Application of Copulas in Hydrometeorological Drought Risk Analysis Under Climate Change Scenarios- a Case Study

An in-depth understanding of drought frequency analysis in a river basin is possible only with a drought characterisation study. Multiple drought characteristics associated with the drought make it essential to analyse their joint behaviour in drought frequency analysis. Conventional univariate frequency analysis may produce overestimated or underestimated drought risk magnitudes. This study utilised the capability of bivariate copulas to construct the joint dependency amongst four drought characteristics (severity, duration, peak, and interarrival time) derived from the drought indices (Standardised Precipitation Index (SPI), Standardised Precipitation Evapotranspiration Index (SPEI), and Standardized Streamflow Index (SSI)) in a tropical river basin, the Bharathapuzha, India, during the historic period and climate change scenarios (RCP 4.5 and RCP 8.5). Appropriate distributions were selected for modelling the drought characteristics to capture the probabilistic behaviour. The best marginal distribution of each characteristic is obtained from the goodness of fit measures. Various copulas from the Archimedean and Elliptical families were applied to construct the four-dimensional joint distributions. Subsequently, the best-fit copula obtained the joint return periods. The results of joint dependence show that the Clayton and Gaussian copulas best fit with meteorological and hydrological drought, respectively, and the spatial investigation at the median threshold of the joint return period provides the hotspots of drought recurrences in the river basin with return periods in the range of 2 to 8 years during the historic period, greater than four years and greater than six years for RCP 4.5 and 8.5 scenarios.