Rainfall results, 1970-1975: Florida area cumulus experiment

The latest rainfall results of the Florida Area Cumulus Experiment (FACE) are discussed after a review of the background, design, and early results of this experiment. Analysis without the benefit of data stratification and appropriate covariates of the 48 random experimentation days obtained through 1975 provided no evidence that dynamic seeding appreciably altered the rainfall over the fixed target area (1.3 x 10(4) square kilometers). Partitioning of the experimentation days according to whether the convective echoes moved across the Florida peninsula or developed in situ was more informative. Use of this echo motion covariate with five meaningful predictor models of natural rainfall in a stepwise regression program produced persuasive evidence for an effect of seeding in both echo motion categories. For days with moving echoes, there is evidence for a positive, statistically significant treatment effect on the rainfall from the subject clouds (the floating target) and in the overall target area. The results for days with stationary echoes, although considerably more tentative, suggest that seeding produces more rainfall in the floating target but with no net change of the precipitation in the overall target area. The ramifications of this result and a possible explanation are discussed. Corroborative statistical analyses and discussion are presented, including a discussion of the physical bases and history of the echo motion covariate and the meteorological predictors, analysis that is supportive of the rain-gage-adjusted radar measurements of precipitation in FACE and results of relevant cloud physics measurements in Florida.