WMOD00234

• Contract/Permit #: #93-5
• Applicant: Western Kansas Groundwater
• Project Name: Kansas Weather Modification
• Date: 1/1/1993
• Weather Modification - Doc Type: Report

In years like 1993, when hail claim numbers are high despite
our best cloud seeding efforts, there is a tendency to focus either
on our job performance or whether we are having any significant
effect in reducing hail. What can we deduce from the data we have?

I
I
I
I
I
I
I
I
I
I
I
I
I
I
I

Loss costs are used frequently by evaluators in analyzing
whether or not a long-term hail suppression is effective. There are
problems using even these numbers to claim authenticity. No attempt
will be made here to derive statistically significant values with
the data on hand.

To quickly review of what Loss Costs are: Loss Cost is a
percentage found by diving Crop Loss by Crop Liability then
multiplying by 100. For instance, if we lost $50 in crops but had
$1000 of crops insured, then our Loss Cost = 5% .

In Fig. 12 the percentage of crop damage over the past years
has usually varied between 2% and 5%, with the exception of the
drought year of 1988 when it was almost half of one percent (0.5
%). The "average" Loss Cost tends to be centered nearer 3%.
Although Loss Costs and Crop Losses (Figs. 10 and 12) show similar
patterns each year, as do both Liabilities and Farm Values (Fig.
11), Loss Costs don't show any clear correlation with either Farm
Values or Liabilities (Fig. 12) as three of the eight years run
counter to the trend of the both Farm Values and Liabilities.. As
a further example: Loss Costs between 1987 and 1989 gyrated wildly
from year to year due to the 1988 drought between those two years
(Fig. 12). Even between year 1990 and 1991 Loss Costs dropped by
as much as 50%. In fact all of this appears to mean, simply, that
hail varies widely from year to year! From a statistical point of
view, this kind of variability causes problems. Perhaps with
adequate numbers of seeding aircraft, the highs and lows of the
Loss Cost data would settle into a more narrow range---but that
still remains to be seen here. As it is, with all the random ways
in which severe storms could originate, among the ones listed on
page 5, it is difficult to visualize how such natural randomness
could ever average out, or be "normalized" so that statistical
results, with respect to hail suppression effect, could have any
genuine meaning whatsoever.

42