Abstract:
Shidian County, Yunnan Province has complex terrains and landforms, strong tectonic activities, and fragile geological environment conditions. In addition, unreasonable human activities have led to frequent geological disasters such as landslides, mudslides, and collapses. RS and GIS technologies are used as support in order to conduct a comprehensive assessment of geological disaster risks in Shidian County. Seven disaster-prone stability factors, nine disaster-causing risk factors, and five vulnerability factors are selected to firstly study the stability of the disaster environment, the risk of disaster-causing factors, and the vulnerability of the disaster bearing body. Then the stability of the disaster-prone environment in Shidian County is evaluated using a combination of information quantity model plus logistic regression model (
I+LR), and deterministic coefficient model plus logistic regression model (CF+LR). The risk of disaster-causing factors in Shidian County is evaluated using Analytic Hierarchy Process (AHP) and Ordinary Least Squares (OLS) models. The vulnerability of disaster-prone areas in Shidian County is evaluated using Analytic Hierarchy Process (AHP). And finally here is conducted a comprehensive assessment of the geological disaster risk in Shidian County. The results show that the areas with high comprehensive risk in Shidian County are mostly concentrated in the central region of the county, with Dianyang Town, Renhe Town, and Youwang Town being the most obvious. About 10% of the areas in Shidian County are in an unstable disaster-prone environment; about 5% face extremely high risk factors for disasters; and about 12%, high-risk environments for disasters. Meanwhile, a comparative analysis is conducted on the evaluation accuracy and applicability of various comprehensive models, and it is found out that the difference between the evaluation results of the information quantity model and the certainty coefficient model is small, both of which are suitable for evaluating the stability of the disaster-prone environment. The AUC value of the ROC curve of the least squares method evaluation model is 0.756, which is slightly higher than that of the Analytic Hierarchy Process.