United States
Environmental Protection
Agency
Environmental Research
Laboratory
Duluth MN 55804
Research and Development
EPA-600/S3-84-053 May 1984
oEPA Project Summary
Instrumentation, Recording, and
Processing of Meteorological Data
Near Portage, Wisconsin:
Wisconsin Power Plant
Impact Study
Charles R. Stearns, Phillip Falconer, George Weidner,
Brent Bowen, and Leonard Dzamba
As part of the Columbia Power Plant
Impact Study, meteorological data were
collected at a network of monitoring
sites from 1972 through 1977. The data
were the basis for a series of studies
whose purpose was to elucidate the
transport of airborne pollutants.
In a pilot diffusion climatology survey,
local wind structure was interpreted in
the context of synoptic weather pat-
terns. A significant new low wind sta-
tistic was introduced: the number of
crossings of the trace of a variable past
its running mean during a given time in-
terval.
A case study traced the movement of
an elevated level of ozone from the
Southern Plains across the Midwest to
the East Coast. It showed that, although
the relative contributions of transport
vs. local or regional formation to an
episode of elevated ozone (O3) are not
clear, some degree of large scale, long
distance transport is necessary in order
for elevated O, levels to occur in most
parts of the United States. A related
study showed that ambient O3 levels are
reduced sharply in the presence of the
plume from a coal-fired power plant.
The horizontal variation of the wind
field is an important factor in the
transport of atmospheric pollutants in
the range of 10 to 100 km. The wind field
in the study area was shown to be or-
ganized as a function of both wind direc-
tion and wind speed around the Baraboo
Hills. These hills therefore have an in-
fluence on the dispersion of pollutants
from the Columbia power plant.
Finally, two models for estimating
concentrations of SO2 at ground level
were compared. Both used the Gaussian
plume equation, but one estimated the
required dispersion coefficients from
the Hino stability model, while the other
was based on data for horizontal and
vertical wind range. The wind range
model was shown to have simpler data
requirements and to give results which
agreed more closely with observed SO2
values.
This Project Summary was developed
by EPA's Environmental Research
Laboratory, Duluth, MN, to announce
key findings of the research project that
is fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
In 1971, scientists at the University of
Wisconsin-Madison began a long-term study
of environmental changes related to the con-
struction and operation of the Columbia
Generating Station, a 1054-MW coal-fired
power plant located in the floodplain of the
Wisconsin River in south central Wisconsin.
Meteorological data were required for inter-
preting the results of several subprojects,
particularly those involving dispersion of air
pollutants. In addition, the power plant could
have a variety of direct effects on local
-------
climate. The purpose of this meteorology
subproject was to provide baseline informa-
tion on meteorological conditions at the site
and to detect any changes in conditions dur-
ing the course of the subproject. It was
designed to:
1. Study meso-scale airflow in the vicin-
ity of the generating station.
2. Measure atmospheric characteristics
necessary for evaluating the mass bud-
get of chemicals transported through
the atmosphere from the generating
station.
3. Determine the influence of the cooling
pond on local climate, and detect any
other local climatic effects of the power
plant.
In addition to providing baseline meteoro-
logical data, this subproject produced a pilot
diffusion climatology survey of the Colum-
bia Generating Station, a study of depletion
of ambient ozone by the plume from the
generating station, a comparison of two
models for estimating ground level concen-
trations of sulfur dioxide (S02) resulting from
the operation of the generating station, and
a study of the influence of the Baraboo Hills
on the surface wind field at the site. Finally,
national weather data were used to elucidate
the observed long distance, large scale
transport of ozone (03) across the Central
and Eastern United States.
Monitoring
Data used for measurements of atmo-
spheric dispersion must be valid for heights
between the top of the stack and the highest
rise of the plume. Because construction ac-
tivity precluded erecting a tower on the site
of the generating station, monitoring stations
were established at four nearby sites where
measurements could be made at the required
heights (between 152 and 305 m) (Figure 1).
Easterly winds are uncommon in the area.
The Messer site was therefore essentially
unaffected by the plume from the stack, and
the most complete instrumentation was in-
stalled there. An additional advantage of the
Messer site was that, with its 122 m eleva-
tion, it required only a 30 m tower for plac-
ing instruments at the same height as the top
of the stack. The other three sites served
primarily to corroborate observations of wind
speed and direction made at Messer and to
provide insight into variation caused by
topographical differences and differences in
direction relative to the generating station.
