Pith. sign in

REVIEW 15 cited by

Deep Learning for Day Forecasts from Sparse Observations

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2306.06079 v3 pith:FFGCJOS2 submitted 2023-06-06 physics.ao-ph cs.LG

Marcin Andrychowicz , Lasse Espeholt , Di Li , Samier Merchant , Alexander Merose , Fred Zyda , Shreya Agrawal , Nal Kalchbrenner This is my paper · ORCID
classification physics.ao-phcs.LG
keywords metnet-3modelsneuraldataforecastshoursobservationssparse
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Deep neural networks offer an alternative paradigm for modeling weather conditions. The ability of neural models to make a prediction in less than a second once the data is available and to do so with very high temporal and spatial resolution, and the ability to learn directly from atmospheric observations, are just some of these models' unique advantages. Neural models trained using atmospheric observations, the highest fidelity and lowest latency data, have to date achieved good performance only up to twelve hours of lead time when compared with state-of-the-art probabilistic Numerical Weather Prediction models and only for the sole variable of precipitation. In this paper, we present MetNet-3 that extends significantly both the lead time range and the variables that an observation based neural model can predict well. MetNet-3 learns from both dense and sparse data sensors and makes predictions up to 24 hours ahead for precipitation, wind, temperature and dew point. MetNet-3 introduces a key densification technique that implicitly captures data assimilation and produces spatially dense forecasts in spite of the network training on extremely sparse targets. MetNet-3 has a high temporal and spatial resolution of, respectively, up to 2 minutes and 1 km as well as a low operational latency. We find that MetNet-3 is able to outperform the best single- and multi-member NWPs such as HRRR and ENS over the CONUS region for up to 24 hours ahead setting a new performance milestone for observation based neural models. MetNet-3 is operational and its forecasts are served in Google Search in conjunction with other models.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 15 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 41 citations worldwide. Full citation record

  1. Geospatial Diffusion-based Evolution Synthesis (GeoDES) for Storm-Centered Weather Augmentation

    cs.LG 2026-07 conditional novelty 7.0 of 10

    GeoDES generates realistic synthetic cyclone evolutions via 2D-pretrained, temporally-inflated diffusion with correlated noise, beating weather foundation models on storm-kinetics and energy-spectrum metrics.

  2. Global reanalysis from observations alone with machine learning

    physics.ao-ph 2026-07 conditional novelty 7.0 of 10

    Observation-only machine learning can generate multi-decade global atmospheric reanalyses with large-scale skill near ERA5 and surface errors between ERA-Interim and ERA5, in a single day of compute.

  3. Prithvi-Precip: Integrating Satellite Observations into an Atmospheric AI Foundation Model for Precipitation Forecasting

    physics.ao-ph 2026-08 conditional novelty 6.0 of 10

    Training on IMERG satellite precipitation and directly ingesting satellite observations improves medium-range AI precipitation forecasts, with the largest gains at lead times under 40 hours.

  4. FuXi-Nowcast: Environment-conditioned deep learning for severe convection nowcasting

    physics.ao-ph 2025-12 conditional novelty 6.0 of 10

    An environment-conditioned deep-learning system with a convective-signal enhancement module reports higher CSI than CMA-MESO for reflectivity, rainfall, and wind gusts up to 12 h over East China.

  5. Learning from nature: insights into GraphDOP's representations of the Earth System

    physics.ao-ph 2025-08 conditional novelty 6.0 of 10

    GraphDOP, trained purely on meteorological observations, spontaneously forms unified latent representations of Earth system state, viewing effects, and synoptic dynamics.

  6. Forecast error diagnostics in neural weather models

    physics.ao-ph 2025-06 conditional novelty 6.0 of 10

    Correcting the stratosphere or boundary layer toward reanalysis truth substantially improves mid-latitude medium-range forecasts in a neural weather model; correcting the tropics does not.

  7. PEAR: Equal Area Weather Forecasting on the Sphere

    cs.LG 2025-05 conditional novelty 6.0 of 10

    A transformer weather model operating natively on the equal-area HEALPix grid beats an equiangular-grid counterpart at longer lead times with 2.6x fewer parameters.

  8. OMG-HD: A High-Resolution AI Weather Model for End-to-End Forecasts from Observations

    physics.ao-ph 2024-12 conditional novelty 6.0 of 10

    An observation-only AI model trained on the RTMA analysis reports lower 0-12h surface forecast errors than operational HRRR and IFS-HRES over the contiguous US.

  9. GraphDOP: Towards skilful data-driven medium-range weather forecasts learnt and initialised directly from observations

    physics.ao-ph 2024-12 conditional novelty 6.0 of 10

    A graph-neural-network weather model trained only on raw observations produces skillful global forecasts out to five days, with tropical 2-meter temperature forecasts competitive with the operational IFS.

  10. ArchesWeather & ArchesWeatherGen: a deterministic and generative model for efficient ML weather forecasting

    cs.LG 2024-12 conditional novelty 6.0 of 10

    ArchesWeatherGen, a flow-matching model trained on residuals of a deterministic transformer, generates ensemble forecasts that outperform IFS ENS and NeuralGCM on most WeatherBench headline variables at 1.5 degrees re...

  11. ADAF: An Artificial Intelligence Data Assimilation Framework for Weather Forecasting

    physics.ao-ph 2024-11 conditional novelty 6.0 of 10

    An encoder-decoder network trained on RTMA and station data produces kilometer-scale near-surface analyses that beat the operational HRRRDAS benchmark on held-out station observations.

  12. Observation-driven correction of numerical weather prediction for marine winds

    cs.LG 2025-12 conditional novelty 5.0 of 10

    An observation-conditioned transformer (ORCA) reduces GFS 10-m wind forecast RMSE by 45% at 1 h and 13% at 48 h over the North Atlantic, with the largest gains near coasts and shipping lanes.

  13. Data-driven Precipitation Nowcasting Using Satellite Imagery

    cs.CV 2024-12 conditional novelty 5.0 of 10

    A satellite-only neural network, trained on a new Korean geostationary satellite and radar dataset, predicts hourly precipitation up to six hours ahead with 2 km resolution.

  14. Regional Weather Variable Predictions by Machine Learning with Near-Surface Observational and Atmospheric Numerical Data

    physics.ao-ph 2024-12 conditional novelty 5.0 of 10

    An LSTM encoder-decoder trained on station data plus HRRR output gives low short-horizon forecast errors, and a transfer variant extends it to nearby ungauged sites using elevation and borrowed observations.

  15. Leadsee-Precip: A Deep Learning Diagnostic Model for Precipitation

    physics.ao-ph 2024-11 conditional novelty 5.0 of 10

    Leadsee-Precip is a deep learning diagnostic model that converts circulation fields into 6-hour precipitation with a loss weighted toward rare heavy rain, and reports strong hit scores for heavy rain in China.

Pith tools