A methodology for the estimation of short duration design storms in South Africa using a regional approach based on L-moments

The network of recording raingauges in South Africa with long records is relatively sparse and only 55 out of a total of 412 raingauges currently have record lengths longer than 30 years. In addition, much of the digitised rainfall data from numerous stations are viewed as being unreliable, with important events missing and errors in the digitised data. In an attempt to overcome data limitations and to improve reliability, a regional, index storm approach based on L-moments has been applied to estimate short duration (≤24 h) design rainfalls in South Africa. The regionalisation was performed using only site characteristics (latitude, longitude, altitude, concentration of precipitation, mean annual precipitation (MAP), rainfall seasonality and distance from sea) and required a minimum of subjective relocation of stations in order to obtain relatively homogeneous clusters. The at-site data were used for independent testing of the clusters of the stations for homogeneity. Using information for 172 rainfall stations which have at least 10 years of short duration rainfall records, 15 relatively homogeneous clusters were identified in South Africa. The index storm used was the mean of the annual maximum series and a relationship was derived between the index storm and MAP. This methodology enables short duration design storms to be estimated at ungauged sites in South Africa. The accuracy of design storms estimated using this technique was assessed at a site not used in the regionalisation procedure and error bounds were computed for the quantile growth curves for each of the clusters.