Modeling the concurrent impact of extreme rainfall and reservoir storage on Kerala floods 2018: a Copula approach

Abstract

This study investigates the concurrent impact of reservoir storage and extreme rainfall on the devastating Kerala floods 2018 using Copulas. First, the effect of storage of four major reservoirs Idukki, Idamalayar, Kakki and Kallada is investigated by developing storage curves. It is noticed that in the four reservoirs, the storage level was above the decadal mean storage in most of the days of July–August period of 2018. An analysis on the storage levels of the four reservoirs showed that the water level crossed the spillway crest level of the respective reservoirs by July 26, 2018 and the controlled release might have helped in alleviating the flood disasters. In the Copula-based modeling, first, the marginal distribution of monthly storage and rainfall of the four reservoirs are developed and it was found that the generalized Pareto (GP) distribution was fitting best for all the storage data. The exponential fit was found to be the best for rainfall of all reservoirs except that of Kakki. Subsequently, the joint probability of reservoir storage and rainfall is investigated using three popular Copulas of Archimedean family namely Gumbel–Hougaard (GH), Clayton and Frank and the best Copula is identified for each case. The return period curves are developed for both joint and conditional cases using the best fitted Copula. The sensitivity analysis of joint return period showed that the rainfall of Kakki and Idukki at 2018 was extreme and the effect of rainfall is dominant in these reservoirs.