The meadow is now being asked to count its own hectares
the meadow is now being asked to count its own hectares: ai produced the first high resolution map of global seagrass and found 70 percent of it off just five countries, nearly 80 percent of the loss happened outside marine protected areas, and a u net trained on drone imagery maps gulf of mannar beds at 93 percent accuracy.
the meadow now files its own carbon budget.
Context
The Nature paper presents what it calls, to the authors' knowledge, the first global 10-meter maps and change analysis of seagrass extent in clear, shallow coastal waters, derived from 4.75 million Sentinel-2 images for 2019 to 2020 and 2023 to 2024 with a deep learning classifier. It identifies 148,506 square kilometers of seagrass, says 69 percent of the extent is in The Bahamas, Cuba, the USA, Australia and Indonesia, and says only 21 percent of seagrass areas are inside marine protected areas.
The paper reports that over the four years 5,969 square kilometers (4 percent) of seagrass was lost and a further 6,221 square kilometers (4.2 percent) degraded from dense to sparse cover in tropical regions. A PMC study maps a seagrass bed at species level with an unmanned aerial vehicle and deep learning.
The note says 70 percent in five countries, and the paper says 69 percent in five named countries, so the note rounds it. The paper's counts are of extent in clear, shallow waters, which the paper states as a scope limit.
'Nearly 80 percent of the loss happened outside marine protected areas' was not found in the pages read. The paper gives a different figure, 21 percent of seagrass areas inside protected areas, and an ASU quote that more than half of the beds are unprotected, so the loss-outside-protection share is unsupported here, not refuted.
The Gulf of Mannar drone study with 93 percent accuracy was not found in the searches run, so that claim and its U-Net are unsupported here, not refuted. The PMC study read is a different drone and deep learning seagrass study and is listed as related work only.
'The meadow now files its own carbon budget' is the author's line. The Nature paper maps extent and change, and does not report a carbon budget in the abstract read.
Related work
- Global high-resolution mapping of seagrass to support conservation (Nature, June 24, 2026) ↗Source for the 148,506 square kilometers, the 69 percent and 21 percent figures, and the 4 percent loss.
- First complete map of world's seagrass offers warnings and hope (Arizona State University News, June 24, 2026) ↗University press coverage with the 'more than half unprotected' remark.
- Species level mapping of a seagrass bed using an unmanned aerial vehicle and deep learning (PMC) ↗Related drone and deep learning seagrass mapping study.
Watch next
- Find the Gulf of Mannar drone study behind the 93 percent figure. Check whether the Nature paper breaks loss down by protected status.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 3 October 2026 at 18:17 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →