Granular Forest Classification
On October 22, 2026, Google DeepMind announced Forest Typology (ForTy), a global mapping framework capable of classifying every 10x10-meter land parcel worldwide into distinct forest types and tree covers. The tool was engineered to clearly distinguish intact natural ecosystems from commercial tree plantations and timber crop lands.
Differentiating Natural Ecosystems from Timber Plantations
According to Google DeepMind, undisturbed natural forests sequester substantially more carbon and support higher levels of biodiversity than monoculture timber plantations that undergo commercial harvesting cycles every 10 to 20 years. Treating all canopy cover uniformly in environmental reports has historically caused significant discrepancies when evaluating forest quality and net carbon storage.
The ForTy framework categorizes global vegetation data into three primary classes: - Natural forests: Undisturbed, high-biodiversity ecosystems. - Planted forests: Managed forestry and silviculture projects. - Farmed forests and tree-crop types: Agricultural plantations and industrial timber land.
Processed consistently across Earth's surface using a 10-meter grid, this dataset enables granular monitoring of shifts across vegetated landscapes.
Enhancing Deforestation Monitoring
Google DeepMind stated that the primary objective of the ForTy project is to help conservation organizations pinpoint deforestation more accurately and direct resources toward preserving high-value ecological zones. Technical details regarding model architecture, satellite training data, and access mechanisms for the broader research community have not yet been disclosed in the initial release.