Evaluation of two models for the distribution of daily rainfall Amounts in Nigeria

Abstract

The distribution of daily rainfall amounts were modeled to the gamma and exponential distribution functions. Historical rainfall data from fourteen stations covering the vegetative zones of the country were fitted, by month to the two functions to estimate the parameters of the models. The gamma model yielded a better fit to the observed data. Using the parameters derived, rainfall data were stochastically simulated and compared with the observed data employing the non-parametric Mann-Whitney test. The gamma model generally reproduced the distribution of the observed data better than the exponential model except at two stations where they practically produced equally profitable results. This outcome was however not replicated at the other stations in the same vegetative zones as the two stations.

JARD Vol. 3 2004: pp. 61-74