Spatial variability of rainfall: Deciphering flood characteristics and model precision

This study investigates the intricate role of spatial variability in rainfall (SVR) concerning flood characteristics and its impact on refining flood prediction models. A spatial variability index was used to classify rainfall events into two categories: spatially homogenous (Class A) and heterogeneous (Class B). The analysis of historical flood events suggests that the SVR influences flood peaks. This research introduces a novel approach to assess SVR's role in calibrating hydrological models, subsequently improving model selection. By separately calibrating Class A and B events within both lumped and distributed models, the models yield superior results as compared to the conventional approach. For the catchments considered, lumped model demonstrated heightened performance for Class A events, while distributed models outperformed in Class B events. This study not only underscores the influence of SVR on flood dynamics but also underscores the efficacy of event-based classification in refining hydrological models for superior flood prediction accuracy.