Global sensitivity analysis in a complex 1D-2D coupled hydrodynamic model: Flood hazard and resilience perspectives over an urban catchment

Abstract

The increasing frequency of flood-related disasters has led to adopting advanced flood models to improve preparedness and response. To ensure reliable model outputs, conducting a Global Sensitivity Analysis (GSA) of model parameters is crucial. This study proposes a GSA framework for static input parameters in a 1D-2D coupled hydrodynamic flood model. MIKE-11 & MIKE-21 are coupled to simulate urban flooding in Mumbai's Mithi River catchment, considering rainfall and tidal influences. The model is simulated by perturbing various combinations of static input parameters to design test-scenarios. Outputs for an urban model setup are robust, complex, and gridded. Accordingly, GSA of static input parameters is also performed grid-wise to quantify sensitivity in terms of spatial variation of Flood Hazard and Flood Resilience across the catchment. Nonparametric probability density functions of flood depth at different locations are compared to calculate Kullback-Leibler divergence for quantifying sensitivity in the Flood Hazard context. Meanwhile, changes in Flood Resilience due to parameter perturbations are evaluated for resilience-based sensitivity. Results reveal varying impacts of input parameters across floodplain, with grid resolution and land use being most sensitive. The proposed novel GSA framework aligns with Sustainable Development Goal 11, aiming to make cities inclusive, safe, resilient, and sustainable, and equips flood management professionals with insights into key flood drivers, guiding data collection and monitoring. Proposed framework is versatile and can be integrated into any flood modeling software, offering resilient urban planning and risk mitigation strategies, contributing to sustainable urban development and better preparedness for flood risks in urban areas.