Delineation of flood-prone areas using geomorphological approach in the Mekong River Basin

For many river basins, flood maps have not been developed, especially in developing countries. Generally, flood hazard maps are generated by a modeling process based on past flood events and require a large amount of input data for the basin. The present study aims to delineate flood-prone areas in the Mekong River Basin from geomorphological features by using linear binary classifiers and receiver operating characteristics (ROC) analysis. Our method investigates the performance of five single features and six composite indices associated with flood hazards. The results indicate that elevation difference to the nearest river network has the best performance among all single features and composite indices with predicted abilities success rate SR = 63.15% and modified success rate MSR = 78.21%. This study shows that the combination of linear binary classifiers and ROC analysis can be used to detect flood-prone areas not only in small basins but also in a large basin such as the Mekong River Basin. This approach is advantageous for basins that lack observation data because the method does not require a large amount of basin information (i.e. all features and indices are derived from a digital elevation model). The outcomes of this study can provide useful information for identifying areas that are prone to flooding.