泛北极工程活动范围灾害易发性分布（2015-2020）

根据泛北极潜在热融灾害（主要为热融滑坡）诱发因素，包括：气温（冻融环境）、降雨、积雪、土壤类型、地形地貌及地下含冰量等，基于地球大数据资源库提供的基础数据，采用机器学习方法（逻辑回归、随机森林、人工神经网络、支持向量机等），以目前已有解译北半球热融滑坡为训练样本，最终获得了泛北极的热融灾害易发性（发生概率）区划图。根据驱动因素敏感性发现气候因素（气温与降雨）对热融灾害的发生于分布贡献度最大，坡度因素贡献度次之，含冰量与辐射也具有较高的贡献。

Based on the triggering factors of potential Pan-Arctic thermokarst hazards (mainly thermokarst landslides), including air temperature (freeze-thaw environment), rainfall, snow cover, soil type, topography and subsurface ice content, this study adopted basic data from the Earth Big Data Repository, used machine learning methods such as logistic regression, random forest, artificial neural network and support vector machine, took the currently interpreted Northern Hemisphere thermokarst landslides as training samples, and finally obtained the susceptibility (occurrence probability) zoning map of thermokarst hazards across the Pan-Arctic region. Through sensitivity analysis of driving factors, it was found that climatic factors (air temperature and rainfall) contribute the most to the occurrence and distribution of thermokarst hazards, followed by slope factors, while subsurface ice content and radiation also have relatively high contributions.