An early warning system for inundation forecast due to a tropical cyclone along the east coast of India

Abstract

Cyclone-induced coastal inundation along the east coast of India is simulated using the state-of-the-art advanced circulation model coupled with Simulating Waves Nearshore model. Bathymetry of the computational domain is prepared by integrating digital coastal bathymetric chart with global digital elevation model (DEM) ETOPO-2. Topography of the coastal land region is created by merging data from the fine resolution airborne DEM with that from the Indian satellite CARTOSAT-2 derived CARTO2 DEM and Shuttle Radar Topography Mission DEM. The major shortcomings of coastal inundation simulation due to inaccurate representation of intricate networks of river, creeks and lagoons have been overcome by careful description of the critical coastal wetland features influencing the landward inundation of storm surge in the present work. Experimental simulation of coastal inundation is carried out for the Phailin (2013) and Hudhud (2014) cyclones. The computed surge residual for Phailin and Hudhud are compared with available tide-gauge observations, which have resulted in root mean square error of 0.06 m and 0.11 m, respectively. The simulations show realistic coastal inundations, which are compared with the Indian meteorological department records. The model set up is then used to generate forecast of coastal inundation due to Titli (2018) cyclone with a lead period of 48 h. The Weather Research and Forecasting model forecast winds at 5 km spatial resolution are used as the primary forcing parameter. The coastal inundation forecast due to Titli cyclone is validated by assessing the spatial coherency of simulated inundation with the water logged area demarcated from synthetic-aperture radar image taken from Canadian satellite RADARSAT-2. The results clearly indicate acceptability of the model performance with a potential to predict the coastal inundation during a cyclone for operational early warning.