geospatial
Natural Forests of the World 2020 model (MTSViT)
Model behind the 10 m Natural Forests of the World 2020 map (92.2% accuracy vs an independent global validation set) for EUDR deforestation baselines; 2026 maps with a Tree Crops class planned.
Levelmodel
Classgeospatial
Parametersnot in the notes
Contextnot in the notes
Released13 Nov 2025
Opennessunknown
Runs onnot in the notes
Statusnot in the notes
Industry Use Cases
R&D Topics
Identity
- Developer
- Google DeepMind, Google, World Resources Institute (WRI), IIASA
- Family
- not in the notes
- Released
- 13 Nov 2025 sourcesannouncement page
- Weights published
- –
- Record checked
- 10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.
Architecture and Size
- Architecture
- dense transformer Multi-modal temporal-spatial vision transformer on Sentinel-2 seasonal imagery plus elevation, slope and coordinates
- Total parameters
- – sources
- Active parameters
- –
- Checkpoint parameters
- –
- Context length
- not in the notes
- Input
- image
- Output
- not in the notes
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).
Openness and Access
- Openness
- unknown
- Licence
- not in the notes sources
- Access routes
- cloud marketplace
Where this comes from
The licence file governs over registry tags and aggregator labels.
Deployment
- Runs on
- not in the notes
- Footprint
- Map data on Figshare and Earth Engine
- Runtimes
- not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.
Indonesia
- Indonesian support
- not stated
- Regional languages
- not in the notes
- Jakarta hosting
- unknown
- Languages
- not in the notes
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.
Benchmark Results
The notes record no benchmark result for this model.
Lineage
The notes record no lineage for this model.
Links
DevelopersGoogle DeepMindfrontier lab, UK / USA
Googlebig tech, USA
World Resources Institute (WRI)USA
IIASAAustria
Google DeepMindfrontier lab, UK / USAGooglebig tech, USAWorld Resources Institute (WRI)USAIIASAAustriaModelNatural Forests of the World 2020 model (MTSViT)Google DeepMind, geospatial
Natural Forests of the World 2020 model (MTSViT)Google DeepMind, geospatialUse CasesDeforestation Baselines and Supply-Chain Monitoring (EUDR)AgricultureSatellite Monitoring (Land Cover, Floods, Forests)Public Sector
Who Runs Itnone recordedDeployments of these use cases; none names this model.
Use cases 2
Organisations 4
Sources
- research.google/blog/separating-natural-forests-from-other-tr... https://research.google/blog/separating-natural-forests-from-other-tree-cover-with-ai-for-deforestation-free-supply-chains/
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Natural Forests of the World 2020 Model (MTSViT)", https://aurion.semai.id/models/google-deepmind-natural-forests-mtsvit/, records of October 2026.