Headline scorecards (2020)
The scorecards below show the skill (measured by the global root mean squared error) of different physical and ML-based methods relative to ECMWF's IFS HRES, one of the world's best operational weather models, on a number of key variables. For a detailed explanation of the different skill metrics and variables, check out the FAQ.
Scorecard for upper-level variables for the year 2020. Operational models are evaluated against IFS analysis. All other models evaluated against ERA5. Order of ML models reflects publication date. For more detail, visit Deterministic Scores.
Scorecard for surface variables for the year 2020. Operational models are evaluated against IFS analysis. All other models evaluated against ERA5. Order of ML models reflects publication date. Precipitation is evaluated using the SEEPS score using ERA5 ground truth for all models. For more detail, visit Deterministic Scores. See FAQ for details on how climatology is computed.
Probabilistic scorecards (2020)
The scorecard below shows the skill (measured by the continuous ranked probability score = CRPS) of physical and ML-based probabilistic models, relative to ECMWF's IFS ENS model.
Probabilistic scorecard for upper-level variables for the year 2020. Operational models (blue) are evaluated against IFS analysis. All other models evaluated against ERA5. Order of ML models reflects publication date. For more detail, visit Probabilistic Scores.
Probabilistic scorecard for surface variables for the year 2020. Operational models (blue) are evaluated against IFS analysis. All other models evaluated against ERA5. Order of ML models reflects publication date. For more detail, visit Probabilistic Scores.