Dispersion Modeling of Sulfur Dioxide
in the Vicinity of
Two Hawaiian Power Plants
by
Northrop Services, Inc.
Las Vegas, Nevada 89119
Contract No. 68-03-2591
Environmental Monitoring and Support Laboratory
Office of Research and Development
U. S. Environmental Protection Agency
Las Vegas, Nevada 89114
-------
021707
DISPERSION MODELING OF SULFUR DIOXIDE
IN THE VICINITY OF
TWO HAWAIIAN POWER PLANTS
by
Robert C. Koch and Douglas J. Pelton
GEOMET, Incorporated
Gaithersburg, Maryland 20760
in conjunction with
George w. Siple
Northrop Services, Incorporated
Las Vegas, Nevada 89119
Prepared under
Subcontract No. 9460-1100Y
for
Northrop Services, Incorporated
under
EPA Contract No. 68-03-2591
Project Officer
Jeffrey van Ee
Monitoring Operations Division
Environmental MonitoriJig and Support Laboratory
Las Vegas, Nevada 89114
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and
Support Laboratory-Las Vegas, 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.
ii
-------
FOREWORD
Protection of the environment requires effective regulatory actions
which are based on sound technical and scientific information. This informa-
tion must include the quantitative description and linking of pollutant
sources, transport mechanisms, interactions, and resulting effects on man and
his environment. Because of the complexities involved, assessment of specific
pollutants in the environment requires a total systems approach which tran-
scends the media of air, water, and land. The Environmental Monitoring and
Support Laboratory-Las Vegas contributes to the formation and enhancement of
a sound monitoring data base for exposure assessment through programs designed
to:
o develop and optimize systems and strategies for moni-
toring pollutants and their impact on the environment,
o demonstrate new monitoring systems and technologies by
applying them to fulfill special monitoring needs of the
Agency's operating programs.
This report is one of six reports published separately by the U.S.
Environmental Protection Agency (EPA). The series of reports was prepared
by Northrop Services, Incorporated, and the Environmental Monitoring and
Support Laboratory-Las Vegas for EPA Region IX. The reports will serve as a
Technical Support Document to be used in the development of a sulfur dioxide
State Implementation Plan for the State of Hawaii. This document provides
the factual basis for the development of regulations for the State of Hawaii
to avoid exceeding National Ambient Air Quality Standards.
Two oil-fired power plants were identified as major sources of sulfur
dioxide within the State. These souces were monitored to:
o determine if violations of air quality standards occurred
in the vicinity of the plant,
o obtain data for incorporation into a mathematical
dispersion model for the plants.
The data and findings presented in the Technical Support Document should
interest a wide variety of parties. Regulatory agencies and electric utili-
ties directly involved with the effects of a State Implementation Plan have an
immediate and obvious interest in the data presented in this report. The
public has an interest in how the data and findings contained in this report
will be used to develop the State Implementation Plan. The implementation
of the Plan will directly affect the electric companies and indirectly affect
111
-------
the public. Scientists who are interested in the methods used to assess the
impact of a power plant on the air quality in the area of that plant may be
interested in the extensive data base presented in this report.
The 1-year study of both power plants was conducted by the Monitoring
Operations Division of the Environmental Monitoring and Support Laboratory-
Las Vegas.
George B. Morgan
Director
Environmental Monitoring and Support Laboratory
Las Vegas
iv
-------
CONTENTS
Page
Foreword iii
Figures vi
Tables vii
Abbreviations and Symbols ix
1. Introduction 1
2. Conclusions and Recommendations 3
3. Model Input Data 5
Kahului meteorological data 5
Kahului emission data 9
Kahe Point meteorological data 11
Kahe Point emission data 19
4. Air Quality Data 21
Kahului 21
Kahe Point 26
5. Model Selection and Modification . 31
Kahului modifications " • • 31
Verification of model performance at Kahului 36
Kahe Point modifications 39
Verification of model performance at Kahe Point ... 41
6. Modeling Results and Evaluation of Strategies 46
Kahului results 46
Strategies evaluated at Kahului 49
Kahe Point results 51
Strategies evaluated at Kahe Point 61
7. Feasibility of a Supplementary Control System 70
Kahului plant 70
Kahe Point plant 72
References 75
Bibliography 76
-------
FIGURES
Number f-22!
1 Comparison of mixing heights derived from corresponding
Kahului and Hilo radiosondes 7
2 Distribution of GQ observations at sites 2 and 3 near
the Kahului plant 10
3 Distribution of wind direction, Kahe Point monitoring
stations 14
4 Frequency distribution of OQ, Kahe Point monitoring
stations ^
5 Comparison of mixing heights derived from Hilo and Lihue
radiosondes with Wheeler AFB radiosondes 18
6 Comparisons among a relationships 35
7 Receptor locations near the Kahului power plant 47
8 Receptor site locations - Kahe Point 53
-------
TABLES
Number
:
Page
1 Meteorological Data Source for Kahului Plant Study 8
2 Correlation Coefficients of o /x to Observed 03 9
3 Comparison of Stability Classes Estimated Three Different
Ways 12
4 Wind Direction Observed in Tethersonde and Monitoring
Stations During 2-Week Period in January 1978 16
5 Cumulative Frequency Distributions of Observed Hourly
SO2 Concentrations at Kahului 22
6 Cumulative Frequency Distributions of Observed 3-Hour
SO2 Concentrations at Kahului • • • • 23
7 Cumulative Frequency Distributions of Observed 24-Hour
SO2 Concentrations at Kahului 24
8 Average Observed S02 Concentration by Wind Direction
in 10-Degree Sectors 25
9 Cumulative Frequency Distributions of Observed Hourly
SO2 Concentrations at Kahe Point 27
10 Cumulative Frequency Distributions of Observed 3-Hour
SO2 Concentrations at Kahe Point 29
11 Cumulative Frequency Distributions of Observed 24-Hour
S02 Concentrations at Kahe Point 30
12 Comparison of Calculated SO Concentrations Using Model
Adjustments with Corresponding Measured Values at
Kahului During May 14-20, 1977 32
13 Fitted Constants for the McElroy-Pooler Diffusion Param-
eters Based on Turner (1970) Stability Classifications . . 34
14 Comparison of Statistics for Hourly Observed and Calcu-
lated Concentrations from 7 Days in the Intensive
Period 37
15 Comparison of Statistics for Hourly Observed and Calcu-
lated Concentrations from the 6-Month Monitoring
Period 37
VI1
-------
Number
TABLES (continued)
Page
16 Comparison of Statistics for 24-Hour Average Observed
and Calculated Concentrations for 6-Month
Monitoring Period ; 37
17 Mean and Maximum Hourly SO2 Concentrations Computed Using
Various Model Parameters at Three Kahe Monitoring Sites for
the 4-Month Monitoring Period 41
18 Hourly Concentrations - Calculated Base Case 42
19 Ratios of Measured-to-Calculated SO2 Concentrations (Ug/m3)
at the Kahe Point Site
20 Receptor Locations Which May Exceed Air Quality Standards
Without Use of Control Strategies 48
21 Comparison of Calculated Concentration Statistics from
6 Months and 12 Months of Data 50
22 Mean and Standard Deviation of OQ Observed at Monitoring
Station 3 for Corresponding Pasquill Stability Class
Observations at the Kahului Airport 50
23 Expected Concentrations Using Abatement Strategies at the
Kahului Plant 52
24 3-Hour Average Concentrations - Calculated Base Case 55
25 24-Hour Average Concentrations - Calculated Base Case 58
26 3-Hour Average Concentrations, Calculated Stacks 2.5 x
Building Height 62
27 24-Hour Average Concentrations, Calculated Stacks
2.5 x Building Height 65
28 Expected Concentrations Using Abatement Strategies at
Kahe Point Plant 68
29 Range of Meteorological Variables Observed on Days with
Highest Measured Concentrations of SO2 72
30 Range of Wind Observations at Station 2 on Days with
Highest SO Concentrations at Station 2 73
V10.1
-------
ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
AFB
BEH072
BTU
C/M
COSPEC
EPA
FPC
km
kw
LST
m
m/sec
min
MW
yg/m3(ug/cu.m)
NE
NOAA
NNW
NRC
P-G
PTMTP
SE
SSE
Std. Dev.
STL
TECO
UNAMAP
—Air Force Base
—subroutine in PTMTP program
—British thermal units
—ratio of calculated to measured
—correlation spectrometer
—U.S. Environmental Protection Agency
—Federal Power Commission
—kilometer(s)
—kilowatt(s)
—Local Standard Time
—meter(s)
—meter(s) per second
—minute(s)
—megawatt(s)
—microgram(s) per cubic meter
—northeast
—National Oceanic and Atmospheric Administration
—north-northwest
—Nuclear Regulatory Commission
—dispersion data collected by Pasquill and Gifford
—program in UNAMAP library; a Gaussian model
—southeast
—south-southeast
—standard deviation
—dispersion data collected in St. Louis, Missouri, by McElroy
and Pooler
—Thermo Electron Company model 43 sulfur dioxide analyzer
—User's Network for Applied Modeling of Air Pollution
SYMBOLS
A
a
B
b
BH
BW
c
c
X
D
-extremely unstable atmospheric conditions
-crosswind constant dependent on stability
-moderately unstable atmospheric conditions
-vertical constant dependent on stability
-building height, meters
-building width, meters
-slightly unstable atmospheric conditions
-regression coefficient relating Oy to OQ
-ground-level concentration, micrograms per cubic meter
-neutral atmospheric conditions
ix
-------
SYMBOLS (continued)
d —regression coefficient relating oy to oe
d —inside stack diameter, meters
o —degrees of temperature
E —slightly stable atmospheric conditions
F —moderately stable atmospheric conditions
f —coefficient in plume rise equation
fl —coefficient in plume rise equation
f (X) —coefficient relating oy to OQ
HE —total plume height, meters
h —contribution to plume height for downwash effects, meters
hp —plume height when BH is greater than or equal to By, meters
h —stack height
HX —plume height at downwind distance of receptor, meters
K —variable
IB --minimum value between BH and By, meters
N —integer value
p —crosswind constant dependent on stability
% —percent
n —Pi = 3.14159
Q —source emission rate, micrograms per second
q —vertical constant dependent on stability
SO? —sulfur dioxide
Ofl —standard deviation of the horizontal wind angle, degrees
0B —standard deviation of the concentration distribution in the
y horizontal direction, meters
a —standard deviation of the concentration distribution in the
vertical direction, meters
u —wind speed at stack height, meters per second
v —stack gas exit velocity, meters per second
XS —relationship "ay = xo0"
x —distance downwind of source, meters
x —"times" (multiplied by)
x-. —virtual distance, meters
-------
SECTION 1
INTRODUCTION
Regulations for air quality in the State of Hawaii do not cite emission
standards but rely on air quality monitoring data to determine that air
quality standards are met. In order to develop an air quality contro]
strategy for sulfur dioxide (802) that is based on emission limitations, it is
desirable to test how effectively alternative strategies will control present
sources. A first step in making this assessment is to measure the present air
quality around significant sources. The expense of maintaining an array of
monitoring sites for a sufficiently long period of time to assure that air
quality standards are being met can be significantly reduced by combining
monitoring with mathematical modeling. Application of mathematical dispersion
modeling was selected as a practical alternative to determine where and at
what levels ambient concentrations of sulfur dioxide may be expected. More
specifically, two electricity-generating stations were selected to be studied
by both a short-term monitoring-data collection period and dispersion modeling.
The two stations were Maui Electric Company power plant in Kahului, Maui, and
the Hawaiian Electric Company power plant at Kane Point, Oahu.
The Kahului plant is located very near the shoreline of Kahului Bay. The
waters of the bay cover the area north of the plant (except for approximately
100 meters of land between the plant and the shoreline). Populated areas lie
principally south and west of the plant.
The Kane Point plant is located on the southwest shore of the island of
Oahu. The shoreline in the vicinity of the power plant follows a line
oriented approximately north-northwest to south-southeast. Consequently, most
of the area to the west of the plant and downwind of the prevailing northeast
(NE) trade winds (with the exception of a narrow strip of land approximately
200 meters wide) is the Pacific Ocean. Along that land strip between the
plant and the ocean there are a highway and public beach. The terrain rises
very steeply in all directions around the plant except on the shore side. The
244-meter (m) elevation contour is generally within 1.3 kilometers (km) of the
plant in the NE quadrant, while ridges in the southeast quadrant are generally
about 90 m within 0.7 km of the plant.
The monitoring stations were set up in the vicinity of each plant to
monitor air quality and meteorological conditions. Near the Kahului plant
location the three stations were placed downwind of the power plant along a
line oriented with the prevailing onshore winds. Station 1 was located
approximately 0.4 km from the power plant; Station 2 was approximately 1.0 km,
and Station 3 was approximately 1.8 km from the power plant. The azimuth
-------
angles from the power plant to Stations, 1, 2, and 3 were approximately 228
degrees, 209 degrees, and 215 degrees, respectively. Ground elevations from
the power plant to the monitoring stations change less than 10 meters. The
stations were operated for a period of 6 months, from February to July 1977.
Near the Kahe Point plant site the three air quality monitoring stations
were set up and operated for a 4-month period, from October 1977 to January
1978. The monitoring stations were located along the shore as follows: Site
1, 0.5 km on a bearing of 185 degrees from the plant; Site 2, approximately
0.15 km on a bearing of 240 degrees from the plant; Site 3, approximately
1.6 km on a bearing of 315 degrees from the plant.
The Kahului power plant operates using four separate oil-fired boilers
each served by a separate stack to generate a maximum nameplate capacity of
34.93 megawatts (MW). The Kahe Point plant operates five oil-fired boilers,
each with a separate stack, which combine to give a total nameplate generating
capacity of 474.93 MW, according to Federal Power Commissions (FPC) figures.
Utilizing an existing dispersion model, Geomet developed parameters for
each plant location which could be used to represent the atmospheric trans-
port and diffusion process at each site. For a number of reasons, including
the simple terrain influences and the ability to access a satisfactory number
of desirable observation locations, the observations and the model treatment
of the Kahului plant plume are more complete and more satisfactory than the
treatment of the Kahe Point plant plume. For use in the vicinity of both
plants, the model was modified by systematically introducing changes based on
plume characteristics observed during a brief intensive monitoring period of
1 week at Kahului and 2 weeks at Kahe Point. However, the final agreement
between model estimates and observations at the three monitoring stations was
better at the Kahului location. More extensive model development work based
on more detailed modeling of the complex terrain features at the Kahe Point
location may produce better modeling results in the future. Only very simple
modeling concepts have been applied in this study. The data collected may be
useful in testing new and more complex modeling concepts which have not yet
become accepted techniques.
The selected model as modified for each location was used to estimate
the air quality effects which could be expected if any of several emissions
control strategies for sulfur dioxide was adopted. The strategies evaluated
included reduction of fuel sulfur content and utilization of an S02 scrubber,
increasing the stack height to 2.5 times the building height or higher, and
using a supplementary control system by which the plant load would be reduced
during selected unfavorable meteorological conditions.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
Site-specific parameters for use in a Gaussian plume model were developed
to represent the transport and dispersion of the exhaust gases from the power
plants located at Kahului and at Kane Point. The model parameters were
developed using wind observations from 10-meter towers (including direction,
speed, and standard deviation of wind direction), temperatures from a nearby
monitoring site, either solar radiation measurements or airport cloud obser-
vations, and mixing heights derived from twice-daily radiosonde observations.
The relationship of meteorological parameters to the model dispersion
parameter oy was determined from a combination of mobile van measurements of
the plume (either from ground measurements in the plume or indirect overhead
measurements using a correlation spectrometer) and comparisons of model
calculations (using the hypothesized a parameter) with fixed station SO^
measurements. Previously published oz values were adjusted for site
specificity by comparisons of model calculations with fixed-station S02
measurements.
The model as modified for application to the Kahului power plant loca-
tion showed good agreement with concentrations measured at three sites over a
6-month monitoring period. The mean values and the characteristics of the
frequency distribution of calculated and measured hourly and 24-hour S02
concentrations were compared. The mean ratio of calculated values to
measured values averaged over the three sites was close to unity for all
comparison statistics. The mean ratios of calculated-to-measured values were
1.0 and 0.8 for the highest hourly and 24-hour concentrations; 1.3 and 1.1
for the 90th percentile hourly and 24-hour concentrations; and 1.4 for the
mean concentrations. The ratios at individual monitoring sites varied from
0.6 to 2.1 for these same statistics.
Model results for alternative control strategies applied to the Kahului
power plant showed that controlling the sulfur content of fuel not to exceed
0.5 percent and raising stack heights to 2.5 times the building height are
not adequate control strategies. However, the model results showed that
requiring the use of a 90 percent efficient SO2 scrubber, raising the stacks
to a height of 91 meters, or raising the stacks to 2.5 times the building
height and limiting the sulfur content of fuel to 0.5 percent or less are
adequate control strategies which meet Federal ambient air quality standards
for SO2 in the vicinity of the plant.
The model as applied to the Kahe Point power plant location showed less
agreement with concentrations measured at three sites over a 4-month monitor-
ing period than were obtained for the Kahului location. The mean ratios of
-------
the calculated values to measured values averaged over the three sites were
15 14 and 2.0 for the mean, second high 24-hour, and second high 3-hour
values'. The ratios at individual monitoring sites for these same statistics
varied from 0.8 to 3.0. The model, in general, overestimates the measured
concentrations. One of the major sources of uncertainty in the model
estimates is the highly turbulent flow which occurs on the lee side of the
Waianae Mountain Range during prevailing northeasterly trade wind flow. The
plant plume is influenced by the enhanced turbulence and wind deflection
caused by the nearly rugged terrain with both onshore and offshore flow.
