| Summary: | Flood frequency analysis is critical in flood planning and management and hydraulic structures design. While univariate flood frequency analysis (using the peak flow) is still widely employed in developing countries, how does it compare to the robust copula‐based bivariate flood frequency analysis remains unknown. Moreover, there is also a decade‐long critical question whether less data requiring hydrological models can be an alternate to the data‐intensive models in flood prediction, especially in a developing tropical country like India? To answer these questions, this study aims in comparing two types of hydrological models (IHACRES, a less data requiring model, and VIC‐3L, a data‐intensive model) in simulating the peak flows, following which the simulated peak flows are used in a detailed comparison of the univariate and bivariate flood frequency analysis. The results indicate that the data‐intensive fully distributed hydrological model performs poorly relative to the conceptually lumped IHACRES model at the study catchment in simulating the peak flows. Moreover, both univariate and copula‐based bivariate flood frequency analyses show similar peak flows for a given return period at the study catchment. Given that bivariate flood frequency analysis accounts for both peak flow and flood volume, it is recommended over the univariate flood frequency analysis since the results are widely applicable for flood planning and hydraulic structure designing the developing countries.
|
|---|