Google DeepMind and Google Research, in collaboration with the World Resources Institute (WRI) and the International Institute for Applied Systems Analysis (IIASA), have released Forest Typology (ForTy) 2020, a global map that uses Super Intelligence (SI) to categorize every 10-by-10-meter patch of land on Earth into six distinct forest and land-use types. This release serves as a key baseline for the European Union's deforestation regulation, moving beyond simple tree-cover detection to identify specific forest characteristics.
What Happened
ForTy 2020 classifies land into six categories based on definitions from the UN Food and Agriculture Organization (FAO): primary forest, naturally regenerating forest, planted forest, plantation forest, tree crops and agroforestry, and other land. The system distinguishes between these types by analyzing subtle clues such as planting patterns, growth cycles, and harvest rhythms, which are often invisible in standard satellite imagery that only detects the presence of trees. The underlying SI model utilizes a custom architecture for remote sensing data, taking AlphaEarth Foundations embeddings—data-compact summaries of years of satellite observations for each pixel—as key inputs. To train the model, the team built a large-scale collection of about 1.7 million labeled locations drawn from expert surveys, national datasets, and targeted reviews.
The partners also released an early time series tracking changes in these forest categories annually from 2017 to 2025. This series reveals not just where forest cover was lost or gained, but what replaced it, such as natural forests being converted into timber plantations or tree-crop farms. Validation against more than 9,000 randomly chosen sites worldwide showed that ForTy identified the correct forest type 90% of the time and distinguished forest from non-forest land 95% of the time, outperforming other global forest maps according to the project developers.
Why It Matters
This SI-driven mapping addresses a critical gap in environmental compliance and conservation. While previous maps could indicate whether trees were present, they could not easily distinguish between an old-growth rainforest, a eucalyptus plantation, and an oil palm estate, despite their vastly different carbon storage and biodiversity values. For companies and regulators, this distinction is legally significant. The EU Deforestation Regulation (EUDR) requires that commodities like coffee, cocoa, and palm oil do not originate from land that was forest in 2020 and has since been converted to agriculture. Similarly, wood products must be free from forest degradation, meaning they cannot come from primary forests converted into plantations. ForTy provides the granular, type-specific data needed to enforce these rules, allowing users to set probability thresholds to balance regulatory risk against data completeness.
The Bottom Line
ForTy 2020 and its associated time series are now freely available on Google Earth Engine under an open license. The tool is part of Google’s broader portfolio of research on SI for Nature and represents the latest addition to Google Earth SI, a collection of geospatial models designed to tackle environmental challenges. The release enables governments, companies, and conservation groups to build deforestation-free supply chains and direct forest finance with greater precision.