略阳县地处秦巴山区，区内生态环境脆弱，坡陡沟深，构造复杂，加之人为不合理的滥砍乱伐，加剧了地质环境的恶化，造成该区滑坡等地质灾害十分严重，威胁到人类生命财产安全，阻碍了社会经济的发展。因此，开展滑坡灾害风险评价工作对于减轻滑坡灾害造成的损失具有重要意义。本文以略阳县为研究区，开展了栅格单元、地形单元、斜坡单元下基于集成学习模型的滑坡危险性评价；将模糊综合评价法与熵权法结合对略阳县滑坡易损性进行评价，并结合危险性评价结果开展略阳县滑坡风险评价。主要研究成果如下：

（1）在分析略阳县滑坡发育特征、分布规律及形成条件的基础上，选取2处典型滑坡，重点分析滑坡体的基本特征及影响因素。结果表明：区内滑坡以小型浅层堆积层滑坡为主；多沿道路和水系发育；地形地貌及地质环境是滑坡发育的主控因素，人类活动和降雨是滑坡发育的主要诱发因素。

（2）基于区内滑坡发育特征与地质环境，初步选取14个危险性评价因子，分别采用主成分分析法、皮尔森相关系数确定坡度、年均降雨量、地层岩性、距水系距离等11个评价因子构建危险性评价指标体系，采用频率比法对评价因子离散化。从CF模型危险性分区中的低、极低危险区选取与滑坡点等量的非滑坡点构建样本数据集，引入麻雀搜索算法（Sparrow Search Algorithm, SSA）优化随机森林模型（Random Forest, RF）和自适应提升（Adaptive Boosting，AdaBoost）模型参数，构建CF-SSA-RF模型、CF-SSA-AdaBoost模型，与随机抽样选取非滑坡点构建的SSA-RF模型、SSA-AdaBoost模型对比，分析非滑坡点选取方法的不同对滑坡预测结果的影响，绘制了三种评价单元下4种评价模型的滑坡危险性分区图，采用ROC曲线和Kappa系数对评价模型的精度进行检验。结果显示：基于栅格单元的CF-SSA-RF模型预测效果最好，分区效果最显著，预测精度和一致性均达到较高水平（AUC=0.967，k=0.803），表明基于CF模型选取非滑坡点的方式优于随机抽样，且CF-SSA-RF模型具有较好的泛化能力。

（3）根据研究区承灾体特征，选取人口密度、道路密度、建筑物密度、耕地密度作为易损性评价因子，采用熵权法计算评价因子客观权重，作为模糊综合评价法的权重集，避免了人为主观因素的影响，开展了基于模糊综合评价法-熵权法的滑坡易损性评价工作。

（4）基于危险性评价和易损性评价结果，采用风险评价模型对略阳县滑坡风险性进行评价，并对风险评价结果进行分析与总结。结果显示：极高、高风险区主要分布在兴洲街道办等经济较发达、人类工程活动频繁的乡镇，占总面积的22.71%。同时选取3处历史滑坡对评价结果进行验证，结果显示分区结果与滑坡分布相吻合。并在风险评价结果基础上，分别针对各风险区提出相应管控措施和防治建议。

Lueyang County is located in the Qinba Mountain area, with a fragile ecological environment, steep slopes and deep valleys, and complex geological structures. Unreasonable deforestation by humans has exacerbated the deterioration of the geological environment, resulting in severe geological disasters such as landslides in the area, which threaten human life and property safety and hinder the development of the social economy. Therefore, conducting landslide risk assessment is of great significance for mitigating the losses caused by landslides. This study selected Lueyang County as the research area to carry out landslide hazard assessment based on integrated learning models at the grid cell, geological unit, and slope unit levels. The fuzzy comprehensive evaluation method and entropy weight method were combined to evaluate the susceptibility of landslides in Lueyang County, and the landslide risk assessment was conducted based on the hazard assessment results. The main research results are as follows:

(1) Based on the analysis of the developmental characteristics, distribution patterns, and formation conditions of landslides in Lueyang County, two typical landslides were selected, and the basic characteristics and influencing factors of the landslides were analyzed. The results showed that small-scale shallow accumulation layer landslides were the main type of landslides in the area, and they were mostly developed along roads and water systems. The topography, landform, and geological environment were the main controlling factors for the development of landslides, human activities and rainfall were the main inducing factors for landslides.

