摘要
地质灾害易发性评价是防灾减灾工作部署、灾害预警以及风险管控的基础,针对评价因子在量纲、性质等方面的差异以及评价模型的适用性、准确性等问题,以陕西省子长市作为研究区,引入信息量(information value,INF)模型与遗传算法(Genetic Algorithm,GA),提出一种信息量融入GA优化支持向量机(Support Vector Ma-chines,SVM)模型的地质灾害易发性评价方法。首先,从研究区的地形地貌、地质环境、生态环境三个方面选取坡度、坡向、河流距离、土地利用类型等9个评价因子,利用INF模型将各评价因子量纲进行统一,构建样本数据集;然后,利用GA迭代求解SVM关键参数c和g的最优值;最后,将全区点集属性数据代入训练好的模型中求解并输出子长市地质灾害易发性指数值,将该值代入ArcGIS软件得到地质灾害易发性区划图,并采用受试者工作特征(Receiver Operating Characteristic,ROC)曲线对模型预测结果的精度进行检验。结果表明:采用信息量融入GA优化SVM模型得到的子长市地质灾害易发性评价结果的准确率为93%,优于INF模型和INF-SVM模型;研究区内地质灾害极高、高易发区集中分布于研究区中南部、西部,地质灾害中易发区主要沿部分道路及支流呈树枝状散布,地质灾害低、极低易发区在整个研究区内分布最广。该研究结果可为同等地质条件下的地质灾害易发性评价提供一定的参考,同时也可为研究区内地质灾害防治工作提供理论依据。
Geological hazard susceptibility assessment is the basis for disaster prevention and mitigation work deployment,disaster early warning and risk management and control.In view of the differences in di-mensions and properties of evaluation factors,as well as the applicability and accuracy of evaluation mo-dels,taking Zichang City of Shaanxi Province as the study area,by introducing the information value(INF)model and Genetic Algorithm(GA),this paper proposes a geological hazard susceptibility assessment me-thod that incorporates information value into GA optimized SVM model.Firstly,based on the topography,geological environment,and ecological environment of the study area,9evaluation factors are selected in-cluding slope,aspect,river distance,land use type,etc,and the INF model is used to unify the dimensions of each factor layer to construct a sample Data set.Then,GA algorithm is used to iteratively solve the optimal values of key parameters c and gof SVM.Finally,the attribute data of the whole area points set are substi-tuted into the trained model and return the value of geological hazard susceptibility index of Zichang City of Shaanxi Province,which is substituted into ArcGIS to get the susceptibility zoning map,and the accuracy of the results is tested by using Receiver Operating Characteristic(ROC)curve.The result shows that the ac-curacy of the susceptibility evaluation results obtained by incorporating information value into GA opti-mized SVM model is 0.930,which is better than the traditional INF model and INF-SVM model.The zones with extremely high and high susceptibility to geological hazards are mainly distributed in the central south and western parts of the study area.The moderate-susceptibility zones are dendritic along some roads and tributaries.The zones with low and extremely low susceptibility are most widely distributed in the whole study area.The research results can provide some reference for the evaluation of geological hazard suscepti-bility under the same geological conditions,and they also can be used in preventing and controlling geolo-gical disasters in the study area.
作者
杨康
薛喜成
李识博
YANG Kang;XUE Xicheng;LI Shibo(College of Geology and Environment,Xi’an University of Science and Technology,Xi’an710054,China)
出处
《安全与环境工程》
CAS
CSCD
北大核心
2022年第3期109-118,共10页
Safety and Environmental Engineering
基金
国家自然科学基金项目(41907255)。
关键词
地质灾害
易发性评价
支持向量机
遗传算法
ROC曲线
子长市
geological hazard
susceptibility assessment
Support Vector Machine(SVM)
Genetic Algorithm(GA)
ROC curve
Zichang City