Model results for alternative control strategies applied to the Kahe
Point power plant showed that requiring the sulfur content of fuel to be
limited to 0.5 percent, reducing emissions by 70 percent, and raising the
stack heights to 115 meters are not by themselves adequate control strategies.
The model estimates of SO2 air quality levels are above Federal levels.
Because of the uncertainty of the model results and the demonstrated tendency
to overestimate, this conclusion might be altered by more detailed study.
Control strategies based on 90 percent reduction of present emissions or
based on limiting the sulfur content of fuel to 1.0 percent and requiring the
stack to be 2.5 times the building height are shown by the model results to
be adequate to meet Federal standards for SO2-
It is recommended that the influence of complex terrain on the dispersion
of plumes from sources on the lee side of major terrain obstacles be studied
in more detail. The results presented in this report for the Kahe Point
plant reflect a relatively large degree of uncertainty regarding the model
estimates. There are many potential plant sites in Hawaii which are on the
lee side of major terrain obstacles during the prevailing trade wind influence.
The nature of this effect on plume dispersion needs to be more carefully
delineated by observation and modeling studies. Data obtained at the Kahe
Point site provide a useful starting point for such a study.
-------
SECTION 3
MODEL INPUT DATA
Meteorological and S02 emission data were obtained for each plant site.
The emission data obtained for each plant are broken down by stack and include
hourly variations. The meteorological data vary in detail and duration of
observation at each site.
The Kahului plant is within 3 km of an airport where a record of mete-
orological observations, representative of winds in the vicinity of the plant
location over several years, is available. At Kahe Point, the plant is in a
bowl-shaped canyon which faces west and backs up against the southern end of
the Waianae Mountain Range. The only meteorological data which can be
considered reliable and representative of wind flow at the plant location were
collected during the 4-month U.S. Environmental Protection Agency (EPA)
monitoring study. Although there also is a strip chart record of wind
observations made at the plant site covering an additional 3-year period,
these data were not adequately maintained and annotated. A careful and
detailed review by McElroy (1977) has developed some useful climatological
estimates of seasonal variations in the frequency of occurrence of weather
regimes (i.e., trade, sea breeze, and Kona winds). However, these data were
judged by us in a conference with McElroy to be unsatisfactory for hour-to-
hour simulation of plume transport.
KAHULUI
Meteorological Data
Meteorological data were available in two distinct forms: that from the
monitoring stations operated by an EPA contractor and that obtained from the
National Oceanic and Atmospheric Administration (NOAA) weather station at the
Kahului Airport. The Kahului Airport data provide surface observations at 3-
hour intervals from which sky condition, temperature, wind direction, and
wind speed data were coded and stored in a computer file. The meteorological
data from three monitoring stations included temperature, wind speed, wind
direction, the standard deviation of wind direction, and, at one station,
solar intensity. The meteorological data from the monitoring stations were
available for each hour during the 6-month monitoring period, February through
July 1977. To demonstrate the validity of the air quality model, the hourly
data from the monitoring stations were carefully reviewed and the most appro-
priate measurements selected for use in the model. The wind direction and
speed from Station 2 and the standard deviation of wind direction from
Station 3 were found to be the most useful meteorological measurements.
-------
Stability class was determined using sky condition information from the
Kahului Airport and the method described by Turner (1964). Mixing height
input to the model was determined using the radiosonde data obtained from the
soundings made at Hilo, Hawaii. Upper air data for Hilo were obtained from
the National Climatic Center.
Mixing depth information was determined from the radiosonde data at two
times each day corresponding to the soundings made at 0000 and 1200 hours
Greenwich time (1400 and 0200 Alaska-Hawaii time zones respectively). Hourly
values of the mixing depth were approximated by a scheme which utilized the
0200 Local Standard Time (LST) sounding in the interval from 0200 to 0600
hours and then made a linear interpolation of the 0200 and 1400 mixing heights
for the hours between 0600 and 1400 LST. The mixing depth from the 1400 LST
sounding was used as a constant until the subsequent 0200 LST sounding.
This interpolation scheme is based on the fact that the mixing height
rises gradually in the morning due to solar heating of the ground after
sunrise and mechanical turbulence induced by surface friction. Linear
interpolation with time has been found to give as good an estimate of the
rate of rise as more sophisticated techniques. Therefore, linear inter-
polation is most frequently used because it is computationally convenient.
In the evening a different process occurs. The boundary layer slowly
stabilizes with the afternoon mixing height gradually weakening, and the
temperature discontinuity which marks the mixing height at the top of the
boundary layer gradually weakens but remains at the same height. A new
layer forms underneath the old one due to a combination of radiative cooling
and mechanically induced turbulent mixing from surface friction. A new
mixing height is established by the time of the 0200 sounding. An alter-
native assumption frequently applied is that the mixing height decreases
linearly in time between the two observations. Under most conditions the
mixing height is high enough even at night to have little impact on the
ground-level concentrations estimated by the model. If the 0200 mixing
height is very low, there may be a period of 2 to 3 hours preceding 0200 when
the two assumptions produce differing results.
Mixing height data were obtained from the Hilo and Lihue stations.
However, the Hilo data were used because (1) the data record on magnetic tape
from the Lihue station contained several significant sections from which data
were missing or garbled, (2) the Hilo data approximated the radiosonde data
for Kahului quite well, and (3) the Hilo data provided a record for the
entire Kahului modeling period whereas the special Kahului radiosonde
information was available for a 6-day period only. Figure 1 shows the mixing
height as determined from the special soundings made during the Kahului
intensive period compared to the mixing height derived from the Hilo data for
corresponding times. In plotting the data in Figure 1, an arbitrary height
of 1,500 m was assigned to the Kahului soundings when no mixing height was
apparent from the soundings. Except for two cases when very low mixing
height was indicated on Hilo soundings in comparison to Kahului soundings,
the agreement between the two estimates is good and very satisfactory for
dispersion modeling purposes.
-------
2400 i-
ZZiK)
2OOO
1800 -
1000
'SJ H00
X
CP
C 12OO
•O
X
•H
2
9
iH
3
1000
8UO
bOO -
4(10
ZOO
i i i i—I I I I—I—UJ
200 400 fiOO 800 1000 1200 140U IbOO 1800 2000
Hilo Mixing Height (m)
*
Figure 1. Comparison of mixing heights derived from corresponding Kahului and Hilo radiosondes
-------
For the Kahului plant, hourly arithmetic mean wind speed and vector mean
wind direction values from Station 2 were used for model input during the
validation study. After a brief study of the data from all three monitoring
sites, the consistency between wind speed, direction, and patterns measured
at Station 2 and at the NOAA station at the airport indicated that the Station
2 wind data were the most representative data.
Airport meteorological data were used for model input when running the
model for the 12-month period.
Table 1 is presented to summarize the availability of meteorological
data with respect to source and time period.
Temperature data which were available from the monitoring stations for a
period of time less than the 6-month monitoring period were easily supple-
mented by airport data. Also, during the intensive study period, hourly
observations from the airport were available in addition to the normal 3-hour
data.
During the intensive study period plume width measurements were made
using mobile-mounted S02 detection instruments, the Thermo Electron Company
(TECO) SOj analyzer, and a correlation spectrometer (COSPEC). Frequently
both the TECO and the COSPEC units were used simultaneously, resulting in two
sets of data which were provided to Geomet for use in the modeling analysis.
A value of Cy (standard deviation of the concentration distribution in the
horizontal cross-wind direction) was calculated for each occasion when there
were three or more plume width measurements completed within a 60-minute
period at a relatively constant distance from the power plant using either
the TECO or the COSPEC unit. There were 22 values from the COSPEC unit and
26 values from the TECO unit. Subsequently, a regression analysis was made
between the observed Oy values normalized with respect to downwind distances,
x and O0 (standard deviation of wind direction). The correlation coefficients
from that analysis are presented in Table 2.
TABLE 1. METEOROLOGICAL DATA SOURCE FOR KAHULUI PLANT STUDY
.
Wind Speed
Wind Direction
Std Dev. of Wind Dir
Sky Condition
Temperature
Mixing Depth
Monitoring
Stations
1
1
I
_
-
-
Kahului
Airport
2
2
-
2
2
3
Hilo Weather
Station
-
-
-
-
-
4
1 - Hourly data available for 6-month period - February- July 1977
2 - 3-hourly data available for 12-month period - August 1976-July 1977
3 -Upper air soundings four times a day - May 14-20, 1977
4 -Upper air soundings twice a day, 12-month period - August 1976-July 1977
-------
TABLE 2. CORRELATION COEFFICIENTS OF Oy/X TO OBSERVED O0
°0
Observed
at Station
1
2
3
TECO
0.273
0.439
0 690
Oy Values
COSPEC
-
0 66S
0 656
Dele-mined from
TECO 01
-
0. 539(1)
0 656(1)
COSPEC
-
0.642(2)
0 656(2)
(1) On the occasions when a value was available from both the TECO and COSPEC
for the same time and same downwind distance t.ie TECO observation was used
(2) On the occasions when a value was available from both the TECO and the
COSPEC for tr.e same nme and at the same downwind distance the COSPEC
wss used.
The best correlation occurred between oy observed using the TECO analyzer
and og observed at Station 3. Based on the fact that the best correlation
was found between the Og values at Station 3 and the oy/x values from the
TECO data set, and that the model performance was improved by calculating oy
by using the coefficients from the regression line of the TECO and Og from
Station 3, the model was adapted to routinely calculate oy based on the
observed 09 of each hour in the data set. There is no obvious reason why the
correlations with measurements at Stations 2 and 3 are different. The fre-
quency distributions of ag at the two stations over the monitoring period are
shown in Figure 2 to be very similar. It is concluded that these two measure-
ments are essentially interchangeable. When the Station 3 og observation was
missing, Station 2 observations were used.
Emission Data
The emission rate data for dispersion model input were calculated using
reported generator output information, fuel sulfur content, and the emission
factor for oil-fired boilers from EPA's (1977) published compilation of
emission factors.
Hourly generation load records were provided for the intensive study
period, and daily generation load data were provided for the monitoring
period. For the 5 months of August 1976 through January 1977, only the fuel
consumption data on a monthly basis were available to estimate emissions.
-------
100 --
so ..
8
§
o
s-
V
3
er
£
U.
0)
U
60 --
40 --
20 - -
20
60
Og , degrees
Figure 2. Distribution of CQ observations at Sites 2 and 3 near
the Kahului plant.
10
-------
Hour-by-hour estimates of the SO emission from each stack were calcu-
lated from hourly generator loads using means of the fuel sulfur content,*
fuel heat content, and the plant operating efficiency (in terms of heat input
per unit of load output).
The power plant reports the fuel sulfur content and heat content to the
FPC on a monthly basis. The fuel information used in this study included
data provided directly from the plant and data obtained from the FPC.
The hourly emission rates were computed for the intensive study period,
and from that information an hourly variation factor was computed by deter-
mining the average fraction of daily emissions which is emitted each hour in
the intensive study period. The hourly variation factor was applied to the
average daily emission rates which were computed from the daily generator
loads available for the monitoring period and to the average monthly emission
rates computed for the 6 months before the monitoring period. The hourly
variation factor was applied to reflect the very consistent load cycle
apparent in the hourly generation data provided.
KAHE POINT
Meteorological Data^
Meteorological data for model input were available from the three moni-
toring stations located near the power plant. Hourly averages of meteoro-
logical (and air quality) parameters which were measured for the 4-month
period from October 1977 through January 1978 were delivered to Geomet by
Northrop Services, Inc., on magnetic tape. The hourly data included the
arithmetic mean wind speed, vector mean wind speed, vector mean wind direction,
and standard deviation of the wind direction from all three stations. The
temperature and pyranometer data were also available from Station 1. Intensive
monitoring data, including vertical wind and temperature data from a tethered
balloon, were available for a few hours each day during a 2-week period.
The vertical temperature gradient from the tethersonde observations was
investigated to determine how well surface pyranometer measurements represent
stability classes defined by vertical temperature gradient using the scheme
recommended by the Atomic Energy Commission Safety Guide 23 (1972) for the
change of temperature from 10 to 100 meters above ground level. Stability
class determined using solar radiation data alone matched with the vertical
temperature classes, as well as solar radiation data stratified by wind
speed. A comparison of three schemes for characterizing stability is shown
in Table 3. The vertical temperature gradient shown for each hour is an
average of the gradient from 10 to 100 meters during up and down descents
Fuel sulfur content (wt. %) data from Federal Power Commission records were
as follows:
1976 - August,1.9 November,1.8 1977 - January,1.98 May,1.90
September,1.8 December,I.8 February,1.98 June,1.96
October,1.8 March,1.98 July,2.05
April,1.86
11
-------
TABLE 3. COMPARISON OF STABILITY CLASSES ESTIMATED THREE DIFFERENT WAYS
Date
Jan 13
14
16
IS
23
24
25
26
27
Hour
of
Day
12
14
16
18
19
S
10
10
10
13
15
17
IS
19
8
9
10
16
17
S
9
10
14
16
17
11
13
10
12
Vertical
Temperature
Gradient
10 to 100 m,
°C/100m
-1.2
-1.0
-1.2
-1.8
-1.0
-0.9
-1.0
-1.4
-0.3
-1.1
-1.0
-1. 1
-1 0
-1 2
-0 5
-0.8
-1.6
-1.1
-1.2
-0.4
-0.8
-1.8
-1.6
-1.1
-1.0
-1.3
-1.9
-0.9
-1.2
NRC
Stabilit>
Class
D
D
D
B
D
D
D
D
£,
D
D
D
D
D
D-E
D
C
D
D
E
D
B
C
D
D
D
B
D
D
Station 1
Pyranometer
Reading
(No units)
3.3
2.7
1. 1
0.0
0 0
0 5
1. 5
1.4
2.8
1.0
1.5
0.3
0 0
0.0
0.7
2. 1
2.4
0. 5
0.0
0:6
M t
M
2. 1
1. 1
0.3
2.7
2.9
2.8
3.5
Wind
Speed
m/sec
0.6
2.3
2 2
3.5
3. 1
2.9
2.9
3.9
1.6
1.7
1.5
2.6
2.4
3. 1
2.4
4.7
3.7
3.6
4.5
4.0
M
M
1.9
1.4
3.5
3. S
3.6
3.4
4.0
Approximate
Pasquill
Stability
Class*
A
B
C
D
D
C
B-C
D
A-B
B
A-B
C
D
D
C
B-C
B-C
C
C
C
-
-
A-B
B
C
B-C
B-C
B-C
B
•
Model Input
Stabibtv
Class
B
B
D
D
D
C
D
B
D
C
D
D
D
D
C
C
D
D
D
D
-
-
D
D
B
B
B
B
A
* See
text.
f Missing
12
-------
closest to the indicated hour of the day. The model input stability class is
the pyranometer reading rounded (i.e., <. 3.5 for D, 1.5 to 2.4 for C, 2.5 to
3 4 for B, and >. 3.5 for A). The approximate Pasquill stability categories
were determined using Pasquill's (1974) scheme assuming that pyranometer
readings of less than 1.5 correspond to slight insolation, 1.5 to 3.0 to
moderate, and greater than 3.0 to strong insolation. These comparisons
suggest that a stability classification scheme based on the pyranometer
reading alone does as well or better than one taking wind speed into account
in estimating stability classifications based on vertical temperature gradient.
Since there is some uncertainty regarding the representativeness of the
surface wind speed measurements at each of the three monitoring sites, it
was decided to use a scheme which did not depend on wind speed.
The hourly averages of wind direction from the three monitoring sites
showed the characteristic wind direction was different from site to site.
Figure 3 shows the frequency distribution of hourly average directions from
the three sites. Figure 4 shows the frequency distribution of oe values from
each site.
The monitoring sites were all located along the shoreline, all within 2
km of one another. For running the model for the 4-month period, the Station
2 wind data were used. Station 2 is located between Stations 1 and 3. The
wind directions show a pronounced change from strongly northeast at Station 1
to almost uniform at Station 3. The strong influence of the very pronounced
Waianae Mountain Range in breaking up the northeast trade winds along the
west coast is clearly shown at Station 3. It was hypothesized that Station 2
would be most representative of the three sites in representing the wind
direction of the plant plume, which is estimated to be the most critical
meteorological parameter to be determined. This was tested by comparing
hourly mean wind direction at the three sites with tethersonde wind »»•
ments taken in the plume. Table 4 presents the comparison Stations 1 and
2 do about equally well in predicitng the wind direct ion ^™* ***?***
sonde during the time the tethersonde was at plume height. Station 2 is
judged best in 14 of 30 comparisons, Station 1 in 12 of 30 comparisons and
Station 3 in 4 of 30 comparisons. Data from Station 2 were selected for use
in the model because of Station 2's closer proximity to the plant. All
three s!tes were influenced to some extent by surrounding obstacles, especially
trees.
Meteorological data required by the model for each hour also include
mixing height and a stability class indicator. • At the time the modeling
effor? was to be performed, the standard twice-daily upper air •»«£«*
for the monitoring period were not available; consequently, the sensitivity
of the model to this parameter and the impact of using monthly average data
instead of daily values were tested. The test of the model sensitivity to
mixing height input consisted of running the model using day-to-day variations
of the mixing height for a 6-month period and running the model again using
monthly averages of the mixing height for the same period. The results of
the test showed that only minor changes occurred in the model-calculated
concentrations. There were 30 receptor locations included in the sensitivity
test. The overall mean from a total of 2,256 hours in the test data set was
changed by less than 1 microgram per cubic meter (ug/m ) at 13 receptors.
