
w <"
43 T3
o e
M-( n)
0 ^.
-1 I 1
"*" "Oo^
bo iJ t^
d t oo
fl O CM
0 2- ">
O W '"
"~" M O 0
s rt
0.
S K -2
>— o p C
« K> a
" »*H flj
u. «_
g
tH
8 c
s fc
ct
o
tc
§rt
>
> IA M
3
.O
XI
O
a
_o
I
(U
o
a
o
o
O
£"
CM^
§1
oo
cr. to
E S.
Is
181
-------
to
X
o
MH
O
«)
a
4-1
M)
fi
3
a
M
t-~
i^
CTl
*V.
H
00
a
5
aj
O MM
•a °
4)
s
fi
w
a
CJ
(M
S1
rt
a
o
•43
I
O
oo
•S
rt
•w "S1
b£ fi
c a
0 ~
-T3 w
o
00
rt
i—i
a
2 oo
182
-------
100
S
•fc
a.
50
0900
100
6
3.
&
«
3
to
"SB
0900
1000
~T
1200
~r
1300
1100 1200 1300 1400
Ending time of observation (EST)
1500
~r
1600
1
1700
Figure 101. SO (dotted line) and SO concentrations at Hockey for August 23, 1977.
(Bearing to plant is 287°)
I
1000
I
1100
1200
1300
I
1400
I
1500
I
1600
1700
Ending time of observation (EST)
Figure 102. SO (dotted line) and SO concentrations at Johnson for August 23, 1977.
(Bearing to plant is 11°)
183
-------
a
•s
V
s
o
13
4-1
3
60
I
O
•-B
rt
O
W5
T3
ti
rt
T3 O
0) CO
6 •*
- (M
-8
m S
2 g,
u .«
B ^
3 rt
6C -
a.
CO
CM
£. o
fl
0)
o
a
8
T30
8J
d,
ui/8ri '
184
-------
•§ c-
a
o
73
US a
tt
a
3
01
•3
o
u
u
o
c
*J T3 O
oo $ 55
I 35
a — -1
W (M-H
LT? a
2 M
« -B
a §
CO oj
p
t/5
w
a
1
O)
fe
u
JS
o
•*-!
0
4)
E
T3
bo
a
=3
^
O
t— *
•M
rt
s
o
•43
n)
is
r]
a
8
0^
to
•d
C
rt
^^
o>
a
i— i
T3
S 0~~
*J >-*
*^ w
(N«
0 g
W J
a
185
-------
a
60
a
tM
tt
I
O
^.nir o
rH
- 5
2
— §
t-i
— g
"*
i
i
1 o
2 — 3
^_^
to
Si
a
0
*4J
a
c*
1
o
"o
V
g
u>
_B
W
0
o
•o
a
B
1)
O
fi
8
T:
o
in
-o
c
rt
V
a
i — *
"O <
£
2.
_
o
•
o
(U
i-,
•i-t
i<
-\
^
%
-4J
rt
i — i
O.
o
w
_c
rt
V
03
§
*""~S
o
— s
0
p
co
B
O
7!
01
jj
I
n
o
•c
1
B
O
B
8
T
o
1/5
B
a!
E
TJ
I
'O
a
-ao
S I
OH
00 3
? ff
2 -c
art
0)
r^ 03
186
-------
SECTION 10
S02 DISPERSION ANALYSIS
The Clinch River pollutant and meteorological data sets offer an oppor-
tunity for development and validation of dispersion models appropriate for
complex terrain. The major goal of this portion of the study was to deter-
mine the modifications to a flat terrain Gaussian model which would produce
estimates of SO- concentrations that would best compare with measured S0?
values at the eight fixed stations.
The flat terrain model chosen for the study was the PTMTP model from the
EPA UNAMAP (User's Network for Applied Modeling of Air Pollution) system.
Assumptions made in this model are:
1. Meteorological conditions are steady-state each hour.
2. Dispersion parameters are those presented in the Workbook
of Atmospheric Dispersion Estimates.
3. Sources and receptors exist in either flat or gently rolling
terrain.
4. No aerodynamic dowriwash occurs.
5 No wind direction or wind speed shear occurs.
6. The given stability exists from ground-level to well above
the top of the plume.
Modifications were made to this model to allow a continuous stream of hourly
input values, including wind speed and direction, stability class, mixing
height, plume height, and emissions data, and to allow the output of a file
of estimated S02 concentrations appropriate for the eight stations. The
most important modification was the use of the plume heights determined from
sounding data described in Section 6 rather than the use of the Briggs1 sub-
routine BEH072.
Wind speed and direction data from the Hockey 30 m level were used as
model inputs. Estimated S0? emissions from Units 1 and 2 of the plant
were used along with the measured S0? emissions from Unit 3. When the
Unit 3 measurements were missing, the estimated values were utilized (see
Section 5).
187
-------
There is general agreement that stability classifications which must
be determined without a measurement of the boundary layer height should
be based on the vertical gradient of potential temperature (A3) and the
wind speed (u) in terms of the ratio AG/U^ (Hanna et al., 1977). Although
some T-sonde observations were made and temperature measurements were
available from several heights, there was only occasionally sufficient
data to define the depth of the boundary layer. As an alternative, we
have used the bulk Richardson number (B), which includes the ratio A0/u2
in it, as a measure of stability. Because the scale of this parameter
is very much dependent on how it is measured, we decided to compare measure-
ments of B with the Pasquill stability classifications determined by
Turner's rules (1964) using cloud measurements from Tri-Cities Airport
and wind speeds from the Tower site. The bulk Richardson number was deter-
mined using Tower site measurements as follows:
g
= g(5T/3z + y)z
Tu2
In this formulation, the vertical temperature gradient aT/az was measured
over the 30 m to 0.5 m layer, z was taken as 14.75 m, the 4 m temperature
was used at T, and u was the 10 m wind speed. T is the dry adiabatic lapse
rate and equal to -0.00976'C/m. Stability classes were determined for both
methods for about 250 hours. The distribution of B values for each stability
class was analyzed in order to establish ranges of B that could be reasonably
associated with each class. The results in Table 61 show how measurements
of B compare with corresponding measures of Turner-Pasquill stability classes.
Although the joint distribution does not show mutually consistent
stratifications for the two stability measures, there is a general trend among
major clusters in the table which shows that more often than not the two
distributors agree as to how stable or unstable any given hour is. Out of 255
observations included in the comparison, 121 observations, or about half, are
mutually consistent under the selected classifications system shown below.
Each of the two methods has advantages, and one cannot be certain for any given
hour which methods best characterizes what is influencing the exhaust plume
from the plant. The Richardson number is measured closer to the source and may
reflect the effect of valley fog and other conditions which are not present at
the Tri-Cities Airport. However, the tower is low and within the influence of
roughness elements so that it does not represent area-wide effects as well as
the airport operations. We have concluded that the onsite Richardson number
better represents stability during nighttime and early morning conditions than
the remote Airport measurements and have used the premeasured to characterize
stability. It may be noted that different ranges of Bulk Richardson would be
obtained for different tower heights, and the values given here can only be
considered representative of this particular tower height. One should not use
the criteria listed below to classify measurements from other towers with
different measurement heights.
188
-------
TABLE 61. DISTRIBUTION OF THE JOINT OCCURRENCES
OF CLASSES OF BULK RICHARDSON NUMBER AND PASQUILL STABILITY
Pasquill Stability Class
Bulk Richardson No.
<-0.085
-0.085 to <-0.065
-0.065 to <-0.045
-0.045 to <-0.025
-0.025 to <-0.015
-0.015 to <-0.005
-0.005 to O.005
0.005 to <0.015
0.015 to <0.025
0.025 to <0.045
0.045 to <0.095
>0.095
A
9
2
5
5
2
3
0
0
0
0
1
8
B
6
6
6
8
2
4
2
1
0
0
0
5
C
6
0
1
4
5
5
3
0
1
0
0
4
D
0
0
1
3
4
10
8
13
3
0
4
6
E
0
0
0
1
1
1
1
7
2
0
0
1
F
0
0
0
2
2
3
4
2
0
2
5
65
Based on the results in Table 61, the following ranges of B were
established for determining an hourly stability class for model input:
Class Range of Bulk Richardson Number
A B _< -0.07
B -0.07 < B < -0.03
'.' -().().i • I! • (I.()()S
D -0.005 -, B ^ 0.01
E 0.01 < B £ 0.10
F B > 0.10
189
-------
The parcel method was used to obtain hourly mixing heights from the
temperature profiles described in Section 7. A profile of mixing ratio
(g/k§) from the same rawinsonde station as the temperature profile was also
available for each hour for use in determining the mixing condensation level.
Below the mixing condensation level the parcel is lifted dry adiabatically
and above this level the parcel is lifted moist adiabatically. When the
parcel, lifted from the surface becomes 1°C colder than the profile tempera-
ture, the mixing height has been determined.
The period modeled extended from Julian Day 287 of 1976 through Julian
Day 273 of 1977. The first model run was made with no further modification
to the program (i.e., this was a flat terrain model run). The next five
model runs each involved the implementation of a plume height adjustment over
terrain elevated above the plant base. The five models employed the follow-
ing five plume height adjustments:
1. CRSTER adjustment (USEPA 1977) - plume height is decreased
by the full amount of the terrain elevation difference
between plant and receptor for all stability classes.
2. NOAA adjustment (Van der Hoven et al. 1972) - plume height
is decreased by the full amount of the elevation difference
for stable conditions only. No adjustment for neutral and
unstable conditions.
3. VALLEY adjustment (Burt 1977) - plume height is decreased
by the full amount of the elevation difference for stable
conditions but the plume is not allowed to approach the
terrain any closer than 10 m. No adjustment for neutral
and unstable conditions.
4. ERT-LAPPES adjustment (Egan 1975 and Slowik et al. 1977) -
plume height is decreased by half the elevation difference
for unstable and neutral conditions and by the amount Fz
for stable conditions, where z is the full elevation differ-
ence and F = 1 for plume heights (H) > 1.7z, F = 0.65 for
plume heights (H) < z, and F = 0.65 + H-z for z < H < 1.7z.
~2T
5. GEOMET adjustment (Koch 1978) - plume height is
reduced by 60 percent of the elevation difference for
stable conditions, but the plume height is never allowed
to become less than 40 percent of the unadjusted value.
No adjustment for neutral and unstable conditions.
After each model run a statistics program was employed to produce a
cumulative frequency distribution of the estimated concentrations for each
station. These distributions could then be compared with the distributions
190
-------
of measured values. The program also computed paired statistics such as the
correlation coefficient, slope and intercept resulting from linear regression
and the root-mean-square error. The highest and second highest measured and
computed concentrations at each station were also compared. All statistics
except the cumulative frequency distributions were computed separately for
stable, neutral and unstable conditions as well as for all stabilities com-
bined.
The plume height for each stack was computed using a multiple stack
adjustment, which was found to be the best technique. The LAPPES o adjust-
ment (Slowik et al. 1977) involves multiplying the appropriate Pasqtnll-
Gifford value by a factor^of 1.43 for unstable and neutral conditions. A
multiplier of (1.43/0.4 u ' } is used for stable conditions, where u is
the wind speed. These factors result from the LAPPES program conducted in
Pennsylvania from 1967 to 1971. The GEOMET o adjustment (Koch 1978)
involves the use of a multiplier of 3.0 for neutral and stable conditions and
a multiplier of 2.0 for slightly unstable conditions. No adjustments are made
for unstable conditions. The values were largely derived from plume measure-
ments in the vicinity of Maryland power plants. The Clinch River helicopter
data also provide a basis for little or no modification to a for unstable
conditions. A model using these parameters along with the GtOMET plume height
adjustment has been validated against measured concentrations at several
Pennsylvania power plants in complex terrain (Koch 1977).
