Indian Journal of Ecology

In the classical method of estimation, the GEV distribution was identified as the most appropriate model for estimating peak flood heights at twelve study sites in the Mahanadi River Basin. This paper presents that Bayesian parametric estimates of the GEV distribution are better compared to maximum likelihood estimates in estimating peak flood heights and their return periods at these sites. To arrive at this target Markov Chain Monte Carlo Bayesian technique is utilized to acquire parameters of GEV distribution. The estimates of Bayesian approach for peak flood heights and their return periods at the sites showed better predicted flood peak return periods with the Bayesian method. These were shorter than estimates obtained using the maximum likelihood method for all the sites.