LANDSLIDE RISK ASSESSMENT OF GONGLIU COUNTY IN XINJIANG BASED ON MULTIPLE COMBINATION MODELS

Freeze-thaw rainfall landslide geological disasters are widely developed in Gongliu County, Xinjiang. The existing research in Gongliu County mostly considers the influence of precipitation but lacks on the influence of temperature. Therefore, the temperature factor was added to evaluate the risk of landslide disaster in Gongliu County in this study. Based on the 682 landslide disaster points that occurred in Gongliu County, nine evaluation factors are selected. They are slope angle, undulation, slope direction, curvature, temperature, distance from fault, distance from river, distance from road and engineering geological rock group. The information model(I), the deterministic coefficient model(CF), the information model+logistic regression model(I+LR) and the deterministic coefficient model+logistic regression model(CF+LR)are used to evaluate the landslide risk in Gongliu County. The risk is divided into four risk levels: extremely high, high, medium and low. The accuracy test and field verification are carried out. The results show that: (1)Temperature has a great triggering effect on landslide. (2)The extremely high and high risk zoning area of the coupled model is significantly lower than that of the single model. The extremely high and high risk zoning area of the CF+LR model is the smallest, and the low risk zoning area is the largest. (3)The AUC values of ROC accuracy test of the four models are 0.889, 0.893, 0.895 and 0.9, respectively, which can objectively evaluate the landslide risk of Gongliu County. The results also show that the CF+LR model has the highest accuracy, and the evaluation results of the CF+LR model are most consistent with the actual situation after field inspection in some areas. The research results have certain reference significance for the prevention and control of landslide geological disasters in Gongliu County, Xinjiang.