13
-------
Station )
0 20 40 00 SO 100 12O HO IOO 18O iOO 220 24O 2otl 28O 3OO }2O 340 }t>0
M.ition 1
Wind Direction, Degrees
Figure 3. Distribution of wind direction, Kahe Point monitoring stations.
-------
Station 3
20 40 60 30 100 120
10
Station 2
Station 1
Figure 4. Frequency distribution of OQ, Kahe Point
monitoring stations.
15
-------
TABLE 4. WIND DIRECTION OBSERVED IN TETHERSONDE AND MONITORING
STATIONS DURING 2-WEEK PERIOD IN JANUARY 1978
Tethenonde
Hour
D ate of
(1978) Day Flight
January 13 14
16
14 S
10
11
IB 10
11
19 8
10
11
13
23 11
24 S
9
10
11
16
18
25 9
10
11
14
15
17
26 11
13
14
27 10
11
12
3
4
7
8
8
12
12
13
14
14
15
IS
24
25
26
26
27
12
S
60
12
39
21
IS
16
48
18
50
13
8
Duration
(mm)
11
9
26
41
22
34
17
25
57
19
6
15
6
7
60
15
16
23
30
31
31
32
32
34
35
36
36
37
38
38
Height
(ID )
190
180
170
150
ISO
200
210
160
200
140
140
130
200
210
150
140
240
ISO
150
100
90
100
120
330
200
150
150
130
100
120
Station 1
Wind Wind
Direction Direction
fDenrees) (Deqrees)
200-260
190-210
200-240
210-240
230-260
20-80
20-110
50-70
10-100
40-SO
30-60
340-70
40-60
40-60
20-70
40-70
290-340
30-50
40-60
10-70
30-60
360-50
270-310
360-90
20-90
10-80
30-90
20-100
60-100
40-110
202
irs
1S4
227
229
51
40
18
42
31
355
244
35
23
35
32
20
35
M
M
37
276
260
9
101
27
2S
47
60
35
Station 2 Station 3
Wind Wind
Direction Direction
(Degrees) (Deqrees)
228
205
204
259
265
M *
115
19
99
39
357
275
72
60
S4
SI
55
72
73
79
34
294
276
32
M
67
S8
115
95
84
226
206
205
256
241
303
297
17
131
258
300
M
M
M
M
M
300
353
359
IS
6
M
M
M
35
M
M
M
M
M
* Missing
16
-------
At the remaining 17 receptors, the mean increased in the l-to-5 jjg/m range
at 10 receptors, by 6 to 10 yg/m^ at 5 receptors, and by less than 15 pg/m^
at 2 receptors. None of the means was affected enough to change the status
of any of the receptor locations from being within compliance to being out of
compliance. In all instances, whenever the distribution of calculated
concentrations was different due to a change in the mixing height estimate,
the concentrations calculated using the monthly average mixing height were
higher. Monthly mean afternoon and morning mixing heights were obtained for
the Hilo radiosonde data for .the periods of October 1976 to January 1977.
The mixing heights derived from Hilo soundings were higher, in general
(see Figure 5), than the mixing heights from nine comparable soundings at
Wheeler Air Force Base (AFB). This result indicates that the Hilo data may
provide an overestimate of the mixing height at the Kane Point plant.
However, this estimate is not expected to have any effect on the maximum S02
concentrations which occur very close to the plant. Figure 5 shows the
mixing heights from Wheeler AFB on Oahu for approximately comparable times
using Hilo and Lihue data. The Lihue data indicate a more favorable comparison
but the Hilo data are close enough that no significant impact would be found
in the modeling results.
Tethersonde flights ranged up to 600 meters above ground level on some
occasions. Mixing height data from the available radiosonde stations in
Hawaii (Hilo, Lihue, Wheeler AFB) indicate mixing heights are generally
higher than the maximum altitude of the tethersonde, especially during midday
when most of the tethersonde data were collected. The tethersonde data
provided excellent detail on the wind and temperature profile in the atmos-
phere below 600 m. The tethersonde temperature data did indicate low-level
temperature inversions which would not show up from the radiosonde data.
The site-specific data were valuable in assessing the utility of
dispersion parameters used in the model, for providing the type of meteoro-
logical data needed for input to the model, and for verification of the
calculated concentrations from the model. More data, especially of the type
collected in the intensive monitoring, would have extended the usefulness of
that data considerably.
Plume width measurements were made during the intensive period by remote
sensing using a mobile mounted COSPEC. The COSPEC data provided estimates of
the plume width at locations along the coastal highway where the plume
crossed the highway. The plume width data were used to evaluate the applica-
bility of conventional methods for calculating crosswind concentration
parameter, ay.
In modeling the power plant at Kahului, a reasonable correlation was
found between the standard deviation of the observed plume width (oy) and
the standard deviation of the wind direction (09)• The data from the Kahe
Point plant observations did not have as good correlation between oy and O0,
and consequently considerable effort was given to determining how the oy
value should be determined. One of the main reasons that the Kahului relation-
ship did not correlate well at the Kahe Point location is that the measured
plume dispersion in terms of ay is mostly affected by the separation of stacks
17
-------
CO
2 b
2 -I
2 2
I 2-°
^*-
•P i a
•H
-------
and the orientation of the wind relative to the line of stacks. This as duo
to the fact that the observations were made on a road only 100 meters from
the plant. This road was the only location from which the COSPEC measure-
ments of the plume could be made. Attempts to make angular sitings through
the side of the plume at some distance away were not successful. Only
vertical observations made from underneath the plume produced useful data.
It might be possible to get a significant correlation between OQ values
determined from the tethersonde observations and COSPEC oy measurements.
However, since the measurements mostly represent one travel distance because
of the influence of the wind relative to the line of stacks, this does not
seem like a promising line of analysis. Perhaps future investigations can be
made into this question. Such future investigations should also consider the
relationship of OQ measured aloft to OQ measured at the surface or some other
parameter, 'in order to find data for use in longer-period modeling considera-
tions .
Since the measured plume dimensions mostly showed the influence of the
wind orientation relative to the stacks and showed very little about the rate
of dispersion, the model was adjusted by the cut-and-try method. The plume
height, initial volume, and oy and OQ relationships were adjusted to produce
the best agreement between calculated and measured hourly mean SO2 concen-
trations. It was decided to relate oy estimates to OQ measurements, as the
meteorological variable, since these are generally accepted as well correlated.
The Kahului relationship was used as the basic starting point. A description
of the modifications adopted is given in the model validation discussion.
The coefficient for vertical dispersion, oz, was determined using the
values of the pyranometer to describe the basic meteorological influence.
Hourly average values of the pyranometer data ranged from 0.0 to 4.0. The
stability class determined from the pyranometer is described below. These
stability classes agreed reasonably well with vertical temperature gradients
observed by tethersonde data.
Pyranometer Stability Class
>3.5 A
2.5-3.4 B
1.5-2.4 C
£1.5 D
Based on the modeler's experience, stability class D was further broken
down to more stable classes E and F. If the wind speed was less than 5.4
meters per second (m/sec) and OQ was between 10 and 20 degrees, stability
class E was assigned. If the wind speed was less than 5.4 m/sec and a., was
less than 10 degrees, stability class F was asssigned.
Emission Data
The emission rate data input to the model were calculated using generator
output information, fuel content, and the emission factor for oil-fired
boilers taken from EPA1s published compilation of emission factors (EPA,
1977). Hourly generator output data were provided for the entire 4-month
19
-------
modeling period. The hourly emission rate from each stack was determined by
calculating the fuel required to generate a kilowatt of electricity, then
subsequently applying the'emission factor and the average sulfur content of
the fuel during the corresponding time period.
Emission rate data input to the model were provided for each hour for
each of the five stacks. Specifically, the SO^ emission rate was calculated
using a monthly average fuel sulfur content which was 1.7 percent in October,
1.6 percent in November, 1.5 percent in December, and 1.3 percent in January.
Data obtained from the FPC were used to calculate the average number of
British thermal units (BTU) required to generate a kilowatt. The value of
9,970 BTU/kw was used for the emission rate calculation; the data available
indicated that monthly averages would be within 0.5 percent of the value
(9,970) used.
20
-------
SECTION 4
AIR QUALITY DATA
KAHULUI
Ambient concentrations of sulfur dioxide were continuously measured from
February 7, 1977, to July 18, 1977, at the three monitoring stations south-
west of the Kahului power plant. Hourly average concentration data were
computed by Northrop Services, Inc., from instantaneous measurements made
every 2 minutes. These data were used to modify and validate the model. The
distributions of observed S02 concentrations were tabulated for each hour of
the day at each station by 10-degree wind sectors. The resulting tables
showed that concentrations of SO, are very dependent on wind direction. The
data collected during this study are presented in a report by Siple and
Freberg (1978).
Statistical summaries of the observed SO2 concentrations for the 6-month
period are presented in Tables 5, 6, and 7. Comparison of calculated concen-
tration to observed concentration is facilitated by these statistical
summaries. Monitoring station 1 was located at 156°28'06" West, 20053'34"
North and about 0.4 km from the power plant; monitoring station 2 was located
at 156°28'12" West, 20°53'18" North, and about 1.0 km from the power plant;
monitoring station 3 was located at 156°28'33" West, 20°53'06" North, and
about 1.7 km from the power plant.
The predominance of the Kahului power plant as a source of SO2 at the
three monitoring stations is indicated in Table 8. The average concentration
for each 10-degree sector of hourly mean wind direction is shown for each
monitoring station in Table 8. The data shown are the averages of the
observed concentrations when the wind was from the sector indicated. It is
obvious that high concentrations are observed when the wind is from the
north, within the range from 330 to 070 degrees. Especially noticeable are
the elevated concentrations at Station 1 when the wind is from the 41 to 50
and 51 to 60 degree sectors; at Station 2 elevated concentrations are most
noticeable with wind from the 21 to 30 and 31 to 40 degree sectors, and at
Station 3 when the wind is from the 31 to 40 and 41 to 50 degree sectors.
The influence of other SO2 sources is also indicated by the data in Table
8. At Station 1 a second high is prominent when the wind is from the 1 to
10 degree sector. The wind from due north would carry emissions from ships
in Kahului Harbor to Station 1. A decrease in average concentrations is
apparent as the wind shifts towards the east until the power plant and Station
1 are aligned. The apparent influence of emissions from a pineapple plant
21
-------
TABLE 5. CUMULATIVE FREQUENCY DISTRIBUTIONS OF OBSERVED HOURLY
SO CONCENTRATIONS AT KAHULUI
RCENTILE
HIGH 4090.000
•- - 99.5 -3567.351
99.0 3370.000
95.0 2603.499
90.0 -2J40.000
80.0 1570.000
70.0 1110,000
60.0---741.000
50.0 414.000
40.0 152.000
30,0 — 34.100
20.0 7.860
10.0 2.620
"50 0 000
l.o o.ooo
0.5 0.000
Statior
— - 2
941.000
623.351
500.350
244.700
--141.000-.
49.800
21.000
-• -10.500 •
5.240
2.620
- - 2.620
2.620
0.000
0 000-
-— - • - — \j 9 \j v */ — -
0.000
0.000
1
3
786.000
-370.676
320.000
212.000
152.000
R9.JOO
52.400
31.400
15.700
10.500
7.860
2.620
0.000
0 000
0.000
0.000
-LOW —0.000- 0.000 0.000
22
-------
TABLE 6. CUMULATIVE FREQUENCY DISTRIBUTIONS OF OBSERVED 3-HOUR
£02 CONCENTRATIONS AT KAHULUI
(pg/m3)
PER3EMTILE
HIGH 3680.000
99.5- 3217.451
99.0 2994.700
95.0 2404.500
- 90. 0-1980. 000--
80.0 1450.000
70.0 1090.000
705.000
516.000
446.490
215.450
•1-32.000
62.080
32.300
--- 60.0— -790.000 17.500-
50.0 520.000
40.0 290.400
30.0 81.260 -
20.0 20.100
10.0 2.620
5.0 0.000-
\.0 0.000
0.5 0.000
LOif.L . 0.000-
9.620
5.240
2.870
2.620
2.620
— 0.000
0.000
0.000
o.ooo
432.000
296.726
266.000
1P3.000
-137.000
88.200
60.100
- 41-.-000-
25.300
14.000
8.740
4.370
0.000
- ---0.000
0.000
0.000
0.000
23
-------
TABLE 7. CUMULATIVE FREQUENCY DISTRIBUTIONS OF OBSERVED 24-HOUR
SO CONCENTRATIONS AT KAHULUI
(ug/m3)
PERCENTIliE
HIGH
99.5
- -- 99.0
95.0
90.0
.. .. _fiO.O
70.0
60.0
... -50.0
40.0
30.0
.... 20.0
10.0
5.0
- 1.0
0.5
LOW
Station
l
2360.000
2284.000
2119.599
1754.000
1584.000
1336.000
1108.000
887.000
- 678.000
520.400
324.400
- 150.600-
14.520
3.516
- - 0.000
0.000
0.000
2
311.000
235.761
-209.400
ISO. 200
115.000
79.220
49.840
36.860
• 24.400
16.700
10.860
... 6.103
3.544
2.216
- 0.000
0.000
0.000
3
193.000
1R6.160
169.440 -
125.400
105.800*
- - 88.360
67.880
4 9 . 5 2 0
--•37.800-
28.060
22.200
14.580
1.986
0.000
0.000 -
n.ooo
0.000
24
-------
TABLE 8. AVERAGE OBSERVED S02 CONCENTRATION BY WIND DIRECTION
IN 10-DEGREE SECTORS
wind Sector
(Degrees)
181-190
191-200
201-210
211-220
221-230
231-240
241-250
251-260
261-270
271-280
281-290
291-300
301-310
311-320
321-330
331-340
341-350
351-360
001-010
011-020
021-030
031-040
041-050
051-060
061-070
071-080
081-090
091-100
010-110
111-120
121-130
131-140
141-150
151-160
161-170
171-180
S02 Concentration
(yg/m3)
Station
123
9
7
7
7
6
3
6
3
5
3
4
6
8
8
12
23
28
25
538
67
62
479
1335
1538
668
217
42
4
1
4
7
9
3
3
8
8
2
6
4
6
3
0
2
2
2
2
1
2
3
2
3
8
8
9
29
47
133
250
62
9
5
4
8
1
0
1
3
2
1
2
1
2
9
7
6
7
7
1
3
6
3
4
3
4
3
8
7
29
62
24
53
33
40
111
126
29
10
10
8
7
1
2
3
3
1
2
2
8
Observed at Station 2.
25
-------
just north of Station 3 on observed concentrations at Station 3 is noticed
when the wind is from the 341 to 350 degree sector.
The average concentrations for the entire 6-month monitoring period at
Stations 1, 2, and 3 are 768, 45, and 50 yg/m3, respectively. If the concen-
tration at Station 1 is reduced by the apparent anomolous contribution from
the harbor area when the wind is from the 1 to 10 degree sector, the average
concentration would be approximately 755 pg/m3. This would lower the average
by 13 ug/m3. In the same manner, if the average concentration at Station 3
is calculated with apparent anomolous contributions from the pineapple plant
when the wind is from the 341 to 350 degree sector, the overall average would
be lowered by 1 ug/m3. The changes in the average concentrations which would
occur if the apparent contributions from sources other than the power plant
are removed are not significant and would not alter the conclusions of this
study.
Assuming that 6 months of data may be used to approximate the annual
average, the mean values show that concentrations at Station 1 are in serious
violation of the primary air quality standard. Stations 2 and 3 are below
the primary and secondary Federal air quality standards for an annual mean.
The distributions of 3-hour and 24-hour concentrations shown in Tables 6 and
7 respectively, show similar results for these stations relative to Federal
standards, in that Station 1 data indicated violations while Station 2 and 3
data are below the corresponding standards.
KAHE POINT
Sulfur dioxide was continuously monitored for the 4-month period October
1, 1977, through January 30, 1978, at three sites along the coast to the west
of the Kane Point power plant. Hourly averages of the observed data were
computed from the instantaneous values measured every 2 minutes. The hourly
SO, concentrations were compared with the model calculations to indicate the
validity of the model. The data collected during this study are presented in
a report by Siple and Freberg (1978). Monitoring Station 1 was located at
158°07'57" West 21°21'23" North, and about 0.5 km from the power plant;
monitoring Station 2 was located at 158°07'58" West, 21-21-34" North, and
about 0.15 km from the power plant; monitoring Station 3 was located at
158°08'20" West, 21e22'24" North, and about 1.6 km from the power plant.
Statistical representation of the observed hourly average S^ concen-
trations from all three stations is presented in Table 9. The data shown
represent approximately 2,650 individual cases at each station. The means of
the hourly averages from Stations 1, 2, and 3 are 80, 402, and 20 ug/m ,
respectively.
Using the mean of all hourly averages to represent an annual mean,* the
pattern of excursions above the Federal standards is similar to the results
for shorter term concentrations. The primary standard for the annual average
*The 4-month monicoring period is considered representative of annual
conditions because it includes both a period of persistent trade winds and
a period of variable winds with a high frequency of the irregular Kona
winds.
