Global 30-m spatial distribution of forest cover in 2020(GFC30_2020)

“Forest” refers to forested land with an area greater than 0.5 hectares, a height exceeding 5 meters, and a canopy cover greater than 10%, or land that meets these criteria, excluding primarily agricultural and urban land (FAO, 2020). This data product utilizes imagery from the 2020 global forest vegetation growing season, including satellite imagery from the Landsat series, China’s GF-1, GF-6, and other satellites. When data quality fails to meet requirements due to cloud cover or other factors, data from a similar timeframe is selected whenever possible. Using satellite imagery data, combined with global ecogeographic zonation and crowdsourced sample data, a stratified random sampling method was employed to obtain samples from crowdsourced data sources such as global flux stations and the Global Forest Monitoring Network’s global forest plot data, resulting in a total of 61,653 sample points. Machine learning algorithms were used to extract global forest cover at a 30-meter resolution, which was then validated using high-resolution satellite imagery from GF-1, GF-2, GF-6, and QuickBird.The data product reveals that the spatial distribution of global forests is uneven, with significant differences in forest distribution across different climate zones. Furthermore, there are marked differences in forest cover across the six continents (excluding Antarctica). Among them, Asia has the largest land area and the largest forest area, but ranks fourth in forest cover among the six continents; South America has the second-largest forest area among the six continents but the highest forest cover.The data products utilize the WGS84 coordinate system and a latitude-longitude projection (EPSG:4326). They are output in 10°×10° latitude-longitude grids with a resolution of approximately 30 m. The output format is GeoTIFF, with each grid cell measuring approximately 40,000 pixels × 40,000 pixels, totaling 504 grid cells. Each image tile is coded based on the coordinates in the upper-left corner, with latitude first and longitude second. Latitude consists of a 2-digit number preceded by a South/North (S/N) identifier, while longitude consists of a 3-digit number preceded by an East/ W) identifier prefix, where N is used for 0° latitude and E is used for 0° longitude. Each tile file contains a single layer, where a value of 1 represents forest and a value of 0 represents non-forest.