Adataset of glacier inventory on the northern slope of the Tianshan Mountains from 1990 to 2020

Glaciers are sensitive indicators of climate change, and studying their dynamics is of great significance in revealing global climate change. Influenced by global warming, glaciers on the Northern Tianshan Mountains are retreating rapidly. However, existing glacier inventories for the region primarily dates from around 2010, and the timeliness doesn’t meet the current research needs. This dataset is based on remote sensing images from the glacier ablation period of 1990 Landsat TM and 2020 Landsat OLI, Sentinel-2A/2B MSI. The glacier vector boundaries were determined using the band ratio method (Red/SWIR), combined with interactive manual methods and the Otsu algorithm to identify the optimal segmentation threshold. The binary images obtained from threshold segmentation were processed with median filtering and then converted into vector polygons using morphological opening and closing operations. The glacier vector boundaries were manually visually revised and data quality was verified using Google Earth and glacier cataloging information. This dataset includes data from both 2020 and 1990. The results show that the number and area of glaciers were 3,253 and 1,670.55 km2 in 2020, respectively, while the number and area of glaciers were 3,270 and 2,003.19 km2 in 1990. The errors in the glacier areas of the two periods due to the spatial resolution of the remote sensing images were 162.14 km2 (8.09%) for 1990 and 71.55 km2 (4.28%) for 2020. This dataset comprehensively reflects the status and changes of glacier resources the northern slope of the Tianshan Mountains between 1990 and 2020, providing fundamental data for research on water resource carrying capacity and water security in the region.

冰川是气候变化的敏感指示器，研究其动态变化对于揭示全球气候变化规律具有重要意义。受全球变暖影响，北天山山脉的冰川正快速退缩。但该区域现有冰川编目主要基于2010年左右的观测数据，时效性已无法满足当前研究需求。本数据集基于1990年冰川消融期的陆地卫星专题制图仪（Landsat TM）、2020年陆地卫星陆地成像仪（Landsat OLI）以及哨兵2A/2B多光谱仪器（Sentinel-2A/2B MSI）获取的遥感影像构建。冰川矢量边界采用红光/短波红外（Red/SWIR）波段比值法，并结合交互式人工解译与大津算法（Otsu）识别最优分割阈值进行提取。经阈值分割得到的二值图像先通过中值滤波处理，再经形态学开闭运算转换为矢量多边形。随后对冰川矢量边界开展人工目视修正，并通过谷歌地球（Google Earth）及冰川编目资料完成数据质量核查。本数据集涵盖1990年与2020年两个时段的冰川数据。结果显示，2020年该区域冰川数量为3253条，总面积1670.55平方千米；1990年冰川数量为3270条，总面积2003.19平方千米。受遥感影像空间分辨率限制，两个时段的冰川面积提取误差分别为：1990年162.14平方千米（占比8.09%），2020年71.55平方千米（占比4.28%）。本数据集全面反映了1990年至2020年间天山北坡冰川资源的分布现状与变化特征，可为该区域水资源承载力及水安全相关研究提供基础数据支撑。