Metrics

ClimateLearn provides the following metrics for deterministic predictions.

• Mean squared error (MSE)

• Mean squared error skill score (MSESS)

• Mean absolute error (MAE)

• Root mean squared error (RMSE)

• Anomaly correlation coefficient (ACC)

• Pearson’s correlation coefficient (Pearson)

• Mean bias

• Normalized root mean squared error (NRMSEs)

• Normalized root mean squared error in global mean (NRMSEg)

For probabilistic forecasts, the library provides the following metrics.

• Gaussian continuous ranked probability score (Gaussian CRPS)

• Gaussian spread

• Gaussian spread-skill ratio (Gaussian SSR)

We refer to the following sources for the definitions and motivations for these metrics:

• https://geo.libretexts.org/Bookshelves/Meteorology_and_Climate_Science/Practical_Meteorology_(Stull)/20%3A_Numerical_Weather_Prediction_(NWP)/20.7%3A_Forecast_Quality_and_Verfication

• https://repository.library.noaa.gov/view/noaa/48746

• https://arxiv.org/abs/2205.00865