Hidden High-Altitude Highway Governs How El Niño Shapes Asia’s Winters

Every winter, the East Asian Winter Monsoon delivers frigid Siberian air across China, the Korean Peninsula and Japan, and when it swings to extremes the consequences ripple through energy grids, agriculture and transport networks across one of the most densely populated regions on Earth. A new study published in Climate Dynamics by Reina Sakamoto of the University of Tsukuba and Weathernews Inc., together with colleagues at the Meteorological Research Institute of the Japan Meteorological Agency and the University of Tsukuba, has taken one of the most systematic looks yet at how well the world’s leading seasonal forecast systems can anticipate these monsoon swings. The answer is sobering but instructive: even the best operational models achieve only moderate skill, and the reason lies in a high-altitude atmospheric pathway that many models fail to capture.

The research team assembled seasonal hindcasts, retrospective forecasts of past winters, from eight state-of-the-art operational prediction systems. Hindcasting is the standard proving ground for seasonal prediction: models are initialized before known winters and their output is scored against observations, allowing scientists to quantify skill without the benefit of hindsight. The team evaluated how well each system reproduced the dominant mode of East Asian Winter Monsoon variability that is tied to the El Niño–Southern Oscillation, the planet’s most influential year-to-year climate fluctuation, which originates in the tropical Pacific through coupled interactions between sea surface temperatures and atmospheric convection.

The headline number from the assessment is a correlation skill of roughly 0.4 for the best-performing model. In seasonal forecasting, where the atmosphere’s own memory fades within about two weeks, any correlation above zero represents genuine predictive information, and values near 0.4 indicate that a meaningful but limited fraction of monsoon variability is captured. The spread among the eight systems was substantial, and that spread became the central puzzle of the study. If all models see the same tropical sea surface temperatures, why do some translate them into East Asian winter circulation far better than others?

The researchers’ diagnostics point to two distinct teleconnection pathways through which ENSO influences the monsoon, one in the lower troposphere and one in the upper troposphere. The lower-tropospheric pathway is the classical picture: ENSO-related heating anomalies modify the western Pacific subtropical circulation, the Philippine Sea anticyclone and the low-level winds that carry cold air southward. This pathway has long been the focus of model evaluation. The upper-tropospheric pathway, by contrast, operates roughly ten kilometers above the surface and has received far less attention in the seasonal prediction literature, despite the new evidence that it may be the decisive factor separating skillful models from unskillful ones.

The study names this upper-level feature the Southeast Asia–Japan pattern. In observations, tropical convective anomalies clustered around the Maritime Continent, the archipelago-rich region spanning Indonesia, Malaysia and surrounding seas, generate a divergence anomaly in the upper troposphere over the southeastern Tibetan Plateau. That divergence acts as a source of stationary Rossby wave activity, exciting a wave train that propagates northeastward toward Japan. The physics is rooted in the vorticity budget of the tropical upper troposphere: divergent outflow from deep convection generates rotational flow through the well-known mechanism described by Sardeshmukh and Hoskins, and the resulting wave energy travels along the westerly jet, guided by the propagation characteristics of the background flow.

When this wave train arrives over Japan, it establishes a barotropic circulation anomaly, meaning a coherent signal through a deep layer of the atmosphere, which then modulates the East Asian trough in the mid-troposphere and the meridional winds near the surface. The East Asian trough is a keystone of the winter circulation, anchoring the storm track and controlling the southward penetration of cold surges. By shifting and reshaping the trough, the upper-tropospheric wave train ultimately influences surface temperature, snowfall and cold-air outbreak frequency across East Asia. The pathway therefore links tropical convection to East Asian winter climate through a chain that runs from divergence over the Tibetan Plateau, through a northeastward-propagating wave, to midlatitude circulation and finally to surface weather.

The crucial finding is that the Southeast Asia–Japan pattern depends sensitively on the longitudinal position and amplitude of the ENSO-related tropical convective anomalies. Models that place their convective responses in the right location, with realistic amplitude, reproduce the divergence over the southeastern Tibetan Plateau and hence the observed wave train and its downstream impacts. Models in which the convective response is shifted too far eastward, or is too weak, displace the divergence source and consequently fail to reproduce the teleconnection. The wave train is not simply absent in these models; it is mislocated, and a mislocated wave train lands its circulation anomalies in the wrong place, degrading the forecast precisely where users need it most.

This diagnosis reframes the challenge of East Asian winter seasonal prediction. Traditional evaluation has emphasized the lower-tropospheric circulation, such as the subtropical high and low-level monsoon winds, as the primary conduit of ENSO influence. The new multi-model assessment demonstrates that an accurate representation of the upper-tropospheric teleconnection is at least as important, and possibly more so, for determining whether a prediction system can anticipate the monsoon’s ENSO-related mode. For model developers, the practical implication is that improving the simulation of tropical convection over the Maritime Continent, a notoriously difficult problem because convection is parameterized rather than explicitly resolved, should pay dividends far downstream over the midlatitude Pacific rim.

The study also carries a broader message about the limits and promise of seasonal forecasting. The modest skill ceiling, even among eight leading operational systems, underscores the inherent difficulty of predicting the East Asian Winter Monsoon, whose variability is shaped not only by ENSO but also by the Arctic Oscillation, Siberian snow cover, Arctic sea ice and internal atmospheric dynamics that are largely unpredictable at seasonal range. Yet by identifying the specific process, the Maritime Continent convection feeding the Southeast Asia–Japan wave train, where model errors concentrate, the research converts a vague skill deficit into a concrete target. The work was supported by the Japan Society for the Promotion of Science and the Environmental Restoration and Conservation Agency, and its diagnostics draw on established reanalysis products including JRA-3Q and precipitation data from the Global Precipitation Climatology Project. For forecasters bracing for the next severe cold surge, the path to better warnings may run ten kilometers overhead, through the divergent outflow of tropical thunderstorms and the Rossby waves they send racing toward Japan.

Subject of Research: Seasonal prediction of the East Asian Winter Monsoon and the upper- and lower-tropospheric pathways of ENSO influence

Article Title: Multi-model assessment of East Asian Winter Monsoon predictions: upper- and lower-tropospheric pathways of ENSO impacts

Article References: Sakamoto, R., Takaya, Y., Hirahara, S., Naoe, H., & Ueda, H. (2026). Multi-model assessment of East Asian Winter Monsoon predictions: upper- and lower-tropospheric pathways of ENSO impacts. Climate Dynamics, 64(10), Article 437. https://doi.org/10.1007/s00382-026-08367-x

Image Credits: AI Generated

DOI: 10.1007/s00382-026-08367-x

Keywords: East Asian Winter Monsoon, ENSO, seasonal prediction, Rossby wave train, Maritime Continent, tropical convection, upper troposphere, Southeast Asia–Japan pattern, East Asian trough, teleconnections, Climate Dynamics, multi-model assessment