(2) Based on the developmental characteristics and geological environment of landslides in the area, 14 hazard assessment factors were initially selected. The hazard assessment index system was constructed using principal component analysis, Pearson correlation coefficient, and frequency ratio method to determine 11 assessment factors such as slope, annual rainfall, lithology, and distance to the water system. The evaluation factors were discretized, and a sample data set of non-landslide points equal to the number of landslide points was constructed from the low and extremely low hazard areas in the CF model hazard zone. The sparrow search algorithm (SSA) was introduced to optimize the parameters of the random forest model (RF) and the adaptive boosting model (AdaBoost), and the CF-SSA-RF model and CF-SSA-AdaBoost model were constructed. The results were compared with the SSA-RF model and SSA-AdaBoost model constructed using randomly sampled non-landslide points to analyze the impact of different non-landslide point selection methods on landslide prediction results. The landslide hazard zoning maps of four evaluation models under three evaluation units were drawn, and the accuracy of the evaluation models was tested using ROC curves and Kappa coefficients. The results showed that the CF-SSA-RF model based on the grid cell had the best prediction effect, the most significant zoning effect, and the highest prediction accuracy and consistency (AUC = 0.967, k = 0.803), indicating that the selection of non-landslide points based on the CF model was better than random sampling, and the CF-SSA-RF model had good generalization ability.

(3) According to the characteristics of the vulnerable bodies in the research area, population density, road density, building density, and cultivated land density were selected as susceptibility assessment factors. The entropy weight method was used to calculate the objective weight of the assessment factors, which was used as the weight set of the fuzzy comprehensive evaluation method to avoid the influence of subjective factors. The landslide susceptibility evaluation work based on the fuzzy comprehensive evaluation method-entropy weight method was conducted.

(4) Based on the results of danger assessment and vulnerability assessment, a risk assessment model was used to evaluate the risk of landslides in Lueyang County, and the results of the risk assessment were analyzed and summarized. The results showed that the extremely high and high-risk areas were mainly distributed in economically developed townships with frequent human engineering activities, such as Xingzhou Street, accounting for 22.71% of the total area. At the same time, three historical landslides were selected to verify the evaluation results, and the results showed that the zoning results were consistent with the distribution of landslides. Based on the results of the risk assessment, corresponding control measures and prevention suggestions were proposed for each risk zon.

[1] 韩用顺, 孙湘艳, 刘通, 等. 基于证据权-投影寻踪模型的藏东南地质灾害易发性评价[J]. 山地学报, 2021, 39(05): 672-686.

[2] 苏晓军, 张毅, 贾俊, 等. 基于InSAR技术的秦岭南部略阳县潜在滑坡灾害识别研究[J]. 山地学报, 2021, 39(01): 59-70.

[3] 朱吉龙. 溪洛渡库区滑坡地质灾害风险评价研究[D]. 成都: 西南石油大学, 2019.

[4] Miller R D. Geology of Franklin, Webster, and Nuckolls Counties, Nebraska[M]. US Government Printing Office, 1964.

[5] Brabb E E, Pampeyan E H, Bonilla M G. Landslide susceptibility in San Mateo County, California[R]. US Geological Survey, 1972.

[6] Cotecchia V, Melidoro G. Some principal geological aspects of the landslides of southern Italy[J]. Bulletin of Engineering Geology and the Environment, 1974, 9(1): 23-32.

[7] Stevenson P C. An empirical method for the evaluation of relative landslip risk[J]. Bulletin of the International Association of Engineering Geology-Bulletin de l'Association Internationale de Géologie de l'Ingénieur, 1977, 16(1): 69-72.

