An enhanced framework for simulating urban pluvial flooding: Integrating nested watersheds and urban areas with spatial heterogeneity

Urban pluvial flooding caused by rainstorms has become an increasingly threat to urban safety due to climate change and rapid urbanization. Coupled hydrological-hydrodynamic models are now widely used for urban flood simulations, enabling improved pluvial flood forecasting. However, existing research often focuses mainly on modeling urbanized surfaces, overlooking the heterogeneity within different functional zones of urban underlying surfaces. Additionally, most simulations treat urban areas in isolation, failing to incorporate upstream basins in a nested framework with the urban environment. This simplification can lead to an incomplete characterization of urban hydrological features and neglect the hydrological connectivity between catchments and urban areas. This research aims to develop a framework Hybrid Heterogeneous Urban-Catchment Flooding Model (HHUCFM) for simulating urban pluvial flooding that incorporates the heterogeneity of underlying surfaces and nested basin runoff modeling, and introduces a lightweight, semi-distributed hydrological model as a surrogate model for urban upstream basin scale. The framework consists of three main components: (1) optimization and uncertainty analysis of heterogeneous parameters based on urban functional zoning and heuristic algorithms; (2) the semi-distributed model FLOWS-Tank, which integrates 4 layers of series-parallel tanks and 2 nonlinear reservoirs to construct differential equation-based control equations, to connect runoff from upstream basins; and (3) a multi-process physically-based urban flood simulation approach incorporating FLOWS-Tank. HHUCFM was applied in Jinan, and the accuracy was validated through observed floods, identifying overflow nodes and surface inundation. The results reveal: First, incorporating zoning-based hydrological characteristics in areas with diverse underlying surfaces, critically highly urbanized, mountainous, and suburban zones, significantly enhances runoff prediction, as evidenced by improved Nash-Sutcliffe efficiency coefficient and increased interpretability of spatially heterogeneous parameters. Second, the proposed framework HHUCFM embedded with FLOWS-Tank, was successfully applied to the study area with hydrologically connected upstream basins and urban zones, demonstrating strong generalizability and interpretability in pluvial flooding simulation in urban scale. The current research provides valuable insights for analyzing urban flooding in regions with typical underlying surface heterogeneity.