Abstract

Flash flood is one of thedangerous devastating natural hazards occurring in almost mountainous riverbasins in the world and especially in Northwestern Vietnam. The multi-layerartificial neural network was applied to establish a flash flood zoning model:a case study of Yen Bai province, which has allowed to establish a flash floodzoning model with high-precision statistical indicators (ROC = 0.960 and AUC =0.951). The results also allow establishing the flash flood zoning model inother areas of Vietnam, making a positive contribution to the planning andconsolidation of the solutions, and responding to climate change.