[8] Carrara A. Multivariate models for landslide hazard evaluation[J]. Journal of the International Association for Mathematical Geology, 1983, 15(3): 403-426.

[9] Atkinson P M, Massari R. Generalised linear modelling of susceptibility to landsliding in the central Apennines, Italy[J]. Computers & Geosciences, 1998, 24(4): 373-385.

[10] Temesgen B, Mohammed M U, Korme T. Natural hazard assessment using GIS and remote sensing methods, with particular reference to the landslides in the Wondogenet area, Ethiopia[J]. Physics and Chemistry of the Earth, Part C: Solar, Terrestrial & Planetary Science, 2001, 26(9): 665-675.

[11] Promper C, Glade T. Multilayer-exposure maps as a basis for a regional vulnerability assessment for landslides: applied in Waidhofen/Ybbs, Austria[J]. Natural Hazards, 2016, 82(1): 111-127.

[12] Shahri A A, Spross J, Johansson F, et al. Landslide susceptibility hazard map in southwest Sweden using artificial neural network[J]. Catena, 2019, 183: 104225.

[13] Gerzsenyi D, Albert G. Geological hazards of the Gerecse Hills (Hungary)[J]. Journal of Maps, 2021, 17(2): 730-740.

[14] Sujatha E R, Sridhar V. Landslide susceptibility analysis: a logistic regression model case study in Coonoor, India[J]. Hydrology, 2021, 8(1): 41-58.

[15] 张业成, 胡景江, 张春山. 中国地质灾害危险性分析与灾变区划[J]. 地质灾害与环境保护, 1995(03): 1-13.

[16] 殷坤龙, 朱良峰. 滑坡灾害空间区划及GIS应用研究[J]. 地学前缘, 2001(02): 279-284.

[17] 石菊松, 张永双, 董诚, 等. 基于GIS技术的巴东新城区滑坡灾害危险性区划[J]. 地球学报, 2005(03): 275-282.

[18] 文海家, 胡东萍, 王桂林. 汶川县地震滑坡易发性LR与NN评价比较研究[J]. 土木工程学报, 2014, 47(S1): 17-23.

[19] 黄发明, 殷坤龙, 蒋水华, 等. 基于聚类分析和支持向量机的滑坡易发性评价[J]. 岩石力学与工程学报, 2018, 37(01): 156-167.

[20] 张曦, 陈丽霞, 徐勇, 等. 两种斜坡单元划分方法对滑坡灾害易发性评价的对比研究[J]. 安全与环境工程, 2018, 25(01): 12-17.

[21] 朱崇浩. 基于BP神经网络及其优化算法的四川省区域滑坡危险性评价方法研究[D]. 成都: 西南交通大学, 2020.

[22] 郑迎凯, 陈建国, 王成彬, 等. 确定性系数与随机森林模型在云南芒市滑坡易发性评价中的应用[J]. 地质科技通报, 2020, 39(06): 131-144.

[23] 赵晓燕, 谈树成, 李永平. 基于斜坡单元与组合赋权法的东川区地质灾害危险性评价[J]. 云南大学学报(自然科学版), 2021, 43(02): 299-305.

[24] 崔成涛, 李丽敏, 王莲霞, 等. 基于组合赋权与物元分析法的滑坡危险性评价[J]. 国外电子测量技术, 2023, 42(02): 23-29.

[25] 蔡向阳, 铁永波. 我国山区城镇地质灾害易损性评价研究现状与趋势[J]. 灾害学, 2016, 31(04): 200-204.

[26] Timmerman P. Vulnerability, Resilience and the Collapse of Society: A Review of Models and Possible Climatic Applications [M]. Toronto, Canada: Institute for Environmental Studies, University of Toronto, 1981.

[27] Hewitt K . Interpretations of Calamity from the viewpoint of human ecology[M]. 1984.

[28] Leone F, Asté J P, Leroi E. Vulnerability assessment of elements exposed to mass-movement: working toward a better risk perception[J]. Landslides-Glissements de Terrain, Balkema, Rotterdam, 1996: 263-270.