In addition to the a adjustments, the °2 formulation suggested by
F.B. Smith (1972) was applied, which for the average roughness of the terrain
in the vicinity of the Clinch River plant is a factor of 1.5 increase over
the PasquiH-Gifford values. An initial plume dilution factor and the use
of a power law to extrapolate the transport wind speed from the Hockey 30
m station elevation to plume height were included in the additional model
runs. The initial dilution factor accounts for the spread of the plume
during the entrainment stage. A method proposed by Pasquill (1976) was
employed which involves modification of the vertical dispersion parameters
as follows:
where °z(l) 1S the original value unmodified by the hot plume entrainment
process and H is the plume rise. The wind speed at plume height was
estimated by extrapolating the Hockey 30 m wind speed to plume height using
a power law with the exponents (p) dependent on stability class:
191
-------
where Up is the wind speed at plume height, U^ is the Hockey 30 m wind
speed, H is the plume height and ZH is the elevation of the Hockey 30 m
station above the plant. There is much uncertainty in this procedure as no
set of exponents appropriate for complex terrain exist. The following
exponent values were employed which resulted from the analysis of meteoro-
logical tower data from several flat-to-rolling-terrain sites (see analysis
in Section 7 ).
Pasquill-Turner
Stability Class _P
A 0.10
B 0.10
C 0.10
D (day) 0.12
D (night) 0.21
E 0.35
F 0.49
The final model run employed the recommendations of Pasqulll (1976)
for using measured o^. This method has been described in Section 6; the
method was employed along with the GEOMET plume height adjustment and the
wind extrapolation technique. The standard Pasquill-Gifford oz was
employed.
Two statistics (root-mean-square error and correlation coefficient)
from the results of the aforementioned model runs appear in Tables 62
through 69. Model calculations could only be performed when all necessary
input data were available., Paired comparisons of measured SOp concentrations
and model estimates could only be made when both were available. Table 70
shows the number of comparisons that were performed for each station and
for each stability. Table 71 presents a comparison of the highest and
second highest 1-hour average S02 concentrations produced for each station
by four of the modeling techniques along with the highest and second
highest measured values.
For the purposes of statistical comparison, background S02 concentra-
tions were added to the model estimates. A background value was estimated
for each hour by the following procedure:
1. All available fixed-station S02 observations for each
hour were used to compute a mean and standard deviation.
192
-------
CO
UJ
••H
1—
_J
I
J
^
£
CO
o
CO
5
OJ
c
i-
4J
• 4J
ttj re
E 4-1
•— o> c
^ 1-3
IQ g
O >%
o S "S
rw i: ^
S l 5
>~ JT •*-
3 I •?
s ! !
4J O>
^- ^- OJ— *,
re *c P — I
— 1 re 1'
a -^ ^ si
u JD QJ i5
re • c
i^ OJ *— "re
C -D QJ 1-
— z at a.'
~: ^ 7-
i*. 1— ~" "3
a< a. -M '•-I
ex en
•£ < £ —
Q 2 h-
o -^ tx
z u. u
in
^_
Oi
I
I
c S
II
C —)
o -o
u re
OJ C
-Q n
re i-
i.
°1
— o
Q.—
*D 13
C
4-> O
C U
Ii
in .0
3 re
re c
3
C
re c
i- ^
^ ^
CT c
* 0
>
oj —
- c
-- t-
•Ej. OJ
^ o
= "S
3 X
I
3;
|
C
SI
Q. in
3 3
-T3
OJ (Q
C i-
re OJ
d.
|!
u c
X O
OJ f~
c. ^5
— c
•*-* o
re u
>
OJ ^
— "in
1_ i.
1- O
*J C
*-*
f ^
Ii
5 c
O 4J c
4-" 3 O
ft) f 0,
(J re i—
t. c .0
^c^
§ 4^ §
•S"51
a> c .—
U
CJ OJ 3
m o c
o
J=. C
C" "C
j= E ^
1 i *
a. to t-
-:
§
-S
3*.
S) £ o
re
C OJ «J
re (-> »—
« £1
"8 «."
QJ -?~ m
X •" »
QJ X isj
<— C i. o
> t- OJ E
o ^j JT re
re OJ
E re c 5
-C 0
0 4-> — >,
cn c "O C
5 iis
« *£ i
01 4J TO OJ
H~ in Qj
in o c"
•«- t. re
•-> OJ j=
^ S "c 4-<
a> o 3 c
^ i. re
jT c. in a.
C
fti O O OJ
I t,5^
CL JD T3 in
C "C
*j -r o o
OJ 3 C
X — C 4-1 -O
• £"=?*
< sj ^ ^ e
O S — i-
in § tn ^ «£
I ,
re
P OO
— : c^
&
_J i
I^ bj
r-
uj
r-k O »^
4^^o
XX —
r C
J- O
1C (J
tn at
Ss
>>t-
^f *-*
O)
II
7"
o
_j
1
M
o.
•D
OJ
in £
oj &
C (J
j=
3 O
e c
o
1 ll
* *"£
OJ * ^
re — «
ft) J=^v
« ^- k.
^ i2
J- y
L. ° **"
O U C
*J **~ w
C C k.
O -— *J
U 3 i.
re *— • o
0 ^
4-> -a
-• 58
U.
X ^ 1. fcD
^ OJ O •
"o re c — *
— i- X
a. c «j •
— — in O
*j re
i fe^ «vT^
re 3 CL •—'
wi «o a>
re o re
t, f T
O QJ — '
1 -C OJ
c. t— .c
1 =
o
D •—
— „
E "~~
"• I
1
3
C
e
o
ft!
1
0
c
o
IO
o
X
sj
k, "re
II
Ii
in
:i
j= s-
fi
ji
CL in
c
1 1
c
o
re
1
i
X
c
193
-------
LU
t-
j
CO
1—
-j
•*
ce
£
00
s
>
D-
§
i
LU
S
2
£
g
UJ
G2
fe
z
•—*
LU
ec
ct
0
o 1
VD
i
CO
«:
»—
I
i
i-4c
S£~-
"
"* ° u
5521
saCeS
s *
^=•=1
g^li
**2i
^
»- "s
ii .»"
sa!"
x u^
55
OS
z£
as
£""
^ ^
~?<
s"
* h
% 5
£8
V
lx
1 *~
»,
2 =
*
i=
5
5_
E
C
— • L.
*"
|
2
«,~-,0--,.~
2 S S 2 S S S 5
oeoeeceSe
= Sr-S£2S
gso»^g^s
<-^r->o — — O — «M
oeeddedca
^?i8 — — oCR
vc^^tt-.^.pe
2S5?SSS5
-^ r. 0 C o -. C -
— r^O^^SSri
cooeoooo
-^SSo — 52
COOOOOC-O
OOOOOOOO
dcoodocs
5. >. >S t §
isiiiiii
;
i
o
I—
a
1
UJ
a
CO
z
oc
o
U—
1
h- *
|
K-
Z
OS
s
!£!
o
LU
O
i
S
i
•
S
LU
|—
= v|.
X ° — e
ii~°*
aal =
••^g
ii||
*a"a
= o*£
h. b. "5
SI-*"
u!
s tT
ii
aa
= ='
ll
X
s.f
< &
2E
oe "^
S
IP
UJ
UJ
Q.
ae
UJ
ir
_
5 s
3
^
a s
ac
_c
1 *~
I !
i 5
o^»^«^«T
S52S£S*i2
»*. — rtrf»^tf>^ —
F^>— VOXp — B9t
ir » cw — •w»o-"-o
^>«.^.gJ0 — 09t
ilxSiS^^
— ^OBOrvrf»-c\,
^vOr»tnOgOODm
2 *•" *"* ~ ""
0 ,.,,-«.
£38 — ft«2tK
Ofp^Ln^firy'W'O
-3'-'S~S
333isSE:*K
X — rO^^-^O*^
—
(M^mOWB^O^
SSSSS^**1^
5*-:£-:'<:T2
iiSs?":*S
— ' C
«» >» >t i. O ^
Isllllls
194
-------
to
p
»— 1
o
i
o
LU
^
v>
i§ i
§
u.
z
o
*2
Of
—
f.
8
§
1
g
s
i
lt
o
™»
o
5
LU
Of
ee.
s
s
LU
CO
I—
i*1l
3^ =
*"1 =
a«|I
XB**; g
*• *— ~ «••
IS ||
*„*£
is °"
$3*
^
X 0^"
>— ^
aa
X o"
S!2
ail
x
S£
h. ^
= ^
s°
i*
5S
»» «s
u
^
1 =
=c
Ul
s=
i
a =
5
i =
at
UJ
S*
- —
z t
*-
e
_o
p>jO'--»<^rwO>^
= g = ss = ss
ddddddcjd
£o — SJS^iS —
doddddod
* o * o S 2 ! 8 — t
*
iAr-rv^tO
«,_,_..„,
"~. . "~. ,
° 7 ° ° ° ° °
jn.-gjn-^^
oddo'dddo
1
<^o — oo — o —
doddddod
tsiO — OO — O~-
oaoooooo
S32SS-S2
ddoddddd
(Co^S^Soo
do'dddddd
S32SS-S2
ddoodddd
- =
41 >i >, u 5
z
o
1—
Q
u
g
2
z
o£
o
u.
i
J2
s
i
0
^~
i
Ul
to
s
0
^
_J
LU
§
„_
O
LU
i
ui
^
OQ
*~
| !
*?ii
Ss _
« o
::is
||i|
31
X.*»i S
"' — 5
ii||
xe>«
5;S .
s»^
u.*
"^ o
ll
* 0
11
ii"
51
£
z
% IK
v5
K
s
^x
«1
DC
bJ
ix
T
OX
o z
ac
1*
».—
<« -e
— V-
a.
e
e
eof*.aj"-sa — o»'-
S -"«-'" = «»
^^o-^ = »-
r °, -. ~. ". ». ™ *!
.y.M
|-8RSis = £
»»«>«s~o»-
^i^rMftjpieo"^
« ff. » Ki cy — »
V^dmjji^^ig
2 ^ kn v — ^
»*0-«)
S-S«- = gS
»*OK.*-'« — »
S" = "" = 22
S* = »°; = S2
»«rf>«r*ul,«1n
s»gssg~g
^-s
x»_»^.-22
— ' =
AJ >t >i i- S
i I !
I 1 I I
195
-------
i
i
g
o
t/1
g
CQ
_»-sg
i:il
*°=g
in II
8i3Jt a
XO'S
??
§<
X ^
li
3 O O O
LO IO » U1 rf> 1*1 »**
d d d d d d d
oooooooo
oooooooo
O O O O O O O
3>>£vcC*rrv^*r
oodddooo
oooooooo
4 |ddodddda{
dcsdcJdcJad
—-^if^ — i^^wi*i|
ddooddddh
ddddddoc
« = I
g
»•"?
III!
*°=g
iili
0% C6 9*
X 3
9> m r*.
O* •»• <-* O 03 trt O
*
a
I I I
5 - J^IiJlIJ,-
196
-------
£
CO
o
o
Ol
*O
LU
=-4
_,"*
C ^. — '-.^r(7ii-)O»
- »-»--i»-c^»
TJ oooooooo
81""
5S-S3-35
dddododd
dddddddd
£* » « - m (Sd O - »
So oooooooo
dddddddo*
< £
cs — 5 o — ~
o o o o o o
O fv CO rg ^ ^
-------
TABLE 70. NUMBER OF MODEL TO MEASUREMENT COMPARISONS
Station
Tower
Castle
Hockey
Munsey
Lambert
Johnson
Kents
Nashs
Unstable
476
73
502
506
398
452
499
481
Neutral
899
611
859
932
835
730
924
810
Stable
2865
1708
2834
2972
2544
2284
2880
2692
All
4240
2392
4195
4410
3777
3466
4303
3983
198
-------
j
H- 00
OO UJ
uJ ID
:n o-
^"* ^
3i o:
o
O UJ
z: t—
LJ O3
1/1 •— >
-J
O UJ
ac o
<0
t—
(/i z
UJ UJ
3= UJ
*g
O l/)
z
Z 0
o •—
t/1 1—
ii
§1 1 1
LU
— i ce
CO 3
ai
r~
01
3:
^
O
O
01
• — *
Q.
