WMOD00473

• Title: Evaluation of Bridger Range Winter Cloud Seeding Experiment Using Control Gages
• Date: 12/12/1983
• Weather Modification - Doc Type: Report

2008

JOURNAL OF CLIMATE AND APPLIED METEOROLOGY

VOLUME 22

during the BRE yielded a 700 mb (near plume top)
wind direction between 180 and 240 deg. Thus, any
impact of seeding on the Crazy Mountain Range
courses is believed minimal.
. In addition to the ten possible control courses noted
above within 60 km of S 1 and S2, five additional
courses were considered at greater range in an attempt
to select upwind courses and additional courses to the
north (six were available within 60 km to the south).
As shown in Table 12, three were -90 km WSW in
the closest upwind mountains, the Tobacco Root
Range. The final two courses were the closest in a
northerly direction, being -75 and 90 km distant from
S 1 and S2. All other snow courses from a westerly
through northerly direction of S 1 and S2 were more
than 110 km away and were not considered.
Table 12 lists the snow courses considered and shows
the correlation each had with the March 1 mean water
equivalent at S 1 and S2 for. 13 nonseeded seasons.
March 1 was used rather than April 1 to reduce errors
due to early spring snowmelt and because the 1970-
71 seeding was suspended on March 1, one month
earlier than planned, due to well above normal snow-
pack conditions. The means of S 1 and S2 were used
because, as might be expected, higher associations were
found with other courses than when each of these two
snow courses was considered separately. The data used
were the entire period of record (1967-83), with the
four winters of any seeding in the Bridgers excluded
(1969-72). It is believed that no other cloud seeding
took place within the time period and region consid-
ered.

~ 30

~
c
.

I
71.
I
/ (/>8
I
/
~2~

/ e67
/ 76 9
/~74 I
I 79 I
, 75 I
/ /
/ /
/ 2 /
/ /
/ 7. tJ3/
/ /
/ /
/ 81 880 /
" /6.77
/ /
/ /
/ /
/
/
/
/

/
/
fiB /
"0,'
/
/
/
/
/
/

N
~

25

~
~
c
u
~
c
>0
;!. 20

~ 15
c
~
'"
~
c
'"

, 10

o
o

Reoress;on line for the 13
nonseeded 'II; nters:
Y . -1. 97 + 1. 57 X
R . .920

20

25

10

15

x = Sacajawea (53) (inches)

FIG. 5. Plot of I March mean water equivalents
for indicated snow courses.

~ 30
~
u
c

~
~ 25
~
c
.
C
>0
!'. 20

25

30

/
/
/
/

~
.

l 15
c
~
'"
~
c
'"
... 10
0
c
~
"

I
/
I
/
/
/
/
I
I
I
. I
I

Regression line for the 13
nonseeded winters:
Y . 6.92 + 1. 52 X
R. .965

o
o

15

20

10

x = Means of New World, Sacajawea, Shower Falls and
South Fork Shields (inches)

FIG. 6. Plot of I March mean water equivalents
for indicated snow courses.

30

It is seen that six courses yielded relatively high
correlation coefficients from 0.88 to 0.92 so that any
one of them would explain more than 75% of the total
variance in the mean target snow course data. All other
courses yielded lower correlation coefficients, from 0.65
to 0.84. These snow courses were not considered fur-
ther.
The most straightforward target-control analysis is
to use the nearby Sacajawea (S3) course as the sole
control with the realization that it may have been oc-
casionally affected by seeding. Fig. 5 is a plot of March
1 snow course water equivalent data for individual
seasons. The least-squares regression line of the form
y = a + bx is shown for the 13 nonseeded winters
while the dashed lines show :t 1.96 standard errors of
estimate calculated for the same winters. (It would be
expected that 95% of a nonseeded population would
lie between the dashed lines.) It is seen that the 1968-
69 winter is on the regression line and the 1969-70
winter is below it. It will be recalled that only low-
elevation generators were used the former winter and
that these were relocated after discovery of a persistent
stable layer. Only the southern seeding site was op-
erational in time for the latter winter and gages with
low probabilities were confined to the southern portions
of the three zones as shown in Fig. 4.
The observations from the 1970-71 and 1971-72
winters, which. are the period of the main statistical
analysis herein, are seen to depart markedly from the
regression line. On an absolute basis, these data points
depart farther from the line than any other points in
either direction, (4.3 and 3.6 inches water equivalent