WMOD00572 - Laserfiche WebLink

Home Browse Search

Printed January 30, 1990 Ground-based generators on windward slopes were added for Israel II (Gagin and Neumann, 1981). The AMS (1984) statement noted that the apparently favorable statistical results, amounting to roughly a 15 percent increase in target-area rainfall, enjoyed "general acceptance." However, recent reanalyses of the Israeli data by Gabriel and Rosenfeld (1989) have raised serious questions about the statistical analyses. Their concern is that, while both projects were designed as randomized crossover experiments, the analysis of Israel II results by Gagin and Neumann (1981) used a target-control design and was applied to the north target area only. Statistical analysis of Israel II results according to the original design does not confirm the favorable results: of IsraelI. Furthermore, Rangno (1988) has pointed out that, contrary to the claims made in Gagin and Neumann (1974), natural rain does fall in Israel from clouds with top temperatures warmer than _lOoC, suggesting that natural precipitation processes are quite efficient, at least in some storms. The best known randomized experiments in this country to test effects of seeding of convective clouds upon area rainfall were the Florida Area Cumulus Experiments (FACE-l and FACE-2). The experimental design emphasized dynamic effects, which were to be produced by heavy seeding with silver iodide pyrotechnics dropped from aircraft. Typically, 10 to 20 pyrotechnics, each containing 50 g or more of silver iodide, were dropped into each treated cloud. Although much effort went into the experimental design of FACE, including numerical cloud modeling and laboratory testing of the silver iodide pyrotechnics used, and some cloud physics data were collected, the main emphasis remained on statistical analysis of precipitation data. FACE-1 gave some promising indications of areal rainfall increases, but FACE-2 failed to confirm for either the fixed target area or for smaller floating targets incorporating the clouds actually seeded each day (Woodley et a1., 1983). Its failure to obtain definitive answers is an indication of the need for a more comprehensive, engineering approach to convective cloud seeding for precipitation augmentation. One of the first experiments on convective clouds to take full advantage of the particle-measuring probes was HIPLEX-l. HIPLEX-l was conducted in southeastern Montana. It was built around a specific, detailed hypothesis of the sequence of events that should follow the introduction of dry ice into carefully chosen, supercooled cumulus congestus clouds of the northern Great Plains (Smith et aI., 1984). According to the hypothesis, the dry ice would produce numerous ice crystals, some of which would grow into graupel by deposition of water vapor followed by riming. The graupel particles would then melt into raindrops during their fall from cloud base to the ground. Because the target clouds had low concentrations of supercooled water, dynamic effects were expected to be negligible. Following some preliminary trials, randomized seeding was carried out during the summers of 1979 and 1980, with 20 test clouds being recorded. Test clouds were selected on the basis of cloud physics data collected on a pre-treatment pass by an instrumented aircraft, analyzed by an on- board computer, and displayed to the flight scientist. It was anticipated that ice crystals would be detectable within 2 minutes of seeding, that some rimed crystals would appear in about 5 minutes, and that graupel particles would be present 8 to 10 minutes after seeding. Response variables, derived from aircraft and radar observations following treatment, were objectively defined to reflect the expected hydrometeor development (Smith et a1., 1984). 8