Geological hazard risk assessment and rural settlement site selection using GIS and random forest algorithm

Abstract

It has been over 20 years since the world entered the 21st century, and the economy and science and technology of various countries have developed rapidly. However, the threat and losses of natural disasters to humanity have never diminished. Global disaster risk is increasing year by year in new forms and scales. Sichuan Province, as one of the provinces with the highest number of geological disasters in China, and its rural settlements, as a high-risk area for geological disasters, are the key and difficult points for disaster prevention in China. This article selects rural settlements in Yibin City, Sichuan Province as the research area. By analyzing historical disasters and local geographical characteristics, statistical methods are used to screen and construct a geological hazard risk assessment model for rural settlements in Yibin City. A quantitative and qualitative analysis was conducted on the key indicators of geological hazard risk in rural settlements in Yibin City. Multiple models were used to comprehensively evaluate the risk of geological hazards in rural settlements in Yibin City and a zoning study was conducted. Then, the accuracy of different models was compared, and the random forest model was selected as the computation model for disaster avoidance site selection in rural settlements in Yibin City. Thus obtaining the zoning map of geological hazard risk level in Yibin City. Finally, we conducted a site selection analysis for Rural settlements with different levels of geological hazard risk. And propose strategies for optimizing site selection.

Keywords

• Geological hazards
• Machine learning
• Risk assessment
• Random forests
• Rural settlements
• Site selection