2022 ECMWF-ESA workshop report: current status, progress and opportunities in machine learning for Earth System observation and prediction

Bonavita, Massimo; Schneider, Rochelle; Arcucci, Rossella; Chantry, Matthew; Chrust, Marcin; Geer, Alan; Saux, Bertrand Le; Vitolo, Claudia

doi:10.1038/s41612-023-00387-2

Cited by 1 publication

(1 citation statement)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As in other areas of applied science and engineering, Machine Learning (ML) methods are having a profound impact in NWP and Climate monitoring and prediction (e.g., Bonavita et al., 2023; Schneider et al., 2022, for recent overviews). Researchers have sought to deploy ML algorithms in specific parts of the NWP and Climate prediction workflow (e.g., Krasnopolsky, 2023, for a recent review), aiming to take advantage of the extremely low computational cost of the trained ML models and the fact that the ML algorithms can be effective at learning complex, nonlinear mappings if large and accurate training data sets are available.…”

Section: Introductionmentioning

confidence: 99%

On Some Limitations of Current Machine Learning Weather Prediction Models

Bonavita

2024

Geophysical Research Letters

Self Cite

View full text Add to dashboard Cite

Machine Learning (ML) is having a profound impact in the domain of Weather and Climate Prediction. A recent development in this area has been the emergence of fully data‐driven ML prediction models which routinely claim superior performance to that of traditional physics‐based models. We examine some aspects of the forecasts produced by three of the leading current ML models, Pangu‐Weather, FourCastNet and GraphCast, with a focus on their fidelity and physical consistency. The main conclusion is that these ML models are not able to properly reproduce sub‐synoptic and mesoscale weather phenomena and lack the fidelity and physical consistency of physics‐based models and this has impacts on the interpretation of their forecasts and their perceived skill. Balancing forecast skill and physical realism will be an important consideration for future ML models.

show abstract

Section: Introductionmentioning

confidence: 99%