Q.
4j
VI
OJ
Ol
TJ
X C
a: <: r
o s-
LJ X
S II LL
o
1— XC
O D C
>a
m o ts
S-
Jj LU X
ST 21 LU
O O
LU LU "O
O C3 C
3
ct
LU
1—
ce
CJ
c
4-1 *O
-0 1_
i— 1-
U_ d)
'
I/I
ai
s
x a
3: <: t3
o ^~
LU X
s: ii uj
o
LU >SO
C3 D C
3
--0
^C o'S
i— i— i
LU LU X
io^
LU LU -C
O O C
3
ce
LU
t—
on
a:
c
, >>
•— co a)
4J Jrf VI
VI U C
CM
O
CM
CO
CO
CO
1^
cu
JQ
?9
-j
Ol
01
01
CO
CO
ao
en
CM
UD
s
(^
In
VD
CVJ
CO
CO
1
CM
ro
in
U3
Ol
ro
o
v>
c
f"
o
1-3
o
CO
, —
r—
O
CM
ai
0
^j.
CO
0
o
CVJ
1
CO
CM
O
O
CM
*'
ro
CM
VI
4-1
c
OJ
:>^
Ol
CM
CM
CVJ
r—
o
Ol
CO
^
CM
in
o
CM
0
CM
Ol
in
CM
CO
VO
CO
CO
VI
I/I
f)
z
199
-------
2. All concentrations greater than the mean plus one stan-
dard deviation for the hour were assumed to represent
impact of the plume from the power plant. These values
were therefore eliminated from further consideration in
the background computation.
3. The mean of the remaining concentrations for each hour
was computed and used as the background for that hour.
Hourly background concentrations ranged from 0 ppb to 74 ppb with a mean
of 3 ppb.
The model to measurement comparisons presented in Tables 64 to 71 show
the effects of adopting various changes to the basic flat terrain Gaussian
plume model as it is represented by the PTMTP model. The results show the
performance changes at different monitoring sites and for different stability
conditions.
The flat terrain model run showed very large root-mean-square errors
(RMSE) for unstable conditions (Table 64) with much smaller values for
neutral (Table 66) and stable conditions (Table 68). For stable conditions
the model showed linear correlation coefficients (Table 69) greater than
0.6 at the Munsey, Lambert and Nash's stations. Corresponding RMSEs for'
these three stations and for Kent's (Table 68) were the four lowest.
Particularly poor correlation between measured and model estimated values
occurred at Castlewood and Kent's for unstable conditions (Table 65), yet
Castlewood showed the best correlation for neutral stability (Table 67).
Looking at the correlation coefficients and the RMSE, the flat terrain
model performed surprisingly well against the measured Clinch River data.
However, the highest and second highest concentrations were overpredicted
at all stations except Kent's. Dilution of the plume due to enhanced
turbulence caused by the complex terrain may have produced actual con-
centrations much lower than the flat terrain model indicates. Also, the
uncertainty in location of the plume is likely to have been greater at
farther distances from the plant.
The plume height adjustments for elevated terrain resulted in very
high concentrations, particularly at the stations substantially elevated
above the plant. For example, use of the CRSTER plume height adjustment
produced the following frequencies of estimated concentrations greater
than or equal to 500 ppb:
Station Cases
Tower 64
Hockey 50
Munsey 50
Johnson 21
Lambert 12
Kents 8
Nashs 1
Castlewood 0
200
-------
Most of these high concentration estimates occurred with stable conditions.
An analysis of the distribution of the arithmetic error for the CRSTER
height adjustment run shows that for stable conditions the model overpre-
dicted by more than 500 ppb with the following frequencies:
Station Cases
Tower 27
Hockey 34
Munsey 12
Johnson 16
Lambert 9
Kents 7
Nashs 1
Castlewood 0
Slight improvement in the RMSE occurred with the NOAA height adjust-
ment was used, but this model allows direct impact of the plume centerline
on elevated terrain in stable conditions. Another slight improvement in
RMSE occurred at three stations with the use of the VALLEY plume height
adjustment which allows the plume to approach elevated terrain to within
10 m under stable conditions. Vast improvement in RMSE was shown with the
use of the ERT-LAPPES adjustment which maintains the plume above the
terrain at all times, but lowers the plume height above elevated terrain
by various fractional amounts for unstable, neutral, and stable conditions.
The best results were obtained with the GEOMET adjustment which makes no
reduction of the plume height for neutral and unstable conditions and also
does not allow the centerline to approach terrain to less than 40% of the
initial plume height for stable conditions. The ERT-LAPPES plume height
adjustment showed a significant improvement over the CRSTER, VALLEY, and
NOAA adjustments for stable conditions, but the GEOMET adjustment proved
to show an even greater reduction in the RMSE. The ERT-LAPPES height
adjustment was only a very slight improvement over the CRSTER formulation
for neutral and unstable conditions. The VALLEY, NOAA, and GEOMET plume
height adjustment results are equivalent for neutral and unstable conditions
since all three formulations do not alter the plume height for these con-
ditions.
The GEOMET and ERT-LAPPES plume height adjustments were each utilized
in conjunction with both the LAPPES and GEOMET enhancement factors. As
mentioned, without oy adjustments, the model runs involving the GEOMET
plume height adjustment showed lower RMSE values than those involving the
ERT-LAPPES adjustment. Very little difference is apparent in comparing
results of the LAPPES oy adjustment with the GEOMET ay modification over
all stabilities. However, for neutral conditions the GEOMET Oy adjust-
ment had the highest correlation coefficient for all stations except
Castlewood and showed significantly better correlations than with use of
the LAPPES ay adjustment at Munsey, Lambert and Nash°s.
201
-------
Two additional modeling techniques (the F.B. Smith oz adjustment and
the Pasquill initial dilution factor) did not produce any improvement over
the combination of the GEOMET plume height adjustment and ay modifications.,
However, the use of the power law to extrapolate the Upper Hockey wind
speed to plume height significantly improved both the correlations and
RMSE.
Table 71 compares results for the two model adjustments which gave
the best RMSE and correlation results with measured values, with the
original model and with the original incorporating CRSTER plume height
adjustment., The results are for the highest and second highest 1-hour S02
concentration at each of the 8 monitoring sites. It should be noted that
the values are not paired values occurring for the same hours. They are
the highest and second highest values over the data period.
Joint frequency tables of estimated concentrations versus measured
concentrations were produced for each station and for each stability
using the hourly output from one of the model runs which for convenience
has been abbreviated to SSCTM for Single Source Complex Terrain Model.
The model used for these comparisons includes the GEOMET H, GEOMET oy, and
wind extrapolation adjustments. Seven categories of concentration were
used:
Concentration
Category Range (ppb)
1 0-25
2 25-50
3 50-100
4 100-150
5 150-200
6 200-300
7 > 300
Tables 72 through 75 present these distributions for all stabilities
combined. The major deficiency of the SSCTM model is the underprediction
of S02 concentrations in the 25-50 ppb and 50-100 ppb ranges. This problem
was very evident at all stations except Lambert and Kent's. It appears
that overprediction was also rather common at several sites.
When results for all stabilities are reviewed (Table 62) the RMSE
is lowest at 5 stations (i.e., Tower, Castle, Munsey, Lambert and Johnson)
when the Pasquill ay is computed from a/\ (Table 64). The most substantial
improvement occurred with unstable conditions where, for example, the
RMSE at Castlewood was reduced by 19 ppb from the SSCTM run. Other signif-
icant RMSE reductions occurred at Tower, Munsey and Johnson., Table 71
indicates that the use of °^ in predicting ay also results in the best
prediction of the highest and second highest values. Very little change in
RMSE and correlations were observed between the SSCTM with the GEOMET cy
and with the use of ^ for neutral and stable conditions. Johnson showed
a significant improvement in RMSE for stable conditions and Lambert's
correlation improved significantly for stable conditions also.
202
-------
TABLE 72. JOINT FREQUENCIES OF SEVEN CLASSES OF SSCTM
CALCULATED AND OBSERVED CONCENTRATIONS OF S02 AT TOWER AND CASTLE SITES
Tower Observed
Calculated
1
2
3
4
5
6
7
1
3553
23
18
5
2
1
3
2
292
30
7
1
0
1
1
3
182
27
1
6
6
0
1
4
44
9
3
1
0
1
0
5
13
3
0
0
0
0
0
6
4
3
0
0
0
0
0
7
0
0
0
0
0
0
0
Calculated
1
2
3
4
5
6
7
i
1
2248
2
2
0
0
0
1
2
90
16
0
0
0
0
0
Castle
3
13
7
1
0
0
0
0
Observed
4
5
3
0
0
0
0
0
5
1
1
0
0
0
0
0
6
2
0
0
0
0
0
0
7
0
0
0
0
0
0
0
203
-------
TABLE 730 JOINT FREQUENCIES OF SEVEN CLASSES OF SSCTM
CALCULATED AND OBSERVED CONCENTRATIONS OF S02 AT LAMBERT AND JOHNSON SITES
Calculated
1
2
3
4
5
6
7
1
3651
51
15
5
1
1
0
2
35
9
1
1
0
0
0
Lambert
3
4
2
1
0
0
0
0
Observed
4
0
0
0
0
0
0
0
5
0
0
0
0
0
0
0
6
0
0
0
0
0
0
0
7
0
0
0
0
0
0
0
Calculated
1
2
3
4
5
6
7
, 1
3205
32
8
5
3
6
3
2
115
14
2
2
0
0
0
Johnson
3
56
1
1
1
0
0
0
Observed
4
8
0
1
0
0
0
0
5
2
1
0
0
0
0
0
6
0
0
0
0
0
0
0
7
0
0
0
0
0
0
0
204
-------
TABLE 74. JOINT FREQUENCIES OF SEVEN CLASSES OF SSCTM
CALCULATED AND OBSERVED CONCENTRATIONS OF S02 AT HOCKEY AND MUNSEY SITES
Calculated
1
2
3
4
5
6
7
1
3829
48
36
25
6
7
3
2
158
25
3
1
1
0
0
Hockey
3
33
5
0
0
0
0
0
Observed
4
6
3
4
0
0
0
0
5
2
0
0
0
0
0
0
6
2
0
0
0
0
0
0
7
1
0
0
0
0
0
0
Calculated
1
2
3
4
5
6
7
1
4168
29
24
11
6
4
4
2
98
26
2
0
0
0
0
Munsey
3
18
10
2
0
1
0
0
Observed
4
4
0
0
1
0
0
0
5
1
0
1
0
0
0
0
6
0
0
0
0
0
0
0
7
0
0
0
0
0
0
0
205
-------
TABLE 75. JOINT FREQUENCIES OF SEVEN CLASSES OF SSCTM
CALCULATED AND OBSERVED CONCENTRATIONS OF SOg AT KENTS AND NASHS SITES
Calculated
1
2
3
4
5
- 6
7
1
4131
86
20
2
0
1
0
2
35
12
1
0
0
0
0
Kents
3
10
2
1
0
0
0
0
Observed
4
1
0
0
0
0
0
0
5
0
0
0
0
0
0
0
6
1
0
0
0
0
0
0
7
0
0
0
0
0
0
0
Calculated
1
2
3
4
5
6
7
1
3853
40
8
1
0
1
0
2
43
16
2
0
0
0
0
Nashs
3
6
10
2
0
0
I)
0
Observed
4
1
0
0
0
I)
0
0
5
0
0
0
0
u
0
0
6
0
0
0
0
0
(i
0
7
0
0
0
0
0
0
0
206
-------
Cumulative frequency distributions of both the measured SO? concen-
trations and model predicted values were plotted on log probability paper
for comparison. The graphs for the eight stations, showing the distribution
results for the measured data along with those for three models*. appear in
Figures 109 to 116, Comparing the CRSTER, SSCTM, and Pasquill c?A_GEOMET
plume height results to the measured S02 data at Tower show that the CRSTER
model most accurately predicted the lower concentration values, while the
Pasquill a/\-GEOMET plume height model performed the best at higher concen-
trations. An examination of the frequency distribution graphs for the other
stations reveals a general pattern of the Pasquill ^ model best repre-
Scfnting the measured distribution at high concentrations, both the Pasquill
o/\ and SSCTM models performing the best at moderate concentration levels,
and only a small difference between models existing at low levels of SO^o
An exception to this pattern is the close match of the CRSTER distribution
to the measured distribucion at moderate SOg levels at Johnson.
