In this dataset, 42 medium and high resolution remote sensing images from 1964 to 2020 were used to complete a comprehensive survey of glaciers and ice lakes in the entire Tama sub basin in the middle of the the Himalayas and collate them into a dataset. The glaciers in 1964, 1980, 1990, 2000, 2010, 2018 and 2020 and the ice lakes in 1964 and 1980 were demarcated manually, and the ice lakes in 1987-2020 were demarcated automatically in the data format of. shp. Among them, maps of 271 glaciers and 196 glacial lakes in the entire basin in 2020 were drawn, covering an area of 329.2 ± 1.9 square kilometers and 14.4 ± 0.3 square kilometers, respectively. According to statistics, in the past 56 years, the total area of glaciers in the basin has decreased by 26.2 ± 3.2 square kilometers (0.13% a-1), with an average thinning of about 20 meters. The average speed has gradually slowed down from 5.3 meters a-1from 1999 to 2003 to 4.0 meters a-1from 2013 to 2015. The total area of glacial lakes has increased by 9.2 ± 0.4 square kilometers (about 180%), with the expansion rate of pre glacial lakes in contact with ice (about 204%) being much higher than other lakes.
The remote sensing images used in this study were obtained from the United States Geological Survey (USGS), https://earthexplorer.usgs.gov/ ）Geospatial data cloud（ http://www.gscloud.cn/ ）And PlanetLabs（ https://www.planet.com/ ）The time span is from 1964 to 2020. A total of 6 KH-4A and KH-9 Keyhole (KH) images (2.7-9 meters), 36 Landsat (TM/ETM+/OLI) surface reflectance images (30 meters), and 4 PlanetScope (PL) orthogonal tile images (3.125 meters) were used. The images are mainly collected after the monsoon and near the end of the melting season (August to December), with less cloud cover and seasonal snow accumulation. The PL product with orthogonal correction is used as the base image for KH image registration.
By using the Python API interactive interface, it is possible to adjust the classification threshold more accurately based on the screen detection results of the original image than the Javascript API interface. All Landsat and PlanetScope images are products that have undergone atmospheric and terrain correction. In addition, an error of ± 0.5 pixels was estimated to calculate the uncertainty of the area (i.e. linear error multiplied by perimeter, for example, the linear error of Landsat image is 15 meters, and the linear error of KH-4A image is about 1.35 meters). For super glacial lakes, only the basal lakes that store water year-round are retained, while the lakes with significant seasonal variations are excluded.
Remote sensing is an important tool for rapidly monitoring glacier dynamics and assessing disaster risks. Although detailed information on the changes in glacial lake area has been obtained in this study, basic data on glacial lakes (such as lake basin topography, dam building material types, internal structure of glacial moraine dams, etc.) are still lacking and require field investigation.
This work is licensed under
CC BY 4.0 (Creative Commons Attribution 4.0 International License).
CSTR:11738.11.NCDC.ZENODO.DB4166.2024
10.5281/zenodo.5496959
