River ice monitoring is of great significance to the study of river runoff process, regional water cycle, water resources management, and ice flood disaster prevention and control. Remote sensing technology has become an important means of obtaining river ice information in high and cold mountain regions where the climate is very cold and field observation is very difficult. At present, the published data on the distribution extent of river ice in the high and cold mountain regions are very rare, and no similar data can be compared. This data set is a long time series of 30 m river ice distribution extent data in the Three Rivers Source Region. The median image was synthesized year by year from images with less than 20% cloud cover during the river ice development period from January to April in the study area based on the long time-series Landsat series surface reflectivity dataset provided by the Google Earth Engine (GEE) platform. The river ice extent and snow extent were extracted using Relative Difference River Ice (RDRI) and Normalized Difference Snow Index (NDSI) algorithms, respectively, and the snow extent extracted by NDSI algorithms was used to supplement when the river ice was missed due to thicker snow cover. Combined with satellite images, DEM, river, lake extent, and glacier extent data, the obtained river ice extent data were manually inspected, misjudged glaciers, lake ice, and other ice bodies were deleted, and areas with obvious omissions of river ice were manually patched, and finally, a total of 34 periods of river ice extent data of the Three Rivers Source Region were obtained for the period of 1990-2023. The dataset was stored in shp file format, which was divided into three areas of Yellow River Source Region, Yangzi River Source Region, and Lancang River Source Region for convenience of use, and the naming rule was: "area+riverice+sensor+year+Revised.shp", for example: yellow_basin_riverice_ETM_2012_Revised. This dataset can provide direct data for the study of the spatial and temporal distribution characteristics of river ice in the Three Rivers Source Region, and can also be used as a reference for the monitoring and mapping of river ice in other high and cold mountain regions.
The "LANDSAT/LT05/C01/T1_SR", "LANDSAT/LE07/C01/T1_SR", "LANDSAT/LC08/C01/T1_SR", and "LANDSAT/LC08/C02/T1_L2" surface reflectivity datasets were used to extract the river ice extent on Google Earth Engine (GEE) platform, and a total of 5,850 images with less than 20% cloud cover were selected. The DEM data of SRTM downloaded from the Geospatial Data Cloud Platform can be used as a reference to correct the river ice extent data and to analyze the parameters such as elevation, slope, and aspect of the study area.
The images with less than 20% cloud cover during the river ice development period from January to April were synthesized into yearly median images, and then the relative difference river ice (RDRI) and normalized difference snow index (NDSI) algorithms were used to extract the river ice extent and snow extent, respectively. When the river ice is covered by thicker snow, it is necessary to use the snow extent extracted by the NDSI algorithm to supplement. Combined with satellite images, DEM, river, lake extent, and glacier extent data, the river ice extent data were manually checked, misjudged glaciers, lake ice, and other ice bodies were deleted, and areas with obvious river ice omissions were manually patched, and finally the river ice extent data of the Three Rivers Source Region for a total of 34 periods from 1990-2023 were obtained and stored in the shp file format.
The overall accuracy of the dataset in the entire Three Rivers Source Region was 97.96%, with a kappa coefficient of 0.8628. Sub-regional evaluations showed that the overall accuracy of the dataset on the distribution extent of river ice in the Yellow River Source Region, the Yangzi River Source Region, and the Lancang River Source Region was 98.28%, 98.21%, and 97.34%, respectively, and the average kappa coefficients were 0.860, 0.903, and 0.824, respectively.
| # | number | name | type |
| 1 | 2021kf10 | The Open Foundation from National Cryosphere Desert Data Center (Grant no. 2021kf10) | other |
This work is licensed under a
Creative
Commons Attribution 4.0 International License.
CSTR:11738.11.NCDC.RIVER_ICE.DB6532.2024
10.12072/ncdc.river_ice.db6532.2024