The lake is now being asked to keep its own level ledger
the lake is now being asked to keep its own level ledger: an explainable network reads altimetry and climate drivers to forecast tibetan plateau lakes decades ahead, recurrent models with augmented records track lake titicaca’s daily water level, and a deep framework stitches legacy altimetry with swot to give 9,184 river reaches continuous elevation back to 2008.
the basin now files its own storage report.
Context
The Tibetan Plateau study (Water, 2025) pairs Sentinel-3A radar altimetry with bias-corrected CMIP6 projections under SSP2-4.5 and SSP5-8.5 to forecast levels of eight lakes. A feed-forward network tuned with Optuna reaches a mean validation R squared of 0.861, and SHAP analysis attributes drivers. It reports a north to south split: glacier-fed northern lakes such as Migriggyangzham rise by 13.18 plus or minus 0.56 m under SSP5-8.5.
The Lake Titicaca paper (Frontiers in Water, 2026) compares LSTM, BiLSTM, GRU and BiGRU models with data augmentation. GRU did best, with an RMSE of 0.0182 m and an R squared of 0.9986.
The river dataset (EarthArXiv, April 2026) uses a deep learning framework to combine legacy altimetry and SWOT into daily water surface elevation from January 2008 to May 2025, covering 9,184 river reaches, 5,926 rivers and 1,342 basins. Median correlation is 0.46 against Hydroweb and 0.62 against independent SWOT observations.
The note's 'decades ahead' fits the Tibetan study's climate scenarios, but those are scenario projections under two emissions pathways, not forecasts, and they cover eight lakes with small samples. Titicaca's R squared of 0.9986 is for a forecast of a slowly varying level, so it is easy to score high, and the note's 'daily' fits the paper only loosely. The river dataset's median correlation of 0.46 against calibration data is modest next to the note's wording of 'continuous elevation back to 2008': the paper describes continuous daily estimates with median bias near zero, but the correlations are the skill statement. It is a preprint, not peer reviewed in the page retrieved.
Related work
- Forecasting Tibetan Plateau Lake Level Responses to Climate Change: An Explainable Deep Learning Approach Using Altimetry and Climate Models (Water) ↗Source for the eight lake projections, SHAP drivers and R squared of 0.861.
- Lake Titicaca water level forecasting using data augmentation and recurrent neural networks (Frontiers in Water) ↗Source for the GRU, BiGRU, LSTM comparison and RMSE of 0.0182 m.
- Continuous Water Surface Elevation Estimates Using Deep Learning with Legacy Altimetry and Surface Water and Ocean Topography (EarthArXiv) ↗Source for the 9,184 reaches, the 2008 start and the correlation numbers.
Watch next
- Check how the Tibetan lake projections compare with observed level trends over the altimetry record. Check river reaches where the correlation is lowest and why.
Sources
- Forecasting Tibetan Plateau Lake Level Responses to Climate Change: An Explainable Deep Learning Approach Using Altimetry and Climate Models (Water)mdpi.com
- Lake Titicaca water level forecasting using data augmentation and recurrent neural networks (Frontiers in Water)frontiersin.org
- Continuous Water Surface Elevation Estimates Using Deep Learning with Legacy Altimetry and Surface Water and Ocean Topography (EarthArXiv)eartharxiv.org
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 01:17 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 →