Machine Learning Techniques and hydrological Modeling for Flood Susceptibility and Inundation Mapping: Case study VGTB River Basin, Vietnam

Vietnam has experienced many natural disasters, particularly typhoons. This study aims to examine three machine learning (ML) approaches—random forest (RF), LightGBM, and CatBoost—for flooding susceptibility maps (FSMs) in the Vu Gia-Thu Bon (VGTB) River Basin of Vietnam. The results of ML are compared with those of the rainfall-runoff model. Ten independent factors that influence the FSMs in the study area, namely, aspect, rainfall, curvature, DEM, horizontal distance from the river, geology, hillshade, land use, slope, and stream power index, are assessed. An inventory map that includes approximately 850 flooding sites is considered based on several post-flood surveys. The inventory dataset is randomly divided into two sets: training (70%), and testing (30%). The AUC-ROC results are 97.9%, 99.5%, 99.5% for CatBoost, LightGBM, and RF, respectively. The FSMs developed by the ML methods show good agreement in terms of extension with flood inundation maps developed using the rainfall-runoff model. The FSMs show that downstream areas (both urbanized and agricultural) are under "high" and "very high" levels of susceptibility. The developed FSMs for such typhoon-prone regions can be used by decision-makers and planners in Vietnam to propose effective mitigation measures for community resilience and development.