Stratospheric precursor induces wintertime phase reversal of the "warm Arctic-cold Eurasia" pattern.

Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Nat Commun . 2026 Feb 27;17:3284. doi: 10.1038/s41467-026-70100-3 Search in PMC Search in PubMed View in NLM Catalog Add to search Stratospheric precursor induces wintertime phase reversal of the “warm Arctic-cold Eurasia” pattern Yijia Zhang Yijia Zhang 1 State Key Laboratory of Climate System Prediction and Risk Management, Nanjing University of Information Science and Technology, Nanjing, China 2 School of Atmospheric Sciences, Nanjing University of Information Science and Technology, Nanjing, China Find articles by Yijia Zhang 1, 2, # , Zhicong Yin Zhicong Yin 1 State Key Laboratory of Climate System Prediction and Risk Management, Nanjing University of Information Science and Technology, Nanjing, China 2 School of Atmospheric Sciences, Nanjing University of Information Science and Technology, Nanjing, China Find articles by Zhicong Yin 1, 2, ✉, # , Wenshou Tian Wenshou Tian 3 Key Laboratory for Semi-Arid Climate Change of the Ministry of Education, College of Atmospheric Sciences, Lanzhou University, Lanzhou, China Find articles by Wenshou Tian 3 , Shengping He Shengping He 4 Geophysical Institute, University of Bergen and Bjerknes Centre for Climate Research, Bergen, Norway Find articles by Shengping He 4 , Pangchi Hsu Pangchi Hsu 1 State Key Laboratory of Climate System Prediction and Risk Management, Nanjing University of Information Science and Technology, Nanjing, China 2 School of Atmospheric Sciences, Nanjing University of Information Science and Technology, Nanjing, China Find articles by Pangchi Hsu 1, 2 Author information Article notes Copyright and License information 1 State Key Laboratory of Climate System Prediction and Risk Management, Nanjing University of Information Science and Technology, Nanjing, China 2 School of Atmospheric Sciences, Nanjing University of Information Science and Technology, Nanjing, China 3 Key Laboratory for Semi-Arid Climate Change of the Ministry of Education, College of Atmospheric Sciences, Lanzhou University, Lanzhou, China 4 Geophysical Institute, University of Bergen and Bjerknes Centre for Climate Research, Bergen, Norway ✉ Corresponding author. # Contributed equally. Received 2025 Sep 15; Accepted 2026 Feb 18; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13065799  PMID: 41760659 Abstract Rapid Arctic warming has reached 2–4 times the global average, contrasting with the cooling trend in mid-high latitude Eurasia during the 1990s–2010s. A notable phenomenon of the “warm Arctic-cold Eurasia” (WACE) pattern is its frequent phase reversal between early and late winter seen in the last decade. As one of the important drivers of climate variability, the role of the stratosphere in WACE reversal remains unclear. Here, we find that approximately 25 days before WACE reversal, the morphology of the stratospheric polar vortex over North America-North Atlantic undergoes a transition between stretching and contraction. Through vertical wave coupling and downward propagation, this stratospheric transition effectively modulates the key atmospheric circulation responsible for the WACE reversal. CMIP6 models including complete stratospheric processes can successfully simulate the WACE reversal induced by the stratospheric precursor, while low-top models fail to capture this linkage. Our findings deepen the understanding of cold-warm transition events from a stratosphere-troposphere coupling perspective. Subject terms: Climate change, Atmospheric science This study shows that a transition of the stratospheric polar vortex morphology is an effective precursor of the phase reversal of the “warm Arctic-cold Eurasia” pattern in winter and that CMIP6 models with complete stratospheric processes can successfully capture this physical link. Introduction The seesaw linkages between Arctic and Eurasian winter temperatures have garnered considerable attention 1 – 3 . Explorations mainly focus on both the long-term trend of “Arctic warming-Eurasia cooling” and the interannual-decadal variability of the “warm Arctic-cold Eurasia” (WACE) pattern 4 , 5 . In the recent decade, a notable phenomenon of the WACE pattern is its frequent phase reversal between early and late winter 6 . That is to say, a significant reversal between the WACE phase and “cold Arctic-warm Eurasia” (CAWE) phase occurs in a single winter, and the phase in both early and late winter is pronounced and persistent (Supplementary Fig. 1 ). Such phase reversal was observed in seven winters since 2012, indicating a notable increase in subseasonal variability. The opposite anomalies between early and late winter cancel each other out on the winter-mean scale, thus presenting a weakened “Arctic warming-Eurasia cooling” trend 6 , 7 . The WACE reversal causes a dramatic change in the meridional gradient of surface air temperature (SAT) between Arctic and Eurasia 8 , and substantially enhances the likelihood of extreme events, including the winter abrupt cold-warm transition in East Asia and the severe spring super sandstorms in North China 9 , 10 . Understanding the physical processes and mechanisms underlying the WACE reversal