Fixing the Double Penalty in Data-Driven Weather Forecasting Through a Modified Spherical Harmonic Loss Function
This addresses a key limitation for meteorologists and forecasters by improving the accuracy and resolution of weather predictions, though it is an incremental improvement over existing models.
The paper tackled the double penalty problem in data-driven weather forecasting by modifying the loss function to separate decorrelation from spectral amplitude errors, resulting in sharp deterministic forecasts and an increase in effective resolution from 1,250km to 160km.
Recent advancements in data-driven weather forecasting models have delivered deterministic models that outperform the leading operational forecast systems based on traditional, physics-based models. However, these data-driven models are typically trained with a mean squared error loss function, which causes smoothing of fine scales through a "double penalty" effect. We develop a simple, parameter-free modification to this loss function that avoids this problem by separating the loss attributable to decorrelation from the loss attributable to spectral amplitude errors. Fine-tuning the GraphCast model with this new loss function results in sharp deterministic weather forecasts, an increase of the model's effective resolution from 1,250km to 160km, improvements to ensemble spread, and improvements to predictions of tropical cyclone strength and surface wind extremes.