26
-------
TABLE 9. CUMULATIVE FREQUENCY DISTRIBUTIONS OF OBSERVED HOURLY
S02 CONCENTRATIONS AT KAHE POINT
(ug/m3)
0. STA™N
1 9. 3
HIGH 2050.000 3680.000 529.000
2ND 1773.000 3568.000 448.000
99.5 1098.950 3005.400 247.710
Q9.0 158.641 7R07.700 2*7.710
9S.O 447.850 1800.000 81.ROO
90.0 183.000 1320.000 3".300
80.0 68.100 74^.000 18.300
70.0 39.300 414.000 11.100
60.0 26.200 189.000 10.500
50.0 18.300 67.<500 7.860
40.0 13.100 76.700 tj.?4»
30.0 10.500 12.840 5.240
20.0 7.860 S.240 2.620
10.0 5.240 7.620 2.670
5.0 2.620 2.620 2.620
1.0 2.620 2.620 2.670
0.5 2.620 2.620 2.670
LOW 2.620 2.620 2.620
27
-------
of 80 yg/n»3 is exceeded at Station 2; Station 1 is exactly at the standard
and Station 3 is well below the standard.
Statistics of the 3-hour average concentrations are presented in Table
10. The data in Table 10 indicate that no violations of the 3-hour secondary
standard of 1,300 yg/m3 were observed at Station 3. At Station 1 only the
highest value exceeded the 3-hour standard. At Station 2, however, the
distribution shows that 10 percent of the 3-hour averages exceeded the
standard, indicating the standard would be exceeded about 265 times over a 1-
year period.
The hourly averages within a calendar day were averaged to compute the
24-hour averages. The distribution of the 24-hour averages is shown in Table
11. Comparison of the 24-hour standard to the observed 24-hour averages
shows a situation similar to the comparison of the 3-hour data, in that there
are no excursions at Station 3; at least 1 percent of the data at Station 1
exceeds the 365 pg/m standard while at Station 2, 30 percent of the observa-
tions (about 35 days) exceed the standard.
28
-------
TABLE 10. CUMULATIVE FREQUENCY DISTRIBUTIONS OF OBSERVED 3-HOUR
S02 CONCENTRATIONS AT KAHE POINT
(pg/m3)
rcentile
HIG^
2 MO
99.5
99.0
95.0
90.0
bO.O
70. D
60.0
50.3
40.0
30.0
20.0
10.0
5.0
1.0
0.5
LD/I
1
1399.080
1236.640
1007.521
900.678
422.344
184.291
78.338
46.374
33.274
23.580
16.239
11.345
7.360
5.240
3.485
2.620
2.520
2.620
Station
2
3537.000
3222.600
2689.432
2211.245
1743.086
1335. Ib2
723.644
472.648
217.303
80.434
32.540
15.720
7.860
4.721
7.620
2.620
2.620
2.620
3
372.040
35^.700
231.071
203. 323
74.408
44.540
21 .875
15.720
10.4RO
7.«bO
6.105
4.375
3.4«5
2.620
2.620
2.620
2.620
2.620
29
-------
TABLE 11. CUMULATIVE FREQUENCY DISTRIBUTIONS OF OBSERVED 24-HOUR
S02 CONCENTRATIONS AT KAHE POINT
(yg/m3)
Percentile
HIGH
2ND
99.5
99.0
95.0
90.0
80.0
70.0
bO.O
50.0
40.0
30.0
20.0
10.0
5.3
1.0
0.5
LDW
1
702.150
581.640
629.848
578.496
327.500
188.640
99.560
62.800
41.920
31.440
26.200
20.960
13.100
7.860
5.240
3.144
2.620
2.620
Station
2
2143.160
2054.080
2091.048
2028.692
1530.473
1226.422
799.100
465.836
243.136
121.830
49.780
26.462
15.720
7.860
5.240
2.620
2.620
2.620
3
104.800
102.180
103.307
100.346
59.212
38.252
25.200
20.960
18.340
13.100
10.4RO
7.860
5.240
5.240
2.620
2.620
2.620
2.620
30
-------
SECTION 5
MODEL SELECTION AND MODIFICATION
The Gaussian plume dispersion model as implemented in the PTMTP program
of EPA's UNAMAP system is the basic model format which was used in this
study. However, extensive changes were introduced to the program to enable
it to represent source characteristics found in the vicinity of the two
plants being studied and to accept the types of data which were available
for this study. The changes produced substantial improvements in the model-
to-measurement comparisons which were used to select or reject various model
modifications. The model modifications are site-specific in that different
modifications were used to obtain the best agreement between model calcula-
tions and measurements at the two power plants. Some of these changes are
relatively general in concept and may be applicable to other areas. Other
changes include site-specific influences which are not obvious but which
indirectly impact on meteorological observations. The relationships between
meteorological variables and SOo measurements derived for the vicinity of
each power plant may be different from the relationships between the same
variables which are observed at other locations. Because of this indirect
site-specific influence, it is recommended that any use of these relationships
with the model be validated using the data selected as inputs before applying
them to another location.
KAHULUI MODIFICATIONS
Initial trial runs were made with the PTMTP program using data from the
7-day intensive monitoring periods. The correspondence between observed and
calculated values is shown in Table 12. Several modifications to the model
were proposed. Two principal modifications to the model included a method
for treating downwash effects from the power plant building and a different
method of determining the horizontal dispersion coefficient. The result of
applying these two modifications to the intensive monitoring data period is
also shown in Table 12. Several additional steps were made to "tune" the
model after these two major modifications were made.
The provision for downwash effects is based on the widely accepted
hypotheses that airflow is not affected by buildings above 2.5 times the
building height and that below 1.5 times the building height there is an
aerodynamic cavity. The rules are based on empirical observations and wind
tunnel experiments. The only additional information required to apply these
rules are the heights and widths of buildings near the source which are
likely to create downwash influences. The following parameters are used to
estimate downwash:
31
-------
BH = building height, m
Bw = building width, m
hs = stack height, m
d = stack diameter, m
vs = stack gas exit speed, m/sec
u = wind speed at stack height, m/sec
Using these parameters, the following are calculated:
*B = min (BH,BW)
h = h + 2(— 1.5) d
E s u s
(1)
(2)
TABLE 12. COMPARISON OF CALCULATED SO2 CONCENTRATIONS USING
MODEL ADJUSTMENTS WITH CORRESPONDING MEASURED VALUES
AT KAHULUI DURING MAY 14-20, 1977
SO2 Concentrations ((Jg/ni3)
Calculated
Monitoring
Site Statistic
1 Mean
Peak
2 Mean
Peak
3 Mean
Peak
Measured
472
2960
15
178
37
249
Original
Model
37
662
44
622
40
482
Downwash
Added
1070
9644
124
947
70
579
New ay and
Downwash Added
332
2349
52
430
32
378
The measurement of wind speed at stack height was not routinely avail-
able and the wind speed data from the 10-m tower at the monitoring site was
used to represent u. The following rules, presented by (or derived from)
Briggs (1975), are used to modify the ground concentration or the effective
stack height to account for downwash effects:
1. If hE >BH + 1.5£B or hs ^ BH + 0.25fcB, no modification is required.
2. If the downwind distance from source to receptor is less than 3.5&B
and the crosswind distance is less than 0.5(BW + S,B) , then
y = -3**
"V
32
-------
where Q = source emission rate, ug/sec
X = ground level concentration, yg/m
K =
0, hE > B
H
0.5£
B
0.352,B < (hE - BH) 1 0.5JIB
1.5, hE <_ 0.35£B + BH
3. If BH 2. BW compute plume height (hp) as follows:
2hE - BH - 1.5BW, hE > BH
hE - 1.5BW, hE _< BH
= 0, if hp < 0.5BW.
hp =
Use hp
If BH < BW, compute plume height as follows:
' 2hE - 2.5BH, hE > 1.5BH
0, hE <_ 1.5BH
hF
Frequently rule 4 above was used to determine plume height in the down-
wash provision. The downwash provision is sensitive to the value used for
BH. The actual BH of the power plant building is 22.56 meters. Little
effect was shown on the calculated SO2 concentrations when BH was entered as
22.56 m, but using BH = 28.0 m caused the effect to be too great. Several
iterations of the program with variations of BH determined that BH = 24.7 m
provided the best approximation of the observed SO2 concentrations. Hour-
by-hour comparisons of the calculated concentrations with measured values
indicated that downwash effects should not be included in calculations during
stable atmospheric conditions, i.e., when stability class was category E or
F. The downwash provision is bypassed for stability category E or F, based
on model-to-measurement comparisons and the fact that there are no stability
criteria for downwash effects.
The method of determining the value of oy, the horizontal dispersion
coefficient, was modified for model validation in the intensive period and
the monitoring period by taking advantage of the availability of 09, the
standard deviation of wind direction, which was recorded at the monitoring
sites.
Based on the recommendaitons of the "AMS Workshop on Stability Classi-
fication Schemes and Sigma Curves" (Hanna et al., 1977), the model was modi
fied to compute oy = 09 x f(x) with f(x) defined by Pasquill (1974). Incor-
porating this method of determining ay improved the correlation of the
calculated with the observed SO2 concentrations over the correlation obtained
using values from the Pasquill-Gifford curves for ay as presented in the EPA
Workbook (Turner 1970). Further improvement in the correlation of calculated
and observed values occurred when the regression coefficients of the rela-
tionship of a /x to 00 observed at Station 3 were used in the equation for
determining Oy. The final version of the model calculated oy using the
following equation whenever a value of 09 was available:
33
-------
Oy = (0.0441 + 0.0104o0)x; x in meters, og in degrees (3)
If the numeric value of 09 was measured as less than 5 degrees, it was set
equal to 5 degrees when used in this equation to prevent negative values of
V
On those occasions when Og was not available, ov was computed as: °y =
a xP with appropriate values of a and p from Table 13. These values were
developed from the results of tracer experiments performed in St. Louis
(McElroy and Pooler, 1968) and are recommended by an AMS Workshop on Sigma
Curves (Hanna et al., 1977) for urban areas. The St. Louis values are more
representative than the commonly used Pasquill-Gifford values in that they
were measured over 1-hour periods in an urban area, while the Pasquill-
Gif ford values were mostly measured in rural areas over periods of 3 to 10
minutes. A comparison of these relationships is shown in Figure 6. The
value of the vertical dispersion coefficient, ay, in the original version of
the program was determined using the Pasquill-Gifford curves (Turner, 1970).
The modified version of the model calculated oz as: az = bx9 with appropriate
values of b and q from Table 13.
TABLE 13. FITTED CONSTANTS FOR THE McELROY-POOLER DIFFUSION
PARAMETERS BASED ON TURNER (1970) STABILITY CLASSIFICATIONS
Stability
Index
A, B
C
D
E
Crosswind
Constants
a
1.42
1.26
1.13
0.992
P
0.745
0.730
0.710
0.650
(2)
Constants for Vertical Diffusion Parameter
x <600
b
0.0926
0. 0891
0. 0835
0.0777
q
1.18
1.11
1.08
0.955
x>600
b
0. 0720
0.169
1.07
1.01
q
1.22
1.01
0.682
0.554
(1)0 = ax , where x is Downwind Distance from Source;Q and x are in meters.
q
(2)(j = bx ; Q and z are in meters.
z z
The model was also modified to treat the initial plume dispersion,
resulting from the enhanced turbulence of the building and the hot stack
gases, by adding a virtual distance to the actual source-to-receptor distance.
The cross-section of the initial plume from each stack is modestly assumed to
be given by the square of one-half the building height and to conform to the
following elliptical shape equation:
ir (2oy) (2az) = (-^)2 (4)
34
-------
STL - Relationships from McF.lroy and Pooler
p_O - Relationships from Pasquill and Cifford
Site - Relationships derived from Kahului monitoring data
1600-
1200 - -
800- -
400 - -
00 = 40°
. 08=15°
=15°
1.0
2.0 3.0
Distance, km
4.0
Figure 6. Comparisons among c relationships.
35
-------
This equation can be solved for a virtual distance by using the selected
relationships for oy and oz as functions of x. The virtual distance, xo, was
determined as follows:
or
R 2
BH
16nb(c+dag)
1
1+q
xo
16ir ab
p+q
if OQ was missing.
(5)
(6)
The constants c and d are regression coefficients of the relationship
of oy to 09 with values of -0.0441 and 0.0104, respectively. BH is the
height of the power plant building as used in this model (24.7 m).
VERIFICATION OF MODEL PERFORMANCE AT KAHULUI
The objective of the modeling work is to evaluate the effectiveness
of several strategies designed to reduce ground level concentrations of
S02- The confidence which can be placed in the estimated effectiveness of
the strategies examined depends on the confidence that can be attributed to
the ability of the model to accurately calculate the concentrations which
would be observed at selected receptor locations.
Verification of the model performance for the Kahului power plant is
based primarily on the comparison of the statistical summaries from observed
and calculated concentrations for the monitoring period using hourly values.
The results of those comparisons are shown in Tables 14, 15, and 16.
To arrive at an annual average we have run the model for an entire 12-
month period. The model input data for the 12-month period were not as
comprehensive as for the 6-month monitoring period, i.e., the meteorological
data are only available from the airport at 3-hour intervals, and the infor-
mation to calculate the emission rate is available only from fuel consumption
data provided on a monthly basis. Due to the difference in the quality and
quantity of the input data, it was necessary to compare the calculated
concentrations for comparable time periods using the model output resulting
from the various types of input. The following comparisons were made in
order to estimate the validity of calculated concentrations made using only
airport data at 3-hour increments.
1. The validity of the model was established by comparison of
observed hourly, 3-hour, and 24-hour average concentrations
with corresponding calculated values.
2. The comparison of 24-hour average concentrations, calculated
using 24 values from the hourly calculations, with 24-hour
average concentrations, calculated using only every third hour
from the airport meteorological records, was used to establish
the validity of more limited meteorological data on 24-hour means.
36
-------
TABLE 14. COMPARISON OF STATISTICS FOR HOURLY OBSERVED AND CALCULATED
CONCENTRATIONS FROM 7 DAYS IN THE INTENSIVE PERIOD
Station
1 Observed
Calculated
2 Observed
Calculated
3 Observed
Calculated
Mean.
(uq/ffl3)
472
332
IS
52
37
32
Standard
Deviation
686
515
32
85
51
51
Highest 3
Value fuq/m )
2960
2349
178
430
249
378
Concentration
90th Percentlle
1608
1051
39
163
108
97
(ua/n3) at
70th Percentlle
905
489
5
43
33
41
TABLE 15. COMPARISON OF STATISTICS FOR HOURLY OBSERVED AND CALCULATED
CONCENTRATIONS FROM THE 6-MONTH MONITORING PERIOD
Station
1 Observed
Calculated
2 Observed
Calculated
3 Observed
Calculated
Mean.
(ug/m3)
768
531
45
96
50
63
Standard
Deviation
889
615
99
135
73
78
Highest ,
Value (uq/mj)
4090
4478
941
1302
786
518
Concentration
90th Percentlle
2140
1413
141
286
152
168
(yq/m ) at
70th Percentlle
1110
754
21
123
52
83
TABLE 16. COMPARISON OF STATISTICS FOR 24-HOUR AVERAGE OBSERVED AND
CALCULATED CONCENTRATIONS FOR 6-MONTH MONITORING PERIOD
Station
1 Observed
Calculated
2 Observed
Calculated
3 Observed
Calculated
Mean
(ug/m3)
767
526
45
95
50
63
Standard
Deviation
586
362
45
69
50
40
Highest ,
Value (ug/nT)
2360
1470
311
312
193
158
Concentration (uq/m ) Percentlle at
90th
1584
1010
115
189
106
116
70th
1108
716
50
120
68
83
30th
324
310
11
44
22
33
37
-------
3. A comparison of concentrations calculated at 3-hour intervals,
using only airport meteorological data, with the corresponding
3-hour concentrations using monitoring station meteorological data,
was used to establish the validity of limited meteorological data
on 3-hour means.
By comparing the statistical summaries from each of these three modes, it was
determined that the model's ability to calculate air quality levels over a
12-month period using limited meteorological data was satisfactory.
Model verification results based on comparison of the hourly concentra-
tions for 7 days in the intensive period are shown in Table 14.
Table 15 shows a similar comparison of hourly concentrations from the 6-
month monitoring period. The data presented in Table 16 show comparison of
observed versus calculated statistics for 24-hour average concentrations in
the 6-month monitoring period.
The comparisons shown in Tables 14, 15, and 16 indicate the model has
been modified to produce results which correspond very well with the observed
data. The ratio of calculated-to-observed concentrations for any of the
comparisons shown in the tables for Stations 1 or 3 stay within the range of
0.6 to 1.6. At Station 2 the comparisons are not quite as good because the
model calculations are consistently higher than the observed data. It should
be noted that the comparisons made in Tables 14, 15, and 16 do not exclude
other sources of SC>2 in the observed S(>2 data. The frequency with which the
wind direction is from a direction which adds SC>2 from the major sources
other than the power plant is low (~2%) so that the overall impact on both
the means and the concentration distribution is small. No explanation of the
difference between the comparisons for Stations 1 and 3 with Station 2 has
been made, but it is interesting to note that the mean values of the observed
concentrations are lower at Station 2 than at Stations 1 or 3.
A possible explanation may be that the plume is above Station 2 as it
moves through that region due to heat buildup and release by the buildings in
the zone between the power plant and Station 2. A conceptual possibility is
that the taller buildings in the vicinity of the power plant may cause the
downwash which accounts for the high concentrations at Station 1. As the
plume moves toward Station 2, the heat released from the lower buildings
could cause the plume to be slightly elevated as it moves across the area
where Station 2 was located; then by the time the plume has moved to the
vicinity of Station 3 the turbulence from the heat introduced by the resi-
dential and commercial buildings has become effective in bringing the plume
to ground level again. The mean calculated and the mean observed concentra-
tions over the 6-month period differ by a factor of 2 which is still a
reasonably good comparison. The comparison of concentrations at the various
percentile levels shown differs by an increasing ratio when examining
comparisons towards the lower concentrations, but is obviously in better
agreement when comparing the higher concentrations.
