Another Family of Distributions for Describing and Analyzing Precipitation Data

Abstract

A positively skewed two-parameterer family of distributions is introduced as an alternative description of precipitation amounts. Each distribution of this family is associated with a simple two-sample non-parametric test which, for large samples, is optimum in detecting scale changes induced, say, by cloud seeding. Since the nonpararmetric test in question here is specified by the shape parameter of this family, two procedures are given for estimating parameters of this family from observed precipitation data. In addition, specific comparisons are made between this new family and corresponding gamma and one-sided t families.