Progressive Risk Evaluation Model of Flash Flood Disaster Based on Watershed Reach Village and Its Application
Flood disasters in riverside villages of hilly regions are often triggered by heavy rainfalls,influenced by wa-tersheds,river reaches,and human activities.Risk assessment is crucial for prevention.This paper proposes a"water-shed-river reach-village"integrated risk assessment unit,and selects 16 indicators from these three aspects to form a risk indicator system.A progressive risk assessment model is established using random forest algorithm.The sample data of historical flash flood disaster were used for training.The Gini index was used to analyze the importance of each index.The 1 105 villages along the river in Weifang City were carried out application practice.The results show that the accura-cy of the risk assessment model by the random forest algorithm reaches 82.47%,indicating high precision;The contribu-tion rates of indicators at the watershed,river reach,and village levels are 30.79%,43.40%,and 25.81%,respective-ly,indicating that each level of the risk process has a certain impact.
万方数据