is therefore essential for interpreting Arctic-Eurasia climate variability and its implications for extreme climate. The formation mechanisms of the WACE pattern have been widely discussed, with more insights revealed on the interannual-decadal scales. Climate forcing factors, such as the autumn Arctic sea ice melting 11 , 12 , La Niña 13 , and the Pacific decadal oscillation 14 , can contribute to the WACE pattern through Rossby wave and stratosphere–troposphere interactions. In the atmospheric internal variability, the weakened stratospheric polar vortex (SPV) and the deep Arctic warming are also conducive to the formation of the WACE pattern 15 , 16 . In recent years, increasing attention has been devoted to the phase reversal of WACE between early and late winter, and its underlying physical mechanisms are being progressively uncovered. The coordination of the preceding tropical-subtropical sea surface temperature anomalies in the North Atlantic and Indian Ocean can effectively modulate the reversal between the WACE and CAWE phase 6 . In addition, the reversal of Ural blocking anomalies in late December is also a potential factor, which can affect the upward propagation of planetary-scale waves to promote the contribution of the SPV to the SAT reversal in Arctic-Eurasia regions 17 . Previous studies provided physical explanations for the WACE reversal by focusing on the air–sea interaction and the variations of tropospheric atmospheric circulations. However, as a key driver of extreme climate, it remains unclear whether the stratospheric variability can impact the phase reversal of the WACE pattern between early and late winter. The SPV is a critical bridge linking the stratospheric variability and the tropospheric response, and plays a crucial role in climate changes over the mid-high latitudes 18 , 19 . The SPV anomalies can affect the phase of Arctic oscillation/North Atlantic oscillation (NAO) through various mechanisms, such as the downward propagation of zonal-mean anomalies, vertical wave coupling, surface amplification 20 , 21 , thereby modulating the frequency of tropospheric blockings in the North Atlantic-Eurasia and the intensity of East Asian winter monsoon 22 , 23 . The weakened SPV contributes to an increasing number of cold air outbreaks in the Northern Hemisphere, especially over northern Eurasia and eastern America 24 , 25 . A stretched SPV, characterized by zonally asymmetric anomalies, can also lead to cold anomalies over Eurasia through stratosphere–troposphere coupling 26 , 27 . The stratosphere exhibits a longer characteristic timescale compared to the troposphere, thus serving as an important source of subseasonal-seasonal predictability for the extreme climate 28 , 29 . In several winters, the SPV undergoes a significant phase transition, such as the 1987/1988 winter. The SPV exhibited a prominent transition from weak to strong, accompanied by an out-of-phase tropospheric zonal wind anomaly between early and late winter 30 . This suggests that the changes in the strength or morphology of the SPV have a potential effect on the SAT reversal in the Arctic-Eurasia region, although this mechanism has not yet been revealed. Given the close linkages between the SPV and troposphere-surface climate variability, we propose exploring the role of the stratosphere in the WACE reversal. We first examine the characteristics and lead time of the SPV associated with the WACE reversal, and then investigate the pivotal processes by which the SPV modulates the phase reversal between WACE and CAWE. The key linkages and processes proposed are further demonstrated in the large ensemble simulations, and the reasons for the differences in simulation abilities are discussed. Results Linkages between SPV morphology transition and WACE reversal The WACE pattern is usually expressed by the SAT difference between the Barents-Kara seas (70°–85°N, 30°–100°E) and Eurasia (40°–60°N, 60°–120°E) after detrending, defined as the WACE T index (the former minus the latter) 5 , 6 . The positive WACE T represents the WACE phase, while the negative value represents the CAWE phase. As illustrated by the daily WACE T index in the winters of 1979/80–2023/24, the WACE T reversal from early to late winter occurred in 16 winters (see “Methods”). Among these, 8 winters experienced a reversal from WACE to CAWE, and 8 winters underwent a reversal from CAWE to WACE. Most of the reversal points are around January 10th, manifesting as the opposite WACE T between periods December 1st to January 10th (Dec1 Jan10, defined as early winter) and January 11th to February 28th (Jan11 Feb28, defined as late winter) (Supplementary Fig. 1 ). The SAT anomalies in the Arctic and Eurasian centers both show significant cold/warm transition within a single winter and remain consistent and persistent in each half-winter. In the composite of these WACE T reversal years, a significant feature of stratospheric anomalies is the transition of the SPV morphology between contraction and stretching over North America-North Atlantic (Fig. 1a, b ). Specifically, when the WACE T reverses from the CAWE phase in early winter to the WACE phase in late winter, the SPV undergoes a variation from contraction to a stretched shape