Based on the favorable comparisons which are noted above, the model was
deemed to be acceptably accurate and the strategy evaluations are based on
the model as described.
38
-------
KAHE POINT MODIFICATIONS
The PTMTP model had been selected and modified for analysis of the
Kahului plant. The version of the model as it existed in the modified form
for the Kahului study was the logical starting point for modeling the Kane
Point plant. It was recognized that the plant sites are significantly
different. However, the differences should be reflected in the meteorological
and plant parameters. The initial run with the modified model resulted in
calculated concentrations excessively higher than the observed S02 concen-
trations at the three monitoring stations. An analysis of those results
indicated the downwash provision was contributing to the overestimation of
the concentrations. Other program changes were made to the model so that the
highest calculated concentrations nearest the plant, especially at monitoring
Station 2, came closer to agreement with the observed concentrations. The
degree of verification achieved was based on how well the hourly calculations
approximated the corresponding observed SO2 concentrations.
In the Kahe Point modeling study, the actual building height was input
to the model. This was slightly different from the process in the Kahului
plant study. For Kahului, the model verification included some adjustment to
the building height value in order to obtain better correlation between the
observed and calculated SO2 concentrations.
Other modifications were made to the model to include assumptions which
improved the agreement between the measured and observed distribution of
hourly concentrations at all three monitoring stations, e.g., representing
the plume height and the dispersion parameters, oy and az- The plume height
for evaluating downwash effects was determined by the following empirically
derived method:
*hE = (HX + 2hE)/3 (7)
where HX equals the plume height (at the downwind distance of the receptor)
calculated using Briggs1 equation as in the BEH072 subprogram of the UNAMAP
programs, and hE is the Briggs formula for sources affected by downwash. The
value hE is calculated as follows:
vs
hE = hs + 2(— - 1.5)ds (2)
where: hs = physical stack height, m
v_ = stack gas exit velocity, m/sec
s
dc = stack inside diameter, m *
a
u = wind speed at stack height, m/sec.
If no downwash effect was present, the standard plume rise equations of
the UNAMAP BEH072 subroutine based on the recommendations of Briggs were
applied.
The location of the power plant is such that the plume is frequently
subject to turbulence from the surrounding ridges. The initial plume
39
-------
dimensions are enhanced in such conditions. One of the adjustments which was
made to the model was to multiply the oz value by 2 (oz is estimated using
McElroy -Pooler coefficients; see discussion for Kahului site) . It would be
desirable to have vertical plume measurements to get the most appropriate
value of az. It was noted by test model runs that increasing oz had the
effect of increasing the means computed for the monitoring stations while
having very little effect on peak concentrations. A factor of three times
the standard oz was too big an increase, so a factor of two was adopted.
The method used to calculate oy at Kahului was replaced by the following
scheme, which gave more spread to the plume and better agreement between
measured and calculated values:
where x = downwind distance, meters, and 00 = standard deviation of wind
direction (Station 2), radians.
Many variations of the model were investigated in the process of obtain
ing the best agreement between the calculated and observed SO2 concentrations.
The three changes described above for downwash plume height, oz and oy, gave
the best results based on a comparison of the frequency distribution of
calculated values and the calculated means with the corresponding observed
data. A listing of some of the major options that were tested, along with
what the results were in terms of the mean and peak SO2 concentrations
estimated at the three monitoring stations, is shown in Table 17. Although
more options for representing the plume height and dimensions could be tested,
it was felt that in view of uncertainty about what the actual plume charac-
teristics are, the final results obtained are as representative as can
be reasonably expected. The relatively large difference between model
estimates and measured concentrations, which differ by as much as a factor of
three, indicates the nature of the uncertainty inherent in the model estimates.
The model variations which were tested in this study represent only a
few of the many variations which could be tested. Although model calculations
are generally not sensitive to wind profile parameters, power laws which are
a function of stability are often introduced in the Gaussian model to increase
the wind speed used to estimate plume transport and downwash effects. The
t ether sonde data could be used to suggest profile parameters which are
particularly related to the monitoring site winds in estimating wind speeds
at plume height. The tethersonde data suggest, at a quick glance, that the
wind speed is much more variable at plume height than at ground level.
Future study of this relationship might lead to improved model parameters.
The McElroy-Pooler parameters were used for az because these values were
determined for 1-hour averages over rougher terrain than the Pasquill-Gif ford
or Brookhaven parameters. As more data from dispersion tests in complex
terrain become available, other sets of oz parameters may be proposed which
could also be tested. There is room for improvement in the model parameters
as the results obtained so far clearly show. Also, since each stack was
represented as a separate virtual point source using the equation stated in
the Kahului model development to estimate the virtual distance, alternative
parameters could be considered. The number of model modifications which
40
-------
TABLE 17. MEAN AND MAXIMUM HOURLY S02 CONCENTRATIONS COMPUTED
USING VARIOUS MODEL PARAMETERS AT THREE KAHE MONITORING SITES
FOR THE 4-MONTH MONITORING PERIOD
Model Calculations (yg/m3)
Monitoring
Site
1
2
3
Statistic
Mean
Peak
Mean
Peak
Mean
Peak
Parameters
Oy =
N =
Measured fj =
80
2050
402
3680
20
529
K
1
1
0
352
17,130
1,006
94,320
48
3,469
K
1
0
1
10
1,736
6
3,453
39
5,135
X
1
.75
.25
35
3,783
253
14,490
39
4,705
X
3
.75
.25
166
4,687
922
9,514
45
2,305
X
2
.75
.25
119
5,041
652
9,766
42
2,434
X
2
.67
.33
97
3,306
496
8,488
42
2,459
Notes on parameters:
Oy: K means the Kahului relationay =(-0.0441 +0.0104O.)x
X means CFy =x p.
Oz = N b xl (see Table 13)
Plume height for downwash effect is
H = f!(HE) + f2(HX)
where HE = the Briggs plume height for downwash effects
HX = the Briggs plume height as given by the BEHO72 subroutine
could be tested is almost endless. The variations selected in this study are
based on the experience of the modelers who performed the work in knowing how
various parameters will affect the model calculations. The most opportune
parameters were selected somewhat intuitively and somewhat with an eye on how
the model results were responding.
VERIFICATION OF MODEL PERFORMANCE AT KAHE POINT
Many variations of the model were investigated to determine the model
parameters which produce the best agreement between measured and observed S0_
concentrations. The 1-hour-average observed data have been presented in
Table 9. The comparable calculated values are presented in Table 18. The
calculated means for the three monitoring site locations are 97, 496, and 42
yg/m3 (Stations 1, 2, and 3 respectively). These values compare with the
observed means of 80, 402, and 20 ug/m3. The calculated minus measured
difference as a percent of the calculated mean was 18 percent for Station 1,
23 percent for Station 2, and 52 percent for Station 3 for an average over-
estimate of 31 percent.
41
-------
TABLE 18. HOURLY CONCENTRATIONS - CALCULATED BASE CASE
CUMULATIVE FREQUENCY DISTRIBUTION OF COMMUTED S02 CONCENTRATIONS (JJG/CU.M)
TH.E RECEPTOR
HIGH
2ND
99. 5
9y.O
95.0
90.0
UO.O
10.0
60. 0
50.0
40.0
10.0
20.0
10.0
S . 0
t .0
0.5
LOrt
1
3306.000
3145.000
1334.309
976.762
565.760
314.700
106.800
1 .647
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1
U48B.OOO
8483.000
6260.860
5682.021
3469.401
1983.000
377.400
12.950
0.356
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
3
2459.000
2265.000
992.967
646.428
245.670
143. 160
11.720
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
u.ooo
0.000
0.000
4
4738.000
4293.000
2355.898
1490.865
589.950
300.860
32.736
0.036
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
5
1875.000
16H4.000
748.206
473.870
159.830
87.774
6.582
0.000
0.000
0.000
0.000
0.000
0.000
' 0.000
0.000
0.000
0.000
o.ooo
6
1631.000
1383.000
603.273
383.642
133.900
74.504
5.588
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1310
1085
461
292
104
59
4
0
0
0
0
0
0
0
0
0
0
0
7
.000
.000
.245
.068
.360
.296
.082
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
8
3202.000
1887.000
1107.385
672.868
120.760
23.718
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
9
3202.000
1882.000
1107.385
672.868
120.760
23.718
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
10
2718.000
2697.000
1361.853
789.897
271.580
84.794
1 .557
0.000
0.000
0.000
0.000
0.000
0.000
o.ooo
0.000
0.000
0.000
0.000
(CONTINUED)
-------
TABLE 18. (CONTINUED)
CUMULATIVE KHCOUENCY DISTRIBUTION OF COMPUTED S02 CONCENTRATIONS (JJG/CU.M)
riLE
11
12
13
14
RECEPTOR
15 16
17
HIGH 2279.000 1868.000 2255.000 1910.000 1470.000 1215.000 935.000
2NO 2056.000 1755.000 1936.000 1653.000 1409.000 1063.000 822.800
99. b 1151.299 964.228 747.673 596.663 617.604 376.003 279.136
99.0 609.078 509.854 404,069 337.195 354.168 195.844 159.248
95.0 190.120 162.520 131.900 104.330 92.557 61.747 49.289
90.0 50.736 42.900 61.754 50.034 28.376 36.672 29.628
80.0 0.199 0.021 1.790 1.423 0.000 1.697 1.124
70.0
60.0
50. 0
40.0
30.0
20 .0
10.0
5.0
1.0
o.s
[.OH
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
o.ouo
0.000
0.000
0.000
0.000 .
0.000
0.000
0.000
o.ooo
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
18
19
20
635.500 728.100 6506.000
550.300 708.900 5248.000
215.154 273.565 3063.722
128.272 172.592 2012.000
42.944 53.901 931.740
18.800 16.982 356.340
0.016 0.000 19.380
0.000 " ""
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
O.UOl
0.000
o.ooo
0 .000
0.000
0.000
0 . 000
0.000
0.000
0. 000
0.000
(CONTINUED)
-------
TABLE 18. (CONCLUDED)
CUMULATIVE FREQUENCY DISTRIBUTION OF COMPUTED S02 CONCENTRATIONS yjG/CU.M)
TIL
HIG
E
H
210
99.
99.
9b.
90.
BO.
70.
60.
50.
40.
30.
?0.
10.
5.
1 .
0.
LJ
S
0
0
0
0
0
0
0
0
0
0
0
0
0
S
III
78B4.
4893.
3935.
3384.
Ib03.
610.
26.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
21
00011500
00010230
656
244
101
120
610
002
000
000
000
000
000
000
000
000
000
000
832B
7475
3735
881
5
0
0
0
0
0
0
0
0
0
0
0
22
.00016410.
.00010480.
.038
.301
.102
.558
.439
.008
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
3086.
2695.
92 1.
26.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
23
000
000
347
042
180
838
093
000
000
000
000
000
000
000
000
000
000
000
5388.
5237.
3821.
3256.
1813,
1007,
65,
1.
0,
0,
0,
0,
0,
0,
0,
0
0,
0,
24
,000
,000
,627
,440
,501
,200
.832
,475
.001
.000
.000
.000
.000
.000
.000
.000
.000
.000
3780
3388
2164
1757
893
487
84
2
0
0
0
0
0
0
0
0
0
0
RECEPTOR
25
.000
.000
.433
.467
.690
.680
.654
.609
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
-------
The distribution of the hourly values can be compared in Tables 18 and
19 The highest calculated concentrations at Stations 1 and 2 as compared to
measured concentrations are high by factors of 1.6 and 2.3, respectively. At
Station 3 the highest calculated value is too high by a factor of 4.6. The
calculated-to-observed ratios are maintained fairly well though the distribu-
tion goes down to the 70th percentile where the calculated values drop off
sharply. Therefore control measures designed to meet criteria based on these
results will favor the safe side from a health-effects point of view, i.e.,will
overestimate. The comparison of calculated-to-observed concentrations
throughout the distribution, from highest to lowest, was considered when the
model modifications were being made. Attempts to distribute the calculated
concentrations over a wider range by altering the model parameters were not
successful. The distributions as shown were considered the best possible
within the period of model experimentation which was available.
The model performance capabilities of the Kane Point location are
summarized by the results presented in Table 19. It may be seen that the
calculated and measured values for the mean and second highest 3-hour and 24-
hour concentrations show good agreement at Station 1, in that the ratio of
calculated-to-measured values ranges from 0.75 to 1.21. At Station 2 the
ratio of calculated-to-measured concentration of SO2 is 1.23 and 1.09 for the
mean and 24-hour periods, respectively. For the 3-hour period the ratio is
2 00 At Station 3 the ratios are higher varying from 2.10 for the mean to
2 47 for the second highest 24-hour period to 2.99 for the second highest 3-
hour period. It may be noted that the mean ratio of calculated-to-measured
values is near 1.5 averaged over the three stations for the mean and the
second highest 24-hour estimates. For the second highest 3-hour estimate,
the mean ratio for the three stations is about 2.0. On the average it may be
expected that the model overestimates the air quality standards of interest
by a factor of 1.5 to 2. Under some conditions, it may underestimate slightly
or overestimate by as much as a factor of three based on the results observed
at the three monitoring stations.
TABLE 19. RATIOS OF MEASURED-TO-CALCULATED S02 CONCENTRATIONS (yg/m3)
AT THE KAHE POINT SITE
Second Highest
Monitoring
Site Measured
1 80
2 402
3 20
Mean 167
Mean
Calculated
97
496
42
212
C/M
1.21
1.23
2.10
1.51
Measured
582
2054
102
913
24 -Hour
Calculated
435
2240
252
976
C/M
0.75
1 09
2.47
1.44
Second Highest
Measured
1237
3223
354
1605
3 -Hour
Calculated
1320
6450
1060
2943
C/M
1.07
2.00
2.99
2.02
45
-------
SECTION 6
MODELING RESULTS AND EVALUATION OF STRATEGIES
KAHULUI RESULTS
The computations made for validation indicate the results from the
model are generally reliable estimates of the expected concentrations at the
monitoring station sites. When running the model, 30 receptor locations were
selected to represent the effects of the power plant emissions. Receptor
locations were concentrated near the power plant (mostly within 2 km), but a
few were spotted up to 7 km away. The configuration of the receptor array is
indicated in Figure 7. Receptor locations 16, 17, and 18 represent the
locations of monitoring Stations 1, 2, and 3, respectively.
The receptor locations were selected to represent the maximum concen-
trations which would be expected under different stability conditions and
wind directions. Some receptors represent locations where exposure is
particularly sensitive, e.g., the nearby shopping center and residential
areas.
The model computations show the mean values of the hourly, 3-hourly, and
24-hourly averages of SO2 over the 6-month data period. The highest mean
values were estimated for receptor locations 16, 19, 17, and 18, in that
order. Table 20 lists the receptors where the model computations indicate
the air quality standards are exceeded. Locations 17 and 18 (which correspond
to monitoring Stations 2 and 3) have also been included.
The highest 3-hour and 24-hour concentrations and the estimated fre-
quencies with which the 3-hour and 24-hour Federal SC>2 standards are exceeded
are listed in Table 20. The percent frequency was determined from calculations
for the 156-day monitoring period.
The mean of the hourly values and the mean of the values calculated
every third hour from the same data set correspond very well. The highest
value in the distribution of the 3-hour averages was always lower than the
highest value from the every-third-hour calculations. The root mean absolute
difference between the highest 3-hour mean and the highest third-hour value
at the six locations listed in Table 20 is 874 pg/m3 or 33 percent of the
six-station mean calculated for every third hour of 2,620 yg/m3. This
difference is representative of the difference of the second-highest estimate
as well as the highest estimate. The mean absolute difference between highest
24-hour concentrations calculated using every-hour and every-third-hour data
for the same six stations is 97 ug/m3 or 12 percent of the six-station mean
46
-------
Kahului Airport
10
1 km
Figure 7. Receptor locations near the Kahului power plant.
-------
TABLE 20. RECEPTOR LOCATIONS WHICH MAY EXCEED AIR QUALITY
STANDARDS WITHOUT USE OF CONTROL STRATEGIES
Receptor
Number
5
16
17
18
19
20
Federal
Standard
3 -Hour
Average
14
531
96
63
411
52
80
/m3)
Every
3rd
Hour
12
538
93
61
402
46
80
Highest Value (yg/irr)
-Hour
Average
1850
2850
575
361
2770
2070
1300
Every
3rd
Hour
2926
3584
944
504
4470
3295
_
24 -Hour
Estimated Frequency
of Federal Standard
Average Based Exceedances (°o)
24 -Hour
Average
379
1*70
312
iSs
1620
544
365
on ara nour
Data
632
1530
352
177
1830
413
_
3 -Hour
(1)
9
0
0
7
(1)
_
24 -Hour
(1)
68
0
0
55
(1)
_
(1) Only the highest value exceeded air quality standard.
calculated for every third hour of 822 yg/m. Although the use of every-
third-hour values results in over estimation of the high 3-hour concentra-
tions, there was reasonable agreement between 24-hour concentrations using
every-hour and every-third-hour data. The results show that the 24-hour
concentrations are four times the standard of 365 pg/m3 while the 3-hour
concentrations are only two times the standard of 1,300 yg/m3. Since the 24-
hour concentration can be accurately estimated by every-third-hour data, and
since the 24-hour concentration is the critical one to limit by controls, it
was prudent for expediency and economy to make additional model runs for
evaluating control strategies using every-third-hour data. Strategies which
could be evaluated using emission alterations to modify the file of present
hourly concentrations were evaluated using hourly data. New model runs were
required to evalute strategies based on new stack heights. The control
strategy evaluation results proved that the 24-hour concentration was indeed
the critical limiting condition requiring the greatest degree of control.