The investigators also used data collected
by the Wisconsin Power and Light Company
(WPL) at sites shown in Figure 1, and by the
National Weather Service at Truax Field in
Madison, Wisconsin.
Portage
7 A University of Wisconsin
+ O Wisconsin Power and Light
Lake George-250m
O
Columbia
Generating
* Station
Wyocena-245m
Lake-24Sm
Dekorra-245m
Figure 1. Monitoring sites for meteorological data.
Monitoring Procedures
Measurements necessary for determining
the power plant's impact on climate differ
in type and duration from those required for
dispersion studies. Long-term measurements
of rainfall and soil temperature are needed
to document climatic change. Dispersion
studies require seven types of continuous
measurements at stack height: horizontal
wind speed, horizontal wind direction, ver-
tical wind speed, vertical wind direction, ver-
tical air temperature gradient, total incoming
solar radiation, and net radiation. Table 1
shows the instrumentation used to monitor
these variables, the sites at which each
variable was monitored, and the dates be-
tween which data were collected. Gaps exist
in the data because of equipment failures
and seasonal monitoring of certain variables.
Data collection began at the Messer site in
1972 and at the other sites in 1973 or 1974.
Processing data recorded on charts in-
volved the following general steps:
1. Timing the charts to mark hours and
days.
2. Digitizing strip chart data.
3. Transferring chart data to 80-column
IBM cards.
4. Checking data for errors.
5. Converting data to scientific units (e.g.,
m/sec) and averaging if necessary.
A Pilot Diffusion Climatology
Survey for the Columbia
Generating Station
The purpose of this study was to identify
easily measured meteorological variables
which indicate characteristics of atmospheric
dispersion, and to relate them to synoptic
weather patterns. Many climatological stud-
ies of the dispersion of air pollutants have
relied on oversimplified models. This study,
rather than smoothing the spectrum of
hourly weather variations, sought to identify
extremes during which adverse conditions
for dispersion exist.
The study period was from August 15 to
November 1, 1972. Twenty-one days were
identified as potentially poor dispersion days
on the basis of minimal wind speeds (gen-
erally < 4 m/sec) recorded during the early
morning hours of 1100 Z to 1300 Z. For com-
parison of mesoscale analysis with mac-
roscale patterns, daily 1200 Z synoptic
weather charts were obtained from the Na-
tional Oceanographic and Atmospheric Ad-
ministration.
The statistics used in the wind structure
analysis were the range of the wind speed
and direction variables during each 10-min
interval between 1100 Z and 1300 Z and the
number of crossings made by the continuous
trace of the variable past its running mean
for the interval. In the final analysis, the ratio
-------
Horizontal
wind speed
Horizontal wind
direction
Wind angle, vector
wind speeds
Vertical wind
speed
Precipitation
Soil temperature
Air temperature
Air-soil temperature
gradient
Solar radiation
Site
All sites
Messer
All sites
Messer
Messer
All sites
Messer
Messer, Dekorra
Messer, Dekorra
Messer
Sensor
Stewart 4-cup anemometer
Beckman-Whitley Type F
3-cup anemometer
U.W. vane
Gill 01002 bivane
anemometer
Gill 27100 vertical
anemometer
Tipping bucket
Bi-metal thermometer
Thermocouple
Thermocouple
Eppley 10-junction
Recorder
Esterline-Angus
event marker
Esterline-Angus pen
Esterline-Angus pen
Daystrom
Daystrom
Weather Measure P521
event recorder
Observed, recorded by
investigator
0-2.5 mv Brown
0-2.5 mv Brown
Brown
Dates
8/72-12/77
8/72-12/77
8/72-12/77
8/72-11/77
Apr-Oct
12/73-12/77
8/72-10/77
Apr-Oct
8/72-12/77
8/72-12/77
8/72-12/77
8/72-12/77
Net radiation
Inversion layer
Messer
Cemetery
pyranometer
Ventilated Suomi Brown
net radiometer
Aerovironment 300 acoustic sounder system
8/72-12/77
1975-1977
of the total 10-min wind passage for each
component direction to the number of cross-
ing during the same period yielded a measure
of the component eddy lengths.
A classification of gustiness types for poor
dispersion periods was developed from
analysis of the range and appearance of
horizontal wind direction traces. The five
gustiness types were related to atmospheric
stability, cloud cover, and eddy lengths
derived from wind passage and trace cross-
ing data. When eddy lengths are stratified
according to gustiness types, recognizable
patterns appear in the resulting histograms.