over North America-North Atlantic. The transition time of the SPV morphology is earlier than the WACE T reversal, mainly manifested as an evident transition between November 1st–December 15th (Nov1 Dec15) and December 16th–February 10th (Dec16 Feb10). During Nov1 Dec15, the stratospheric anomalies show a zonal asymmetry, with negative potential vorticity (PV) anomalies over North America-North Atlantic and positive anomalies over Eurasia on the 530 K isentropic surface (Fig. 1a ). However, the PV anomalies turn to a uniform positive pattern in Dec16 Feb10, presenting an enhanced SPV, with its center more biased towards North America-North Atlantic (Fig. 1b ). The SPV changes in these two periods do not exhibit a completely out-of-phase variation, so the zonal mean stratospheric anomalies cannot capture the evident features well (Supplementary Fig. 2 ). In both periods, the contraction and stretching of the SPV edge relative to the climate state are most significant over North America-North Atlantic, and the local stratospheric anomalies show significant phase reversal. Therefore, we focus on the linkages between the SPV morphology transition over North America-North Atlantic and the WACE T phase reversal. Fig. 1. Linkages between the precursory stratospheric polar vortex and the phase reversal of “warm Arctic-cold Eurasia”. Open in a new tab Composites of the potential vorticity on the 530 K isentropic surface (shadings) in ( a ) November 1st to December 15th and ( b ) December 16th to February 10th in the “warm Arctic-cold Eurasia” (WACE) pattern reversal years during 1979/80 to 2023/24 (cold Arctic-warm Eurasia (CAWE) to WACE years minus WACE to CAWE years). The green, pink and blue lines represent the composited polar vortex edge in climate mean, CAWE to WACE years and WACE to CAWE years, respectively. The edge of stratospheric polar vortex is determined by 58PVU on the 530 K isentropic surface. The white dots indicate that the results are significant above the 95% confidence level. c The upper part: Daily variations of the stratospheric polar vortex morphology over North America-North Atlantic (NANA-SPV IPV ; red solid line) and the WACE pattern (WACE T ; black solid line) composited in the WACE T reversal years (CAWE to WACE years minus WACE to CAWE years), and the composited variations after 45-day low-pass filtering are represented by the smooth lines. The thicker lines indicate that the results are significant above the 95% confidence level. The red shadings represent the Nov1 Dec15 and Dec16 Feb10 periods associated with NANA-SPV IPV transitions, and the gray shadings represent the early and late winter associated with WACE T reversal. The lower part: 45-day moving t-test of the NANA-SPV IPV and WACE T during November to February in each WACE T reversal years (shadings) and their means (lines), and the most significant points are represented by the markers. The linear trend is removed. The morphology and shifting of SPV are represented by the area-weighted mean of potential vorticity on the 530 K (approximately 44 hPa) over North America-North Atlantic (0–120°W, 60–75°N) 31 , defined as the NANA-SPV IPV index. The positive NANA-SPV IPV indicates that the SPV shifts and stretches toward North America-North Atlantic, while a negative NANA-SPV IPV indicates that the SPV moves away from and contracts over North America-North Atlantic (Supplementary Fig. 3 ). When the SAT pattern changes from CAWE to WACE, the composited NANA-SPV IPV is persistently negative in Nov1 Dec15, then turns to positive in mid-December and maintains until February (Fig. 1c ). The SPV edge determined by 58PVU on the 530 K isentropic surface 32 also clearly reflects the SPV area over North America-North Atlantic transits from contraction to expansion (Supplementary Fig. 4 ). That is to say, a contracted SPV over North America-North Atlantic in Nov1 Dec15 tends to be followed by a CAWE phase in early winter, and a stretched SPV towards North America-North Atlantic in Dec16 Feb10 tends to be followed by a WACE phase in late winter. Conversely, the SPV morphology transition from stretching to contracting corresponds to a precursor to the reversal from WACE to CAWE. Both NANA-SPV IPV and WACE T experience significant phase reversal, and the reversal time of NANA-SPV IPV is approximately 25 days prior to that of WACE T (Fig. 1c ). The NANA-SPV IPV and WACE T processed with a 45-day low-pass filtering can focus more on the phase changes (Fig. 1c ). The filtered variations highlight the pronounced transition of the SPV morphology over North America-North Atlantic at mid-December, and more clearly reflect the lead-lag relationship between the NANA-SPV IPV transition and the WACE T reversal. The moving t-test of NANA-SPV IPV and WACE T variations can detect the mean-value shift in subsequences (see “Methods”), which also confirms their phase reversal and corresponding time points in the WACE T reversal years (Fig. 1c ). Further analysis on information flow (see “Methods”) shows that the PV anomalies at 530 K over North America-North Atlantic exceed the 95% confidence level for causality (Supplementary Fig. 5 ), indicating that the NANA-SPV IPV transition plays a pivotal role in modulating the WACE T