热融湖塘是受自然或人为因素影响地下冰或多年冻土发生局部融化，引起地面沉陷并在凹陷处积水的现象，是多年冻土退化形成的热喀斯特地貌类型之一。近年来，受气候变暖影响，多年冻土稳定性降低，热融湖塘的时空动态变化显著增快，不仅引发区域生态环境变化，更重要的是对周边居民点、交通和管道等基础设施产生极为不利的影响。然而，目前对热融湖塘的认识不足，制约了较大区域热融湖塘分布及变化规律的研究，不利于探明热融湖塘的变化趋势，难以进行合理风险防范。因此，本研究以阿拉斯加北极地区为研究区，依托谷歌地球引擎、ArcGIS平台进行热融湖塘的识别，探究其分布现状及分布规律，分析热融湖塘的时空变化及影响因素。研究得出的主要结论如下：

（1）构建了热融湖塘的识别方法，建立了研究区2000、2010、2020年的热融湖塘数据集。利用Landsat遥感影像，基于地物波谱特性，通过修正的归一化差异水体指数、归一化植被指数、增强型植被指数之间的指数关系，并结合现有热融湖塘的调查研究，依据海拔、冻土、冰川、河流等数据辅助以人工目视解译识别研究区内热融湖塘，并检验了热融湖塘的识别效果。2000、2010、2020年热融湖塘的总数量分别为66,011个、61,735个、62,744个，总面积分别为9,610.52 km²、9,509.22 km²、9,580.21 km²。

（2）发现了热融湖塘的总体分布特征。热融湖塘的数量在不同纬度间差异较大，不同经度间差异较小。热融湖塘的总数量主要由面积较小的热融湖塘贡献，总面积主要由面积较大的热融湖塘贡献。

（3）研究了热融湖塘的个体形态特征。研究区内分布的面积大于等于14,400 m²的36,913个热融湖塘的形态分析表明，热融湖塘的面积与周长存在显著正相关关系，主轴方向以NNW、EEN方向为主，圆度指数分布于0.02~0.79之间，伸长指数分布在1.00~15.84之间。

（4）分析了热融湖塘的空间分布特征及影响因素。热融湖塘的分布受地形、冻土、土壤、地质、气象因素影响。在研究区内，地形湿度指数是影响热融湖塘分布最重要的影响因素，其次是海拔、降水、地层、坡度、土壤温度、地下冰含量等，多年冻土类型对热融湖塘分布的影响程度最小。

（5）总结了热融湖塘的时空变化情况。在2000~2020年，研究区热融湖塘的数量和面积先减少后增加，整体上减少。热融湖塘的时空变化受气象因素影响明显。从空间上看，在2000~2020年热融湖塘的网格空间净数量变化无明显规律，网格空间净面积变化情况与研究区气温、土壤温度的变化趋势相似。研究区内热融湖塘的总数量和总面积变化情况呈现一定的一致性，其在2000~2010年、2010~2020年的变化情况与降水、土壤湿度的变化趋势具有相似性。

Thermokarst lake is a phenomenon caused by the local melting of underground ice or permafrost affected by natural or human factors, which causes ground subsidence and water accumulation in the depression. It is one of the types of thermokarst landforms formed by permafrost degradation. In recent years, affected by climate warming, the stability of permafrost has decreased, and the dynamic changes of thermokarst lakes have increased significantly. It not only leads to a change in the regional ecological environment, but more importantly, it has extremely adverse impacts on surrounding residential areas, transportation, pipelines, and other infrastructure. However, the lack of understanding of thermokarst lakes restricts the research on the distribution and variation of thermokarst lakes in large areas. It is not conducive to ascertaining the changing trend of thermokarst lakes, and it is difficult to prevent reasonable risks. Therefore, in this study, Arctic Alaska is taken as the research area, and Google Earth Engine and ArcGIS platforms are used to identify thermokarst lakes, explore their distribution status and regularities, and analyze the temporal and spatial changes as well as influence factors of thermokarst lakes in different years. The main conclusions are as follows:

(1) The identification method of thermokarst lakes is constructed. The dataset of thermokarst lakes in the study area in 2000, 2010, and 2020 is established. Using Landsat remote sensing images based on the spectral characteristics of ground objects, through modified Normalized Difference Water Index, Normalized Difference Vegetation Index, and Enhanced Vegetation Index. Combined with the investigation and research of existing thermokarst lakes, the artificial visual interpretation is used to identify thermokarst lakes in the study area based on the data of elevation, frozen soil, glaciers, rivers, and other data, and the identification effect of thermokarst lakes are tested. In 2000, 2010, and 2020, the total number of thermokarst lakes is 66011, 61735, and 62744 respectively, and the total area is 9610.52 km², 9509.22 km², and 9580.21 km² respectively.

(2) The general distribution characteristics of thermokarst lakes are found. The number of thermokarst lakes varies greatly in latitudes and little in longitudes. The smaller lakes contribute more to the number of thermokarst lakes, and the larger lakes contribute more to the area of thermokarst lakes.

(3) The individual morphological characteristics of thermokarst lakes are studied. The morphological analysis of 36,913 thermokarst lakes with an area larger or equal to 14,400 m² shows that there is a significant positive correlation between the area perimeter of thermokarst lakes. The main axis directions are mainly NNW and EEN directions, the roundness index ranges from 0.02 to 0.79, and the ratio between short and long axes ranges from 1.00 to 15.84.

(4) The spatial distribution characteristics and influencing factors of thermokarst lakes are analyzed. The distribution of thermokarst lakes is affected by topographic, frozen soil, soil, geology, and meteorological factors. In the study area, topographic wetness index is the most important factor affecting the distribution of thermokarst lakes, followed by elevation, precipitation, strata, slope, soil temperature, ice content of frozen soil, etc. In the study area, the type of permafrost has the least influence on the distribution of thermokarst lakes.

(5) The temporal and spatial changes of thermokarst lakes are summarized. From 2000 to 2020, the number and area of thermokarst lakes in the study area decreases first and then increases, and decrease overall. The temporal and spatial variation of thermokarst lakes is affected by meteorological factors. From the perspective of space, there is no obvious rule in the net number of grid space in thermokarst lakes from 2000 to 2020, and the change in net grid space area is similar to the changes in air temperature and soil temperature in the study area. The changes in the total number and total area of thermokarst lakes in the study area are consistent to a certain extent, and their changes from 2000 to 2010 and from 2010 to 2020 are similar to the trends of precipitation and soil moisture.