Flood Susceptibility Mapping Using Image-Based 2D-CNN Deep Learning: Overview and Case Study Application Using Multiparametric Spatial Data in Data-Scarce Urban Environments

This study presents results for urban flood susceptibility mapping (FSM) using image-based 2D-convolutional neural networks (2D-CNN). The model input multiparametric spatial data comprised of land-useland-cover (LULC), digital elevation (DEM), and the topographic hydrologic conditioning derivatives, precipitation, soil types. implemented dropout regularization 2D-CNN with ReLU activation function, categorical cross-entropy loss AdaGrad optimizer produced case area FSM overall accuracy (OA) 82.5%. outperformed multilayer perceptron (MLP) network by 18.4% in terms corresponding lower MAE higher F1-measures 10.9% 0.989, as compared to 25.6% 0.877, respectively, MLP-ANN results. that map efficiency were evaluated under ROC curve (AUC) respective success prediction rates 0.827 0.809. Using 2D-CNN, 27% 247.7 km2 studied was mapped a high risk flooding, overestimating degree 4.7%. Based on gain ratio index analysis factors (FCFs), most significant FCFs LULC (18.5%), precipitation (14.9%), proximity river (13.3%), (12.4%). Soil types contributed 8.6%, slope 9.1%, DEM-derived hodological indicators 23.2%. demonstrate areas scarce hydrological monitoring networks, use can produce high-quality maps management environments.