Results for 3-hour concentrations based on every-third-hour input data can be
considered to be on the order of 33 percent too high based on the comparisons
between 3-hour and third-hour maximum concentrations. Subsequent calculations
to evalute the effectiveness of extended stack height control strategies were
made using every-third-hour data from the 6-month data.
The model was also run using only the data available from the airport
observations for meteorological input. This was done to evaluate how well
the 6-month monitoring period represented conditions over a full year. The
evaluation compared the annual average and distribution of air quality
levels at 30 sites based on 12 months of airport data with the same charac-
teristics based on 6 months of airport data. The 6-month period corresponded
to the monitoring period, so that the comparison to 12-month data could be
easily related to the other sets of input data available for the same period.
48
-------
The means and the distributions of the calculations using 6 months of airport
data correspond very closely to the calculations using 12 months of input.
It can be assumed that the closeness occurs because the 6-month data period
represents the full range of meteorological conditions which occur over the
course of a full year. A comparison between the statistics (mean and
frequency distribution) of hourly concentrations calculated using 6 months of
airport meteorological data is shown in Table 21 for six locations. About 60
percent of the compared concentrations are within 10 percent of each other.
The largest percentage differences occur when the hourly concentrations are
below 80 yg/m3. A comparison was made between calculations using airport
meteorological data and calculations using monitoring site data for the same
6-month period. The concentrations calculated using the airport data were
generally lower by a factor of two. The reasons for these differences lie in
differences between the two data sets since the model is the same in both
cases. An examination of the data sets over selected periods showed that
the wind speed observed at the airport generally averaged 50 percent greater
than the values observed at Station 2 (the value selected from the monitoring
stations). Station 2 data still showed better agreement with airport wind
data than the data from Stations 1 and 3, which appeared by visual comparison
of measurements to differ by an even greater amount. Another difference is
due to the method by which the standard deviation of wind direction (oe) is
estimated from airport data. Table 22 lists the mean and the standard
deviation of 0e observed at Station 3 when each of the six Pasquill stability
classes was observed at the airport. The mean value found for each class was
used as the model input value for O0 corresponding to that class. The fact
that the actual range of oe associated with each stability class is quite
large (e.g., the standard deviation is 50 percent or more of the mean as
shown in Table 22) indicates that the use of an average value introduced an
underestimate of calculated concentrations.*
\
STRATEGIES EVALUATED AT KAHULUI
To determine what abatement measures might be employed to bring the
power plant emissions down to a level which would not cause ambient concen-
trations to exceed the S02 air quality standards, the following eight
abatement strategies were evaluated:
1. Reduce sulfur emissions by (1) 50%, (2) 75%, and (3) 90%;
2. Increase the stack heights to (4) 38m, (5) 57 m, and (6) 91 m;
3. Increase stack height to 57 m (2.5 times building height) and
limit fuel sulfur content to (7) 1.0% or (8) 0.5%.
* This is because OQ is in the denominator of the equation used to calculate
concentration. The average of concentrations calculated using
-------
TABLE 21. COMPARISONS OF CALCULATED CONCENTRATION STATISTICS
FROM 6 MONTHS AND 12 MONTHS OF DATA
Receptor
5
16
17
18
19
20
Months
in
Data
Period
6
12
6
12
6
12
6
12
6
12
6
12
Mean
(ug/m3
9
7
275
273
34
45
31
36
156
192
24
23
Standard
Deviation
(yg/m3)
66
77
325
344
81
92
92
91
287
321
171
ISO
Highest
Concen-
tration
(nq/m3)
1273
1723
2110
2499
731
839
1879
1672
2353
2858
3344
3325
Concentration (yg/m ) at Percents'.e
99.5
460
355
1345
1540
537
491
445
551
1434
1531
1116
1179
99
161
162
1172
1312
337
441
339
436
1223
1321
764
663
95
0
0
930
947
194
230
144
159
7S8
368
6
35
90
0
0
742
769
117
153
34
102
52S
670
0
0
SO
0
0
531
572
41
78
38
54
266
374
0
0
70
0
0
337
394
11
24
14
23
113
164
0
0
60
0
0
269
242
3
5
3
5
52
71
0
0
TABLE 22. MEAN AND STANDARD DEVIATION OF OQ OBSERVED AT MONITORING
STATION 3 FOR CORRESPONDING PASQUILL STABILITY CLASS
OBSERVATIONS AT THE KAHULUI AIRPORT
Stability
Class
A
B
C
D
E
F
All
Co (degrees)
Mean.,
(UQ/m3)
18.1
29.8
19.7
19.2
23.9
32.6
21.5
Standard Deviation
(M
3.8
17.8
11.6
10.2
16.5
25.0
14.4
Number of
Cases
3
96
765
2039
475
378
3761
50
-------
Sulfur emissions can be reduced by reducing the sulfur content of the
fuel or by using a scrubber. A 50-percent reduction in sulfur emissions
would correspond to reducing fuel sulfur to 1 percent or using a SO-porrent
efficient scrubber; a 75-percent reduction in sulfur emissions would corres-
pond to using 0.5 percent sulfur fuel or a 75-percent efficient scrubber.
Reducing emissions by 90 percent was assumed to be possible using a 90-percent
efficient scrubber.
The stack heights were selected on the following bases:
1. 38m- to raise the stack to more than 1.5 times the building
height in order to assure that the plume is above the aerodynamic
cavity caused by the power plant building.
2. 56 m - to raise the stack to 2.5 times the building height to
avoid all downwash effects associated with the building.
3. 91 m - to raise the stack to a level high enough to prevent
significant ground-level concentrations in the vicinity
of the stack.
Table 23 shows the results of the strategies evaluated. As previously
mentioned, the strategies were evaluated using meteorological data from the
monitoring stations which cover a 6-month period. The highest 1-hour calcu-
lations do not relate directly to an air quality standard but are displayed
to indicate the effectiveness of the control strategies. The concentration
at the 99.5 percentile is indicative of the probability that the standard
will be exceeded more than once in the course of a year. It is an interpo-
lated value which, for 24-hour concentrations, is higher than the second-
highest concentration, since the data period consists of 156 days. The
annual mean is represented by the mean concentrations over the 6-month period.
The data presented in Table 20 indicate that if air quality standards at
receptor 16 (monitoring Station 1) and receptor 19 can be met, the standards
would be met at any of the receptor sites modeled. The strategies which
bring the calculated concentrations below the primary air quality standards
at receptors 16 and 19 are strategy 3 (90-percent reduction of sulfur
emissions), strategy 6 (stack 91 m tall), and strategy 8 (stack 57 m tall or
2.5 times the building height and fuel sulfur limited to 0.5 percent).
At receptors 16 and 19, the annual mean is above the primary standard
for strategies 1, 2, 4, and 5. Strategy 7 results in an annual mean above
the primary standard at receptor 16. The highest 3-hour average (estimated
by a single concentration calculated every third hour for strategies 4, 5, 6,
7, and 8) exceeds the Federal standard at the same receptors using strategies
1 and 4. Strategy 2 exceeds the 24-hour standard at receptor 19, and strategy
5 exceeds the 24-hour standard at receptor 16.
KAHE POINT RESULTS
Receptor coordinates were specified for locations in the vicinity of the
Kahe Point plant. The receptor locations are indicated in Figure 8 with the
51
-------
TABLE 23. EXPECTED CONCENTRATIONS USING ABATEMENT STRATEGIES
AT THE KAHULUI PLANT
An mini
Receptor -au' -gy Moan
luc'm1)
1 1
2
3
4
5
G
7
H
16 1
2
3
4
5
G
7
8
17 1
2
3
4
5
G
7
8
8 1
2
3
4
5
G
7
n
19 1
2
3
4
5
r.
7
R
20 1
2
•)
4
5
G
7
H
_
3
1
3
1
21
1
0
2G5
U3
51
?94
1G9
9
84
42
4H
2-1
9
77
511
4
29
14
31
15
C,
5G
44
4
2f
1 1
206
103
41
I'll
ID1
r,
51
jr.
2(<
\ i
r,
If.
9
1
r>
2
,-on±trn,
2511
1255
502
1 095
5r.H
58
27')
12')
22 19
1 1JO
4-17
;.'42
I77C,
HOG
RHR
444
G51
3 >r>
1 10
7H!I
717
675
36H
mi
25')
1 )0
52
550
r,4f.
.159
.71
llf.
2411
1 !OG
4H1
K.70
1 1 fiH
Ml1.
5H4
292
209r,
1047
41')
I4'.4
101M
CiH4
r.09
254
190
9'i
Ill
1 w
70
12
15
17
7 l'i
1,18
1 17
Hr,5
f. !rj
JHH
11 '
I5(,
15f.
,'4
U
>r.1
2 It
279
IK.
58
79
19
If.
Ill
14"
1 > 1
7 1
17
>t|0
40ri
11. '
f,7')
lit')
I'M)
!44
I.1 >
!74
1 17
'.4
2H'
141
f.H
7(1
tr>
1-1 lull i
9!5
4G2
1111.
-
-
-
-
-
14 >r.
/I '
-115
1 17H
1010
4111
515
'OH
>HH
. j.|
5H
1 lr.
l')4
114
197
98
1HO
911
11.
11')
1 1 '
•47
If.l.
ni
1 Wj
fi')2
' '7
1 l-lu
'i ' t
17;
11 !
!(>(,
.'74
MH
'07
H5I
(.(,4
4(.7
1 1 '
|(.(,
i nm.uiitration
(..•l/m1)
1-Houi 24-Moiir
2112
1056
422
-
>53
-
127
G1
.'102
I0r:l
4 '0
1H74
Kif.7
717
834
417
472
2 If.
94
(.51
5HD
454
.'"0
1-15
252
12(.
50
142
498
3>f.
.'49
1?4
2>15
HIM
447
1 r.f.9
1 1 40
r.r.5
•570
>H5
1035
nr.9
1411
1-1 ".
•)HO
r. :i
4')n
/45
IK)
5'T
23
-
t2
-
Id
).
7 '(I
U.r.
!4d
'114
Vifl
.'82
J99
150
145
73
'9
243
209
K,2
114
52
77
1H
15
490
1 17
1 14
i. 8
34
7_'0
1GO
144
f.21"'
111.
1 59
;on
104
17 W
1 II.
.(•
1 17
105
(.0
'<-•
il-
l-Hour
800
400
160
-
-
-
-
-
1395
G9H
279
1 !70
934
4R3
4G7
234
277
Mf!
55
403
168
305
]H4
92
175
R«
35
!95
278
241
119
70
1 125
Mi 2
165
1110
800
tfi5
400
JOO
21J
495
198
774
5f,7
320
2H4
14 '
Concont ration
.iL 99.5 Porccntilu *
1-Hour 24-Hour
327
164
65
98
31
1
15
R
1377
GH9
277
1431
1048
541
524
262
331
1GG
Gfi
485
443
34fi
222
111
194
97
39
1.10
283
233
142
71
1405
70.1
-'81
1198
904
424
45J
7 -'6
9'10
4T.7
187
7r,4
57(,
3JG
J8H
144
132
06
26
91
40
18
20
r
731
365
14G
823
604
283
302
151
147
4
29
245
214
166
107
"4
77
39
15
144
139
166
70
35
740
37C
14S
i.',7
41.'
IGG
21C-
108
J-10
I-V
48
15')
11 »
GJ
50
2H
3-Hour
304
152
61
-
-
-
-
-
1334
567
22?
1142
858
447
429
214
234
117
47
362
318
258
159
-0
145
72
29
244
213
174
106
53
1150
575
230
"••»
714
3J1
157
17t»
MS
349
140
48(<
144
!08
172
KG
52
-------
Figure 8. Receptor site locations - Kahe Point
53
-------
corresponding identification numbers in Tables 24 and 25. Receptors 1, 2,
and 3 correspond to monitoring Stations 1, 2, and 3. The receptor locations
were selected at places where maximum concentrations would be expected and
that were representative of residential areas. Receptors 8 and 9 represent
results for locations on the ridge behind the plant. The effectiveness of
the model in representing the impact of the plume on this ridge has not been
validated, since no monitoring data were collected on the ridge and attempts
to observe the plume remotely with the COSPEC instrument at positions other
than directly above the highway west of the plant were not successful. To a
lesser extent, results computed for the receptors to the southeast of the
plant also represent the unvalidated influences of flow over complex terrain.
The concentrations estimated for receptors 10 through 19 are in this category.
The effects of the plume over the water just off the coast were not estimated,
although technically this also could be considered ambient air. The emphasis
is on locations where people are likely to be exposed for significant periods
of 3 to 24 hours.
The calculated 3-hour average concentrations shown in Table 24 indicate
that the 1,300 yg/m3 standard for a 3-hour average would be exceeded by the
second highest 3-hour period at 10 locations, all of which are within
approximately 1 km of the power plant.
The calculated 24-hour average concentrations are presented in Table 25.
The 24-hour average standard of 365 pg/m3 is exceeded by the second-highest
value at 11 of the 25 receptor sites listed. Again, the excursions all occur
at the receptors within 1 km of the power plant. It can be expected that the
actual occurrences of excursions may not be as drastic as indicated in Tables
24 and 25 due to the fact that the verification results showed the model
overestimates these short-term concentrations.
All of the locations where standards are exceeded are along the coast to
the west of the plant or on the ridge above the plant. The locations not
exceeding Federal standards are those to the north of location 4 and those to
the south and southeast of location 10. All of the stations near the coast,
except station 10, are within or on the edge of the bowl-shaped terrain which
encompasses the plant. The maximum impact of the plume is estimated at
locations 2, 21, 22, and 24. These locations are almost directly across the
street from the plant. Locations 2 and 22 are directly downwind of the low
stacks when they are affected by the prevailing trade winds. The extent to
which locations 20 and 4 are also sites of major effects from the plant is
partly dependent on the extent to which the frequency of wind directions
observed at location 2 represents the actual frequency of wind directions of
the plume. The 2 weeks of daytime tethersonde wind data collected in January
suggest that there is a pronounced tendency of the northeasterly trade winds
to turn northward around the southern tip of the Waianae Mountain Range. The
wind direction observations at Station 2, which were used in the model
estimates, strongly reflect this tendency. In any case, the most critical
estimates which are predicted for locations 2 and 22 due to the effect of the
northeast trade winds on the four lowest stacks of the plant are reasonable.
The reduction of these estimates by control strategies provides a reasonable
control objective.
54
-------
TABLE 24. 3-HOUR AVERAGE CONCENTRATIONS - CALCULATED BASE CASE
CUMULATIVE FREQUENCY DISTRIBUTION OF COMPUTED 502 CONCENTRATION'S (JJG/CU.M)
PERCEHTIbE RECEPTOR
123456789 10
OT
HI3H
2ND
99.5
99.0
95.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
5.0
1 .0
0.5
10.4
1470.000
1320.000
1110.151
868.883
484.000
316.900
176.600
70.370
15.200
0.037
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
7240
6450
5364
4513
2849
1BBO
705
176
8
0
0
0
0
0
0
0
0
0
.000
.000
.351
.506
.000
.000
.400
.300
.514
.377
.006
.000
.000
.000
.000
.000
.000
.000
1330.000
1060.000
778.480
542.740
227. 3bO
129.600
51.280
7.311
0.017
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
2200.000
2120.000
1665.801
1440.000
556.450
297.300
96.800
19.370
0.600
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
973.000
835.000
595.351
372.671
148.450
79.330
31.300
5.U87
0.003
0.000
0.000
0.000
0.000
0.000
0.000
o.ooo
0.000
0.000
823.000
708.000
486.450
312.290
125.000
68.340
28.260
3.839
0.001
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
642.000
561.000
369.290
248.251
98.695
54.000
21.560
2.865
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1520.000
1300.000
719. 382
519.701
169.600
53.6BO
1.372
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1520.000
1300.000
719.382
519.701
169.800
53.680
1.372
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1720.000
1450.000
1054.351
784.454
245.050
116.900
29.040
3.621
0.013
0.000
0.000
0.000
0.000
O.UOO
0.000
0.000
0.000
0.000
(CONTINUED)
-------
TABLE 24. (CONTINUED)
CUMULATIVE FREQUENCE DISTRIBUTION OK COMPUTED SOI CONCENTRATIONS (JJG/CU.M)
Ui
TILE RECEPTOR
HIGH
2 -JO
99. 5
99.0
95.0
90.0
80.0
70.0
60.0
SO.O
40.0
JO.O
20.0
10.0
S.O
I .0
O.S
LO.-J
11
1310.000
1 1 10.000
128.725
602.502
lui .000
B2.890
15.220
1 .067
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
O.OOQ
12
1130.000
879.000
687.740
503.320
154,000
70.900
11.260
0.091
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
13
13RO.OOO
use. ooo
610.206
461 .580
118.450
61.630
23.180
3.224
0.010
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
14
1 150.000
743.000
487.221
350.771
94.370
48.640
19.800
2.794
0.005
0.000
0.000
0.000
0.000
0.000
0.000
• o.ooo
0.000
0.000
15
1010.000
630.000
435.221
326. 450
90.900
42.090
6.014
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
16
724.000
479.000
288.610
192.931
59.470
31.860
13.260
2.407
0.003
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
547.
365.
218.
151.
46.
24.
10.
