The warming ocean is now being asked to keep its own heat ledger
the warming ocean is now being asked to keep its own heat ledger: a concept bottleneck model predicts mixed layer heat content through named physical concepts instead of black box numbers, a transformer based ocean model makes seasonal to decadal forecasts from cmip6 simulations, and a global emulator reconstructs temperature salinity and currents down to 643 meters at 1/12 of a degree.
the sea now files its own heat content report.
Context
OceanCBM routes mixed-layer heat-content predictions through four prescribed concepts: heat-flux entrainment, vertical shear, buoyancy frequency and mixed-layer-depth tendency. A free concept carries residual structure. The May 2026 preprint tests whether mixed supervision makes these internal representations consistent across model initializations.
ORCA-DL uses a transformer architecture and CMIP6 simulations for seasonal-to-decadal ocean forecasts. WenHai instead forecasts daily sea-surface height, temperature, salinity and horizontal currents in the upper 643 metres on a 1/12-degree grid, trained on GLORYS reanalysis. These are separate models and forecast scales.
A named intermediate concept is an interpretability constraint, not proof of a complete causal heat budget. OceanCBM retains a free residual channel. ORCA-DL uses a 1-degree grid and develops a cold bias at longer leads in its decadal test; some dynamical models retain better correlation. WenHai is a short-term forecast system, not a decadal reconstruction. The note brings together explainability, long-range prediction and eddy-resolving prediction rather than one ocean-wide ledger.
Related work
- Forecasting the eddying ocean with a deep neural network ↗WenHai supplies the exact 643-metre and 1/12-degree match, but its output is a forecast rather than a historical reconstruction.
Watch next
- Independent observations for concept fidelity, skill through marine heatwaves and errors in residual heat transport. For long-range models, compare drift, conservation and skill at the same lead time rather than comparing grid resolution alone.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 04:17 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 →