The combination of Pasquill's method of estimating Oy from tf/
the GEOMET plume height adjustment produced the highest correlations
and che lowest RMSE when the results were compared with the measured S02
data. In addition, this model also performed the best in predicting the
highest and second highest concentrations. This model was therefore
judged to have validated the best against the Clinch River data. However,
for the purpose of common usage a model that requires a/\ as input is not
practical since ^ data are scarce and not normally observed at standard
weather stations. When a/\ data are not available, other means of specifying
dispersion parameters must be utilized., Therefore, the Single Source Com-
plex Terrain Model (the combination of the GEOMET plume height and ay
adjustments and wind extrapolation techniques), which performed almost as
well as the Pasquill aA model, against the Clinch River data should be
recommended for use. However, it has been demonstrated that this model
shows an under-predicticn of S02 concentrations in the 25-100 ppb range.
The remaining alternative for air quality prediction in complex terrain
is the use of a numerical ooundary layer flow model. Several such models
were reviewed; the feature of these models that is most readily adaptable
for inclusion in a Gaussian model is the use of a curvilinear plume trajectory.
It has also been demonstrated at Clinch River that a curvilinear trajectory
best matches the recorded i lurne position (at least within a few kilometers
or the Plant) in the frequent high turbulence conditions associated with
complex terrain (see Sections 7 and 8). For several hours during two days
occurring during the July, 1977 helicopter monitoring period, curvilinear
trajectories were proouced using the method described in Section 7e Hand
computations of S02 concentrations resulting from the use of curvilinear
trajectories were performed using the Gaussian plume equation,. Downwind
an.i cross/n'nd distances were determined by plotting the trajectories on
maps. The resulting SO^ predictions, along with the measured hourly average
and peak S02 concentrations, are presented in Table 76. Improvement over
the straight-line trajectory was found for hours 0900 EST of July 24 and
207
-------
999 9«J
:n
i
005 010? 05 1 2 5
} _ .
figure 109. Cumulative fr*.niency dlsti tbut Ion of SO^ tm Tciwe
208
-------
10 ' I Li -4L LL-LLULUL1 11._. _L J. JJ
001 005010? 05 I ? b
I 2O 30 4U 5O 6O /O W)
frobablllty of Concentration
-------
100 I — -
HJ OOI 005 0102 05 I 2
|0 20 W -10 SO 60 70 ftO
Probability of Concentration < Valti
90 95 98 99
figure 111. Cumulative frequency distribution
210
-------
loooo ?!??_ _ -,n~—,—,-,-.-,- * _ -^--, V-
001 005 0102051 2 5
10 ?0»40S06070BO
Pro iaMHty of ronrcntratlnn
-------
in 99 99 99 9 99 8 99 98 9& 91) MO /O bO bO 41) -IU .'0 10 '. / ! 0 ') 0 ,' U i 0 K> 0 ill
1000*"
H
;
k
4
2
o. a
8 <>
&
4
3
I
"'f
I
r
"T '
.
-
- - -
ir
i
1
i_
r
-
"T
- ••
-
- -
-- -
:"•{"[".: n T. r:
!
. , «r
li • me
< . ' CR
S3
'' • n
- '
..
-
h
- -f • j -:•-
|_ f - -;;•-;- -
• B Jl ej
pt«
liji 11 i(
- " 'J
. _ ...
- . . .
-
i'."L ' dj "1 J 1
"- " ' ' ' h
I- "
1 i " •
if -
! t
--
i , . /
/ /
1 if/'
i
- 1
j
'
.,
j
'
(i
>A .
s [
i i 7
•I/
i. i
h —
1
"~7\
j'
.
•"-]"
~~
j
I
_ |
|
.
1
/
//
rt
r/
r
..
ff
i -
i
i
A
*" //
/
^- -
.*'""
t
--
.-
i
1000J0050I02051 2 *> 10 ?0 304O5060/0 BO 90 9S «W 99 998999 9999
Probability of concentr at ton v" V« lue
Figure 113 CumuUtivt- f re juem y d lntr ibut Ian of SO fnr 1-in
212
-------
99
-------
10 001 OO50I02 05 I 2 6
20 10 4O 50 GO 70 BO 90 96
Probability of concentration < Value
US. Cumulative frequency distribution of S0? for Hocfc«y
98 9"J 998
-------
3 998 99
9'* 90 ao m eo w
i
f-
1
OOl 005 0 1 02 05 1 i
5 10 ?O TO 40 50 6Q 70 80 90
Probability of concentration
-------
and for 1200 EST on July 26, but no general trend toward improvement in
this small set of test cases was noted. Two alternatives may have potential
for the improvement of model estimates: (1) use the trajectories produced
for every two minutes to make a model calculation every two minutes, followed
by averaging of these values for the hour; or (2) the computation of the
average position of the plume during the hour using the individual two-
minute trajectories followed by a single model calculation.
TABLE 76. COMPARISON OF LINEAR AND TRANSPORT CURVILINEAR MODELS
Date
7/24/77
7/26/77
Time
0800
0900
1000
1100
1200
1300
1100
1200
S09
(ppb)
Linear
0
0
154
39
39
184
3
9
S02 (ppb)
curvil inear
0
9
225
1
57
390
4
54
Measured SO,,
(ppb) hourly)
average/peak
31/135
31/85
38/175
73/165
21/85
15/34
33/82
45/82
Stat ion
Tower
Tower
Tower
Tower
Tower
Tower
Johnson
Johnson
SUMMARY
The effectiveness of adjustments to the basic Gaussian plume model in
improving its capability to estimate ground level concentrations in complex
terrain was investigated. Thirteen modifications or combinations of modifi-
cations were introduced to the UNAMAP PTMTP program; including adjustments
to treat the height of the plume above terrain (6 variations including the
standard and 5 adjustments), the width (4 variations) and vertical thickness
(3 variations) of the plume, and wind speed (2 variations). Tne model
results were compared with measured values to determine correlation co-
efficients and root-mean-square-error (RMSE). Based on the model to
measurement statistics, the highest calculated and measured values and
plots of frequency distributions, the effectiveness of each model adjust-
ment was determined at each of the eight monitoring sites, for unstable,
neutral and stable conditions and for all stability conditions combined.
The greatest improvement over the unadjusted flat, terrain model occurred
with the use of a moderate (60 percent of terrain rise) plume height
reduction under stable conditions, plume width based on hourly wind
direction fluctuations (o^) as proposed by Pasquill (1976) and wind speed
216
-------
extrapolated from near surface observations to plume height by a stability
related power law function. Without wind direction fluctuation data the
next best adjustment consisted of the same plume height and wind speed
adjustments, but empirically increasing the Pasquill-Gifford plume width
function (ay) by a factor of 2 for moderately unstable conditions and by
a factor OT 3 for neutral and stable conditions. These were the two best
results of 13 adjustments evaluated. The evaluation presented here is by
no means exhaustive of ways of adjusting the Gaussian plume model. One
area of investigation which particularly needs to be pursued is the
discrepancy between the hours of measured and calculated high SC>2 con-
centrations. Although we have shown that the frequency of occurrence
of high values is reasonably well predicted over the course of a year,
using the relationships we tested, the calculated and measured high values
seldom occur during the same hour. There is a need to determine whether
this is due to inadequate model inputs, such as wind direction, or whether
there is a failure to represent the basic dispersion process; although both
types of errors may occur, further investigation of the high value cases
may identify which type of error is more improtant. A review of a few
selected cases in Section 8.0 suggests that both types of situations
occur. It appears that more detailed modeling approaches than the single
empirically adjusted plume are required for a significant number of cases.
217
-------
SECTION 11
APPLICATION OF DISPERSION MODEL AT ANOTHER LOCATION
Although the use of a model which employed measured ^ data validated
the best against the measured S0£ data at Clinch River, Virginia, such a
model cannot be used routinely due to a lack of standard CTA observations
at most locations. Therefore, for the purposes of testing a model at
another site, the SSCTM without the use of ^ values was chosen. This
model utilizes the GEOMET plume height adjustment, the GEOMET Oy adjust-
ment and a wind extrapolation technique.
The region chosen for the test was the area surrounding the Homer City,
Keystone, Seward and Conemaugh power plants in the vicinity of Johnstown,
Pennsylvania,, A network of 17 SO;? monitoring stations exists in the vicinity
of these plants. The major terrain features in the area are Chestnut Ridge
and Laurel Ridge, both oriented in a southwest to northeast direction,,
Chestnut Ridge lies between the Homer City and the Conemaugh and Seward
plants. Laurel Ridge is located to the southeast of all the plants (see
Figure 117). Laurel Ridge extends as much as 500 m above the Homer City
plant base* The terrain is generally hilly in all directions.
Table 77 summarizes the plant operating characteristics for each of
the four plants. Hourly values of S02 emission rate and stack gas exit
temperature and velocity were estimated from recorded hourly measurements
of generator load.
S02 monitoring data for the year 1975 were available. Meteorological
data for this same period were prepared for model input. Wind direction
and speed were observed from the top of a 91 m tower located 10 km south
of the Homer City plant. The elevation of the base of the tower is 244 m
above plant base, putting the wind observation at 335 m above the plant
base quite close to the normal plume height,, As a result, the measured
wind speed was taken as representative of wind speed at plume height and
the wind extrapolation technique of the SSCTM was not employed,,
Atmospheric stability was characterized as Pasquill Classes 1 to 6
using the vertical temperature gradient measured on the tower from 12 to
46 m. The following criteria were used (U.S. AEC, 1972):
Class (°C/100 m)
A < -1.9
B -1.9 to -1.7
C -1,7 to -1.5
D -1.5 to -0.5
E -Oo5 to +1.5
F > +1.5
218
-------
«lltl--
f
r
V
NL
NO
N
NC
-\
«ND
N7
N3
Figure 117. Chestnut Ridge Monitoring Network.
V
a
S*
S
219
-------
TABLE 77 PLANT OPERATING CHARACTERISTICS
Stack Height
(m)
Diameter
(m)
Exit Velocity
Homer
1
242.6
7.3
23.7
City
2
242.6
7.3
23.7
Conemaugh
1 & 2 identical
304.8
8.3
23.1
Keystone
1 & 2 identical
244. 1
9.1
21.3
Seward
1 2
68.7
4.9
13.7
65.2
2.0
23.4
(m/sec) 100% Load
Exit Temperature
(°K) 100% Load
413 413
Typical SO2 Emission 2043 2043
Rate 100% Load
(g/sec)
Generator Capacity
AW 100% Load
600 600
405
3499
890
404
3300
850
417 421
135 90
220
-------
Mixing heights were estimated from twice daily rawinsonde observations
from Pittsburgh by the DeNardo and MacFarland Weather Service. These
estimates were then interpolated to give hourly values. The interpolation
is based on the following concept of diurnal mixing height variations.