Attempts to correlate synoptic conditions
with local wind structure analyses suggested
that the resolution of isobaric analyses on
daily surface weather maps is not good
enough to provide an adequate picture of
local pressure gradients and mesoscale flow.
Regardless of the large scale distribution of
cyclonic and anticyclonic pressure centers
and macroscale pressure gradients, local
dispersion remains to some extent a func-
tion of generally undetected local pressure
variations. A researcher who is familiar with
local circulation under a wide range of syn-
optic conditions might be able to predict
local gradients from larger scale patterns.
In summary, this pilot study introduces the
number of crossings of the trace of a variable
past its running mean as a significant low
wind statistic. It suggests that a similar study
over a longer observation period, or perhaps
another approach such as photography or
tracer analysis, might provide a meteoro-
logical basis for a local diffusion climatology
that would be of value in prediction and
analysis of environmental impact.
Study of the Large Scale
Transport of Low Level Ozone
Across the Central and
Eastern United States
Recent evidence shows that ozone (03)
can be transported over long distances, and
elevated levels of 03 are often found in areas
remote from sources of pollution. This sec-
tion of the report presents the findings of a
case study of the distribution and transport
of low level O3 east of the Rocky Mountains
from September 4 to September 10, 1976.
The production, chemistry, and photochem-
istry of O3 were reviewed. Possible relation-
ships between high 03 levels and weather
conditions such as visibility were in-
vestigated. Backward trajectories were con-
structed to trace air movements in the
boundary layer.
All available data for all states east of the
Rocky Mountains were used to construct
daily maximum 03 maps. Daily maxima ex-
ceeding 70 and 100 ppb were analyzed for
the Central and Eastern United States. Mete-
orological data from the National Weather
Service included daily surface maps show-
ing sea level pressure and surface winds,
visibility, and dew point, an 850 mb contour
map, and a maximum surface temperature
map.
At the beginning of the case study, a large
area with O3 levels >70 ppb formed under
an 850 mb ridge in the Southern Plains. This
area spread northward into the Northern
Plains and then eastward across the northern
Midwest, as the 850 mb ridge moved toward
the East Coast. A smaller area with 03 levels
>100 ppb developed within the larger area
over the northern Midwest. Another area
with levels >70 ppb formed over the Ohio
Valley. These two areas subsequently
merged over the Great Lakes, and then
spread to the East Coast. Ozone levels sub-
sided sharply as a cold front swept eastward
across the United States.
Trajectories averaged through the
300-2000 m layer agreed well with the move-
ment of elevated ozone areas. Surface
winds, however, were not necessarily in the
same direction as the direction of O3
transport.
The relative contributions of 03 transport
and local or regional formation of O3 to an
episode of elevated 03 levels are not clear.
On the basis of this study, however, large
scale transport seems necessary in order for
elevated 03 levels to occur in most parts of
the United States. More extensive and
strategic monitoring of 03 is needed for bet-
ter understanding of this phenomenon.
-------
Depletion of Ambient Ozone
by the Plume from the
Columbia Generating Station
Coal-fired power plants contribute sulfur
dioxide (S02), nitrogen oxides (NO and NO2,
or NO,), and many other chemicals to the
atmosphere. The plume from such a plant
can deplete 03 to a distance of 15 km or
more downwind, primarily through the reac-
tion: 03 + NO -* IM02 + 02. This effect was
studied at the Columbia Generating Station
during the summer of 1976.
The meteorological monitoring network
supplied data on wind speed and direction
at stack height and on solar radiation. S02,
03, and NO, were monitored 10 km north of
the power plant. High hourly averages of
these gases indicated times when the plant
plume was drifting over the site. The
simultaneous, instantaneous extremes of the
levels of SO2, 03, and NO, were then ana-
lyzed. Ambient levels were subtracted from
the total, to focus the study on the effects
of the plume.
Ozone levels were reduced sharply when
the plume was present. The amount of 03
depletion depended on ambient O3 levels and
on levels of S02 and NO, associated with the
plume. The amount of N02 from the plume
was proportional to the amount of 03 re-
moved from the ambient air. This relation-
ship can be summarized: [N02] + [03] =
constant.
Influence of the Baraboo Hills
on Surface Winds Near the
Columbia Generating Station
The horizontal variation of the wind field
is an important factor in the transport of at-
mospheric pollutants in the range of 10 to
100 km. Wind data collected from the moni-
toring network (Figure 1) provides strong
evidence that the Baraboo Hills have a
significant influence on surface winds in the
area. This influence is reflected in the linear
wind field and in the first order properties of
divergence and vorticity as calculated by the
Bellamy Triangle Technique. For wind direc-
tions resulting in convergence, continuity
arguments imply a compensating upward
vertical motion, while divergence requires a
downward motion. Such motions would be
important in the transport and diffusion of
atmospheric pollutants.
