global-soil-background-albedo

Global QRF Albedo Q50 Dataset Overview This package contains global Quantile Random Forest (QRF) median (q50) albedo predictions for the visible (VIS) and near-infrared (NIR) spectral domains. The q50 product is the 50th-percentile (median) prediction, not an uncertainty layer. File naming convention Files follow the pattern GQRF_Albedo_Q50_<BAND>_WGS84_v1.0.tif. GQRF: Global Quantile Random Forest product identifier. Albedo: predicted variable. Q50: 50th-percentile (median) prediction. <BAND>: VIS for visible albedo or NIR for near-infrared albedo. WGS84: geographic WGS 84 coordinate reference system. v1.0: dataset version. The release contains: GQRF_Albedo_Q50_VIS_WGS84_v1.0.tif GQRF_Albedo_Q50_NIR_WGS84_v1.0.tif GQRF_Albedo_Q50_WGS84_v1.0_metadata.csv checksums_md5.txt Data format and spatial reference Each prediction file is a single-band GeoTIFF with Float32 values. Coordinate reference system: WGS 84 geographic coordinate system (EPSG:4326). Grid size: 20626 columns x 7948 rows. Spatial extent: -179.978352466000 to 180.013258891000 degrees longitude; -55.990162396000 to 82.728606615000 degrees latitude. Pixel size: 0.0174532925122176 degrees longitude x 0.0174532925278057 degrees latitude. Approximate pixel spacing at the Equator: 1.94 km. The east-west ground distance decreases with latitude because the grid is geographic. NoData value: -9999. Reading the data The GeoTIFF files can be opened directly in GIS software such as QGIS, ArcGIS Pro, and GRASS GIS. The CRS is embedded as WGS 84. If a GIS application asks for a CRS, define the raster as WGS 84 / EPSG:4326; do not reproject it merely to define the CRS. R (terra) library(terra) vis_q50 <- rast('GQRF_Albedo_Q50_VIS_WGS84_v1.0.tif') nir_q50 <- rast('GQRF_Albedo_Q50_NIR_WGS84_v1.0.tif') plot(vis_q50) Python (rasterio) import rasterio with rasterio.open('GQRF_Albedo_Q50_VIS_WGS84_v1.0.tif') as src: vis_q50 = src.read(1, masked=True) print(src.crs, src.res, src.bounds) Notes The publication files are metadata-complete copies of the original q50 products. The operation assigns EPSG:4326 and standardizes NoData to -9999; it does not resample, reproject, or alter valid prediction values. The accompanying CSV records the exact grid metadata for this release.