The mixing height rises from its early morning minimum to an afternoon
maximum. The afternoon maximum remains constant but weakens during the
evening. A new mixing height is formed under the weakening old height and
becomes established by midnight at a height equal to the height indicated
by the next morning's sounding,, The interpolation procedure assigns the
morning mixing height to the hours from 0000 EST to 0600 EST. The after-
noon mixing height is assigned to 1400 through 2300 EST. The morning
and afternoon mixing heights are linearly interpolated in time between
0600 and 1400 EST. Plume heights were estimated using the UNAMAP sub-
routine BEH072.
The same method of estimating background concentrations was employed
for the Pennsylvania network as was used at Clinch River. The mean SO?
background was much higher in the Chestnut Ridge area (17 ppb) compared
with Clinch River (3 ppb).
A set of 8760 hourly concentration estimates of S02 were computed
for each of the 17 monitoring stations. Each hourly concentration includes
a contribution from each stack of each of the four plants lying within the
Chestnut Ridge Monitoring Network. In addition a background concentration
was added to the calculated plant contributions.
The results are presented in Table 78 as an analysis of the predictive
ability of the model for the highest and second highest 1-hour average
values over all stabilities,, The highest and second highest values were
oyerpredicted at 13 stations and underpredicted at 4 stations. Overpre-
dictions were particularly evident at the two ridgetop stations (N3 and NG,
Figure 117). Over the 17 stations the average model/observed ratio for
the highest values was Io23. It is concluded from the limited comparisons
made with the central Pennsylvania data, that the model calculations com-
pare more favorably with observed values than was the case with the Clinch
River observations,, The major shortcoming appears to be a tendency to
overestimate concentrations on ridges. This result is found in the Clinch
River data as well. It appears that more meteorologically comprehensive
models which can more realistically represent flow over ridges will be
required to obtain improved model performance.
221
-------
TABLE 78. MODEL TO MEASUREMENT COMPARISONS OF 1-HOUR AVERAGES
FOR CHESTNUT RIDGE - ALL CONDITIONS DURING 1975 (STATION LOCATIONS
ARE SHOWN IN FIGURE 117)
Station
NA
.\B
NC
ND
NE
NF
NG
Nl
NL
NM
NN
NO
N2
N3
\6
,\7
N8
Highest
Observed
787
607
524
543
612
724
859
987
638
508
838
535
1181
1498
1064
1812
2439
( 9/m3)
SSCTM
1082
428
572
602
803
1015
1845
724
649
777
637
636
1772
2585
1380
1844
2318
Second Highest
Observed
683
607
439
529
484
710
649
798
494
407
630
489
1048
1330
896
1601
2437
( 9/m3)
SSCTM
1080
405
485
560
771
911
1380
676
592
667
544
607
1700
2528
1252
1772
2296
222
-------
REFERENCES
Briggs, G.A., 1969. Plume Rise. USAEC Critical Review Series, TID-25075,
82 pp.
Briggs, G.A., 1974. "Plume Rise from Multiple Sources," ATDL Contribution
No. 91, Atmospheric Turbulence and Diffusion Laboratory, Oak Ridge,
Tennessee.
Briggs, G.A., 1975. "Plume Rise Predictions," Proceedings of the AMS Work-
shop on Meteorology and Environmental Assessment," American Meteorological
Society, 296 pp.
Burt, E.W., 1977. "VALLEY Model User's Guide," EPA Office of Air Quality
Planning and Standards, Research Triangle Park, North Carolina, EPA-450/
2-77-018.
Egan, B.A., 1975. "Turbulent Diffusion in Complex Terrain," Proceedings of
the AMS Workshop on Meteorology and Environmental Assessment, American
Meteorological Society, 296 pp.
Hanna, S.R., G.A. Briggs, J. Deardorff, B.A. Egan, F.A. Gifford and F. Pasquill,
1977. "AMS Workshop on Stability Classification Schemes and Sigma Curves —
Summary of Recommendations," Bulletin of the American Meteorological
Society. 58, 1305-1309.
Holzworth, G.C., 1978. "Estimated Effective Chimney Heights Based on Rawin-
sonde Observations at Selected Sites in the United States," Journal of
Applied Meteorology 17:153-60.
Irwin, J.S., 1979. "Estimating Plume Dispersion - A Recommended Generalized
Scheme," Proceedings of the Fourth Symposium on Turbulence, Diffusion and
Air Pollution of the American Meteorological Society, Reno, Nevada.
Koen, R.C., 1977. "A Study of Ambient S0? from Selected Non-Urban Pennsylvania
Sources," GEOMET Report No. EF-583, Final Report to Pennsylvania Department
of Natural Resources under Contract No. ME-75913, GEOMET, Incorporated,
Gaithersburg, Maryland.
Koch, R.C., W.G. Biggs, P.H. Hwang, I. Leichter, K.E. Pickering, E.R. Sawdey
and J.L. Swift, 1978. "Power Plant Stack Plumes in Complex Terrain, an
Appraisal of Current Research," EPA-600/7-77-020, Contract No. 68-02-2260,
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
223
-------
Koch, R.C., 1978. "Use of Monitoring Data to Modify M^oro'ogical Disper-
sion Models for Point Sources," Presented at th'< 84th Nati-.Aui Meeting of
American Institute of Chemical Engineers, Atlanta, Georgia.
Koch, R.C., W.6. Biggs, D. Cover, H. Rector, P.P. Stenberg and K.E. Pickering,
1979. "Power Plant Stack Plumes in Complex Terrain, Description of an
Aerometric Field Study," EPA-600/7-79-01Qa, Contract No. 68-02-2260,
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina,
Maxwell, C., 1979.. Personal Communication.
Pasqui11, F., 1976. Atmospheric Dispersion Parameters in Gaussian Plume
Modeling, Part II. Possible Requirements for Change in the Turner Workbook
Values. EPA Report Number EPA-600/4-7S-030b, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina,,
Slowik, A.A., J.M. Austin, and G.N. Pica, 1977. "Plume Dispersion Modeling
in Complex Terrain Under Stable Atmospheric Conditions," Presented at the
'.Oth Annual Meeting of the Air Pollution Control Association, Toronto,
Ontario.
Smith, F.B., 1972. "A Scheme for Estimating the. Vertical Dispersion of a
Plume from a Source Near Grojnd Level," Proceedings of the Third Meeting
of the Expert Fane! on Air Pollution Modeling, NATO/CCMS.
Thompson, R., 1979. "Dispei sion of Sulfur Du-x'ide from the Clinch River
Power Plant - A Wind Tunnel Study," Preliminary manuscript, U.S. Environ-
mental Protection Agency, Research Triangle Park, North Carolina.
Timer, D.B., 1970, "Workbook for \tmospheric Dispersion Estimates," U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina.
U.S. Atomic Energy Commissior, 1972. Safety Guide 23, Onsite Meteorological
Programs, February 17, 1972.
U.S. Environmental Protection Agency, 1977. "User's Manual for the Single
Source (CRSTER) Model," EPA Office of Air Quality Planning and Standards,
Research Triangle- Park, North Carolina, EPA-450/2-77-013.
Var, der Hoven, I., G.J. Ferber, P.A. hamphrey, 3.C. Holzwo^th, J.L. Heffter,
-nd K,F. Quiring, 1972. ''Southwest rnergy Study ueport of the Meteorology
lnork Group," NOAA.
Whcley, H., 1974. "The Derivation of Plume Dispersion Parameters: from
Measured Three-Dimensional Data," Atmospheric Environiaent 8:281-90.
224
-------
APPENDIX A
FREQUENCY DISTRIBUTIONS AND MEAN CONCENTRATIONS
OF MOBILE VAN DATA
Two-way classifications of pollutant concentrations, giving the mean
concentration and frequency of occurrence of each table entry, are presented
for the stationary samples taken by the mobile van. The tables are in the
order listed below. All distances from the plant are in kilometers and all
elevations are in feet above mean sea level.
INDEX OF TABLES
Page Two-Way Classification
226 Distance versus elevation for SO
227 Hour versus elevation for S02
228 Distance versus direction for SO
2
229 Distance versus elevation for NO
230 Hour versus elevation for NO
231 Distance versus direction for NO
232 Distance versus elevation for NOp
233 Hour versus elevation for NOp
234 Distance versus direction for NO-
235 Distance versus elevation for NO
236 Hour versus elevation for NO
237 Distance versus direction for NO
238 Distance versus elevation for 0^
239 Hour versus eTevation for Oo
240 Distance versus direction for 0^
225
-------
l/l -~
J\
«r n
226
-------
•M II — II
X
-4 a.
-^ ~4 -DO
->* It -N II
227
-------
r— .n ON
o -> -H •
228
-------
3s M -«
-M -^ O
J1 3D
•n M
,O -N
O
^ II
-* -H 7i r>
> ~i 00
229
-------
I r>| f—• r*
-** It -^
It -N It -4 tl
O
-------
ft It ~«i II
231
-------
•n u -* ii
"M It —* II
;N Ii -H M
232
-------
•3 O
3> 0
1 O
O O
O
O
o o
o
o
:TI -N
o
3D
A -O
"•*
•n
O 3
o
o
o o
o
o
i> o
Cs
o
"4 II -« II
r>
i
JD
•*>
X>
*t
-^
-
)l
2
7>
,0
•N
-*
*n
H
z
•D
*t
-,
(I
^
-,
•n J3
j^
~* II
Z
-s
» -N
O
-N n
z
«. -*
-. j* -«
"N -n
*r H — n
Z 2
^
-0
*^
""
M
2
233
-------
o ~<
,0 *
>* It -• II
O »-* O £> XP-N — f) ~-
-N./1 ^t:
*n ii •*» ii
234
-------
It "1 II -H
-* -*'**
II ~« II
-4 II
z
I -. J> -VJ
"S It <-• II
235
-------
-•* It -^ II
-X It *4 II -t tl
-M II -« II ~H II *•* II —t II
236
-------
O
o
it -* it
O
X
-4-4 O O
237
-------
- yi O •*
>i II
Z
M Cb
^4 a,
z
238
-------
"O II -N II
-• t) -H II -* U
233
239
-------
M ->4 tl
—t SO
II •** tl
240
-------
APPENDIX B
FREQUENCY DISTRIBUTIONS
OF METEOROLOGICAL DATA
This appendix contains 2-way joint frequency distributions of meteoro-
logical variables observed at fixed monitoring sites. The following codes
are used to represent meteorological measurements:
WD1 - Wind direction at 10 meters, deg,,
WD3 - Wind direction at 30 meters, deg.
WS1 - Wind speed at 10 meters, m/sec
WS3 - Wind speed at 30 meters, m/sec
SGI - Standard deviation of wind direction at 10 meters, deg.
SG3 - Standard deivation of wind direction at 30 meters, deg.
