Key points are not available for this paper at this time.
Flood frequency analysis is a challenging but essential hydrologic problem for design of control structures and water resources management. The design flood estimates based on traditional stationary assumption may lead to inaccurate estimation of flood risk because of non-stationarity and the compounding impacts of several drivers in a dynamic environment. Copulas are a useful and adaptable technique for determining the multivariate joint dependency amongst flood variables. This study employed time-varying copula models to investigate the nonstationary dependence structures between two highly correlated flood variables, such as flood peak and flood volume, in order to determine the joint and conditional return periods of the flood events revealed by the 2018 Great Kerala floods. The proposed approach is executed for two potential locations of high flood risk namely, Periyar river basin and Greater Pamba river basin of Kerala, India. The Archimedean copula (Clayton, Frank and Gumbel) parameters were estimated using Maximum likelihood estimation and the optimal copula selection was made using Akaike Information Criterion. The non-stationary joint return time was found to be shorter than the stationary joint return period, suggesting that the extreme flood occurrences happened more frequently in the non-stationary bivariate study. Thus, it can be demonstrated that the extreme flood episodes are underestimated by stationary bivariate flood frequency analysis. The validation of results by comparing the flood magnitude of Neeleswaram station for 2018 flood quantile ascertained the necessity of non-stationary flood risk estimation. The study advocates the conduct of multivariate frequency analysis over the univariate analysis for the risk assessment of hydrological extremes. The results demonstrate that the long-term decision-making methods need to be updated to account for the oddities of the nonstationary climate. This study rendered flood risk assessment indicators as well as a risk-based design approach for hydraulic infrastructures in a non-stationary environment, which is crucial for climate change adaption and water security management.
Adarsh et al. (Fri,) studied this question.