[29] O'Hare G, Rivas S. The landslide hazard and human vulnerability in La Paz City, Bolivia[J]. Geographical Journal, 2005, 171(3): 239-258.

[30] Galli M, Guzzetti F. Landslide vulnerability criteria: a case study from Umbria, Central Italy[J]. Environmental management, 2007, 40(4): 649-665.

[31] Zêzere J L, Oliveira S C, Garcia R A C, et al. Landslide risk analysis in the area North of Lisbon (Portugal): evaluation of direct and indirect costs resulting from a motorway disruption by slope movements[J]. Landslides, 2007, 4: 123-136.

[32] Das I, Stein A, Kerle N, et al. Probabilistic landslide hazard assessment using homogeneous susceptible units (HSU) along a national highway corridor in the northern Himalayas, India[J]. 2011,8(3): 293-308.

[33] Perera E N C, Jayawardana D T. Jayasinghe P, et al. Landslide vulnerability assessment based on entropy method: A case study from Kegalle district, Sri Lanka[J]. Modeling Earth Systems and Environment, 2019, 5: 1635-1649.

[34] 史培军. 区域自然灾害系统研究[C]. 北京: 北京师范大学, 2003.

[35] 张梁. 地质灾害风险评价理论与方法[J]. 中国地质矿产经济, 1996, 9(4): 40-45.

[36] 汪洋, 郭跃, 赵纯勇, 等. 基于3S技术的地质灾害易损性面评价研究[J]. 灾害学, 2003(04): 18-24.

[37] 梅海, 张纪勋. 兰州市地质灾害易损性评价[J]. 南水北调与水利科技, 2010, 8(02): 103-106.

[38] 陈新宇, 卢新中, 王琛. 昆明市地质灾害易损性模糊综合评价[J]. 安全与环境工程, 2012, 19(02): 54-57.

[39] 丁朋朋, 杨宗佶, 游勇, 等. 川藏铁路沿线滑坡灾害易损性评价[J]. 铁道建筑, 2017, 57(08): 133-138.

[40] 张蒙. 顾及灾害等级的多分类滑坡风险评价方法[D]. 阜新: 辽宁工程技术大学, 2021.

[41] 侯明华. 基于SSA优化SVM的地质灾害风险评价应用研究[D]. 石家庄: 河北地质大学, 2022.

[42] 吴越, 向灵均, 吴同情, 等. 基于受灾体空间概率的滑坡灾害财产风险定量评估[J]. 岩石力学与工程学报, 2020, 39(S2): 3464-3474.

[43] Maskrey A. Disaster mitigation: a community based approach[J]. Oxford England Oxfam, 1989.

[44] UNDHA. Internationally agreed glossary of basic terms related to disaster management[J]. UN DHA (United Nations Department of Humanitarian Affairs), Geneva, 1992.

[45] Smith K. Environmental hazards: assessing risk and reducing disaster[J]. 1996.

[46] Ragozin A L. Landslides in research, theory and practice[J]. vulnerability and risk assessment. Thomas Telford, London, 2000:1257-1262.

[47] Arnous M O, Green D R. GIS and remote sensing as tools for conducting geo-hazards risk assessment along Gulf of Aqaba coastal zone, Egypt[J]. Journal of Coastal Conservation, 2011, 15(4): 457-475.

[48] Talaei R. Landslide risk assessment using a multi-method approach in Hashtchin region (NW of Iran)[J]. Acta Geodaetica et Geophysica, 2014, 49(3): 381-401.

[49] Psomiadis E, Charizopoulos N, Efthimiou N, et al. Earth observation and GIS-based analysis for landslide susceptibility and risk assessment[J]. ISPRS International Journal of Geo-Information, 2020, 9(9): 552-578.

[50] Biçer Ç T, Ercanoglu M. A semi-quantitative landslide risk assessment of central Kahramanmaraş City in the Eastern Mediterranean region of Turkey[J]. Arabian Journal of Geosciences, 2020, 13(15): 1-26.