DTI - Vertical temperature change from 0.5 to 4 m, °C
DT3 - Vertical temperature change from 0.5 to 30 m, °C
HOUR - Hour of the day representing ending of 1-hour average
INDEX OF TABLES
Page
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
Station
Parameters
Hockey
Hockey
Hockey
Hockey
Lambert
Lambert
Tower
Tower
Tower
Tower
Munsey
Munsey
Castle
Castle
Nashs
Nashs
Johnson
Johnson
Hockey
Hockey
Lambert
Tower
WD1
WD1
WD3
WD3
WD1
WD1
WD1
WD1
WD3
WD3
WD1
WD1
WD1
WD1
WD1
WD1
WD1
WD1
WS1
WS3
WS1
WS1
(10 to
(190 to
(10 to
(190 to
(10 to
(190 to
(10 to
(190 to
(10 to
(190 to
(10 to
(190 to
(10 to
(190 to
(10 to
(190 to
(10 to
(190 to
(1 to 1
(1 to 1
(1 to 1
(1 to 1
180) vs. HOUR
360) vs. HOUR
180) vs. HOUR
360) vs. HOUR
180) vs. HOUR
360) vs. HOUR
180) vs. HOUR
360) vs. HOUR
180) vs. HOUR
360) vs. HOUR
180) vs. HOUR
360) vs. HOUR
180) vs. HOUR
360) vs. HOUR
180) vs. HOUR
360) vs. HOUR
180) vs. HOUR
360) vs. HOUR
9) vs. HOUR
9) vs. HOUR
9) vs. HOUR
9) vs. HOUR
241
-------
Page
Station
Parameters
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
Tower
Munsey
Castle
Nashs
Johnson
Hockey
Hockey
Lambert
Tower
Tower
Munsey
Castle
Nashs
Johnson
Tower
Tower
Tower
Tower
Tower
Tower
Tower
WS3 (1 to
WS1 (1 to
WS1 (1
(1
(1
WS1
WS1
SGI
S63
SGI
SGI
SG3
SGI
SGI
SGI
SGI
DTI
DT3
DTI
SET
SE3
SGI
to
to
to
10 to
10 to
10 to
(10 to
(10 to
(10 to
(10 to
(10 to
(10 to
(-2 to
(-2.1
(-2 to
(3 to
(3 to
(10 to
vs
vs
vs
19) vs.
19) vs.
19) vs.
19) vs.
19) vs.
120
120
120
120) vs
120) vs
120) vs
120) vs
120) vs
120) vs
1.6) vs
to 3.3)
1.6) vs
SE3 (3 to
48) vs.
48) vs.
120) vs
48) vs.
HOUR
HOUR
HOUR
HOUR
HOUR
. HOUR
. HOUR
. HOUR
. HOUR
. HOUR
. HOUR
. HOUR
. HOUR
. HOUR
. HOUR
vs. HOUR
. DT3
HOUR
HOUR
. DTI
SG3
242
-------
O U1 —i
c— • 01 >r rs ,-<
CM •- »- —
v:
U
O *-« (N »- fN •»• —
a C7 i* •-
o o c c c
ococcc rsc-tr
243
-------
•-« e o
*M_ ^H^.fNjmOMf^*-'-'
or-o
— CT C tf r- C , r- O
a kC C f\ »- — C
c r- cv c CT
coco
r«. cc ? " r- *c
— oc c.
244
-------
« ••-» in o
-------
^ o- r-
o u*> «^ o
IT * o
o — — . — «-
— rvrvcv-.cs — — ^.
s:
UIL «
C -I
z C
.
r-O.
•
CT-CCNCC ^
r*. rv :>
246
-------
«• f »- f" C
c "• cc c
O1
-------
1C
c
o
0
— z
c <-<
*-i at tr.
248
-------
•- *~ 0
C — \£
^ — C
C
-------
— -- r-
c •-
c
(N £S. — — —•
— v- fM
250
-------
rv c ••-
c.-»-fvc'**rNir
T- r^ r-
^, ^. If.
C — CM
251
-------
v- — a
*. c- —
252
-------
ccr-c
r*«-oo~ ora r*-
— fir,
f- a:
y or
fN ~- CN
^ \c ac
C O t-5
O O O C
C-rTTC T-COOC-V
r^crn o •— ^v •*) •* ia
253
-------
f- —
o «
~ -* O Q
~-.-rvr\ — rsrN.-^.^-
«? r* rv rv
C- 'J, -si
254
-------
C ^
2 C
f On o. *-• —> *- ^
*
x a.
c «- «• o ir
255
-------
X
X
o
u
O *H •-
LO
5 ~
j-
tr. >•
< (j
r, r
o c. o o
256
-------
*•* O —•
«- .- r< rj r*. rx
*"
IT *~ C
c c c a
257
-------
— fN O C
tt*\£ r-o^cNC*xin«tt r- fs.
Uu
O
C
z
u:
c cr vc j- "• f
C ITp —
vc-irir «•
— CO"
r~ a i—
c- — —
— IT cr o c.
258
-------
C "Z
O -i
c o
—ex
u:
e — c
— — -HO — CrC— '
259
-------
O X
j;
m u;
c
H IT
O
«^ 2
O >H
O O
• «r
o a;
_-.- — c- x
— ir x a- c
260
-------
(-.
O
oooooooooe ooooooooooooooo
C- COOOOC OCO
oooooocooc cocooocoooocooc
O
w
C
occoccccoc
OCCOCCCCCC
~" — oooceooeorv
— CCCCOOCOOOC-H
a x
u.
c*-ococoocc coc ccr^ccc coc
— — c —
c cooco*-»—
— c(%c; r* c »-i»-»—
c
IT O
IT i-- f , — -
— cc(y
.- >- c
CT "^ X
261
-------
CO
o
ooooocooooooocoo oooooeooo
oooocco oooceoocc- c-ooooococ
— — OOOCOOOOCOCO-^OOOOCOC
ooocoooooooooooo ccooccooc
OO—'^OCOOOOCOCCCCN CCC O C — O O U"
c
X
1C
C (J
X 12
fs *i
•er
C
'"S V.
3:
c c
» T— *£ ^- a fv^»— cxiri'*'fs^-vcfvoectc c *^ a" --o-cr cv
IT
a
o
c c
c x
IT tr
C = QC
262
-------
ooooooooooooooooooocooooo
oeooooococooocccooooooooo
oooooooooooooooo coooocooo
ocoooooooc oooooo oocoooooc
ccoooooooc ccccoc. cococc ccc-
cocococecc occcc-ooooccocoo
c »- o —' o c c ococc-cocsococccc c
cc-coccccooocccrsa-
CCOOOCOCCC COCC^CrC —t O O »- C C C
CCCCCCiTU."
— orv ^-
c o r-
263
-------
o
m
c
•* .- OO>
ocoooocooGocccoocoooooooo
cococ occcoc-ocoooeooocccoc
OOOOOOCOOOCOCOOOOOOGOCOOC
ooooooooooocoococcoccoooo
ooc e o c c o o c ococoooccc ooccc
ooc oocoooo — c — ece o c c o c coco.
oeoooococc
^-»- — occooocco
ccoococcc — •«-«-CGC''r>»-*-oc;cGcc eta
UJ -
u tJ
z c
coocooco —
c,coc*oc —
cocooo —
c-co— '
It
a
x c
u*
X >•
A
C
e
— u:
c —
— *~ — — O- j.
c-
-------
h-
coooooooccoooooooocooeooo
ccoocccooeoooocoooc oocooc
oocooooooccooooooc ocoooo
cocoocoooococoooooocooooo
cococcoccoccococccccooccc
o u
O T
^ — cccc ccccir
c
H
COOOOCCCOO — — OC*-*-C^-C:OCC:CCiri
c occ — c-co — ccof*-c-r- — rvcc ccooc
r^ 3E
«, O
C — c O O
\t if v a
<* — c c —
O C C LT CN
265
-------
-J
o
ooocoooccoooooooooooooooo
ocoococcocoocoooooococcoo
oooc ooooooocc o o c. o o o ooocc c
ococoocococcc oooocccooooo
ccoc occooocoocoocoocococc
c u
I >—'
a- i
u
OCCCCOOC-OCOCCO—lOCCC OC
ceoocccoccccccooccccc c coo
CCOOOOCCCCCCCCCCSCOCOCOOCCO'
O
u
ecocccccccoocccooooc-cooco
c
X
cococococococcoococ c oc cccii
OCCOOCCOCCC-OOCCCCOCOCCCCOb.
cococcccoccoc — rvotr cccc
— e o e c o c
c
f-
c «- o — c
(N*ii£r-crr^i--r-r^-i«-«-o — ^»^rr
or y.
Or — X >
X <-
y;
266
-------
o oooooooooooooooooeoooooo
cococccooccroccocooococcc
o coooooocooooooooococoooo
o coooooooococooccococooc
o cocooocoococoooccc cooccc
c oooooc-oce coc crcccococcc c c
c c
O fa.
coccococcccoocccoirccccccc
c cccoccccccc-occocoeccc c
c ccQCOcceoeecccccoccceccc *t
ocoooccccococ ococccococc-cu
ccocccccccocreccc ccccc
ecu.
c ooooooooooooccocooccoocrc:
c ccococococr-cc — CCCCOCCTCC —
^- c o c-
c —
-------
« *C
fr- (C
— o- r.» in
o* cr a r* •*
o
ir:
ooooooooooooooco cooocoooc
ooooocoocoocc oooccooooocc
oooocoooooooo ooc coooocooo
ooooooooooocooco coooccooo
ococs-ccrooocoocooocoooccoec-
ccccc ooocococ
c u:
X
cccccococccccccc cccccococ a
«- 3:
1C O
cccooocccc ccc c ccccoccoccc «
X
n:
u.
ooocoooccoc cc-cocr ccoccocc-cc
a-
<
If.
oooc oc-cooc cc*-occ>ccccococ — u,
r
•a
=>
C-OCOCCiC — CCCrrcfSOCCCOOceCC^
—
IT,
S
C
c-ccoocoe»-O'*xfvc**-ir»-oc ccoc~ccj^
*-
O — oc-co^-cc — rst«»-\c arsCr^^'rv — — — «-cCtr
v
C'fNOJr*ifNr^cp«Cv«^ac'3^Lr*3 c*~<*''ir»-»-«— --u"
fN fS Pn t^ •- (N p^^
c
— ^••e^^iru'xrr-fv^- f
tt
-tf
T
tf
268
-------
o o o o o c o o c o ooooooooo ooooco
OOO OOOOOOOOOOOO OOOOOO
ooooooo ooooooocoooooooooo
ooooooocoo ooooeooc-o o o c o o o
cocooooocooooococcc oococc
ococ c-oc ocoocooc-coooccccoc
oooc ccc oocoooooccoooc oocc-
CCOOOCC COCOOOCCCCOC OCC
cooccocoooocrooocrwcc oec-ocfv
COOOOOOOOOOCOO^H—«occ occ
COOCCOOC1 — C-OC^-CXO—
-------
o
2
tL
O
-------
o
fr-
X — C O C C O
O C O O C COCCOCCOOCOC.CCOC COO
ccccoo
oococc c c c c
o-^cococ c
or — >*i CD
271
-------
cccc-ccc-cccc-cccc
«- c: c
c e tr
272
-------
U> J
Z C
*J ~
r* £ ^ X1 y
—. x
— j:
X O^ C —i ^4
273
-------
tl.
a
to
o
t- K
f Z
c *-*
o c
• *X
O LL
at cr- e?
