CS(AR)-1/2002-03 : Regional Flood Frequency Modelling for river Parvati (M.P.)

Abstract

Determination of magnitude and the probability of likely occurring floods is of great importance for the solution of variety of water resources problems such as design of various hydraulic structures, urban drainage system, flood plain zoning, economic evaluation of flood protection works, flood insurance computation etc. Flood frequency analysis is a tool used to estimate the frequencies of likely occurrence of future floods. In this approach the sample data are used to fit frequency or probability distributions, which in turn are used to extrapolate from recorded events to design events.

In the present study, flood frequency analysis has been applied for river Parvati in Madhya Pradesh. The river Parvati is an important tributary of Yamuna river system, which rises in the Vindhyan Mountains near Ashta village in Sehore district of Madhya Pradesh. The annual flood series analysis has been carried out to estimate the flood quantiles at different return period at Pilukhedi site of river Parvati. The regional parameters for basin have also been evaluated using limited data of two other sites in the basin. The regional approach provides a significant advantage of estimation of floods at any sites in the homogeneous region with very less or no data.

In the at-site analysis of annual flood series the Normal, Log normal, Pearson type III, Log Pearson type III, EV- I and Log EV distributions were applied using method of moments. The Wakeby IV, Wakeby V, EV- I, GEV distributions have been applied using probability weighted moments. The L-moments- have been used for estimation of parameters of EV-I, GEV, Logistic, Generalised logistic, Generalised Pareto, Normal and Log normal distribution. The standardized PWMs and L-moments approach have been used to estimate regional parameters in the basin. The L-moment ratio diagram and other tests based on simulation technique were used to evaluate the best-fit distribution. The regional parameters may be used to compute the flood frequency estimates at any ungauged sites in the region

From the analysis of different goodness of fit tests, it has been found that the GEV distribution with L-moments as parameter estimation found the best-fit distribution for Pilukhedi and other sites in the region. . It is recommended that the regional parameters for Parvati basin may be used only for primary estimation of flood and should be reviewed when more regional data available.