Researchers have released a new global dataset that tracks land use and land cover from 1960 through 2100.  

Existing datasets that combine historic and future land-use data tend to have coarse resolution. They also don't account for realistic spatial patterns when projecting the future. This matters because the arrangement of land, not just the total area of each type, affects things like fire spread, species movement, and water flow. 

The new dataset resolves this by downscaling projections from the global land-use model LandSyMM. The researchers used LandScaleR to convert coarse projections into detailed 1km maps. They calibrated this method against two decades of historic data. This ensured the future maps show realistic patterns. 

The result spans 1960 to 2100 at 1km resolution, with yearly updates. It covers five future scenarios – from strong climate action to а fossil-fuel-driven future. The dataset shows major differences depending on the region. Cropland expands most under scenarios with high climate change and limited cooperation between countries. Tropical forests face the biggest losses in several scenarios.  

The authors argue their dataset fills several gaps at once. It has finer spatial and temporal resolution than earlier products and uses a consistent modelling framework across all scenarios. It also accounts for local variation in how landscapes change, rather than assuming the same patterns everywhere. 

The dataset will be useful for many environmental models – such as tools for predicting fire risk, hydrology, biodiversity loss, and carbon emissions. It could support policymakers designing land-use strategies that are tailored to local conditions rather than applied uniformly across regions. 

Read the full study.