COOOCCC
CC.OCCCOCCC CCCCrC
ecu:
c
CB
X sC V
C IT IT ITI
c: CM •-
r*>£ — ^r*>£ — XC C IT ^ XC
o a a:
T — j* ij-
^* rv **> tr j^ £^3
274
-------
o
(-
o c
» a
c jt
WD
J
OVV^V^^^ ^ *J- ^ V^^t'^^V^^'^O1*!'^ «9 ^
t-
c
o •
• »- O C *~ ^-^.T-^-C — OOC CC.C OOC CrC OCCr-O
c r^
r* •-
c s
c»-^-fNcVi'^-rvr^. (MX
o I*
C' r- *
o W-
CT — *-T
3*
C
c c
r- ^
»f\e-vTvCLr^troxcififfl^)'-vT '"vcv^c-tfCco-Tr^r. ? •»• r*
i/- r* C—
c c-
— Z
C X
c
• ^•nrf-'xxac—^ffCf*"Ciru*ira't/^otNiru^'w'a -sCO
T cr r-
rv rv rv
275
-------
,J
>
«£
(X
u
X
c o
C X
r«
*. b:
*-« ^
o »-.
o o
CCCC
— COOC CCCCCOCCCCOCrCC — C
— 3.
o o
c — c
— orcCT-ri cr
— rv «• a- "-. c
-r rv j-
occ- or^o^oc^ crooc-
OT-ccrc-ocotr
276
-------
C — O — CCC —
rr, a i
«- o- r- i :*•
r- CT j~ c
c o — — -n —
277
-------
u
f-
2
Q
Z
U,
X
e o
—i a
c o
v: s
c o
2 O
£-<
O
O •
• o ccoccc ocoocccoococ-oococc-e
rv
z
O ~" " — X
*-• s-
or
O IiC
o *" — m a
o — «
o in
a " •»• ~J
«r
5"
O
cr ^T :'
L1:
C Cr
r- •£
c
«C IT
IT. •£• ~
C;
T CT- 3. — •—
f J~ £> J-
O rs c
278
-------
coooooccco cocccoooooooooc
ocococcoec *-< o c o c
COOOOOOCOC OC--OOO OCOOO —
CCCO— i —
— OOOCOCCCOC — — OlT
\o c
*4 X
-- — cccoc — ccccoccc
U
c z
O t-<
o u- — IT
ccrc.c;CC— — IT
c C. C
coo
~ c r- c o c rr
279
-------
-- — — OC-«-*OOOOOOOOCrslf»>
ct
It
— — OOOOOC
o
2
— ^OOOOOCOOIN—.
u:
cn
c* -* —
c r- * --
r* ^- *
fr- C
o a
at r- \c ^c in a a
c *» o- x ac
O
X
C C — C
C C O O
coo
a c c- c c
r- — coccca i
a:
b?
*
O
cccccc-cc
occrc~ co-oocoooooe ococ oc
— '%m«*ir*cr-icy'C — r*i **^ ^ »r* xi r^ cr (T o —
280
-------
ooo
ooooc ooocooooococ
U-
o
y,
c.
X
VC f C
c o
c- E- K
O G Z
00*-
OJ \£) O
• * «
C< CN la.1
I I X
^OCGCOOGCOCOOGCOOC"-'
OOOCOOCOO — OOOO«OOOOC
-------
> l/>
« J
Z 6-"
t~-i
O
u, -0000000000000000000000000
n CD
^
05
o
2 C
U" • CCOCOOOCCOCOCOOOCOCOOOCC'C:
U"»
x -*
c u
05 X O
•e •ooooocococcooocccococccoc
u.1 rs
£X «•
A
o
t- « o
u .ccccoooocc — o — occcocococco.
— 2 O-
o »-* ^
o o
• « c:
c u: c •
i »ooccccce^occ^o — ccccocccc^-a
vC 9
^ **>
r
o j c ss
ZO •CCCCCCCC"-r^rvfs,^-«cc^C — CCC — C'-r.Kt
* u <*" — x
o: f fr.
4-
a
— x c ^ ,^_h
« a o r* cc
m
or v — ..f^»-.-._,.__fv,-. w ^. »jc^
O r>; rs. tu
O . \r.
X
O C
CN ^
u;. c
u.
t-
t- £t C
f- ~-
IC
^^ W
a c c
c cj
6- 2
282
-------
OOOOOOOOOOOO^tOOOOOOOCCO
ccca
^- c: ^ c o c .-
COO — O — — CC-CCCC
— c o c c o •-
r<* * et
rs c —
ID iT —
o a c.
— 00 —
283
-------
—• o z
O O t-
C CX O
—< — C O
C et
COCCGCGGCGCGCGCCCOGGC'-
Z 3Z O
rf « C)
a. cc.
O *- C
w. c a:
2 Z
c occccoa*«
— 3: T
p_ H
a a:
f>. C G O G C. C. cs- U U
(f OL a.
** c ooc
a: u; J.
284
-------
oococooooo ooo
G C C
QE> O X
-« u.
O-"OOOO«->C>GO o o
o o c — e
*- c m o
IT, rv p-
C O r-
CC *- G
-co
p- — »-
— CC ID •" ^f Ct
,c -H o ir>
o o a
285
-------
APPENDIX C
FREQUENCY DISTRIBUTIONS OF S02 AND NOX VS.
OBSERVED WINDS
This appendix presents 2-way joint frequencies of occurrence of classes
of S02 or NOX concentrations with wind directions observed at the Hockey
site, 30 meter level (WD3)0
INDEX OF TABLES
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
Station
Hockey
Hockey
Lambert
Lambert
Tower
Tower
Munsey
Munsey
Castle
Castle
Nashs
Nashs
Johnson
Johnson
Kents
Kents
Hockey
Hockey
Castle
Castle
Nashs
Nashs
Tower
Tower
Munsey
Munsey
Kents
Kents
Parameters
WD3 (10 to 180) vs. S02
WD3 (190 to 360) vs. S02
WD3 (10 to 180) vs. S02
WD3 (190 to 360) vs. S02
WD3 (10 to 180) vs. S02
WD3 (190 to 360) vs. S02
WD3 (10 to 180) vs. S02
WD3 (190 to 360) vs. S02
WD3 (10 to 180) vs. S02
WD3 (190 to 360) vs. S02
WD3 (10 to 180) vs. S02
WD3 (190 to 360) vs. S02
WD3 (190 to 360) vs» SO?
WD3 '
(10 to 180) vs. S02
WD3 (10 to 180) vs. SO?
WD3 (190 to 360) vs "
WD3 (10 to 180) vs.
WD3 (190 to 360) vs
WD3 (10 to 180) vs
WD3 (190 to 360) vs.
WD3 (10 to 180) vs.
WD3 (190 to 360) vs,
WD3 (10 to 180) vs.
WD3 (190 to 360) vs,
WD3 (10 to 180) vs.
WD3 (190 to 360) vs,
S02
NOX
NOX
NOX
NOX
NOX
NOX
NOv
NfL
NOXX
NO
WD3 (10 to 180) vs. NOX
WD3 (190 to 360) vs. NOX
286
-------
o o o
cox
a c
— ir £*;
c o
s- t- w
u
C O 2
O C —
COO
mecooooooooooooo rn
LTttooococcccccoo'^i
-HCCCOC OCCCOC -r-*
O»C'— CC — COOOOO
CO CC
ooc eccccc
fS fs f^ •** IT,
-"CCCCCOX
c e c cr
c. c o ~ o
a- c IT o c:
287
-------
—• w o* in •«— —4 —•
oc-eofNooooo
m CM o* vt •—
—• o o o o o o
C O I
VD O
m »r u:
C O Z
coo
X O IT
(»•) fT, ^H —
rv — — c coo
C C rr,
— V «• (S (Ni (N-^1 — C^-»-COC^-
O O
*c c
rr-O--rMC —.«-.fs — C-CCCO
C O C
s to a
D O' m
— r- *-> tr o
-- e CCO
k£ — —-
— c—'
C'-'^CCCOCC
H X
C: cc
rr CN
c -c- c o
c c
c c
r*. c —
c c c
o c- o o o
288
-------
*- — T C C
c c *-
ceo
c o o
c o c
C C X
(T C
— u- u.
C C C C
c c c. c- :r
C O C1
& is. a.
— — coccoeocccecr-oe
X K JT
iTCCOCCOCOC-CCCC;
If ±3 ~
oc oocococoocoo
ccccocca
rr, CO
c c
c c
'^ — COC OOC
O. C C O
oc f^ o o
289
-------
— C: C:
O O C *n
ceo
c e- c %c
c c- o c
c c
{-, f_ I
I
e c
C C (
c c
C — rr
C C O *•-
r-«-~oec; — c c
— e ?»>
o o o e c w
o e
o c-
C: O I
X 10 i
o c — c c
X -t
t- f-
u: E*'
o c;
X ff.
c c c o
o o o c. o o^
^•c<-ir — ccccocr*.
3E l/x
o c
c — (*•
c —- c .- c. c-
— c or r*
o o c c
— cr at
•*?•— — OOOOCO'V
O CD —
•^•^COOOCC-O —
— CO
c o o
C C1
•*! G
290
-------
O cr «-
*~ M, w c-(NiCCCCCCCCO
C C CCCCCCC
TT — OCC. — C C —
f- E- tr.
u
c c ^
~ c •—
c c c
• « «c
ecu
— — c c — c — cccccc
crccccc
cc
ce c
— IT
— CCCCOCCCC
U3 I*.
OC
a c:
CT E r
o c
2 ,
cccocccccccc
ace — cccccc ccc c
_ -••• c c e c ccc
c c
c c
CT fN — — CCCCCC CC
(V ,- C »« — COOCCO
291
-------
— .- c
c c c
C C ~
c c ?
^ c. »-
c o c
z
u. t- -
cr tt *s
Z E C
c o c
* ic a
tr or
U- u.
a a
< *t
j >i
1C K
u: b.
— »
> >
c c
*- o
c o
it C
•— a
cr b.
>C — —
fN> — —«
— tC i
CN u; i
292
-------
cc--cccccc.cccc.rv
D-.-C c-cc^cccccccc
c c a:
cc c
•- IT b
C — C C CCOCCCCCCT
C C F1
C C i-
C, C C:
O C O O C C —
c .- c c c
c c c-
Zi Z C
i -a t,
c- tx
ceo-ccc — ccccir «
ccccccccn-
o a a
C-IT--C.CCC ccccccccir
c rr ~
cccoerc ccoerc cccr
> :>
c r
r- C
c •— c cccccccc^ec
--ccc cccccoccsc
O X *
x — c^ ~~ — c >
rs
C C
c- c
IT CtrClTCC'CCOCC'C-OC'ff
f-
tr- c rv
293
-------
— C r* \f u- c:
r. a ir>
ccccco
.- — c c c — cc
C C
• I
c c
•- c -c
c c
*- *- v
c
c c *
•- c ^~
c c. c
I • <
C C it
~- *- c, ccc
t' C -
SEC
w fx, 3
CCC
•« K ft.
a a.
u; ti.
t5C
a a
oc K
b." b
r- r*. ~ c c
o O
•X IT.
C C
fs. — .-
c —
o
C C
C C
a. c
>- a:
cr u.
-^ — — C
X X
fe- F-
u: m
if tft
u- i»'
j ^
C C
3C «
IT O if C
294
-------
c- c c c
— cf.cc-ccc-sccrc
ccococccccccir
----cccc-ccccccecc
a- — —
i*. u. —
C U J
r E c
cr*-cc:c ccc-ccr-
ts c r^
CCCCCCCCC
c c
o c
— IT
f. to
X 3:
IT *r •- — c — ccceoccoor-
tr
u
f-
i/
I ;
ir
>
f
o
•y
>• r-
u. rs,
tJ r-
C
r c
t-
ti.
o ^
a
z r*
C vC
>— r-
^
DC O
H- Jl
X U-
f-
ir
>•
c-
LX
rr
O
C
a
c
,-.'
r-
c
c
\c
c
t
m O
if
^
C
•^ c
^
c
(
Cr
pr
C
C
fN
0
-------
C -- fM C.
— — c
— CCClT
cccccc c ca
^ — c
cere cccocac
CO — C C C CO.
C C c c
«- if U.
fl:
£
C C
{- t- V.
c
C C 7
-- O H-
c c o
I * tf
ecu,
a x
(*• C
at •-
— — c
— cc«---cccr-
C C C C C r-
u: s
c ^
* C
f CN 3:
C C C
u- U.
istj
a: a
oco «- ceo
CTCO.--O. — cccocooca
>• >•
C O
«-
-------
• me occocc — c
e: »-
o o o o c i
CCCCCCCOCOOr"1
c eccococctf
C O Z
C C t->
ceo
*r — co-^ec-cc-c;
c o o c
• CO CCCOOOCO
c c —
rv —
.
c e
* •*
*? >
o o o o o o —>
o o o o
irO mCiTOCOOGC-C
^-^-r\fv'w"rfirvcoC'OU"
297
-------
or e IT
e ccocccc
w.
c
c c c
cox
— w rs -H p- c ccccccrooco
e — f^
c — *- c c c
c c c c —
c c
(- f- K
U
C O Z
— c —
c c c
.-• cc — —
»-»-cxrs»-»-cccc:c.eir
c c
c c
I X
H f-
a or
ii Lt
u o
a a.
