赣南离子型稀土矿区滑坡灾害易发性评价——以定南县为例

Evaluation of Landslide Disaster Susceptibility in Ionic Rare Earth Mining Area in Southern Jiangxi Province——A Case Study of Dingnan County

赣南是我国离子吸附型稀土矿床密集分布区,稀土开采为当地经济社会发展做出了重要贡献,但同时也造成了边坡稳定性减弱等一系列生态环境问题。以离子型稀土高强度开采区江西省定南县为研究区,基于538个历史滑坡灾害点数据,选取高程、坡度、坡向、归一化植被指数、地形湿度指数、距水系距离、距道路距离、地层岩性和土地利用类型共9项影响因子构建滑坡易发性评价指标体系,运用加权信息量模型进行滑坡易发性分区与评价,通过ROC曲线检验和比较模型的预测能力。结果表明:坡度、地层岩性、距道路距离和距水系距离是产生滑坡的主要因素。信息量模型和加权信息量模型均具有较高的预测准确性,极高和高易发区与历史滑坡灾害点的实际分布情况吻合程度较好。信息量模型的极高和高易发区占总面积的33.80%,包含86.24%的滑坡灾害点,灾害点密度为2.55个/km^(2)。加权信息量模型的极高和高易发区占总面积的25.50%,包含81.97%的滑坡灾害点,灾害点密度为3.21个/km^(2)。加权信息量模型的预测精度优于信息量模型,AUC值增加了0.042,更适用于离子型稀土矿区的滑坡易发性评价。

Southern Jiangxi Province is an area of intensive distribution of ionic-adsorption rare earth deposits in China.While rare earth mining has made important contributions to local economic and social development,it has also caused a series of ecological and environmental problems,such as slope stability weakening.Dingnan County of Jiangxi Province,an ionic rare earth high-intensity mining area,was selected as the study area.Based on 538 historical landslide disaster points data,nine impact factors including elevation,slope,aspect,normalized difference vegetation index(NDVI),topographic wetness index(TWI),distance to rivers,distance to roads,stratum lithology and land-use type were selected to construct the landslide susceptibility evaluation index system.The weighted information value model was used to divide and evaluate the landslide susceptibility,and the ROC curve was used to test and compare the prediction ability of the model.The results show that slope,stratum lithology,distance to roads and distance to rivers are the dominant factors for landslides.Both the information value model and the weighted information value model exhibit high prediction accuracy,with the extremely high and high susceptible areas showing good agreement with the actual distribution of historical landslide disaster points.The extremely high and high susceptible areas of the information value model account for 33.80%of the total area,including 86.24%landslide disaster points,with a density of 2.55 disaster points/km^(2).The extremely high and high susceptible areas of the weighted information value model account for 25.50%of the total area,including 81.97%landslide disaster points,with a density of 3.21 disaster points/km^(2).The prediction accuracy of the weighted information value model is better than that of the information value model,and the AUC value is increased by O.042,which is more suitable for the evaluation of landslide susceptibility in ionic rare earth mining areas.