Skillful subseasonal forecasts are crucial for various sectors of society but pose a grand scientific challenge. Recently, machine learning-based weather forecasting models outperform the most successful numerical weather predictions generated by the European Centre for Medium-Range Weather Forecasts (ECMWF), but have not yet surpassed conventional models at subseasonal timescales. This paper introduces FuXi Subseasonal-to-Seasonal (FuXi-S2S), a machine learning model that provides global daily mean forecasts up to 42 days, encompassing five upper-air atmospheric variables at 13 pressure levels and 11 surface variables. FuXi-S2S, trained on 72 years of daily statistics from ECMWF ERA5 reanalysis data, outperforms the ECMWF’s state-of-the-art Subseasonal-to-Seasonal model in ensemble mean and ensemble forecasts for total precipitation and outgoing longwave radiation, notably enhancing global precipitation forecast. The improved performance of FuXi-S2S can be primarily attributed to its superior capability to capture forecast uncertainty and accurately predict the Madden-Julian Oscillation (MJO), extending the skillful MJO prediction from 30 days to 36 days. Moreover, FuXi-S2S not only captures realistic teleconnections associated with the MJO but also emerges as a valuable tool for discovering precursor signals, offering researchers insights and potentially establishing a new paradigm in Earth system science research.

An extreme positive Indian Ocean Dipole (IOD) event occurred in 2019 boreal autumn, which has induced severe climate impacts around the Indian Ocean basin. In this study, the cause for 2019 IOD event and the related mechanism are explored. We find that the remarkable strengthening of Australian high and weakening of sea level pressure over South China Sea/Philippine Sea have been evidently visible since May 2019. Such a record‐breaking interhemispheric pressure gradient (IHPG) induced northward cross‐equatorial flow over the western Maritime Continent, which triggered strong wind‐evaporation‐SST and thermocline feedbacks in 2019. In addition to 2019 case, historical IOD events are highly correlated with IHPG from boreal late spring to summer. We show that skillful IHPG prediction can be made in March by European Centre for Medium‐Range Weather Forecasts seasonal forecast system, which makes it quite possible for the early warning of extreme IOD events by two to three seasons ahead.

Using the reanalysis data and local weather station data, the relationship between El Niño and Southern Oscillation (ENSO) and the summer precipitation over Northwest China (SPNWC) is investigated. A pronounced rainfall response is observed in post‐ENSO summer, especially over North Xinjiang and Yili River valley. A possible mechanism is proposed for such a seasonal lagged impact. El Niño could induce Indian Ocean warming that persists into boreal spring and summer although El Niño itself has dissipated. The warmer Indian Ocean sea surface temperature (SST) would heat the tropospheric atmosphere and elevate the tropical geopotential height at the upper level. As a result, the South Asian High (SAH) exhibits a southward extension due to the increasing geopotential height in its southern flank. A southward displacement of subtropical jet is evident associated with such SAH shift, which further induces a barotropic low pressure anomaly over central Asia. Meanwhile, associated with the southward displacement of subtropical jet, the water vapor from tropical Indian Ocean tends to transport across Iranian Plateau and bring wet and warm air into central Asia. The resultant convergence of water vapor with cold and dry air from high latitude leads to the observed rainfall anomaly to the south and east of Balkhash Lake, explaining the lagged relationship between ENSO and SPNWC. Our results have potential applications on the seasonal prediction of SPNWC.

Land surface temperature (LST) is an important parameter that affects the water cycle, environmental changes, and energy balance at global and regional scales. Herein, a time series analysis was conducted to estimate the monthly, seasonal, and interannual variations in LST during 2001–2019 in the Tarim Basin, China. Based on Moderate Resolution Imaging Spectroradiometer (MODIS) LST, air temperature, air pressure, relative humidity, wind speed, precipitation, elevation, and land-cover type data, we analyzed the spatio-temporal change characteristics of LST and the influencing factors. High LSTs occurred in the desert and plains and low LSTs occurred in surrounding mountain regions. The highest LST was recorded in July (25.1 °C) and the lowest was in January (−9.5 °C). On a seasonal scale, LST decreased in the order: summer > spring > autumn > winter. Annual LST showed an increasing trend of 0.2 °C/10 a in the desert and mountain areas, while the plains indicated a decreasing trend. In spring and autumn, western regions were dominated by a downward trend, whereas in winter a downward trend occurred in eastern regions. In summer, areas covered by vegetation were dominated by a downward trend, and desert and bare lands were dominated by an upward trend. Random forest (RF) model analysis showed that elevation was the most significant influencing factor (22.1%), followed by mean air temperature (20.1%). Correlation analysis showed that the main climatic factors air temperature, relative humidity, and elevation have a good correlation with the LST. Land-cover type also affected LST; during February–December the lowest LST was observed for permanent glacier snow and the highest was observed in the desert. El Nino and La Nina greatly influenced the LST variations. The North Atlantic Oscillation and Pacific Decadal Oscillation indices were consistent with the mean LST anomaly, indicating their considerable influence on LST variations.