Abstract
The data obtained from the Florida Area Cumulus Experiment in the years 1970–75 are analyzed statistically. Specifically a set of empirically derived predictors for both seeded and unseeded rainfall is identified. First the experiment is briefly described and the data given. The concept of echo motion categories is presented. The responses to be predicted and the variables used as predictors are listed and described and the methods for obtaining the prediction models are given. Next comes a listing of the model equations obtained by those methods, along with some commentary on their possible physical meaning. Examples illustrate the use of some of these prediction models for estimating seeding effects and possible bias in selection of experimental days. A discussion of the echo motion covariate and the basic predictor variables, their histories, rationales and some theoretical indications of their importance completes the main body of the paper.
