// NATURE NEWS — SPAZIO & SCIENZA
Advancing subseasonal forecasting with machine learning
Nature
(2026) Cite this article
Decision-makers rely on weather forecasts to plant crops, manage wildfires, allocate water and energy and prepare for weather extremes1,2,3. Today, such forecasts enjoy unprecedented accuracy as far as 2 weeks thanks to steady advances in physics-based dynamical models and data-driven artificial intelligence models4,5,6,7,8,9,10. However, model skill drops precipitously at subseasonal timescales (2–6 weeks ahead) due to compounding errors, systemic model biases and the chaotic nature of the atmosphere11,12,13,14,15. To counter this degradation, we introduce probabilistic bias correction, a machine learning framework that substantially reduces systematic error by learning to correct historical probabilistic forecasts. When applied to the leading dynamical and artificial intelligence models from the European Centre for Medium-Range Weather Forecasts (ECMWF), probabilistic bias correction doubles the modest subseasonal skill of the AI Forecasting System16 and improves the skill of the operationally debiased dynamical model17 for 91% of pressure, 92% of temperature and 98% of precipitation targets. We designed probabilistic bias correction for operational deployment, and, in the 2025 real-time forecasting competition of ECMWF18, its global forecasts were the best for all weather variables and lead times, outperforming the dynamical models from six operational forecasting centres, an international dynamical multimodel ensemble, the AI Forecasting System of ECMWF and the forecasting systems of 34 teams worldwide. These probabilistic skill gains translate into more accurate prediction of extreme events and have the potential to improve agricultural planning, energy management and disaster preparedness in vulnerable communities19.
Subseasonal forecasts—weather predictions 2–6 weeks ahead—lie at the heart of agricultural planning, water allocation, disaster preparedness and energy management1,2,3. Yet accurate prediction at this timescale remains difficult. Today, our best estimates of future weather come from dynamical models that numerically simulate how the atmosphere and oceans change over time4,5 and artificial intelligence (AI) models that learn to predict the weather from historical data6,7,8. Because of the steady improvements in modelling, computation and data collection over the past half-century, deterministic forecasting systems can now provide accurate point estimates of the weather 1–2 weeks in advance10. However, the 2-week horizon also marks a transition in predictability: beyond this point, even small differences in initial weather conditions can lead to dramatic differences in outcomes and a single deterministic forecast cannot capture this uncertainty20.
To surmount this predictability barrier, one often turns to probabilistic forecasting, that is, predicting a distribution over future weather rather than a single point estimate21,22,23. Probabilistic forecasting explicitly encodes forecast uncertainty and provides actionable estimates of the likelihood of high-impact extreme events. For both dynamical and AI models, probabilistic forecasts are typically generated by perturbing the inputs, parameters or internal state of a model to produce an ensemble of possible future weather states13,23,24,25. Although these ensembles are routinely skilful at shorter lead times (to 15 days), operational dynamical ensembles and even state-of-the-art probabilistic AI routinely underperform the climatological average at longer subseasonal lead times, especially outside the tropics13.
A core difficulty is that leading subseasonal ensembles are generated iteratively over many time steps and even small per-step modelling errors compound into large systematic errors over several weeks14. This effect is especially pronounced for precipitation, which is sensitive to cascading inaccuracies across model components and highly variable in space and time4,26. Indeed, subseasonal f