Microsoft's Probabilistic Bias Correction improves two-to-six-week weather forecasts

Microsoft Research says its Probabilistic Bias Correction (PBC) method uses machine learning to correct systematic bias in existing weather models, improving forecasts two to six weeks ahead.

The paper, published in Nature, reports better accuracy than ECMWF's physics-based model on 98 percent of precipitation and 92 percent of temperature targets, and doubled the accuracy of ECMWF's AI forecasting system in the study.