Abstract
One approach in the stochastic modelling of rainfall is to derive a distribution function which has physically meaningful parameters. Such a distribution is derived from a Poisson process of rainfall occurrence with an independent exponentially distributed rainfall amount. The distribution appears to fit monthly precipitation totals extremely well.This type of model has many potential applications in climatology and hydrology. The homogenizing of long term rainfall sequences is one of these, and a methodology for this process is presented here. A 130 year record, from Auckland, New Zealand with three changes of site, is used to evaluate the procedure described.The significant advantage in the proposed method is that the homogeneity testing is done within the single rainfall record and the use of comparison stations is not required.

This publication has 9 references indexed in Scilit: