WMOD00534

• Title: Reanalysis of "Generalized Criteria for Seeding Winter Orographic Clouds"
• Date: 5/5/1980
• Weather Modification - Doc Type: Report

MAY 1980

1.1.i'

seed/no-seed ratio was reduced. The number of zero
events was greatly reduced for non-seeded cases
in Table 2b, which undermines the implication stated
by Vardiman and Moore (1978), that seeding effec-
tively initiates precipitation in the cloud-top tem-
perature range between -10 and - 30oC. However,
the trend for fewer zero precipitation events at lower
cloud-top temperatures, regardless of whether they
were seeded or not, is now stronger than before.
Table 3 shows the old and new statistics with no
stratification and without the Central Sierra Research
experiment (CENSARE). The exclusion of CEN-
SARE follows the precedent established in Vardi-
man and Moore (1978). This was done because sig-
nificantly more rawinsondes were launched in seeded
cases than were in non-seeded cases. The corrected
data set shows a seed/no-seed ratio less than 1.00 and
a complete loss in statistical significance. This com-
plete loss of significance shows the relative weight
that the error in the Climax data set had on the over-
all results.
The stratification of data into physically mean-
ingful categories was accomplished in an identical
manner to the stratifications of the original paper.
Tables 4-8 display the crest group statistics. Again,
the set has five projects and the tables represent
areas of both positive and negative seeding effect.
The variables in the tables are defined as follows:

BTI-Barrier trajectory index (time available)
CBWS-Saturated mixing ratio at cloud base (water
available)
LCTTI-Lifted cloud-top temperature with respect
to ice (nuclei available)
EPOS-Positive energy area (mixing available).

A detailed discussion of these variables and their
bounds is in Vardiman and Moore (1978). Table 4
shows the stable cloud with a crest trajectory,
moderate water content and cloud-top temperature
between -10 and -30oC. Significance was lost on

TABLE 7. As in Tab]e 4 except for highly unstable cloud.

Projects

All less CENSARE

Precipitation group
Stratifications

Crest

EPOS > 0.05 J g-I
- 3600 < BTl < 3600 s
CBWS > 3 g kg-1
- 300C < LCTTI < - lOoC
M + N '"' ]26/103
Seeded Nonseeded
M = 57/55 N = 69/48
Zeros =4/4 Zeros = 12/5
Mean = 1.93/2.03 Mean = 1.88/2.25
Ratio (seeded/non seeded) = 1.03/0.90
Test statistic = 1.165/0.417
P value = 0.242/0.874

~.

NOTES

625

TABLE 8. Statistics for unstable clouds with a blow-over trajectory,
low water content and LCTTI < - 30oC.

Projects

All less CENSARE

Precipitation group
Stratification

Crest

EPOS > 0.005 J g-I
BTI<-1800s
CBWS < 5 g kg-I
LCTTI < - 300C
M + N = 58/41
Seeded Nonseeded
!'vI '"' 23/17 N = 35/24
Zeros'"' 3/2 Zeros = 3/1
Mean'"' 1.8711.92 Mean = 4.11/4.22
Ratio (seeded/nonseeded) = 0.46/0.46
Test statistic = -2.378/-2.198
P value = 0.018/0.028

reanalysis and the seed/no-seed ratio became less
than 1.0.
Table 5 shows the statistics for unstable clouds.
Significance was also lost for this stratification and
the seed/no-seed ratio is now 1.0. Tables 6 and 7
represent the division of the unstable cases into
moderate and high instability, respectively, while
holding other conditions constant. The seed/no-seed
ratios for the moderately unstable category remained
positive although significance was lost. The highly
unstable category was not markedly affected by the
reanalysis.
Table 8 shows the original and modified statistics
for the unstable blow-over situation. These condi-
tions had yielded negative seeding effects on the
crest. The reanalysis showed a strong decrease of
54%, which was identical to the original analysis,
with near statistical significance and with a very
small sample size.

4. Conclusions

The results of Vardiman and Moore (1978) were
severly compromised by the inclusion of a large
number of non-experimental cases in the no-seed
population of the Climax experiment. Even though
this problem only occurred in the Climax data set,
it had a major impact on the overall results of the
original Vardiman and Moore study. Repartitioning
of the corrected data set in the windows reported
earlier led to mostly non-significant results. The only
window that remained significant was one indicating
a decrease when unstable clouds with a blow-over
trajectory, low moisture and cloud-top temperatures
less than - 300C are seeded.

5. Further studies

Because of the dominant effect of the renegade
cases the windows reported in the original paper by
Vardiman and Moore (1978) have been found to
poorly characterize the optimum conditions for