[51] Novellino A, Cesarano M, Cappelletti P, et al. Slow-moving landslide risk assessment combining Machine Learning and InSAR techniques[J]. Catena, 2021, 203(5): 105317.

[52] Caleca F, Tofani V, Segoni S, et al. A methodological approach of QRA for slow-moving landslides at a regional scale[J]. Landslides, 2022: 1-23.

[53] 张业成, 张春山, 张梁. 中国地质灾害系统层次分析与综合灾度计算[J]. 中国地质科学院院报, 1993(Z1): 139-154.

[54] 朱良峰, 殷坤龙, 张梁, 等. GIS支持下的地质灾害风险分析[J]. 长江科学院院报, 2002(05): 42-45.

[55] 张春山, 张业成, 马寅生, 等. 区域地质灾害风险评价要素权值计算方法及应用——以黄河上游地区地质灾害风险评价为例[J]. 水文地质工程地质, 2006(06): 84-88.

[56] 吴树仁, 石菊松, 张春山, 等. 地质灾害风险评估技术指南初论[J]. 地质通报, 2009, 28(08): 995-1005.

[57] 薛强, 张茂省, 孙萍萍, 等. 基于多期DEM和滑坡强度的滑坡风险评估[J]. 地质通报, 2013, 32(06): 925-934.

[58] 孙强, 张泰丽, 伍剑波, 等. 基于GIS与层次分析法的龙溪流域滑坡风险评价[J]. 华东地质, 2018, 39(03): 227-233.

[59] 何静. 基于机器学习的滑坡灾害空间预测及风险评估[D]. 成都: 电子科技大学, 2019.

[60] 范亦嵩. 宜君县地质灾害特征及风险评价[D]. 西安: 西北大学, 2020.

[61] 闫成林, 郑德凤, 年廷凯, 等. 基于ANP的模糊贝叶斯网络模型及其在海岸带地质灾害风险评价中的应用[J]. 工程地质学报, 2021, 29(06): 1862-1868.

[62] 陆新卓. 不同尺度的地质灾害定量风险评价与区划[D]. 北京: 中国地质大学(北京), 2021.

[63] 高鹏程. 甘肃省滑坡灾害风险评价模型研究[D]. 兰州: 兰州交通大学, 2022.

[64] 崔阳阳. 基于不同评价单元的滑坡易发性评价方法研究[D]. 西安: 西安科技大学, 2021.

[65] 汤国安, 刘学军, 闾国年. 数字高程模型及地学分析的原理与方法[M]. 科学出版社, 2005.

[66] Li L, Lan H. Integration of spatial probability and size in slope-unit-based landslide susceptibility assessment: A case study[J]. International Journal of Environmental Research and Public Health, 2020, 17(21): 8055.

[67] 田述军, 张珊珊, 唐青松, 等. 基于不同评价单元的滑坡易发性评价对比研究[J]. 自然灾害学报, 2019, 28(06): 137-145.

[68] 吴先谭, 邓辉, 张文江, 等. 基于斜坡单元自动划分的滑坡易发性评价[J]. 山地学报, 2022, 40(04): 542-556.

[69] 王凯, 张少杰, 韦方强. 斜坡单元提取方法研究进展和展望[J]. 长江科学院院报, 2020, 37(06): 85-93.

[70] 郑玲静, 李秀珍, 徐瑞池. 基于斜坡单元的区域滑坡敏感性评价——以云南省小江流域为例[J]. 科学技术与工程, 2021, 21(28): 12322-12329.

[71] 马啸. 府谷县崩-滑地质灾害发育特征及其易发性评价[D]. 西安: 西安科技大学, 2021.

[72] 程斌, 杨根兰, 覃乙根, 等. 基于CF-LR模型的典型红层地区地灾易发性评价——以赤水市为例[J/OL]. 河南理工大学学报(自然科学版): 1-14[2023-21].

[73] 李怡静, 胡奇超, 刘华赞, 等. 耦合信息量和Logistic回归模型的滑坡易发性评价[J]. 人民长江, 2021, 52(06): 95-102.