1.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
17
000
000
160
160
B45
650
580
771
003
000
000
000
000
000
000
000
000
000
18
283.000
249.000
176.091
148.450
41.250
21.580
5.768
0.097
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
19
514.000
382.000
189.015
151.160
54.485
22.320
1.984
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
20
3760.000
2920.000
2217.251
1779.004
891 .900
344.900
106.600
12.760
0.145
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
(CONTINUED)
-------
TABLE 24. (CONCLUDED)
CUMULATIVE: FREQUENCY DISTRIBUTION OF COMPUTED so: CONCENTRATIONS UJG/CU.MI
TILE:
HIGH
IHO
99
99
95
90
eo
70
bO
50
40
30
20
ID
5
1
0
.5
.0
.0
.0
.0
.3
.0
.0
.0
.0
.0
.0
.0
.0
.5
LCM
3S70.
3730.
3270.
2668.
1584.
690.
158.
n.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
21
000
000
000
701
500
900
BOO
540
067
000
000
000
000
000
000
000
000
000
8570
8400
7S65
6864
3033
1440
146
4
0
0
0
0
0
0
0
0
0
0
22
.000
.000
.952
.111
.500
.000
.800
.190
.039
.000
.000
.000
.000
.000
.000
.000
.000
.000
6340.
3490.
2737.
2312.
857.
201).
2.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
93
000
000
705
900
too
900
180
079
001
000
000
000
000
000
000
000
000
000
3620
3600
3071
2468
1479
913
350
52
V
0
0
0
0
0
0
0
0
0
24
.000
.000
.602
.701
.000
.700
.000
.110
.744
.016
.000
.000
.000
.000
.000
.000
.000
.000
1950
1880
1520
1418
70S
493
199
59
6
0
0
0
0
0
0
0
0
0
RECEPTOR
25
.000
.000
.151
.701
.750
.400
.000
.500
.302
.022
.000
.000
.000
.000
.000
.000
.000
.000
-------
TABLE 25. 24-HOUR AVERAGE CONCENTRATIONS - CALCULATED BASE CASE
CUMULATIVE FREQUENCY DISTRIBUTION OF COMPUTED S02 CONCENTRATIONS (JJG/CU.M)
ui
oo
IT1LE RECEPTOR
HIGH
2ND
99.5
44. 3
95.0
90.0
RO.O
70.0
60.0
50.0
40.0
JO.O
20.0
10.0
5.0
1.0
0.5
LOrf
1
5H5.000
436.000
•144.85(1
42b.nO
328.900
236.500
165.600
122.700
89. 120
61 . 100
33.680
13.490
0.677
0.000
0.000
0.000
0.000
0.000
3060
2240
2S63
220B
171B
1522
1076
715
3B1
159
34
2
0
0
0
0
0
0
2
.000
.000
.904
.500
.000
.000
.000
.000
.000
.bOO
.300
.9UO
.220
.002
.000
.000
.000
.000
3
343.000
252.000
2H7.945
2-15.700
189.850
129.700
60.440
33.920
22.460
15.700
10.920
J.931
I .2
-------
TABLE 25. (CONTINUED)
CUMULATIVE FREQUENCY DISTRIBUTION OF COMPUTED S02 CONCENTRATIONS (JJG/CU.H)
ui
VO
TILE RECEPTOR
HIGH
2ND
99.5
99.0
95.0
90.0
80.0
70.0
60.0
bO.O
40.0
30.0
20.3
10.0
5.0
1 .0
0.5
LDrf
11
250.000
230.000
237.900
221.810
146.300
94.020
47.420
30.690
19.200
1 1.950
5.956
1.2H7
0.114
0.000
0.000
0.000
0.000
O.OUO
12
210.000
202.000
205.160
193.390
1 15.9bO
9U.2bO
39.780
22.820
14.300
10.. 3 50
3.016
0.783
0.001
0.000
0.000
0.000
0.000
0.000
13
284.000
174.000
217.450
169.800
95.045
57.230
36.140
24.300
16.460
1 1.800
5.406
2.830
O.B45
0.000
0.000
0.000
0.000
0.000
14
236.000
13U.OOO
176.710
133.800
73.885
46.060
27.980
. 19.600
13.920
9.060
4.544
2.789
0.961
0.000
0.000
0.000
0.000
0.000
15
146.000
129.000
135.715
128.790
79.530
61 .050
24.660
17.500
8.462
3.740
1 . 33d
0.045
0.000
0.000
0.000
0.000
0.000
o.ooo
16
150.000
80.100
107.711
78.588
51.690
29.120
17.500
12.170
8.902
5.940
3.486
2.076
0.576
0.000
0.000
0.000
0.000
0.000
114
60
81
b9
38
22
13
10
6
4
3
I
0
0
0
0
0
0
17
.000
.100
.391
.197
.490
.990
.480
.147
.892
.695
.052
.710
.389
.000
.000
.000
.000
.000
18
43.700
42.600
43.035
42.579
34-.390
25.190
11.660
7.151
4.970
3.555
1 .760
0.603
0.000
0.000
0.000
0.000
0.000
0.000
19
80.300
67.000
72.254
66.055
42.410
29.260
15. 160
8. 180
4.578
1.835
0.374
0.000
0.000
0.000
0.000
0.000
0.000
0.000
20
1010.000
769.000
859.961
760.950
540.400
451.800
212.600
100.800
62.900
38.250
30.440
16.410
2.508
0.000
0.000
0.000
0.000
0.000
(CONTINUED)
-------
TABLE 25. (CONCLUDED)
CUMULATIVE FREOUENCr OISTKIBUTION OF COMPUTED S02 CONCENTRATIONS (JJG/CU.M)
TILE
HIGH
2ND
99.5
99.0
95.0
90. 3
BO.O
70.3
60. 0
50.3
40.0
30.3
20.0
10. 0
5.0
1 .0
0.5
LOW
1870
1260
1500
1278
771
639
410
207
137
87
40
25
3
0
0
0
0
0
21
.000
.000
.953
.500
.250
.800
.BOO
.100
.600
.550
.100
.950
.142
.000
.000
.000
.000
.000
6170
2360
3864
2343
1635
1 124
663
452
290
170
48
6
0
0
0
0
0
0
22
.000
.000
.966
.200
.499
.000
.400
.300
.200
.000
.740
.168
.812
.000
.000
.000
.000
.000
1720
1010
1290
979
701
430
175
72
37
14
2
0
0
0
0
0
0
0
•
•
•
•
•
•
m
m
•
.
•
,
•
•
m
w
m
•
23
000
000
453
130
250
600
600
250
740
150
374
306
008
000
000
000
000
000
1660.
1310.
1448.
1297.
902.
801 .
447 .
303.
18U .
100.
40.
b.
0.
0.
0.
0.
0.
0.
24
000
000
251
400
000
500
200
t>no
600
6SO
•jUO
599
080
000
000
000
000
000
757
740
746
719
431
362
239
165
106
71
32
10
0
0
0
0
0
0
RECEPTOR
25
.000
.000
.715
.210
.950
.900
.800
.600
.800
.800
.540
.830
.428
.000
.000
.000
.000
.000
-------
STRATEGIES EVALUATED AT KAHE POINT
Evaluation of the observed results from Tables 10 and 11 by linear
rollback considerations indicated that the necessary reduction in fuel sulfur
content was 90 percent, i.e., the fuel used would have to be approximately
0.2 percent sulfur. A 90-percent efficient scrubber would allow the plant to
meet the necessary emission reduction and would reduce ambient air concentra-
tions to levels prescribed by the standards.
An alternative strategy was to assume a stack height of 2.5 times the
building height. Tables 26 and 27 show the expected concentrations for the
3-hour and 24-hour averages with the stacks raised to 2.5 times the building
heights. In this strategy, stacks 1 and 2 were modeled as being 87.6 m tall,
stacks 3 and 4 as being 94.0 m tall, and stack 5 as being 115.8 m tall.* with
this strategy the 3-hour secondary standard was exceeded by the second-
highest value at receptors 20, 21, 22, and 24. The 24-hour average standard
was exceeded by the second-highest value at receptor 21. These results
indicate that raising the stacks to 2.5 times the building height is not
sufficient to meet the Federal air quality standards.
Another strategy was evaluated to determine the effect of raising stacks
1 through 4 to 115 m to be equal in height to stack 5 at the Kahe plant.
Significant reductions to the 4-month mean occurred with this stack configura-
tion. The 3-hour and 24-hour averages were also significantly reduced.
However, the second-highest 3-hour concentration estimated at site 21 exceeds
the Federal standard.
One further strategy which was investigated was raising all stacks to a
height 2.5 times the respective building heights combined with limiting the
fuel to a maximum sulfur content of 1.0 percent. This strategy permits all
receptor locations to comply with all Federal standards. Evaluation of
control strategies at Kahe Point is presented in Table 28.
At the onset of this modeling work, data were provided which stated the
height of the stacks to be 42 m for stacks 1 and 2, 45 m for stacks 3
and 4, and 92.7 m for stack 5. These data were input to the model. The
building height data input to the model were derived from the stack
height data by dividing the stack heights by 1.2 for stacks 1 through 4
and by 2 for stack 5, giving building heights of 35 m for stacks 1 and
2, 38 m for stacks 3 and 4, and 46 m for stack 5. If the building
heights for stacks 1 through 4 are all 38 m and these stacks are all 45
m tall, the impact on the model results does not affect the conclusions
of this study.
61
-------
TABLE 26. 3-HOUR AVERAGE CONCENTRATIONS,
CALCULATED STACKS 2.5 x BUILDING HEIGHT
CUMULATIVE FREQUENCY DISTRIBUTION OF COMPUTED SO? CONCENTRATIONS UJG/CU.M)
IMLt RECEPTOR
HI3H
2ND
9y.s
99.0
«5.0
90. 3
60.0
7U.O
oO.O
50.0
40.0
30.0
20.0
10.0
5.0
1 .0
0.3
LQ.1
\
674.000
640.000
b39.045
448.290
225.450
138.800
53.020
15.800
1.590
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
2
861.000
782.000
618.725
416.450
149.200
20.090
2.844
0.416
0.067
0.001
0.000
0.000
0.000
0.000
0.000
0.000
0.000
o.oou
3
642.000
541.000
389.290
283.021
152.000
74.840
2S.840
2.097
0.000
0.000
0.000
0.000
0.000
o.ooo
0.000
0.000
0.000
0.000
4
644.000
633.000
507.457
468.160
268.350
136.600
41.000
4.464
0.036
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
s
588.000
520.000
351.305
241.580
115.700
62.920
24.400
2.067
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
6
535.000
477.000
316.290
214.033
98.340
54.440
21.140
1.6SO
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
o.ooo
0.000
7
457.000
409.000
263.290
175. 480
79.640
44.360
16.260
1.584
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
8
998.000
880.000
541.744
394.061
145.700
39.580
0.717
0.000
0.000
0.000
n.ooo
0.000
0.000
0.000
0.000
0.000
0.000
0.000
9
998.000
8BO.OOO
341 .744
394.061
145.700
39.580
0.717
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
10
737.000
534.000
368.091
291.541
102.350
50.130
10.200
0.367
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
(CONTINUED)
-------
TABLE 26. (CONTINUED)
CUMULATIVE FREQUENCE DISTRIBUTION OF COHPUTLO S02 CONCENTRATIONS (JJG/CU.I4)
ui
rlLE RECEPTOR
1 1 u C*
HIGH
2ND
99.5
y9.o
95. D
90.0
ao. o
70.0
60.0
50.0
40.0
30. U
20.0
10.0
5. 0
1.0
0.5
LDrf
11
617.000
569.000
353.290
317.641
97.435
41.3UO
4.024
0.094
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
12
614.000
4U4.000
372. 305
299.671
105.350
38.370
3.732
0.002
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
906
564
356
251
72
39
12
0
0
0
0
0
0
0
0
0
0
0
13
.000
.000
.106
.701
.835
.430
.020
.869
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
14
820.000
520.000
319.136
215.151
65.185
35.480
10.440
0.674
0.000
0.000
0.000
0.000
0.000
o.ooo
0.000
0.000
0.000
0.000
15
767.000
492.000
314.236
246.740
77.845
36.450
2.846
0.000
0.000
0.000
O.OOU
0.000
0.000
0.000
0.000
0.000
0.000
0.000
16
594.000
3B8.000
232.030
149.030
42. 43S
24.300
9.302
0.701
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
17
469.000
310.000
184.5
-------
TABLE 26. (CONCLUDED)
CUMULATIVE: FREQUENCE DISTRIBUTION or COMPUTED 302 CONCENTRATIONS
en
TILE
HIGH
2ND
99.
«9.
95.
90.
80.
70.
toO.
50.
40.
30.
VO.
10.
5.
1 .
0.
5
5
0
0
0
0
0
0
0
0
0
0
0
0
5
LOrf
1950.
1820.
1610.
1221.
34b.
118.
9.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
21
000
000
000
602
900
700
424
466
001
000
000
000
000
000
000
000
000
QUO
1800
1600
1375
1167
245
62
2
0
0
0
0
0
0
0
0
0
0
0
22
.000
.000
.801
.402
.050
.340
.936
.108
.002
.000
.000
.000
.000
.000
.000
.000
.000
.000
960
896
786
532
67
14
0
0
0
0
0
0
0
0
0
0
0
0
23
.000
.000
.061
.450
.aib
.060
. Ib3
.003
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
1450
1340
714
349
96
32
5
0
0
0
0
0
0
0
0
0
0
0
24
.000
.000
.272
.320
.37b
.370
.302
.453
.021
.000
.000
.000
.000
.000
.000
.000
.1)00
.000
1150
1030
726
4S6
164
82
19
3
0
0
0
0
0
0
0
0
0
0
RECEPTOR
25
.000
.000
.435
.883
.250
. 180
.760
.258
.137
.000
.000
.000
.000
.000
.000
.000
.000
.000
-------
TABLE 27. 24-HOUR AVERAGE CONCENTRATIONS,
CALCULATED STACKS 2.5 X BUILDING HEIGHT
CUMULATIVE FREQUENCY DISTRIBUTION OK COMPUTED S02 CONCENTRATIONS fJJG/CU.M)
TILE RECEPTOR
HIGH
2ND
99. 5
99.0
95.0
90.0
UU.O
7U.O
bO.O
50.0
40.0
30.0
20.3
10.0
i.O
1 .0
0.5
I.U4
1
260.000
194.000
220.070
192.530
1SS.300
113.400
70.460
48. 370
78. 120
16.100
11 .120
1.986
0.055
0.000
0.000
0.000
0.000
0.000
2
202.000
176.000
186.270
167.390
94.330
74.4UO
3D. 060
21.170
10.340
3.020
0.337
0.085
0.013
0.000
0.000
0.000
0.000
0.000
3
171.000
121.000
141.540
121 .000
9->.28!>
67.120
35.120
18.970
14.920
11 .000
7.650
3.251
0.700
0.000
o.oon
0.000
0.000
0.000
4
273.000
201. noo
209.690
199.530
158. 750
126.000
67.800
15.910
27.620
I8.4b0
13.420
7.398
1.296
0.000
0.000
0.000
0.000
0.000
146.
109.
123.
108.
84.
59.
27.
15.
12.
9.
5.
3.
0.
0.
0.
0.
0.
0.
5
000
000
615
790
240
060
360
740
320
670
580
700
905
000
000
000
000
000
f>
in. ooo
97.100
110.491
96.827
71.895
50.340
22.900
14.240
10.720
8.690
5.062
2.BB2
0.606
0.000
0.000
0.000
0.000
0.000
7
110.000
82.500
93.363
82.038
57.090
42.900
19.240
11.840
8.82H
7.080
4.604
2.169
0.404
0.000
0.000
0.000
0.000
0.000
8
277.000
775.000
275.790
244.760
83.455
71.360
35.660.
20.340
7.478
1.34-5
0.123
0.000
0.000
0.000
0.000
0.000
0.000
0.000
9
277.000
275.000
275.790
244.760
83.455
71.360
35.660
20.340
7.478
1.345
0.123
0.000
0.000
0.000
0.000
0.000
0.000
0.000
10
145.000
139.000
141 . 370
137.320
73. 100
41 .060
24.660
17.010
12. OHO
5.6KO
2.990
1 . 375
0.053
0.000
0.000
0.000
0.000
0.000
(CONTINUED)
-------
TABLE 27. (CONTINUED)
CUMULATIVE FHEQlltllCY DISTRIBUTION OF COMPUTED S02 CONCENTRATIONS (JJG/CU.M)
IMLE RECEPTOR
HIGH
2ND
99. 5
99.0
9b.O
90.3
80.0
70.0
60.0
so.n
40.0
30.0
20.0
10.0
•j.o
1.0
0.5
I.J.4
11
142.000
lib. 000
126.270
113. ObO
81.035
47.500
25.540
15.260
1 1.060
6.710
1.774
0.4\ 1
0.026
0.000
0.000
0.000
0.000
0.000
12
129.000
111 .000
118.1 10
107.178
70.660
b4.740
28.180
lb.810
9.136
5.245
0.976
0.162
0.000
0.000
0.000
0.000
0.000
0.000
U
184.000
110.000
139.230
108.110
56.005
35.070
21.680
15.420
10.300
6.935
3.232
1.432
0.3S8
0.000
0.000
0.000
O.bOO
0.000
14
167.000
96.900
124. S90
94.330
53.870
32.300
18.480
13.870
9.472
6.120
2.778
1.456
0.267
0.000
0.000
0.000
0.000
0.000
15
111.000
108.000
109.18!.
