https://watch.eeg.cl.cam.ac.uk/videos/watch/84549d13-4bcb-4c64-a6d8-6d528e5ca448 Abstract: Knowing which ecosystems are where underpins conservation planning, restoration and reporting on biodiversity targets. Yet existing ecosystem maps are coarse, inconsistent across countries, and rarely say how much to trust them. We ask whether Tessera encodes ecosystem identity. We test this against roughly 100,000 vegetation plots harmonised across Latin America and labelled with the IUCN Global Ecosystem Typology. We then train a lightweight classifier on the embeddings and paint a continuous 30 m map of more than 70 ecosystem types across the continent. Bio: James Ball is a postdoc in the Forest Ecology and Conservation Group at the University of Cambridge. His work combines remote sensing, machine learning and ecology to map forests and ecosystems at scale Tue, 27 Oct 2026 09:44:58 GMT https://validator.w3.org/feed/docs/rss2.html PeerTube - https://watch.eeg.cl.cam.ac.uk https://watch.eeg.cl.cam.ac.uk/client/assets/images/icons/icon-1500x1500.png https://watch.eeg.cl.cam.ac.uk/videos/watch/84549d13-4bcb-4c64-a6d8-6d528e5ca448 All rights reserved, unless otherwise specified in the terms specified at https://watch.eeg.cl.cam.ac.uk/about and potential licenses granted by each content's rightholder.