A computationally efficient local-inertial model for simulating storm surges and coastal inundation

Abstract

Tropical cyclones pose significant risks to low-lying coastal regions through storm surges and inundation, making numerical modelling crucial for disaster planning and evacuation. This study introduces IROMS-iS2D (Integrated River Ocean Modelling System-Inertial Surge 2D), a novel local-inertial 2D model that simulates storm surge and inundation within a single framework on unstructured grids. Unlike existing local-inertial models that rely on external surge inputs or constrained by structured grids, IROMS-iS2D dynamically simulates storm surges and subsequent inundation. It achieves a 10–15 times speedup over 2D shallow water models while supporting high-resolution grids (30–200 m). Cyclone forcing is derived from JTWC/IMD best-tracks using the Holland model, ensuring realistic surge inputs. IROMS-iS2D was validated for Bay of Bengal cyclones using INCOIS data, showing good agreement (R² > 0.9, RMSE ≈ 0.2 m, NSE >0.9). For Cyclone Mocha, Sentinel-1 validation showed 4359 km² inundation with 79.6 % accuracy and an F1-score of 61.7 %, demonstrating potential for real-time flood forecasting.