An Evaluation of the Whitetop Cloud Seeding Experiment through a Covariance Analysis

Abstract

The Whitetop experiment concerning cloud seeding was conducted in southern Missouri and northern Arkansas during 1960?64. In general, the early analysis showed the rainfall in the treated areas on days with seeding to be significantly less than the rain in corresponding areas on days without seeding. In the present analysis, the rainfall for an area adjacent to the treated area was taken as a covariant. Using the analysis of covariance, the seeding effect is not statistically significant at the 5% level. At least a portion of the negative effect from treatment previously reported is associated with the high frequency of heavy rain on the days without seeding. From the covariance analysis the differences in precipitation for varying distances from the seeding line were not statistically significant.