Most global maps only show whether trees are present, so an old rainforest and an oil palm estate can look the same from space. The EU’s deforestation regulation needs more than that: it bars products like coffee, cocoa and palm oil from land that was forest in 2020 and was later converted. The model learns from AlphaEarth Foundations satellite summaries to spot planting patterns and harvest cycles. Checked against more than 9,000 random sites reviewed by experts, it also told forest from non-forest land 95% of the time. Instead of a yes-or-no label, it gives a probability for each type, so users can set their own thresholds.
Why it matters
Supply-chain rules need to know what replaced a forest, not just that trees disappeared. A natural forest that is logged and replanted as a plantation looks green again from space, so ordinary tree-cover maps miss the change. Free, open data at this level of detail could make compliance checks cheaper, though the year-by-year series is still an early release and will keep changing.
Original source
Beyond tree cover: mapping the world’s forest types
Google DeepMind
deepmind.google
Summary by Second Order AI. Full reporting is at the original source.