TPRoGI: a comprehensive rock glacier inventory for the Tibetan Plateau using deep learning

An inventory of 44,273 rock glaciers, covering approximately 6,000 km², across the Tibetan Plateau, was compiled using remote sensing and DeepLabv3+, a kind of deep learning model. The file in the format of a shapefile contains the boundary of each rock glacier as vectors in the Coordinate Reference System of EPSG:4326-WGS 84. The associated attribute table includes rock glacier ID, source data, date of mapping, mapper’s name, reviewer’s name, additional information, latitude (units: degrees), longitude (units: degrees), subregion of rock glacier, rock glacier area (units: m²), mean elevation of rock glacier (units: m), median elevation of rock glacier (units: m), minimum elevation of rock glacier (units: m), maximum elevation of rock glacier (units: m), mean slope of rock glacier (units: degrees), median slope of rock glacier (units: degrees), minimum slope of rock glacier (units: degrees), maximum slope of rock glacier (units: degrees), aspect of rock glacier (units: degrees), mean annual air temperature (units: °C), mean annual ground temperature (units: °C), annual precipitation (units:  mm), annual potential incoming solar radiation (units:  kWh/m²). The corresponding names for the table fields are ‘ID’, ‘SOUR_DATA’, ‘MAP_DATE’, ‘MAPPER’, ‘REVIEWER’, ‘ADDI_INF’, ‘LAT’, ‘LON’, ‘SUBREGION’, ‘AREA’, ‘ELE_MEAN’, ‘ELE_MEDIAN’, ‘ELE_MIN’, ‘ELE_MAX’, ‘SLO_MEAN’, ‘SLO_MEDIAN’, ‘SLO_MIN’, ‘SLO_MAX’, ‘ASPECT’, ‘MAAT’, ‘MAGT’, ‘AP’, ‘PISR’.