— •£. rv —• — —
— ccccco.
to t
<"", — C CCCCC
c- r
0 C
cr c
-------
tr, r~ r~
C c c —
c c c c c •£
rv /*• •-
^vCCr^CvCC — lT(s''*'CCC —
C C 2
.. c —
c c c:
o c c a
c^-coccvrcc
Z £ "
>
fw
C"
*
C
c
w
K
•e
X
»-
or
U.
c
a:
«2
tf.
t
i«2
«e
>•
fv
rf-,
K
C
U" If r~
c ^- -• c o c
IT — C
c c
— C
c c
c c
c c c o o
c c
o- "
O O C: C C
O C C C
C LT O C
299
-------
» i- JT a m «-> c c c c
O C *-
a «-
c c: c c o c
c c, c
OCX
«r c
^- IT i».
, rr ~~ CL r- C CC CC
C ^
c c- c c c- —
c e c <
C C
«- f- Li
e c *~
ceo
C C c- c or
3T
!»• u: -
C c j
2T E C
.cwiTir — cc. — c
o r- — —
r^ p" c: c •- Cr c
— c c c o
c c c e c —
C C
a c
w~ IT
IT
f- *-
r+ O. 3
C O C
3t V, (X
- c c c o c -
C C C C C v
C C
c c
o c
it fuCp^mO-^r^ —
tr t/:
w: it
Cu £*-
hJ ^3
•^ iT "" •—
J >D
<
ooceo^oc-o ">ec.or>c-oi
..*••*•...»*. • • o
inciAOt/^ooc oo
-------
— occcc c. ~ e c e c- C- c \r
C CCCCCOCOCO
C C fs
cccccccc
C C
f- f- 1C
e
c c ^
C C H-
c c c
c c
O C.
a c
ccccccccctrc
«3 4
I X
»- r-
— CvCCCCCCCCCCCClf
c- u
a. a
ccocccccccccecc
U. !>'
s r=
Jj *:
•a *r
CCCCCCOCCOC-C
c r
r- c
cccc ccccccocooa
c c
c c
O O
rvCO»-C — CCC
c c c
occcoccc
c c c e —
tr; tr;
IT tr
U U
301
-------
o **• ~-
v£ C IT —t
c c o c
O C O OC
*- O Cr O C C O
Cf^CM — OOC-OC.C
O C C C
irxvc.rs.cvfs — cccoccc
cot'
C C X
f. ir b.
C C
O C Z
— C —
C. C C
CvCtfOOCCC CCC
irv^fsccc-ccoc
c c
o c
2 Z C
a < I)
a er
\cccccc — coe
fe 2
X X
C C C
S « X
C~OC COOC^OO
. .
o c
ix a.
< a
u. r.
^ r*
tj r^
" — f^^c^ce
«n — O O C
c^ o
X l£
(N IT »»1
r- — o .- c
r\ r* rs »•
* c
—' a:
f- r\ ir
— CX C C O C
— o o o
o c- c
o c cr c
c o
r** C
302
-------
A *c
-------
m r- >£> o if.
fN (T> CV C
O "T V
— o o o o «-i o — —>
ooor*
OCX
*c c
(MC O C «\
c o
f- *-> U3
C
O O Z
— o »-*
COD
• . «*
o c u:
O O O
o c
* c .
X E
*n X 36
COG
r- — —
— Ct fit
— « «
.J _;
— — — c
a z
^ H
u: r>
C*- —
^ & '? C O
o ~ o c c
304
-------
• O — VC — f. »~
C — •-
• •- *f o-c~ c r^ c —
c o
c c
c c-
c c
cr C
*- IT
c tr
-, .j
< fS
EL a:
c c c
a
-------
-- — C
— *- _ _ C C _. _ ^
e. e
c c
u.
e
— — c c
— o c c
— — c - c —
ir — c
c «?
fN,
Z ff-
e* c e
fS — O
c c
*- c ^
sC W «3
u: u:
— ~* c c
m x
O 0
» e
tr «
j m
306
-------
CM »- C
cccctr
•C LT J
c c- x
cr c
— IT U
— arc
—• (X a
*- «c »•
cr oocn^
r- rr yt.
c r
^ £
« fN
f C> fS
o c c —
— w-i fN
c o
1 Z
o o
307
-------
CCCC:
*- «- c c
c c- e
c c x
C C 2"
C C C
— C <- C
— »*• ir
2 s cr
•ff « t,
(r or
^ T Z
cr a
u. u.
J- ^ C
fi Cfc CL
-- *I <
IT- C ^
C C C
C C
— c
C £.
C1 r- -
^ j- -r-
- If V
308
-------
r-trct^ — -«r >r ir r* »- c
O X «-
c ,-, _ ccocc-CCa
CCS
X C
.- IT u.
cor:
c c •—
c c c-
z s C
0 Kf ( .
a c.
oca: cr-rvf — a
2 2
C"*!*!
-- 3: x
— f- t-
cr a
i*. U^
r-Oti
c a. a.
r- GL — C —
ccc — cccccr^
*c c c
o c
o c
o c
— c c c c cooc«
cr c c c c ir
-------
— c. c
ecu
C C X
-f C
" IT u.
— o —
c ac o
O C Z
— C «-
c c c
rr X "?
C C C
31 z a
v o c
c c
— c
c c
in CM —
310
-------
— CCOCC
C C C C C
cc — e coerce-
— — c — c c cccc
c e
c c
Ct C
— u*
n- X >
C C C
3: * u
c a »-
ire — — o
c e c c
o
f-
ir
•c
a
b.
Cu
Li
>
f
C
3-
r-
*— —.c-cccocc
IT — «-
c c
c c
c c
~ 3"
•— ac
— CCC
— c c — c cccccr-
X 3:
H I-
•-• — c ccc cccccr —
C C
o e
311
-------
— c c e
C C C
fv „- _ ,-
c c ^
•^ C >-
c c c
C — C C
ff X 2
C C C
3 z a
*- F- t-
c c- e c ir
— C C C C C CT
r^ CN cr c c c
IT C C
^T — C
~ O
ff C
f> El
a u.
tr, tr
CO IT
a.' i
or O
ir ^
O tP
312
-------
—• c c~
c e
-------
c c. ,
c c 2:
c c: c:
c c c
c c
(- f- tr
c
c c -z.
.— c- •—
Ct C C
o c
o c
X X
f- t-
or a
u: u.
a. a.
c r.
o o
C C O
c c
— 00
o- ••*
•z. x
* «
X T
— o c
314
-------
TECHNICAL REPORT DATA
.' read liudi'i'Ctioi , ')h the re,ersc before, completing)
. 1 REPOFT. !\O
3. RECIPIEN PS ACCESSI Or* NO.
4 T ITLF ANT SUB". ITLc
POWER PL.ANT STACK PLUMES IN COMPLEX TEKRAII,
' Data Analysis and Characterization of Plume Behavior
5. REPORT DAT"
January 1980
6 PERTORMING ORGANIZATION CODE
! 7266
1 AUTHOR!?!
Kenneth E. Pickering, Robert H. Woodward,
and Robert C. Koch
U PERFORMING ORGANIZATION REPORT NO.
|9. PERFORMiM ORGANIZATION NAME AND ADDRl.SS
• GEOMF.T incorpo--,ted
s 15 Firs\ifie"o Roaj
! GaithersDirg, .-.av
1 1. PROGRAM ELEMENT NO.
IriC&32B EA-020 (FY-78)
11. roivJTRACT/GRANT KO.
68-02-^60
i
, 12. SPONSORING AGE'\C" NAME AND AD! HESS
] Environmental S-iences Research Lahor.itcry - RIP, 'C _ 1 "nj± 6/1/78 - 8/17/79
I Office of Resea-ch and Development -~:r- -------
U.S. Environment;! Protection Agen
13 TYPL OF RL 'ORT AND PERIOD COVERED
_ ! "pji 6/1/78 - 8/'
4. ".P', ^SOR.NC AGENCY CODE
5 Research Iriancs: ?a"k, North Car jli.-.a 27 AH
I
l p''' '
;1S SUPPLFMFNT^PY NOTES
ie.ABSTRACT
tiioul
nn in?
months in the vii.i.iity of the
1 complex terrain of southwestern
''.oncenti'ations at eight fixed
sjnfil Vc.r;aticns. Ridge sites
-Jt" !-..efined diurnal pattern;
ui, ;,, hypothesized to be caused by
.'-ogicel measurements made from
the plume. P'i -j^e heights calcu-
duhs ca]:ulateo f'"om the standard
Aeromet.ric'data were collected LU "tv 16
•coal-fired Clinch River Power P"1'.^- To;ct^J ' . -h
jVirgim'a. Statistical 'aaiysts o> Su9 !V, a/d iv\
{monitoring sites /i ea.ed significant'"r''jr,-;a"
^showed maximum concentrations ?L rr'yl"' r;' w'
jvalley s'.tes snowjd a well-deified • . - ,..."n\
jplume fumigation. Cross-sectien ,'oTUh ft a,.'1 < -.te>.
!a helicopter were analyzed to shew ihe u mens'ion; o.
jlated from wind dm' temperature pru''i'e ano plume -
deviation of wind direction produce)' improve tents "ver standard estimates. The
'influence of terrain features on tbe . .iape ar,J oath of the plume is clearly shown
when assessing the helicopter data enu selected ^f-se studies of ground monitoring
data. Parameters for a Gaussian piuu model were modified in a stepwise manner by
using sitt specifi^ meteorological data. The study showed Jiat the standard flat
tterrain mode; and modifications frequontly ^sed to represent complex terrain influences
Scan be improved by judicious Uoe ' si: p -specific rlata. The Gaussian plume model can
^provide useful estimates of maximum concentrations, but >t cannot generate reliable
Incur-by-!1 cur concerur (:iors du? ,o ivfli!ences 'mposeo hy a complex terrain setting.
•The datj obtained tor tin C"!inr.h_River s ce ^hou1^ b>^ useful in further diagnosis and
imoijBl_±£f±ing_n.f pi ump beh^'' i or i n comrjl^. te_'ra i n.
,17.
DESCRIPTORS
KEY WORDS A.ND DOCUTVu
Jb.h'E
\N, LvSi£
"" Air po'l. Lien
* Meteoroiogy
* Plumes
, Electric power olarits
J* Atmospheric difrusion
J* Terrain
1 Field tests
+ SulFur dioxide
* Nitrogen oxides
* Data processing
* Mathematical
models
." DSTRIBUTION STATEMENT
RELEASE TO PUBLIC
J i=ND, D TFRMS
rqinia
(This Report)
TFIED
t J Ms page)
IFIFD
C. COSATI 1
13B
04 B
21B
10B
04A
08F
14B
icld/Gioup (
07B ~|
09B !
12A i
I
21 NO. OF PAGES
333
22. PRICE
EPA Form 22J.O-1 (9-73)
3!5
-------
(t
c
CO
o
o
H
O
O
00
O
O
o
oo
3 °
J o O
QJ o ~
^-^ B
3 s o
If
U »
tt) 3" tB
io3
§8 I
TJ
o
00"
III
5 <5 <:
3
Q) 3
- c
O 3
i «
& i
M 2
m J> ~o m ~n "^
Tito g 3 tt> o
5 Q.
------- |