Use of the index-flood method at ungauged sites requires methods for estimation of the index-flood parameter at these sites. This study attempts to relate the mean annual flood to site characteristics of catchments in KwaZulu-Natal, South Africa. The ordinary, weighted and generalised least square methods for estimating model parameters are compared and found to perform equally well, with preference given to the generalised least square model. A separation of KwaZulu-Natal (KZN) into two regions was found to improve predictive ability of the models in the western and north-western parts of the study area. The study also revealed problems with the estimation of the mean annual flood in the coastal areas of the study region.
|Publication status||Published - 2001|