Copula-based joint distribution of flood peak and volume for improved flood risk assessment: a case study of the Surma-Meghna River Basin, Bangladesh

ABSTRACT This study applies a copula-based multivariate approach for assessing flood risk, focusing on the interdependent behaviors of flood peak and volume in the Surma-Meghna River Basin, Bangladesh. Traditional univariate analyses often relied on only flood peak or volume, overlooking correlations among hydrological variables and producing fewer reliable risk estimates. Using 53 years of hydrological data, we explored joint distribution models to capture these dependencies accurately. For univariate frequency analysis of flood return periods, we employed the Weibull formula alongside its corresponding marginal distribution. The flood peak was modeled using a Log Pearson type III distribution, while the flood volume was best described by a Log-normal distribution. Six copula families were assessed, with the Clayton copula showing the best fit, particularly in capturing lower tail dependence, making it effective for regional flood risk analysis. The findings highlight the usefulness of the copula framework in offering a fuller understanding of flood risk by capturing interdependence between key variables – often underrepresented in univariate methods. This multivariate approach provides practical insights for hydraulic design, floodplain management, and extreme events. By enabling more accurate modeling of joint distributions, the study supports improved flood forecasting and the development of more effective resilient flood risk management strategies.