[74] Chen W, Chen Y, Tsangaratos P, et al. Combining evolutionary algorithms and machine learning models in landslide susceptibility assessments[J]. Remote Sensing, 2020, 12(23): 3854.

[75] 刘福臻, 戴天宇, 王军朝, 等. 耦合Random forest与信息量模型的地质灾害易发性评价——以西藏自治区工布江达县为例[J/OL]. 安全与环境学报: 1-12[2023-04-21].

[76] 陈芯宇, 师芸, 温永啸, 等. 基于确定性系数与支持向量机的滑坡易发性评价[J]. 科学技术与工程, 2023, 23(02): 518-527.

[77] 王念秦, 郭有金, 刘铁铭, 等. 基于SVM-LR模型的滑坡易发性评价——以临潼区为例[J]. 科学技术与工程, 2019, 19(30): 62-69.

[78] 周萍, 邓辉, 张文江, 等. 基于信息量模型和机器学习方法的滑坡易发性评价研究——以四川理县为例[J]. 地理科学, 2022, 42(09): 1665-1675.

[79] 田春山, 刘希林, 汪佳. 基于CF和Logistic回归模型的广东省地质灾害易发性评价[J]. 水文地质工程地质, 2016, 43(06): 154-161.

[80] 刘璐瑶, 高惠瑛. 基于证据权与Logistic回归模型耦合的滑坡易发性评价[J]. 工程地质学报, 2023, 31(01): 165-175.

[81] 罗路广, 裴向军, 黄润秋, 等. GIS支持下CF与Logistic回归模型耦合的九寨沟景区滑坡易发性评价[J]. 工程地质学报, 2021, 29(02): 526-535.

[82] 封建民, 杨波, 谢瑞莲, 等. 陕西省略阳县地质灾害危险性区划评估[J]. 地理空间信息, 2021, 19(02): 78-82.

[83] Shortliffe E H, Davis R, Axline S G, et al. Computer-based consultations in clinical therapeutics: explanation and rule acquisition capabilities of the MYCIN system[J]. Computers and Biomedical Research, 1975, 8(4): 303-320.

[84] Heckerman D. Probabilistic interpretations for MYCIN's certainty factors[M]//Machine intelligence and pattern recognition. North-Holland, 1986, 4: 167-196.

[85] Xue J, Shen B. A novel swarm intelligence optimization approach: sparrow search algorithm[J]. Systems science & control engineering, 2020, 8(1): 22-34.

[86] 周萍, 邓辉, 张文江, 等. 基于信息量模型和机器学习方法的滑坡易发性评价研究——以四川理县为例[J]. 地理科学, 2022, 42(09): 1665-1675.

[87] 周超, 殷坤龙, 曹颖, 等. 基于集成学习与径向基神经网络耦合模型的三峡库区滑坡易发性评价[J]. 地球科学, 2020, 45(06): 1865-1876.

[88] Chen W, Pourghasemi H R, Kornejady A, et al. Landslide spatial modeling: Introducing new ensembles of ANN, MaxEnt, and SVM machine learning techniques[J]. Geoderma, 2017, 305: 314-327.

[89] 张晓东. 基于遥感和GIS的宁夏盐池县地质灾害风险评价研究[D]. 北京: 中国地质大学(北京), 2018.

[90] 李云杰, 屈飞行, 杨梓俪. 层次分析法在线路工程地质灾害危险性评价中的应用[J]. 甘肃水利水电技术, 2020, 56(4): 33-36.

[91] 祁于娜, 王磊. 层次分析-熵值定权法应用于山区城镇地质灾害易发性评价[J]. 测绘通报, 2021(06): 112-116.

[92] 张周. 基于AHP-模糊综合评价法的山体生态修复评价研究[D]. 重庆: 重庆大学, 2020.

[93] 程皓阳. 基于GIS的地质灾害风险评价[D]. 成都: 成都理工大学, 2021.

[94] 章瑞. 基于GIS的元谋县地质灾害风险性评价[D]. 昆明: 昆明理工大学, 2021.