Assessment of daily stochastic rainfall generation in Southern Africa

Abstract

This research report sets out to assess daily stochastic rainfall generation applied on Southern African historic rainfall data. The research aims to evaluate the performance of three methods, namely, Transition Probability Matrix (TPM), Weather Generator (WGEN) and Variable Length Bootstrap (VLB). The evaluation was based on the ability of the methods to preserve the daily, monthly, and annual statistics of the observed data, simplicity, and ease of application of the methods. From the evaluation of the performances from all three methods, it can be concluded that all the methods are capable of generating daily data with statistics that preserve the observed historic data, although to different degrees of accuracy. It was also found that utilization of software such as GoldSim makes the application of the methods more convenient and faster.

The evaluations have shown that it is imperative to have a comprehensive understanding of both advantages and disadvantages of the methods and their limitations. Although some of the methods were originally not developed with intention of being applied in South Africa, they can generate data for different climatic conditions that they were not originally developed for.

It is recommended that future research focus more on how the impacts of climate change can be incorporated into the methods. Additional performance assessment for other areas of Southern Africa including the winter rainfall regions of the Western Cape Province and drier areas of South Africa is recommended.