Rainfall Frequency Analysis Using Probability Distribution Models: A Case Study of Imphal West District, Manipur, India
Samananda Keisham *
Department of Environmental Science, Manipur University, Canchipur, Imphal-795003, Manipur, India.
Bindiya Aribam
Department of Environmental Science, Manipur University, Canchipur, Imphal-795003, Manipur, India.
Mohd Kamaluddin Sheikh
Department of Environmental Science, Manipur University, Canchipur, Imphal-795003, Manipur, India.
Wazir Alam
Department of Environmental Science, Manipur University, Canchipur, Imphal-795003, Manipur, India.
*Author to whom correspondence should be addressed.
Abstract
Flood frequency analysis is a crucial step in hydrology for the prediction of flood risks and designing suitable region-specific water management strategies. Selecting the most suitable probability distribution model is essential for accurate flood predictions. GEV is a best-fit model as compared to other distributions due to its flexibility in modelling extreme values. The choice of parameter estimation methods and model selection criteria significantly impacts the identification of the best-fit distribution. The study aimed to analyse the frequency of extreme rainfall events based on precipitation data from 2013 to 2023 for Imphal West District. The district is facing water scarcity issues during the lean season. It is imperative to predict the future rainfall quantiles for sustainable supply and policy inclusion of the water security issues at the district level. Kolmogorov-Smirnov, Anderson-Darling, and Chi-square tests are employed for goodness-of-fit testing (α = 0.05 confidence level), revealing that the Generalised Extreme Value distribution is the best-fit model in all goodness-of-fit tests in terms of ranking, followed by the Log Pearson III model for simulating extreme rainfall events in the district.
Keywords: Frequency analysis, probability distribution, generalised extreme value, log Pearson III, lognormal, goodness of fit