On the Design and Evaluation of Cumulus Modification Experiments

摘要: Abstract Combination of numerical simulation, many simultaneous measurements, and a large assortment statistical tools, employed at all stages, have been found useful in design evaluation modification experiments on cumulus clouds. A randomized sample is essential, although non-random controls supplemented it by providing necessary information natural distributions.Obstacles to definitive estimates treatment effects are huge variability compounded the expense labor involved obtaining an adequately data sample. 26 pair set from dynamic seeding experiment isolated Florida cumuli used here illustrate both problems combined approach overcome them. In this set, rain volumes unmodified single varied three orders magnitude days screened as suitable. The field phase cost above $250,000, requiring instrumented aircraft, calibrated radar, several radiosondes daily.Numerical simulation o...