基于IV、CF、LRIV和LRCF模型的海南昌江县地质灾害易发性评价研究  

Susceptibility assessment of geological hazards in Changjiang County,Hainan Island,based on IV,CF,LRIV,and LRCF models

作　　者：李信 薛桂澄 柳长柱[1,2] 马波 杨永鹏[1,2] 杨峰[1,2] 王晓林 李洁玉[3] LI Xin;XUE Guicheng;LIU Changzhu;MA Bo;YANG Yongpeng;YANG Feng;WANG Xiaolin;LI Jieyu(Hainan Key Laboratory of Marine Geological Resources and Environment,Haikou 570206,Hainan,China;Hainan Geological Survey Institute,Haikou 570206,Hainan,China;Hainan Association of Geology and Geography,Haikou 570206,Hainan,China)

机构地区：[1]海南省海洋地质资源与环境重点实验室,海南海口570206 [2]海南省地质调查院,海南海口570206 [3]海南省地质地理学会,海南海口570206

出　　处：《地震工程学报》2025年第2期319-330,341,共13页China Earthquake Engineering Journal

基　　金：海南省自然科学基金(421RC664);海南省海洋地质资源与环境重点实验室重点课题(22-HNHYDZZYHJKF020)。

摘　　要：海南省昌江县地质条件复杂,地质灾害频发,开展地质灾害易发性研究对该县地质灾害防治与监测预警具有重要的现实意义。以昌江县为研究区,选取岩土体、高程变异系数、坡度、断裂、起伏度、河流、道路、降雨作为易发性评价指标,采用信息量(IV)模型、确定性系数(CF)模型、信息量逻辑回归(LRIV)模型及确定性系数逻辑回归(LRCF)模型进行地质灾害易发性评价研究。评价结果显示,昌江县地质灾害高易发区主要分布于山区公路和河流两侧沿线,而极低易发区基本位于西部平原地区;四种模型的AUC值分别为0.755、0.749、0.766及0.764,表明预测精度满足易发性评价要求;LRIV模型与其他三种模型相比具有更高的可靠性和准确性。研究结果可为海南岛西南山区地质灾害防治与监测预警提供科学依据。In Changjiang County,located in the southwest mountainous area of Hainan Province,geological hazards have occurred frequently in recent years owing to the complex geological conditions and frequent engineering activities in the area.Therefore,studying the susceptibility of geological hazards in this region for the prevention,monitoring,and early warning of geological hazards has great practical significance.To evaluate the susceptibility of geological hazards in Changjiang County,the rock-soil mass,elevation variation coefficient,slope,fracture,undulation,river,road,and rainfall were selected as assessment indicators.The evaluation was conducted using four models:information value,certainty factor,logistic regression of information value(LRIV),and logistic regression of certainty factor.The results indicate that the high-susceptibility areas of geological hazards in Changjiang County are mainly distributed along roads and rivers in the mountainous areas,while the extremely low-susceptibility areas are located in the plains.The AUC values of the four models,0.755,0.749,0.766,and 0.764,respectively,indicate that the prediction accuracy meets the requirement of susceptibility assessment and that the LRIV model is more reliable and accurate than the three other models.The findings of this research can provide a scientific basis for the prevention,monitoring,and early warning of geological hazards in the southwest mountainous areas of Hainan Island.

关 键 词：确定性系数 信息量模型 逻辑回归 易发性 地质灾害 

分 类 号：P642.21[天文地球—工程地质学]