Permafrost Distribution and Probability in Northeast China during 2000–2020

Data Documentation for Permafrost Distribution and Probability (Northeast China, 2000–2020) 1. Overview This dataset presents the spatial distribution and occurrence probability of permafrost in Northeast China from 2000 to 2020, generated using machine learning models driven by multi-source environmental and meteorological variables. The dataset includes two GeoTIFF files at 1 km spatial resolution:- Permafrost_map.tif: Binary permafrost distribution (0 = absence, 1 = presence)- RF_Prob.tif: Permafrost probability values (range 0–1) 2. Data Description 2.1 Variables - Permafrost_map: Binary raster representing presence/absence of permafrost- RF_Prob: Continuous raster representing permafrost occurrence probability 2.2 Spatial Coverage Geographic extent: approx. 115.5°E–135°E, 38.5°N–53.9°NResolution: 1 km × 1 kmCoordinate System: WGS 1984, Albers Equal Area projection 2.3 Format Format: GeoTIFF (.tif)Bit depth: Permafrost_map = integer (byte); RF_Prob = float32 2.4 Data Access Data can be read using common GIS software (e.g., ArcGIS, QGIS, ENVI, or Python libraries such as rasterio and GDAL).- Permafrost_map: 0 = permafrost-free; 1 = permafrost-present- RF_Prob: 0–1 continuous values; higher values indicate higher probability 3. Methodology The model was driven by variables such as land surface temperature, snow cover duration, NDVI, precipitation, soil organic matter, elevation, slope, aspect, and land use. A Random Forest model was used. Validation was based on ground temperature records from meteorological stations, field surveys, and borehole data. Model performance: Area Under the Curve (AUC) = 0.88; Accuracy = 0.81. 4. Applications The dataset supports ecological monitoring, infrastructure planning, geohazard risk management, permafrost carbon estimation, and modeling of future climate scenarios. It is suitable for regional-scale and long-term studies. 5. Citation Please cite this dataset as:Huang, S., Jin, H., Wang, Y., et al. (2025). Machine learning model mapped permafrost distribution in Northeast China during 2000-2020. IEEE Transactions on Geoscience and Remote Sensing. vol. 63, pp. 1-18. https://doi.org/10.1109/TGRS.2025.3569727 6. Contact Contact person: Shuai HuangInstitution: School of Ecology, Northeast Forestry UniversityEmail: [e-mail]