WMOD00528 - Laserfiche WebLink

Home Browse Search

4. DESIGN CONSIDERATIONS FOR A WEATHER MODIFICATION DEMONSTRATION PROGRAM One of the purposes of this study is to provide a preliminary design of a weather modification demonstration program should winter clouds over the Sevier Drainage appear seedable. TIley clearly do appear seedable a large fraction of the time but the effectiveness of current seeding approaches has yet to be demonstrated as discussed in section 3. Accordingly, consideration has been given to an appropriate design of a weather modification demonstration program as will be discus~ed. 4.1 Review of Past Experimental Approaches and Recent Developments Several statistical experiments have been conducted with winter orographic stonns in the West but most produced inconclusive results. Past statistical experiments typically had a number of common characteristics. Hypotheses were stated, with varying degrees of detail, which noted the chain of physical events expected to follow seeding. In broad tenns, the hypothesis would note that SL W in the fonn of tiny droplets would have to exist within the cloud, and seeding would have to convert some of these droplets to ice crystals capable of growing and settling to the surface as snow or melting and falling as rain. When a presumably suitable stonn appeared or was forecast, a random decision was made to seed or to reselVe the event as a nonseeded control case. In either event, the same obselVations were taken. After a number of field seasons, cases were statistically tested for treabnent effects on precipitation. The entire data set might be partitioned into meteorologically similar categories; for example, by ranges of estimated cloud top temperature, and each category would be statistically tested for diffen::nces between seeded and nonseeded precipitation, ObselVations usually were limited to target area precipitation and some general indications of the stonn structure. Such efforts have been referred to as "black box" experiments because, if the statistical testing did not indicate significant differences between populations of seeded and nonseeded experimental units (stonns, days, etc.), insufficient physical obselVations existed to detennine where the hypothesized chain of events failed, Even when statistical testing suggested significant precipitation differences, the exploratory nature of most analyses left doubts about the estimated probabilities that the differences were not due to chance. Furthennore, uncertainties remained about the physical mechanisms involved so that transferability of results to other locations was questionable. Even if statistical testing indicates that a particular seeding program produced significant precipitation increases, there is no assurance that the seeding method employed was highly efficient. Perhaps greater in~reases could be achieved with a different treatlhent approach. Thus, the seeding process cannot be optimized without a thorough physical understanding. Improvements in numerical modeling of winter orographic clouds have significantly aided our understanding of airflow and microphysical processes (e.g., Young, 1974; OaIk, 1977; Cotton et al" 1986; Bruinjtes et al" 1991). Incorporating obselVations from a particular mountain region into a numerical model adapted for that region can markedly increase understanding of the key processes involved and how seeding influences them. Less complex models can be used in real time to guide the conduct of physical seeding experiments (Rauber et al., 1988). Reynolds (1988), in a review of winter snowpack augmentation, showed that a consistent picture is emerging between recent physical studies and suggestions from earlier statistical experiments. For example, Reclamation scientists have provided convincing evidence that the physical seeding hypothesis was correct in a limited number of experiments in recent years, Super and Heimbach (1988) were able 33