An Analysis of a Possible Crop Response to Hail Suppression Seeding: The Nelspruit Hail Suppression Project

Abstract

A commercial hail suppression program has been in operation at Nelspruit, South Africa (25°S, 31°E) since December 1971. The seeding technique employed has been the ?on-top? method, which consists of dropping silver iodide pyrotechnics into the tops of cloud turrets growing on the flanks of mature storms. Daily crop insurance records of hail damage to tobacco are compared for seeded and nonseeded days. The control (unseeded) days consist of historical data and of days during the operational period when storms in the area were definitely not seeded, usually for operational reasons. A ?hail severity ratio?, defined as the ratio of the crop lost to the total crop planted in the areas hit by hail, is derived from the hail-damage assessments and appears to be the parameter most sensitive to the claimed treatment effect. A comparison of seeded days during the initial phases of the program, when propeller aircraft were used, with more recent days when one or two jets were used for the seeding, indicates significantly lower severity ratios for jet-seed days. The analysis of the seed versus the control days for three separate regions in the operational area show, on the average, statistically significant reductions in hail severity on the days on which jets were used for seeding. As a ratio, the severity parameter can be altered by changes in the numerator (total crop damage) and/or the denominator (total crop planted) in the areas hit by hail. An analysis of total damages and areas hit by hail indicates that the numerator appears to be affected most strongly by the treatment, and that there is no indication that the seeding is causing an increase in the area hit by hail. It is unfortunate that the data were not collected within the framework of a randomization scheme planned before the initiation of the seeding; the statistics, however, do possess an internal consistency that lends support to the validity of the conclusions.