DEVELOPMENT OF A GEE-BASED FLOOD INUNDATION MAPPING FRAMEWORK

Abstract

Flooding is a significant threat to the Lower Narmada Basin (LNB) in central India. Accurate flood inundation mapping is necessary for effective risk assessment and mitigation. This study presents an open-source framework that integrates the LISFLOOD-FP hydrodynamic model with the Google Earth Engine (GEE) platform for automated flood inundation mapping in the LNB. The framework leverages open-source tools and automation techniques, streamlining the workflow from daily inflow data retrieval to flood map generation. A Python script automates the process, enhancing efficiency and reducing human error. The LISFLOOD-FP model is parameterized using a 30m digital elevation model (DEM) and flood observations that simulate river flow and inundation dynamics. The model is calibrated and validated against HEC-RAS simulation data for the 1994 flood event, with a maximum discharge of 60,460 cumecs, achieving a fitness value of 0.995 for steady conditions and 0.996 for unsteady flow. Validation against HEC-RAS simulations demonstrates the model's accuracy, with the LISFLOOD-FP-generated flood map closely resembling the HEC-RAS output (541 km² inundation area). Moreover, the study obtained a satisfactory comparison result of flood extent with a Hit rate of 0.92, a False alarm ratio of 0.002, and a Critical success index of 0.97, indicating high agreement between the model and benchmark flood extents. The contingency map analysis reveals regions of agreement and disparity between the model and benchmark, providing a comprehensive understanding of model performance. The open-source framework enables automated daily generation of flood inundation maps and immediate display on the GEE app, enhancing accessibility for a wider audience. Its integration streamlines time-consuming workflows making it error-free, and enabling rapid map dissemination that offers significant advantages over traditional methods. By empowering stakeholders with accessible flood inundation maps, this study contributes to proactive adaptation, disaster mitigation, and resilience improvement in flood-prone areas.