Re-evaluation of area reduction factors and their impact on floods in South Africa

Abstract

The hydraulic design of structures depends on rainfall and runoff records to model and predict flood frequency for design purposes. The amount of runoff (and consequently flooding) from a catchment area has a direct relationship to the amount of rainfall in that catchment area, but as the point rainfall in any given area is likely to be more than the representative average rainfall over the entire area, a factor needs to be derived to transform point rainfall to representative areal rainfall. This factor is known as the Area Reduction Factor (ARF) and has been calculated and researched in many previous independent studies. The traditional methods for estimating ARFs are empirical in nature and include methods such as the UK Flood Studies Report (NERC 1975), the US Weather Bureau (USWB 1957), and Alexander (1990). The research presented in this paper focuses on re-evaluating ARF’s estimates using the UK Flood Studies Report (NERC 1975) and Alexander’s method (Alexander 1990) by following an analytical approach based on Bell’s method (Bell 1976). This research uses statistical methods to evaluate the validity of previously published ARFs. The results presented in this paper indicate that the empirical methods currently in use (specifically Bell’s nomogram) to derive ARFs may result in conservative estimates of flood peaks. There is ample evidence to suggest that modern analytical techniques can be applied to longer rainfall records in order to make a more accurate estimate of the ARFs.