The frozen lake is now being asked to keep its own calendar
the frozen lake is now being asked to keep its own calendar: a random forest reads multi sensor satellite data and hits 94 percent accuracy on ice state across 31,000 thermokarst lakes on the alaska north slope, deep learning reconstructs ice on and ice off for minnesota lakes from 1980 to 2018 with lake specific estimates, and a tiny network segments sea ice on board an fpga with low power.
the lake now files its own freeze diary.
Context
The thermokarst lake study is an EGUsphere preprint, Multi-sensor satellite analysis reveals latitudinal and morphometric controls on ice phenology across 31,000 thermokarst lakes on the Alaska North Slope (2026). It uses Sentinel-1 SAR classified by a random forest trained on Sentinel-2 optical imagery and ERA5 temperature for 2019 to 2023, with 94% accuracy for ice state, phenology retrieved for 97% of lakes and a mean ice-free period of 115 days.
The Minnesota work is a USGS data release and paper, Predicting Minnesota lake ice phenology with deep learning, explainable methods, and a physically based benchmark, 1980-2018. The data release says it hindcasts ice phenology from daily weather drivers and lake ice records across 625 Minnesota lakes (4,359 lake-years) and provides LSTM predictions for 881 lakes. The FPGA network is TinyIceNet (arXiv 2603.03075), a compact segmentation network for on-board Stage of Development sea ice mapping from Sentinel-1 SAR under strict power constraints.
The 94% accuracy is for ice state classification and the 97% is the share of lakes with a retrieved phenology, two different figures. It is a preprint. The Minnesota model is trained on 625 lakes with records and predicts for 881, so lake specific estimates for unmonitored lakes come from the model and not from observations. TinyIceNet maps sea ice stage of development, which is close to but not the same as segmenting sea ice. The excerpt read gives no power figure in watts, so none is claimed.
Watch next
- TinyIceNet power and accuracy figures. Minnesota model error on lakes it was not trained on.
Sources
- Multi-sensor satellite analysis reveals latitudinal and morphometric controls on ice phenology across 31,000 thermokarst lakes on the Alaska North Slope (EGUsphere)egusphere.copernicus.org
- Model application: modeling lake ice phenology in Minnesota, 1980-2018 (USGS)usgs.gov
- Predicting Minnesota lake ice phenology with deep learning, explainable methods, and a physically based benchmark (USGS)pubs.usgs.gov
- TinyIceNet: Low-Power SAR Sea Ice Segmentation for On-Board FPGA Inference (arXiv 2603.03075)arxiv.org
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 09:06 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 →