Oriel — AI & technology worth noticing, every day.

Beyond tree cover: mapping the world’s forest types

Web page · https://deepmind.google/blog/beyond-tree-cover-mapping-the-worlds-forest-types

A global map categorizing every 10-meter patch of land on Earth into distinct forest types.

First seen by Oriel on .

AI summary

Google DeepMind, Google Research, WRI, and IIASA released ForTy, a global 10-meter forest map and 2017–2025 time series distinguishing natural forests, plantations, tree crops, and other land.

  • ForTy classifies land into six FAO-based categories: primary forest, naturally regenerating forest, planted forest, plantation forest, tree crops and agroforestry, and other land.
  • The model was trained on about 1.7 million labeled locations and uses AlphaEarth Foundations embeddings summarizing years of satellite observations.
  • Validation across more than 9,000 sites found 90% accuracy for forest type and 95% accuracy in distinguishing forest from non-forest.
  • Class probability scores let users choose thresholds suited to their applications, balancing strict regulatory confidence against broader forest-extent estimates.
  • The freely available, openly licensed map and early-release time series are intended for governments, companies, conservation groups, and scientists; the time series is still being improved.

#remote sensing #forest mapping #geospatial ai #deforestation monitoring

Generated by gpt-5.6-sol. Check the original for details.

Aliases

SourceExternal IDURL
Google DeepMind Bloghttps://deepmind.google/blog/beyond-tree-cover-mapping-the-worlds-forest-types/https://deepmind.google/blog/beyond-tree-cover-mapping-the-worlds-forest-types/ Read discussion