2026/05/25 by Jiahao Zhu, Guokun Dai
Earth and Planetary Sciences · Environmental Science · #Climate variability and models #Meteorological Phenomena and Simulations #Tree-ring climate responses
paper · doi:10.1175/waf-d-25-0091.1
openalex publication_date 2026/05/25 · openalex created_date 2026/05/26 · openalex updated_date 2026/07/08
Abstract Utilizing the ERA5 reanalysis data, Eurasian extreme cold and heatwave events after 2017 are identified with objective regional extreme event identification techniques. Composite analyses reveal that extreme events are mainly modulated by persistent blockings, intensified subtropical highs, and anomalous midlatitude eddies. Forecast skill for these events is assessed with subseasonal-to-seasonal predictions from the China Meteorological Administration, the European Centre for Medium-Range Weather Forecasts, and the artificial intelligence (AI) model FuXi-S2S. Results from the ensemble mean of surface air temperature anomalies and the extreme forecast index indicate that FuXi-S2S can provide skillful cold event forecasts for 3–4 weeks, whereas dynamical models generally exhibit a shorter effective forecast duration, consistent with difficulties in predicting the persistence of blocking. However, dynamical models exhibit superior forecast skill in cold case 2, as FuXi-S2S has fewer training samples for such atypical circulation events. For heat waves, both the dynamical and AI models exhibit persistently high skill for up to 3 weeks for low-mobility heatwave case 7 when forecasts are initialized 1 week prior to the event onset. Furthermore, the forecast performances are related to the atmospheric circulation systems, indicating that these systems are sources of predictability for extreme events in Eurasia. All the extreme events, except heatwave case 8, are related to blockings. Further diagnosis shows that synoptic eddies modulate blocking onset and behaviors, further influencing the subseasonal forecast skills in dynamical models. However, FuXi-S2S struggles to replicate this eddies–blocking relationship, indicating limited physical interpretability in AI models. Significance Statement This study advances the mechanistic understanding of Eurasian extreme cold waves and heat waves by establishing how persistent atmospheric circulation anomalies drive their occurrence. Through systematic analysis of recent extreme temperature episodes, we demonstrate that nearly all such events are linked to atmospheric blocking, which prolongs and amplifies temperature extremes. Unlike prior studies limited to traditional dynamical models, we compare the extended-range (3–4 weeks) forecast skills of these models to artificial intelligence (AI) model. Both dynamical and AI-based models can skillfully forecast extreme heatwave events with low mobility up to 4 weeks. Crucially, our findings reveal that synoptic-scale eddies modulate the onset and evolution of blocking, thereby impacting the subseasonal forecast skill.