Intercomparison of probability distributions for selecting a best fit for estimation of rainfall

Abstract

Estimation of rainfall for a given duration and return period is considered as one of the important design parameters in hydrological studies for planning, design and management of civil and hydraulic structures. This can be computed by fitting a probability distribution to the annual 1-day maximum rainfall observed at the rain-gauge site located within the vicinity of the study area. This paper presents an analysis of rainfall estimation at Dahanu using an extreme value family of probability distributions, namely, Generalized Extreme Value (GEV), Extreme Value Type-1, Extreme Value Type-2, and 3- parameter Pareto, wherein the parameters are determined using the Method of Moments, Maximum Likelihood Method, and Method of L-Moments (LMO). The adequacy of fitting probability distributions adopted in rainfall analysis is evaluated by quantitative assessment through Goodness-of-Fit (viz., Chi-Square and Kolmogorov-Smirnov) and diagnostic (viz., D-index) tests, and qualitative assessment by using the fitted curves of the estimated rainfall. The study shows that the estimated 1-day maximum rainfall given by GEV (LMO) distribution could be considered as a design parameter while designing the civil and hydraulic structures at Dahanu.