A Statistical Model for Flood Discharges of the Damodar River

Flood flow analysis is an important aspect of any water resources project. Flood being a hydrologic event, can be thought to be the outcome of a stochastic process. Hence, along with frequency analysis, one must also examine the sequential relationship of a flood series.

In the present paper, the annual maximum flood flow series for a period 1934 to 1961 of the river Damodar are analysed. Based on Gumbel's suggestion that the distribution of extreme values was appropriate for the analysis of annual maximum flood flow, the magnitudes of flood discharge for various return periods were calculated. Hershfield technique was applied to calculate the probable maximum flood, but the result obtained by this method was not found to be useful.

The formulation of a transition probability matrix shows that the flood series does not possess a significant serial correlation. Hence, a random number model with a definite range and definite mean was selected and with the help of this model flood flows for future years were generated synthetically.