Assessing Compound Flooding Hazards in Estuaries by Integrating a Climate Emulator and a Hybrid Metamodel

Abstract Compound flooding in estuaries results from the interaction of multiple forcings, including tides, storm surges, waves, precipitation and river discharge. These events present significant challenges for dynamic modeling due to their computational requirements. To address this, we present a hybrid statistical-dynamical framework for efficiently assessing compound flooding hazards. The framework incorporates a new version of the climate-based emulator TESLA, which generates synthetic extreme multivariate events by linking them to weather types that represent atmospheric pressure patterns. A multiscale approach considering Regional Weather Types (RWT) and Local Weather Types (LWT) captures the spatial variability of oceanographic and land drivers. TESLA is coupled with a hydrodynamic metamodel, enabling fast predictions of water levels and inundation under various forcing conditions while significantly reducing computational demands. Using flood proxies based on water volume, the framework identifies events associated with specific return periods. High-resolution simulations of these events provide detailed assessments of compound flooding hazards. Applied to the Asón River estuary in northern Spain, this approach demonstrates its ability to analyze historical climate variability, predict flood risks, and define mitigation strategies. It offers an efficient tool for exploring low-probability, high-impact scenarios in estuarine environments under current and future climate conditions.