基于GIS的白龙江流域舟曲—武都段的滑坡危险性评价

本研究以围绕着白龙江流域的甘肃省南部的宕昌县、舟曲县和武都区部分地区为研究区，根据全国滑坡编目中得到的272个历史滑坡数据以及选取的高程、坡度、坡向、平面曲率、剖面曲率、归一化植被指数（NDVI）、降雨、岩性、距道路距离和距河流距离10种影响因子，利用三种具有代表性的定量方法：信息量模型、以及基于频率比模型的逻辑回归模型和人工神经网络模型对研究区内滑坡灾害危险性进行评价。三种评价结果均显示研究区内滑坡灾害的极高和高危险区主要沿白龙江河谷地区呈带状分布。从危险性分区图可看出，人工神经网络模型得到的分区图较为合理，既表现出沿河谷地区集中分布的趋势，也呈现出对滑坡历史数据较为独立的特征，这一研究结果与前人研究结果一致。根据受试者工作特征曲线（ROC曲线）对三种模型的精度进行检验，检验得到的AUC值分别为0.818、0.829和0.837，说明三种评价结果均具有较高的可靠性，基于频率比模型的人工神经网络模型相比其他两个模型具有更好的评价精度，能更好地进行滑坡危险性的预测和评价，其中高程、降雨、岩性以及距道路距离对评价结果影响更大，这四种影响因子重要性值占比为52.1%。为该地区的城市扩建与灾害预防预测提供了参考。

Landslide Risk Assessment of the Zhouqu-Wudu Section of Bailong River Basin Based on Geographic Information System

Based on the data of 272 historical landslides from the national landslide catalog and 10 influencing factors, 3 representative quantitative methods, namely, information value model, logistic regression model, and artificial neural network based on frequency ratio model, were used in this paper to evaluate the landslide hazard risk in areas of Tangchang County, Zhouqu County and Wudu District in southern Gansu Province surrounding the Bailong River Basin. The results from these three models showed that the extremely high- and high-risk areas of landslide disaster are mainly distributed along the Bailong River valley. From the hazard zoning map, the result obtained by the artificial neural network model is found to be relatively reasonable, showing not only the trend of centralized distribution along the valley area but also the feature relatively independent to the landslide historical data, which is consistent with previous research results. The accuracy of the three models was tested according to the receiver operating characteristic curve, and the AUC values obtained were 0.818, 0.829, and 0.837, respectively, indicating that all three evaluation results have high reliability. Compared with the other two models, the artificial neural network model based on frequency ratio has better evaluation accuracy and can better predict and evaluate the landslide risk. Elevation, rainfall, lithology, and distance from the road are factors that have greater influence on the evaluation results than other factors, and the importance value of these four influencing factors accounts for 52.1%. The results of this study provide a reference for urban expansion and disaster prevention in the study area.