(note: This excerpt is an abridged version, due to the extremely technical nature of the article. For the original, click here: PDF)
Proceedings of the National Academy of Sciences
Vol. 68, No. 1, pp. 147-151, January 1971
Further Studies of the Whitetop Cloud-Seeding Experiment
JEANNE L. LOVASICH, JERZY NEYMAN,
Elizabeth L. Scott, and Marcella A. Wells
The Whitetop cloud-seeding experiment was performed by Braham over five summers (1960-64). The target center was near West Plains, Mo.
The original evaluation showed that within the so-called "Missouri Plume" the average seeded precipitation "per fair hour" was about half that without seeding.
Several subsequent studies were concerned with amounts precipitated over 24 hr in a fixed area up to 180 miles from the target center and, separately, in 6 parts of that area.
It was found that in all six regions the average 24-hr precipitation on the 102 days with seeding was less than those on the 96 days without seeding.
For the entire area, the apparent loss of rain was 21%.
Although none of the apparent effects was found significant by customary standards, further studies appeared justified by the fact that if the 20% apparent loss of rain was actually caused by seeding, this would mean operational (contrasted with experimental) cloud seeding under similar conditions would result in a loss of about 8 million acre-feet of rainwater per summer or, possibly, about $80 million loss of revenue to the local economy.
In the next study, the 198 experimental days were stratified into two categories according to patterns of winds aloft, by three different methods.
One category was supposed to "conform," and the other not to conform, with Braham's rule of determining experimental days.
For "conforming" days, no significant effects of seeding were found.
On the other hand, the apparent effects of seeding on nonconforming days were negative in all the regions and were frequently significant or highly significant.
For the entire area studied, the effects were -39%, -36%, and -46% for the three systems of stratification, respectively.
Thus, there is little doubt that, among the 198 experimental days of the Whitetop Project, there is a category within which the seeded rainfall was roughly 40% less than that without seeding and that this difference could hardly be explained by chance variation.
If we assume that randomization was faultless, the differences must have been caused by seeding. However, specialists in meteorology appear divided on this point.