A multi-scale loss formulation for learning a probabilistic model with proper score optimisation
This work addresses the challenge of scale-aware training for weather forecasting models, offering incremental improvements in probabilistic predictions.
The paper tackled the problem of training probabilistic weather forecasting models by introducing a multi-scale loss formulation, which improved small-scale variability without harming forecast skill in the AIFS-CRPS model.
We assess the impact of a multi-scale loss formulation for training probabilistic machine-learned weather forecasting models. The multi-scale loss is tested in AIFS-CRPS, a machine-learned weather forecasting model developed at the European Centre for Medium-Range Weather Forecasts (ECMWF). AIFS-CRPS is trained by directly optimising the almost fair continuous ranked probability score (afCRPS). The multi-scale loss better constrains small scale variability without negatively impacting forecast skill. This opens up promising directions for future work in scale-aware model training.