The wind field appears to be organized as
a function of both wind direction and wind
speed around the Baraboo Bluffs. This
organization is most apparent when the wind
direction is perpendicular to the bluffs.
Although the data from the Wyocena site
give an adequate estimate of the mean wind
field for the region, deviations at the other
sites must be considered in an accurate
determination of pollutant transport. The
wind direction at Dekorra deviated about 15°
from that at Wyocena. Data from Messer,
strongly influenced by the direction of the
wind relative to the bluffs, provided a poor
representation of the mean wind field in the
valley. The symmetric variation of the
divergence and vorticity fields about the
azimuthal direction of the southern bluffs
support the conclusion that the bluffs are im-
portant in the kinematics of the regional wind
field.
A Comparison of the Hino
Stability Method and the
Wind Range Method for
Estimating Concentrations of
SO2 at Ground Level
Gaussian diffusion models are widely used
in pollution studies. This study emphasizes
the advantages of calculating horizontal and
vertical dispersion coefficients from actual
wind statistics, rather than basing them on
a classification of stability.
Two methods for estimating the dispersion
coefficients used in the Gaussian plume
equation were compared. All other param-
eters in the equation are identical for the two
methods. The Hino stability method esti-
mates the dispersing characteristics of the
lower atmosphere from a net radiation bud-
get for the lower atmosphere. It is an indirect
method based on variables such as cloud
cover and lapse rate. The vertical and
horizontal wind range method uses wind
data to describe the turbulent structure of
the atmosphere and estimate the dispersion
coefficients. The results predicted by the two
methods were compared with concentra-
tions of S02 measured at six monitoring
sites.
Overall, results from the wind range model
were closer to observed values than results
from the stability model, particularly at low
wind speeds. For the 343 hours included in
this study, 64% of the values from the range
method lie within a factor of two of the
observed values, while only 53% of the
results from the stability method fall within
these limits. The discrepancies are still large,
but the range method clearly represents an
improvement.
Table 2 shows the standard deviations
from observed data for both methods, ac-
cording to monitoring site. The accuracy of
both models improves as the distance from
the power plant increases. A partial explana-
tion for this may be that the Gaussian plume
equation assumes a normal vertical distribu-
tion of pollutant — a condition which is
poorly fulfilled in proximity to the power
plant. This explanation is supported by oc-
casional erratic fluctuations in S02 data.
These fluctuations are not readily explained
and are not reflected by the model; but they
are most frequent and most pronounced as
the Dekorra site which is closest to the
power plant.
Discrepancies in the range model may be
due in part to the source of wind data.
Calculations were based on measurements
made at the Messer site, the only place
where a vertical anemometer was installed.
The topographical study showed, however,
that the Messer site was not typical of the
area because of the influence of the Baraboo
Bluffs.
In summary, the vertical and horizontal
wind range method offers two significant im-
provements over the traditional stability
classification method. First, it is far superior
during light wind conditions. Second, it
employs fewer atmospheric variables while
offering a more direct approach to estimating
the dispersing conditions of the atmosphere.
-------
Table 2. Standard Deviations of Predicted SO2a Values from Observed Values, for Range and
Stability Models, at Six Sites
Standard deviation Distance from power
Site Range method Stability method plant, km
Dekorra
Lake George
Messer
Genrich
Bernander
Russell
61
49
37
43
17
27
66
57
32
43
18
21
4.4
5.9
7.3
8.3
14.6
15.5
a Units are
Charles R. Stearns, Phillip Falconer, George Weidner, Brent Bowen, and Leonard
Dzamba are with the University of Wisconsin, Madison, I/I// 53706.
Gary E. Glass is the EPA Project Officer (see below).
The complete report, entitled "Instrumentation, Recording, and Processing of
Meteorological Data Near Portage, Wisconsin: Wisconsin Power Plant Impact
Study," (Order No. PB 84-172 469; Cost: $16.00, subject to change) will be
available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Research Laboratory
U.S. Environmental Protection Agency
Duluth, MN 55804
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
a a h I-. V IK H« t A H '.-i u H i J b I K t fc T
t M I ( A b U I I h 0 O 0 4
A U.S. GOVERNMENT PRINTING OFFICE: 1984-759-102/949
------- |