106.005
67.425
45.540
22.360
13.550
7.556
2.245
O.'j89
0.007
0.000
0.000
0.000
0.000
0.000
0.000
16
123.000
65.500
88.213
62.476
34.410
21 .500
12.320
9. 138
6.796
4.130
2.424
1 .374
0.19)
0.000
0.000
0.000
0.000
0.000
17
97.300
51.400
69.531
48.9b4
26.670
18.240
10.028
7.214
5.412
3.475
2.178
1 . 161
0. lit)
0.000
0.000
0.000
0.000
0.000
18
37.700
37.400
37.519
36.749
30.750
20.840
9.588
5.664
4.126
2.890
1 .158
0.243
0.000
0.000
0.000
0.000
0.000
0.000
19
75.000
59.300
65.502
58.271
40.555
25.U40
13.860
7.043
4.298
1 .275
0.257
0.000
0.000
0.000
0.000
0.000
0.000
0.000
20
3S8.000
327.000
334.245
320.200
240.900
167.900
62.500
37.660
23.720
15.?50
9.992
4.616
0.159
0.000
0.000
0.000
0.000
0.000
(CONTINUED)
-------
TABLE 27. (CONCLUDED)
CUMULATIVE FREOUtN^Y DISTRIBUTION OF COMPUTED SO? CONCENTRATIONS (JJG/CU.M)
PERCMTILE RECEPTOR
21 22 23 24 2b
H13H
2ND
99.5
99.0
95.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
b.O
1 .0
O.S
L3.4
434.000
419.000
424.925
411.230
2U8.650
224.100
98.660
42.870
28.500
12.800
5.900
1.368
0.091
0.000
0.000
0.000
0.000
0.000
365.000
337.000
34H.060
331 .540
224.650
175.30(1
61.200
34.930
19.520
6.200
1 .560
0.291
0.014
0.000
0.000
0.000
0.000
0.000
184.000
173.000
177.345
167.960
101. 61b
70.260
19.880
«. 040
4.824
0.801
0.251
0.018
0.000
0.000
0.000
0.000
0.000
O.OuO
297.000
295.000
295.790
275. 6UO
93.090
IB. 720
27.200
10.336
4.660
2.1U.S
O.f>33
O.OSO
0.000
0.000
0.000
0.000
0.000
0.000
297.000
264.000
277.035
251.400
127.250
HS. 160
44.520
2">.5bO
1 1 .640
5.95b
7.894
O.b83
0.00?
0.000
0.000
0.000
0.000
0.000
-------
TABLE 28. EXPECTED CONCENTRATIONS USING
ABATEMENT STRATEGIES AT KAHE POINT PLANT
'lean Highest
Concentration Concentration
(yg/m3) (uq/m3)
Receptor
1
2
8
10
20
21
22
Strategy* 4-month
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
S
6
62
31
10
40
26
30
314
157
50
22
14
10
13
9
3
22
15
20
29
15
4
17
12
14
35
42
13
51
32
41
141
71
22
60
38
43
268
134
42
46
29
28
24-Hour
377
188
58
260
168
211
1970
985
306
202
131
76
256
128
41
277
178
227
253
126
35
145
110
123
654
327
101
358
218
303
1210
605
187
434
270
331
3980
1990
617
365
218
234
3 -Hour
957
478
147
674
530
646
4260
2130
724
361
506
370
924
462
152
998
608
802
1140
570
172
737
432
464
2420
1210
376
1400
852
1250
1490
1245
387
1950
1188
1600
5520
2760
857
1800
1160
1200
Second Highest
Concentration
(Uq/m3)
24 Hour
281
140
44
194
147
169
1320
660
224
176
114
73
249
124
40
275
169
227
234
117
32
139
88
115
488
244
76
327
212
273
811
406
126
419
264
323
1590
795
236
337
214
224
3 -Hour
947
474
132
640
432
600
3910
1955
645
732
506
326
352
426
130
880
584
711
1010
505
145
534
420
391
1880
940
292
1330
810
1200
2400
1200
373
1820
1110
1480
5420
2710
840
1600
942
1090
99.5 Percentile
Concentration
(uq/m3)
24-Hour
319
160
49
220
155
186
1577
788
256
186
121
74
252
126
40
276
172
227
242
121
33
141
96
118
554
272
36
339
214
235
969
484
150
425
266
326
2534
1267
386
348
216
228
3 -Hour
700
350
111
559
360
507
3265
1632
536
619
371
260
455
223
72
542
333
474
727
364
105
368
248
280
1425
712
222
1064
664
982
2101
1050
327
1610
983
1341
4752
2376
757
1374
848
380
68
(CONTINUED)
-------
TABLE 28. (CONCLUDED)
Highest Second Highest
Concentration Concentration
(pg/m3) (yig/m3)
Receptor Strategy* Concentration 24-Hour 3-Hour 24-Hour 3-Hour
99.5 Percentile
Concentration
(yg/m3)
24-Hour 3-Hour
23
24
25
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
80
40
13
18
11
12
161
81
25
20
13
11
88
44
14
29
19
20
1110
555
172
184
112
122
1070
535
166
297
232
218
489
244
76
197
208
239
3730
1865
634
960
620
664
2400
1200
362
1450
1138
1140
1270
635
195
1150
908
1060
654
327
101
173
108
118
877
438
131
295
192
206
477
238
74
264
192
228
2250
1125
349
896
546
645
2340
1170
360
1340
786
2050
1260
630
188
1030
618
935
834
417
129
177
110
120
953
476
145
296
208
211
482
241
75
277
198
232
1820
910
274
786
484
546
2022
1011
307
714
462
482
1084
542
152
726
468
599
* Strategies:
(1)
(2)
(3)
(4)
(S)
Limit
Limit
fuel to maximum
fuel to maximum
sulfur content of
sulfur content of
1. 0 percent.
0. 5 percent.
Reduce present emissions by 90 percent.
Raise
Raise
stack heights to 2
stack heights to 2
. 5 times building
. S times building
heights.
heights and
limit fuel to
maximum
sulfur content
of
1. 0 percent.
(6)
Raise
all stacks to a height of 115 meters.
69
-------
SECTION 7
FEASIBILITY OF A SUPPLEMENTARY CONTROL SYSTEM
A supplementary control system to control air quality in the vicinity of
a major source of pollution requires that a fairly sophisticated decision
process be instituted at the source which takes measurements of meteorological
conditions and air quality levels into account when determining the operating
level of the source and consequent air pollution emissions. In order for
such a system to be feasible, it must be possible to identify meteorological
conditions which separate acceptable and unacceptable air quality levels for
a given level of emissions.
KAHULUI PLANT
The monitoring data collected over a 6-month period near the Kahului
plant showed that the 3-hour and 24-hour Federal standards for SC^ were
frequently exceeded at Station 1, about 0.4 km from the plant. The other two
stations showed no excursions of the Federal SC>2 standards. The periods of
the highest 24-hour and 3-hour mean concentrations were examined to determine
the nature of the concentration patterns on these days and to identify
significant meteorological parameters which set them apart from other periods.
The 24-hour concentration of S02 at Station 1 exceeded the Federal
standard of 365 pg/m^ on 107 days out of the 156-day monitoring period. Even
if the emissions were reduced by a factor of two so that all measured
concentrations would theoretically be reduced by 50 percent, the standard
would still be exceeded on 80 out of 156 days. If the emissions were reduced
by 75 percent, the measured concentrations were theoretically reduced to a
point where the standard would be exceeded on 23 out of 156 days. This is
perhaps a reasonable number of days on which to reduce emissions further by a
supplementary control system.
In reviewing the pattern of hourly concentrations at Station 1 on the
days when the 24-hour concentration is among the 12 highest values, one finds
a large-to-moderately-high concentration occurs for nearly every hour. This
suggests a strongly persistent wind direction for 24 hours. The range of
wind directions, and also other meteorological parameters (including wind
speed, standard deviation of wind direction, and mixing height) for the 12
days with the highest 24-hour measured SO2 concentrations, are listed in
Table 29. The narrow range of wind directions associated with these days
over a 24-hour period is noteworthy. The range varies from 15 degrees on
March 6 to 53 degrees on July 13; however, 10 of the 12 days had ranges of 33
degrees or less. An example of a possible control strategy based on this
information is to limit the plant load to 60 percent of its maximum capacity
70
-------
on days when the range of hourly mean wind directions is less than 33 degrees.
Assume the plant also burns only 0.5 percent sulfur fuel oil. During the 6-
month monitoring period, there were 55 days out of 156 when the wind direc-
tion range was 33 degrees or less. If the plant emissions were limited to 60
percent of capacity on these days, there would only have been 4 days on which
the monitored 24-hour S(>2 concentrations exceeded 365 yg/m^. On these 4
days the wind direction range exceeded 33 degrees. This is a significant
improvement over 23 days without the load reduction.
A more detailed analysis of meteorological variables is required to
develop sharper criteria than are indicated in the above example for separat-
ing conditions which produce high air quality levels from conditions which do
not. Combinations of meteorological variables may determine conditions in
which the joint occurrence of two or more meteorological variables provides
useful separation categories. For example, none of the days with high SO2
concentrations includes low wind speed or high values of the standard
deviation of wind direction. It is also possible that variables other than
the range of wind direction may better indicate the significance of wind
variability, e.g., variance of wind direction over 24 hours. The results
presented in Table 29 suggest that the meteorological conditions associated
with excursions can be identified.
The ability to forecast the occurrence of meteorological conditions
which result in high SO2 concentrations needs to be tested. The forecast
problem consists of the ability to predict periods when a well-developed,
persistent trade wind will occur with almost no diurnal variation. This
appears to occur about 50 percent of the time. The capability to forecast
situations which have generally persistent winds from situations which have
critically persistent winds and unfavorable dispersion conditions needs to be
developed and proven. The practical operation of a supplementary control
system must provide the plant with an acceptable operating level which does
not require frequent curtailment of operations. A control system requiring
that the plant operate at no more than 60 percent of maximum load on 50
percent of all days might not be an economically desirable system.
In general, the feasibility of a supplementary control system, as demon-
strated to date, can be summarized as follows:
o A model exists which reliably estimates the distribution
of short-term concentrations over a long period of time;
the ability to accurately predict air quality on a day-to-
day basis needs further testing.
o The meteorological conditions associated with high
SC>2 concentration levels are identified; they have an
undesirably high frequency of occurrence. More refined
and limiting conditions are desirable, if they can be identified.
o The emissions of SO2 can be accurately estimated from the plant
load and fuel sulfur content and can be effectively controlled by
reducing the rate of fuel combustion.
71
-------
TABLE 29. RANGE OF METEOROLOGICAL VARIABLES OBSERVED ON DAYS
WITH HIGHEST MEASURED CONCENTRATIONS OF SO2
Date
July 4
July 7
JulyS
June 1
March 4
March 6
June 20
July 9
March?
May 24
July 13
July 11
Wind
Direction
(degrees)
29-56
30-52
34-56
31-52
22-55
41-56
19-63
26-51
31-59
23-51
358-51
23-53
Speed
(m/sec)
3. 4-8. 5
3. 0-8. 4
3.8-7. 1
4. 4-8. 6
3. 0-8. 6
3.8-8.3
4. 3-3. 3
2.9-7.9
2. 9-8. 7
2. 7-7. 1
2.1-6.1
2.1-6.2
Standard
Deviation
of Wind
Direction
(degrees)
10-29
11-20
10-25
10-24
9-23
11-24
11-24
11-22
9-26
10-22
12-24
9-61
Mixing
Height
(m)
341-1588
277-1559
226-1322
100-1893
139-1615
1046-2422
100-1563
102-1943
153-1994
100-2:76
310-1750
176-1400
Observed
S02
(Ug/m3)
2463
2255
1994
1985
1396
1853
1342
1835
1333
1752
1730
1699
o The extent of emissions curtailment necessary in terms of the
frequency of curtailment of the plant operations has not been
established. The frequency may be high enough to make the
supplementary control system an uneconomical alternative in
comparison with other options such as scrubbers and tall stacks.
KAHE POINT PLANT
The monitoring data collected over a 4-month period near the Kahe Point
plant showed that the 3-hour and 24-hour Federal standards for SO2 were
frequently exceeded at Station 2, about 150 m from the plant. The 24-hour
standard was exceeded about 4 percent of the time at Station 1, about 500 m
south of the plant. The periods of the highest 24-hour and 3-hour mean
concentrations were examined to determine the nature of the concentration
patterns on these days and to identify significant meteorological parameters
which set them apart from other periods.
In contrast to the situation at Kahului, the highest SO2 concentrations
occur under a range of observed meteorological conditions. A summary of the
winds which were observed at Station 2 on the 15 days with highest 24-hour
SO2 concentrations is given in Table 30. Although the wind direction is
generally from the northeast, it is extremely variable. The large standard
deviations of wind direction observed also indicate the high degree of
turbulence. Northeast winds with the same degree of turbulence occur on
other days on which much lower SO2 values are measured. The meteorological
differences between high and low SO2 days are not evident from the surface
meteorological conditions. The tethersonde data collected during the inten-
sive monitoring periods do show the presence of large turbulent eddies. It
72
-------
TABLE 30. RANGE OF WIND OBSERVATIONS AT STATION 2 ON DAYS
WITH HIGHEST S02 CONCENTRATIONS AT STATION 2
Date
1977
Nov. 4
Oct. 26
Nov. 5
Nov. 3
Nov. 14
Nov. 6
Nov. 13
Oct. 25
Nov. 2
Oct 27
Oct. 17
Oct. 16
Dec. 16
Dec. 17
Oct. 20
Wind
Direction
(degrees)
26-159
53-95
54-114
29-86
287-108
37-107
10-105
348-178
10-96
10-125
18-174
112-141,
257-276
37-121
272-98
16-153
Wind
Speed
(m/scc)
0.3-2.9
1.2-3.6
1.5-3.1
1.1-3.8
0.4-2.3
0.7-2.6
0.4-1.9
0.5-5.6
1.2-3.4
0.1-2.9
0.2-4.9
0.6-2.9
1.0-3.4
0.4-3.5
0.6-2.8
Stand.urd Deviation
of Wind Direction
(degrees)
41-101
34-75
34-64
32-80
24-79
31-69
27-107
34-103
23-65
30-110
24-105
18-86
28-76
36-102
28-83
24 -II our
Measured,
so, (UP »o
2140
2050
1900
1830
1760
1530
1530
1410
1350
1310
1270
1240
1240
1160
1100
may be possible to make a more detailed analysis of the turbulence spectrum
and find meteorological parameters which distinguish between low and high
S02 concentrations at ground level. This more detailed analysis has not been
performed for this study. It is not clear that it will produce the desired
separation of high and low S02 days.
Some of the high concentrations occur on days with above-average output
loads from the plant and thus higher-than-normal SO2 emission rates. However,
the SO2 concentrations are much higher relative to their mean than are the
emission rates which tend to be very consistent from day to day. It is
possible that a small portion of the variation in SO2 concentrations can be
attributed to S02 emission rates.
Until evidence is developed to indicate the meteorological difference
between high and low SO2 days, it does not seem feasible to consider a supple-
mentary control plan for the Kane Point plant.
73
-------
The concentrations listed in Table 30 all occur during the months of
October, November, and December, and none occur in January. This is because
the listed values are 24-hour concentrations which are highest when the wind
direction is most persistent. During January the trade winds are weakest and
least persistent. This fact accounts for the absence of any high 24-hour SCU
concentrations in January. On the other hand, there is a much greater
variety of complex windflow situations in January. All three types of wind-
flow situations, i.e., trade, sea breeze, and Kona winds, have more nearly
equal frequencies of occurrence in January than at any other time of the
year.
74
-------
REFERENCES
Briggs, G.A. Estimation of Downwash Effects. ATDL Contribution File No.
75/23. Atmospheric Turbulence and Diffusion Laboratory, National
Oceanic and Atmospheric Administration, Oak Ridge, Tennessee (1975).
Environmental Protection Agency, Office of Air Quality Planning and Standards.
Compilation of Air Pollutant Emissions Factors, Second edition,
Supplement 7 (19 7 7).
Hanna, S.R., G.A. Briggs, J. Deardorff, B.A. Egan, F.A. Gifford, and F.
Pasquill. "AMS Workshop on Stability Classification Schemes and
Sigma Curves - Summary of Recommendations." Bull. Amer. Meteoro-
logical Soc. 58(12):1305-09, 1977.
McElroy, J.L. Ambient Air Monitoring Design: Methodology and Illustrative
Examples. Conference Proceedings, Fourth Joint Conference on
Sensing of Environmental Pollutants, American Chemical Society,
Washington, D.C. (1977).
McElroy, J.L., and F. Pooler, Jr. St. Louis Dispersion Study, Vol. II;
Analysis. Publication No. AP-53, National Air Pollution Control
Administration, Raleigh, North Carolina (1968).
Pasquill, F. Atmospheric Diffusion, Second edition. John Wiley & Sons, New
York, New York (1974).
Siple, G.W., and W.A. Freberg. Evaluation of Sulfur Dioxide Near Two Hawaii
Power Plants - Data Appendix. Prepared by Northrop Services,
Incorporated, for U.S. Environmental Protection Agency (1978).
Turner, D.B. Workbook of Atmospheric Dispersion Estimates. Publication No.
999-AP-26, U.S. Department of Health, Education, and Welfare, Public
Health Service (1970).
Turner, D.B. "A Diffusion Model for an Urban Area." J. Applied Meteorology
3(1):83, 1964.
U.S. Atomic Energy Commission. On-Site Meteorology Program, Safety Guide 23.
Washington, D.C. (1972).
75
-------
BIBLIOGRAPHY
Turner, D.B., and A.D. Busse. Users Guide to the Interactive Versions of
Three-Point Source Dispersion Programs; PTMAX, PTDIS, PTMTP. U.S.
Environmental Protection Agency, National Environmental Research
Center, Office of Research and Monitoring, Research Triangle Park,
North Carolina 27711 (1973, preliminary draft).
76
------- |