WMOD00265 - Laserfiche WebLink

Home Browse Search

~ material to affect the rainfall at the surface is 30 minutes, then one might expect the seeding effect to be displaced to the east approximately one grid cell. Again, comparison of the seeding-material pattern and the rainfall pattern shows little relationship. It is possible that a relationship exists if only the rainfall on seed days or only that on days seeded for rain increase (as opposed to hail suppression) were considered. For reasons stated elsewhere, however, this study does not consider rainfall on individual days. The facts that (1) the observed rainfall in the target area during the seed years was only slightly less than that in the controls, (2) the target-area patterns are not materially different from the total study area, and (3) there is little relationship between location of release of seeding material and rainfall amount suggest that the rainfall differences are due to climatic variations; i.e., the 1975-85 seed period was, in general, simply drier than the 1948-74 no-seed period. This conclusion, however, carries with it the implication that the Western Kansas Weather Modification Project had little or no effect on seasonal rainfall. 1bese results are reasonably consistent with those of earlier studies of this project. Hsu and Chen [I] studied monthly and seasonal rainfall for 1975-79. They calculated ratios of mean seasonal (May-August) rainfall in the seed period (1975-79) to that for the 1931-71 period. In general, the ratios were not significantly different from 1.0. They also applied multiple regression and principle component regression techniques, both indicating slight (but not statistically significant) decreases during the seed period. Huff et al. [2] concentrated on daily, rather than monthly, data. They restricted their analyses to May, July, and August; June was omitted because operations were primarily directed toward hail suppression. Days in the other three months on which seeding was exclusively for hail suppression were also omitted. The days were categorized by storm motion, synoptic type, and storm intensity. They found that target rainfall exceeded control rainfall on days when storms moved from a generally southwest direction (SSW through WSW) or from the west-northwest. Control rainfall exceeded target rainfall on days when storms moved from the west and northwest. Target rainfall was greater than control rainfall in squall-line, stationary-front, and air-mass storms, but was le'ss than control rainfall in cold-front, postfrontal, and squall-zone storms. However, these things could well have happened by chance. Huff et al. [2] conclude that, although overall increases in rainfall averaging about 9 percent were found for seed days, ". · · natural rainfall variability and lack of randomization procedures in the seeding operations confounded establishment of firm conclusions from the [Western Kansas Weather Modification Project] operations." 4. EVALUATION OF EFFECfS OF SEEDING UPON HAIL LOSSES 4.1 Organization of Data .. As previously noted, the evaluation of hail losses was carried out using data supplied by the CHIAA (Crop-Hail Insurance Actuarial Association) of Chicago. There was no practical altemative to using crop-hail data, because other records of hail in the target area, say from National Weather Service stations or from newspapers, are very spotty and do not provide a standard measure of the intensity of hailstorms. As the seeding program did not reserve any no-seed cases to serve as controls, it was decided to conduct the analysis by comparing the relative intensity of hail damage in the target and surrounding areas during the years with seeding and during prior years without seeding. This is the same method as that used in the evaluation of the effects of seeding upon rainfall. It was decided to extend the historical or base period back to 1948. This decision represented a compromise between the need for a sufficiently long base period to provide some stability to the 25