The kelp is now being asked to keep its own canopy count
the kelp is now being asked to keep its own canopy count: ucla mapped california’s coast with ai and found heatwave damage invisible to older tools, an open source tool segments sentinel 2 images to track kelp from orbit, and the public kelpwatch maps train a u net on planet dove imagery.
the kelp now files its own coverage.
Context
UCLA's March 18, 2026 release says satellite imagery combined with AI has produced new high-resolution statewide maps that detect changes in kelp forests previously invisible to traditional monitoring tools, showing how kelp survived or declined after the 2014 to 2016 marine heatwave. It says applying deep learning to Planet's Dove satellite constellation gave a map 10 times more detailed than previous standard satellite records.
The SKeMa repository describes a Python tool for classifying kelp in Sentinel-2 images using a deep learning segmentation model. The Communications Earth & Environment paper says the team developed a convolutional neural network, a VGG16-U-Net, to map kelp canopy from Planet Dove imagery and created annual 3 m resolution maps for 2017 to 2024.
The three items match separate sources. The post says open source tool, and SKeMa is on GitHub with a Hugging Face copy; it is a Sentinel-2 tool from a different group than the California Planet Dove work. The post says the public kelpwatch maps train a U-Net, and the paper describes a VGG16-U-Net behind high-resolution maps for California, which Kelpwatch.org hosts.
'Invisible to older tools' is UCLA's wording. The 10 times more detail refers to 3 m against 30 m Landsat records. 'The kelp now files its own coverage' is the author's line.
Related work
- How AI deep learning is helping UCLA scientists protect California's coastal ecosystems (UCLA Newsroom, March 18, 2026) ↗Source for the UCLA statement on heatwave damage and the 10 times more detail.
- High-resolution Planet Dove data identify local drivers of kelp canopy persistence (Communications Earth & Environment, January 6, 2026) ↗Source for the VGG16-U-Net, Planet Dove imagery and 3 m annual maps.
- m5ghanba/skema: Satellite-based Kelp Mapping using Semantic Segmentation on Sentinel-2 imagery (GitHub) ↗Source for the open source Sentinel-2 kelp segmentation tool.
Watch next
- Read the paper's accuracy assessment against field data. Check the SKeMa license and training data.
Sources
- How AI deep learning is helping UCLA scientists protect California's coastal ecosystems (UCLA Newsroom, March 18, 2026)newsroom.ucla.edu
- High-resolution Planet Dove data identify local drivers of kelp canopy persistence (Communications Earth & Environment, January 6, 2026)nature.com
- m5ghanba/skema: Satellite-based Kelp Mapping using Semantic Segmentation on Sentinel-2 imagery (GitHub)github.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 3 October 2026 at 19:07 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 →