Evaluation of Kovacs 1988 Regional Maximum Flood Method

Abstract

Estimation of design flood peaks is required for the design and evaluation of hydraulic structures. Methodologies developed between the 1960s and the late 1980s, such as the Regional Maximum Flood (RMF), are used to estimate extreme flood peaks in South Africa. The RMF method uses envelope curves to estimate the maximum flood that can be expected in a specific region in South Africa and neighbouring countries, and represents an indicative upper flood limit for other flood estimation methods. The method developed by Kovacs for South Africa was last updated in 1988. Using more than 30 years of additional data for analysis, this paper highlights the RMF's perceived shortcomings. In two RMF regions, updated observed flood peaks exceeded the existing envelope curves. In 78% of the catchment areas of the evaluated stations, the station RMF values were at least 50% more than the observed maximum flood peak. When the different parameters from the Kovacs 1988 and present DWS (Department of Water and Sanitation) datasets were compared, 98% of the evaluated stations had different flood peaks recorded/reported, while 33% of the stations logged different catchment area sizes. Kovacs 1988 ratios used to estimate flood peaks at different probabilities of occurrence, using the RMF, were found to generally over-estimate expected flood peaks. It is concluded that the 1988 RMF method needs to be updated to still provide relevant guidance.