Search Results

You are looking at 1 - 1 of 1 items for

  • Author or Editor: Gerald F. Cotton x
  • Refine by Access: All Content x
Clear All Modify Search
Joanne Simpson
William L. Woodley
Alan H. Miller
, and
Gerald F. Cotton


A randomized, single-cloud, dynamic seeding experiment was conducted with airborne pyrotechnics in South Florida in 1968 with results extensively reported. In the first 40 min following seeding, large increass in rainfall (about 150 acre-ft or approximately 100% per seeded cloud) were obtained by analysis with a calibrated 10-cm radar, the accuracy of which had been tested by a raingage comparison. The statistical significance of the rainfall differences was, however, marginal, ranging from 5–20% with a series of two-tailed tests.

In the spring and early summer of 1970 an improved repeat of the experiment was conducted in two phases. Five instrumented aircraft participated in the first phase and only two in the second. Altogether 13 seeded clouds and 16 controls were obtained. All seeded clouds reached cumulonimbus stature as did 10 of the controls. The average difference in vertical growth following seeding of seeded vs control clouds was 6200 ft, significant at the 1% level.

This paper is concerned primarily with the rainfall results of the 1970 experiment and the combined 1968 and 1970 experiments, together with the results of a detailed statistical investigation of their significance. The rainfall analyses are made with the University of Miami's calibrated 10-cm radar by the method developed and tested for the 1968 data. For the first 40 min following seeding, the average seeded minus control rainfall difference is about 100 acre-ft while it is more than 250 acre-ft, or more than 100%, for the entire cloud lifetime. Significance is better than 5% for the whole cloud lifetime for the 1970 data alone and for the 1968 and 1970 data combined; it is better than 5% for the combined data for the first 40 min and better than 10% for the 1970 data alone. When the rainfall data are objectively stratified into fair and rainy days, the fair-day differences are of the order of 350–400 acre-ft and the rainy-day differences are negative. Intraday comparisons are also made, comparing seeded and control clouds on the same day. This analysis, if anything, increases seeded-control differences, which retain high significance. The main result of the statistical analysis is that for all 1968 and 1970 data combined, the positive seeding effect is not only significant but exceeds a factor of 3.

The shortcomings of the radar evaluations are discussed; it is shown that if they could be removed the rainfall conclusions would be strengthened.

Full access