Pan-Arctic Visualization of Landscape Change (2003-2022), Arctic PASSION Permafrost Service

This raster dataset, in Cloud Optimized GeoTIFF format (COG), provides information on land surface changes at the pan-arctic scale. Multispectral Landsat-5 TM, Landsat-7 ETM+, and Landsat-8 OLI imagery (cloud-cover less than 80%, months July and August) was used for detecting disturbance trends (associated with abrupt permafrost degradation) between 2003 and 2022. For each satellite image we calculated the Tasseled Cap multi-spectral index to translate the spectral reflectance signal to the semantic information Brightness, Greenness, and Wetness. In order to characterize change information, we calculated the linear trend of the Brightness, Greenness and Wetness over two decades on the individual pixel level. The final map product therefore contains information on the direction and magnitude of change for all three Tasseled Cap parameters in 30m spatial resolution across the pan-arctic permafrost domain. Features detected include coastal erosion, lake drainage, infrastructure expansion, and fires. The general processing methodology was developed by Fraser et al. 2014 and adapted and expanded by Nitze et al. 2016 and Nitze et al. 2018. Here we upscaled the processing to the circum-arctic permafrost region and the recent 20-year period from 2003 through 2022. The service covers the permafrost region up to 81° North: Alaska (USA), Canada, Greenland, Iceland, Norway, Sweden, Finland, Russia, Mongolia, and China. For Russia and China, regions not containing permafrost were excluded. The data has been processed in Google EarthEngine within the research projects ERC PETA-CARB, ESA CCI+ Permafrost, NSF Permafrost Discovery Gateway, and EU Arctic PASSION. The dataset is a contribution to the 'Panarctic requirements-driven Permafrost Service' of the Arctic PASSION project (see references). Changes in the Tasseled Cap indices Brightness, Greenness, and Wetness are displayed in the image bands red, green, and blue, respectively. Here, coastal erosion (a trend of a land surface transitioning to a water surface) is depicted in dark blue colors, while coastal accretion (a trend of a water surface transitioning to a land surface) is depicted in bright orange colors. Drained lakes appear in bright yellow or orange colors, depending on the soil conditions and vegetation regrowth. Fire scars are a further common feature, which can appear in different colors, depending on the time of the fire and pre-fire land cover. The data can be explored via the Arctic Landscape EXplorer (ALEX, see references) and is available as a public web map service (WMS, see references), both hosted by Alfred Wegener Institute Helmholtz Centre for Polar and Marine Research.本栅格数据集采用云优化GeoTIFF（Cloud Optimized GeoTIFF，COG）格式，提供泛北极尺度下的地表变化相关信息。研究选取云量覆盖率低于80%的7月、8月多光谱Landsat-5 TM、Landsat-7 ETM+及Landsat-8 OLI影像，用于探测2003年至2022年间与突发多年冻土退化相关的扰动趋势。针对每幅卫星影像，我们计算了缨帽多光谱指数（Tasseled Cap），将光谱反射信号转换为亮度（Brightness）、绿度（Greenness）与湿度（Wetness）三类语义信息。为表征变化信息，我们在单个像素级别上计算了20年间亮度、绿度与湿度的线性变化趋势。因此，最终制图产品包含泛北极多年冻土范围内，30米空间分辨率下三类缨帽指数参数的变化方向与变化幅度信息。可探测的地表变化特征包括海岸侵蚀、湖泊疏干、基础设施扩张与火灾。本研究的通用处理方法由Fraser等人于2014年提出，后经Nitze等人于2016年及2018年优化拓展。本研究将该处理流程升级至环北极多年冻土区域，并将研究时段更新为2003年至2022年的近20年周期。本数据集覆盖北纬81°以北的多年冻土区域，包括美国阿拉斯加州、加拿大、格陵兰、冰岛、挪威、瑞典、芬兰、俄罗斯、蒙古国及中国；其中俄罗斯与中国境内不含多年冻土的区域已被排除。本数据基于谷歌地球引擎（Google EarthEngine），在ERC PETA-CARB、ESA CCI+多年冻土、NSF多年冻土发现网关（Permafrost Discovery Gateway）及欧盟北极PASSION等科研项目中完成处理。本数据集是北极PASSION项目“泛北极需求驱动型多年冻土服务”的成果之一（详见参考文献）。缨帽指数的亮度、绿度与湿度变化分别对应影像的红、绿、蓝波段。其中，海岸侵蚀（即地表向水体转变的变化趋势）以深蓝色表示，而海岸淤积（即水体向地表转变的变化趋势）以亮橙色表示。疏干湖泊根据土壤条件与植被恢复情况，以亮黄色或橙色表示。火烧迹地是另一类常见特征，其颜色会根据火灾发生时间与火灾前的土地覆被类型呈现出不同色彩。用户可通过北极景观探索器（Arctic Landscape EXplorer，ALEX，详见参考文献）浏览本数据，同时其以公开网络地图服务（Web Map Service，WMS，详见参考文献）的形式提供，二者均由亥姆霍兹极地与海洋研究中心阿尔弗雷德·魏格纳研究所（Alfred Wegener Institute Helmholtz Centre for Polar and Marine Research）运维。