Search papers, labs, and topics across Lattice.
This paper introduces AIFS-SUBS, a machine-learning model designed to enhance data-driven forecasting at sub-seasonal timescales by addressing challenges such as error accumulation and systematic biases. By adopting a 24-hour autoregressive time step and incorporating additional predictors like stratospheric levels and thermal radiation, AIFS-SUBS achieves comparable probabilistic skill to the operational Integrated Forecasting System (IFS) while extending skilful forecasts of the Madden鈥揓ulian Oscillation (MJO) by eight days. Notably, AIFS-SUBS operates with significantly lower energy consumption, making it a promising candidate for large-scale real-time forecasting applications.
AIFS-SUBS not only matches the IFS in forecasting skill but also extends MJO forecasts by eight days while using 200 times less energy.
Data-driven models now rival numerical weather prediction in the medium range, but extending them to sub-seasonal lead times raises challenges absent at shorter horizons. Errors accumulate over long autoregressive rollouts, systematic biases grow with lead time, and several years of data must be held out for independent verification, even though machine-learning models otherwise benefit from longer training records. To address these challenges, we adapt ECMWF's AIFS-CRPS medium-range model. AIFS-SUBS adopts a 24h autoregressive time step to reduce error accumulation, adds stratospheric levels and top-of-atmosphere thermal radiation as predictors, and reserves 2007--2011 as an independent verification window. We evaluate two config-durations: AIFS-SUBS, fine-tuned on operational analyses, and AIFS-SUBS-ERA5, trained on ERA5 alone. Across weeks 2--6, AIFS-SUBS matches the operational Integrated Forecasting System (IFS) in probabilistic skill while reducing systematic biases. For the convective (OLR) component of the Madden--Julian Oscillation (MJO), AIFS-SUBS extends skilful forecasts (correlation > 0.5) by eight days relative to the IFS, while matching or exceeding the IFS for the full multivariate RMM index. AIFS-SUBS also reproduces the observed MJO modulation of tropical cyclone activity comparably. Stratospheric skill is particularly strong with AIFS-SUBS reproducing sudden stratospheric warming (SSW) frequency and surface impact. In the AI Weather Quest, AIFS-SUBS-ERA5 attains a variable-averaged ranked probability skill score slightly ahead of the IFS at weeks 3 and 4. At inference, AIFS-SUBS uses about 200 times less energy than the IFS, opening the door to much larger real-time ensembles. AIFS-SUBS is ECMWF's first machine-learning model targeted at sub-seasonal time-scales.
