A STOCHASTIC MODEL FOR GENERATION OF DAILY RAINFALL DATA

Abstract

A long term runoff sequence is necessary for planning of any water resource development scheme. Generally long term historical runoff series are not available and the available data are insufficient to identify the desired runoff characteristics. Even if the long term historical series is available, it is not likely to be repeated in .future. Therefore, attempts have been made to generate synthetic runoff sequences using statistical parameters. Observed runoff at a point depends on the hydro—meteorological characteristics and the physiographic features of the basin. The physiographic features of the basin generally change with the time due to natural changes as well as human interferences. Thus, there is likelihood of variations in runoff characteristics with time and generated runoff sequences may not represent a realistic and correct picture. The effect of variations in physiographic charac-teristics on runoff can be avoided by generating rainfall sequences and using the same to compute the runoff series with the help of a deterministic model in which the physio-graphic features can be explained by suitable parameters.