î
Koracin et al.
î
•
•
Meridian (MER)–Recorded data exist from Janu-
ary 1992 to December 1996;
in Koracin et al.2 The main objectives of this study are
•
to evaluate selected EPA-regulatory and EPA-
alternative dispersion models for a base year and
episodes of elevated PM10 concentrations using
estimated emissions; and
Nampa (NAM)–Recorded data exist from July
1993 to December 1996 (not shown in Figure 2;
located 5 km south and 29 km west of FS5); and
Cloverdale (CLO)–Recorded data exist from Au-
gust 1995 to December 1996 (not shown in Fig-
ure 2; located 24 km south and 7 km west of FS5).
•
•
to develop strategies to control emissions by us-
ing the evaluated models and emissions projected
to future years.
A list of recorded PM10 exceedances is shown in Table 1.
Four episodes of elevated PM10 concentration have been
selected for data analysis and modeling:
(1) December 11, 1985–January 1, 1986;
(2) January 7–15, 1986;
DISPERSION MODELS
Based on recommendations from the EPA, Region 10, the
IDHW-DEQ selected the ISCST3 (Industrial Source Com-
plex Short Term 3) model4 for base-year simulations and
the WYNDvalley model5–7 for modeling episodes of el-
evated PM10 concentrations, including days with
exceedances.
(3) January 20–29, 1988; and
(4) January 4–7, 1991.
These episodes were considered the most extreme
cases of the meteorological conditions conducive to the
development of elevated PM10 concentrations. Meteoro-
ISCST3 Model
logical and monitoring conditions for these episodes are
î
ISCST3 is a steady-state, Gaussian dispersion model de-
signed for assessing pollutant concentrations from a vari-
ety of sources. According to the EPA Guideline on Air Quality
Models,4 the ISCST3 model (classified as an “A” group pre-
ferred model) is recommended for industrial source com-
plexes, rural or urban areas, flat or rolling terrain, transport
distances less than 50 km, 1 hr to annual averaging time,
and continuous pollutant emissions. ISCST3 can be used
to model primary pollutants from point, line, area, and
volume sources. ISCST3 contains algorithms for effects
such as building downwash and pollutant removal (wet
scavenging and dry deposition), and is capable of model-
ing either gaseous species or PM.
described in detail by Koracin et al.2 Figure 3 illustrates
specifics of atmospheric and monitoring data during the
one of the episodes (January 1991). The figure empha-
sizes that meteorological conditions characterized by low
temperature (Figure 3a), low wind speeds (Figure 3b), and
high humidity (Figure 3c) were associated with elevated
PM10 concentrations (Figure 3d). An examination of the
nocturnal radiosonde (at 0500 LST) indicates that the
episodes occurred during strong surface inversions with
high atmospheric stability and a vertical potential tem-
perature gradient frequently in excess of 20 K/kmwithin
the first kilometer (Figure 4). During the first episode, the
potential temperature gradient reached as high as 35 K/ km.
This extreme atmospheric stability appears to be a domi-
nant process that prevents vertical mixing and ventila-
tion of the pollutants from the Boise basin.
WYNDvalley Model (Version 3.11)
The basic assumption of Gaussian models is that the in-
tensity of dispersion is inversely proportional to wind
speed. Since increased PM10 concentrations occur during
stable, low-wind conditions, the use of ISCST3 might not
be appropriate. One alternative (or complementary) ap-
proach is to use the WYNDvalley model.5–7 The
WYNDvalley model does not rely on the Gaussian assump-
tion and can be used for low-wind situations. According
to the EPA Guideline on Air Quality Models,4 the
WYNDvalley model (classified as a
To support a State Implementation Plan (SIP), the
Idaho Department of Health and Welfare–Division of En-
vironmental Quality (IDHW-DEQ) initiated observational
and modeling studies, including the present one, to pro-
vide a basis for control of emissions for this area.3 A com-
prehensive PM10 observational and modeling study was
conducted by Desert Research Institute; details are given
“B” group alternative model) is rec-
ommended on a case-by-case basis
Table 1. Daily exceedances of PM10 in Ada County, Idaho, with associated meteorological data.
to estimate concentrations during
Year
Month
Day
Site
Total PM10 Avg. Temp Avg. WS
Avg. WD
Precip.
valley stagnation periods of 24 hr
or longer. WYNDvalley is a five-
layer, Eulerian-grid dispersion
model that permits its users flexibil-
ity in defining the borders around
the areas to be modeled, the bound-
ary conditions at these borders, the
(µg/m3)
(oC)
(mph)
(deg)
(in. per day)
1986
1988
1991
1991
Jan
Jan
Jan
Jan
14
28
7
FS5
FS5
FS5
MVS
314
165
173
164
–10
–2
5.5
5.2
5.9
5.9
175
214
244
244
0.00
0.00
0.09
0.09
–8
7
–8
Volume 50 August 2000
Journal of the Air & Waste Management Association 1337