Nearly 25% of all lakes on earth are located at high latitudes. These lakes are formed by a combination of thermokarst, glacial, and geological processes. Evidence suggests that the origin of periglacial lake formation may be an important factor controlling the likelihood of lakes to drain. However, geospatial data regarding the spatial distribution of these dominant Arctic and subarctic lakes are limited or do not exist. Here, we use lake-specific morphological properties using the Arctic Digital Elevation Model (DEM) and Landsat imagery to develop a Thermokarst lake Settlement Index (TSI), which was used in combination with available geospatial datasets of glacier history and yedoma permafrost extent to classify Arctic and subarctic lakes into Thermokarst (non-yedoma), Yedoma, Glacial, and Maar lakes, respectively. This lake origin dataset was used to evaluate the influence of lake origin on drainage between 1985 and 2019 in northern Alaska. The lake origin map and lake drainage datasets were synthesized using five-year seamless Landsat ETM+ and OLI image composites. Nearly 35,000 lakes and their properties were characterized from Landsat mosaics using an object-based image analysis. Results indicate that the pattern of lake drainage varied by lake origin, and the proportion of lakes that completely drained (i.e., >60% area loss) between 1985 and 2019 in Thermokarst (non-yedoma), Yedoma, Glacial, and Maar lakes were 12.1, 9.5, 8.7, and 0.0%, respectively. The lakes most vulnerable to draining were small thermokarst (non-yedoma) lakes (12.7%) and large yedoma lakes (12.5%), while the most resilient were large and medium-sized glacial lakes (4.9 and 4.1%) and Maar lakes (0.0%). This analysis provides a simple remote sensing approach to estimate the spatial distribution of dominant lake origins across variable physiography and surficial geology, useful for discriminating between vulnerable versus resilient Arctic and subarctic lakes that are likely to change in warmer and wetter climates.

中文翻译：

---

沿冰期湖泊起源影响阿拉斯加北部湖泊排水的可能性

地球上所有湖泊中近25％位于高纬度地区。这些湖泊是由热喀斯特，冰川和地质过程共同形成的。有证据表明，冰川湖形成的起源可能是控制湖泊流失可能性的重要因素。但是，有关这些主要北极和亚北极湖泊的空间分布的地理空间数据有限或不存在。在这里，我们利用北极数字高程模型（DEM）和Landsat影像，利用湖泊特有的形态学特征，开发了一个喀斯特喀斯特湖沉降指数（TSI），该指数与可用的冰川历史地理数据集和多年生永冻土的地理空间数据集结合使用来进行分类北极和亚北极的湖泊分别进入Thermokarst（非yedoma），Yedoma，Glaccial和Maar湖。该湖泊起源数据集用于评估1985年至2019年之间阿拉斯加北部湖泊起源对排水的影响。使用五年的无缝Landsat ETM +和OLI图像合成器合成了湖泊原点图和湖泊排水数据集。使用基于对象的图像分析，通过Landsat马赛克对近35,000个湖泊及其性质进行了描述。结果表明，湖泊排水的类型随湖泊的起源而变化，并且在1985年至2019年之间，在喀斯特（非耶德玛），耶德玛，冰川和马阿尔湖中完全沥干（即面积损失> 60％）的湖泊比例是分别为12.1、9.5、8.7和0.0％。最易排水的湖泊是小型热喀斯特（非耶德玛）湖泊（12.7％）和大型耶德玛湖（12.5％），而最具韧性的是大中型冰川湖（4.9和4）。1％）和马阿尔湖（0.0％）。这项分析提供了一种简单的遥感方法，可以估算在变化的地貌和表层地质条件下主要湖泊起源的空间分布，有助于区分在温暖和潮湿的气候下可能会发生变化的脆弱湖泊和弹性北极和亚北极湖泊。