WMOD00289 (3) - Laserfiche WebLink

Home Browse Search

'1 5 ice crystals by deposition and riming into graupel, and fallout of the graupel particles (perhaps already melted into raindrops) from cloud base (Silver- man, 1980). This physical hypothesis, which involves a specific version of the Bergeron Process, was adopted on the basis of several years' work which showed that Montana clouds did not have characteristics generally considered conducive to success in hygroscopic seeding (e.g., Cotton, 1982), and that dry ice was a satisfactory seeding agent for forming ice crystals in them at temperatures between -100C and OOC (Holroyd et al., 1978). The experimental unit for HIPLEX-1 was a single cumulus congestus cloud. The seeding agent was dry ice, which was dropped from a jet aircraft at a rate of 0.1 kg km-1 on passes through the cloud near the -100 C level, or at cloud top, whichever was lower (in elevation). Test clouds were selected principally on the basis of data collected on a pretreatment pass by a King Air instrumented aircraft operating near the - 80 C level. The target clouds are classified as types AI, A2 and B (Silverman, 1980). The differences in behavior between seeded and unseeded clouds were expected to be most readily detected in type Al clouds and hardest to detect in type B clouds. Type Al clouds were considered unlikely to precipi- tate naturally because of warm tops (- 6 to -120C) and a presumed absence of ice multiplication processes. Type A2 clouds were clouds in the same temperature range but in which an ice multiplication process was considered likely to occur. Type B clouds were considered likely to precipitate natural- ly, with or without ice multiplication, due to their relatively cold tops (most- ly -12 to - 200 C). Type B clouds were selected for experimentation only when no suitable type A clouds were available. The experimental unit was limited in time to 40 min, as observations and numerical modeling results have indicated that the hypothesized physical process should take place in less than 40 min. Flight procedures were worked out to ensure proper coordination between the seeding aircraft and the King Air instrumented aircraft, which systematically worked downward from the - 80 C level to mo- nitor the precipitation process. Details are given in 'The Design of HIPLEX- l' (BuREC, 1979). The statistical design of HIP LEX-1 included double-blind randomization, the selection of ten primary response variables (Table I) to test the steps of the physical hypothesis, and the development of new statistical methods based on MRPP (multi-response permutation procedures) for testing for pos- sible differences between seed and no-seed cases (Dennis et al., 1980). Simu- lation experiments indicated 50 to 100 test cases would be required for con- clusive results to be obtained. Only 20 test cases were recorded during the 2 years, 1979 and 1980, that HIPLEX-l was in the field. However, the experimental design permitted the drawing of some useful conclusions from the truncated data set. The following section, which summarizes present knowledge conceming aspects of augmentation of rainfall from cumulus clouds, draws upon the re- sults of HIPLEX in general and of HIPLEX-1 in particular.