reversal. During the longer period from 1950/51 to 2023/24 (25 WACE T reversal years), the lead-lag linkage between the NANA-SPV IPV transition and the WACE T reversal still exists (Supplementary Fig. 6 ). Thus, the SPV morphology transition over North America-North Atlantic is assumed to be one of the important drivers for the phase reversal between WACE and CAWE. The NANA-SPV IPV can effectively capture the variations in SPV morphology and shift over North America-North Atlantic (Supplementary Fig. 3 ). Based on the analysis of the daily variation of NANA-SPV IPV , the years when the SPV morphology over North America-North Atlantic undergoes a transition are identified. During 1979/80–2023/24, the significant phase transition of NANA-SPV IPV is observed in 18 years (see “Methods”), including two types: the SPV morphology transition over North America-North Atlantic from contraction to stretching (9 years) and from stretching to contraction (9 years) (Supplementary Fig. 7 ). The transition time of NANA-SPV IPV is predominantly concentrated around December 15th. This reversal is characterized by stable yet opposite NANA-SPV IPV anomaly between Nov1 Dec15 and Dec16 Feb10, and the anomaly in Dec16 Feb10 is stronger than that in Nov1 Dec15. Additionally, to assess whether the phenomenon of NANA-SPV IPV phase transition is sensitive to the selection of the NANA-SPV IPV reversal timing, we perturbed the division dates by a few days around December 15th and found the results to be consistent. In 12 of the 18 years with NANA-SPV IPV transition, the WACE T reverses from early to late winter (Supplementary Fig. 8 ), covering up to 75% of the winters with WACE reversal. The stratospheric anomalies associated with WACE T reversal are deep and barotropic on the vertical structure, especially over the North America-North Atlantic (Supplementary Fig. 9 ). Both during Nov1 Dec15 and Dec16 Feb10, the opposite geopotential height anomalies over North America-North Atlantic can extend from the stratosphere down to the near surface. This phenomenon can also be observed in the westerly jet surrounding SPV over North America-North Atlantic. The anomalies of geopotential height and zonal wind in Dec16 Feb10 are stronger and last longer than those in Nov1 Dec15. Correspondingly, the WACE T in Jan11 Feb28, lagging NANA-SPV IPV by about 25 days, is more prominent and persistent (Fig. 1c ), indicating an enhanced impact of NANA-SPV IPV in Dec16 Feb10. The stratospheric anomalies of geopotential height and the zonal wind over Eurasia are opposite with those over North America-North Atlantic in Nov1 Dec15 (Supplementary Fig. 9a ). These anomalies can also extend from the middle stratosphere to the near-surface, indicating the barotropy of the zonal asymmetric pattern. During the Dec16 Feb10 period, the center of the stratospheric signal is more concentrated over North America-North Atlantic, while it is relatively weak over Eurasia and fails to effectively propagate downward (Supplementary Fig. 9b ). Thus, an important stratospheric precursor associated with the WACE T reversal is the SPV morphology transition over North America-North Atlantic. This raises a question of how the SPV morphology transition over North America-North Atlantic drives the WACE pattern to reverse. Cross-continental physical processes modulated by the SPV The SPV variations correlate closely with the tropospheric atmosphere through stratosphere–troposphere coupling processes, thereby affecting the occurrence of extreme weather and climate 20 . When a negative NANA-SPV IPV occurs in Nov1 Dec15, the SPV morphology is contracted over North America-North Atlantic. The geopotential height anomalies at 50  hPa form a zonal asymmetrical pattern, with positive anomalies over North America-North Atlantic and negative anomalies over Western Eurasia, which are consistent with the stratospheric anomalies when WACE T reverses (Fig. 2a ). We use the Plumb wave activity fluxes averaged over 60°–70°N as a function of longitude and pressure to investigate the role of vertical wave coupling 21 . The Plumb fluxes in the lower stratosphere over North America-North Atlantic are predominantly upward (Supplementary Fig. 10a ), contributing to the enhancement and persistence of the anomalous stratospheric high pressure over North America-North Atlantic. The upward and eastward propagating Plumb waves form the anomalous downward wave activity flux over Western Europe, accompanied by the negative geopotential height anomalies that propagate from the stratosphere to the troposphere (Fig. 3a ). Fig. 2. Responses of the atmospheric circulations and surface air temperature to the stratospheric polar vortex morphology transition over North America-North Atlantic. Open in a new tab a Composites of the geopotential height at 50 hPa in Nov1 Dec15 (the top panel), the geopotential height at 500 hPa in Nov1 Dec15 (contours) and in Dec1 Jan10 (shadings), the wind at 850 hPa in Dec1 Jan10 (arrows, the middle panel), and the surface air temperature (SAT) in Dec1 Jan10 (the bottom panel) according to the transition of stratospheric polar vortex morphology over North America-North Atlantic (NANA-SPV IPV ) during 1979/80 to 2023/24 (negative to positive NANA-SPV IPV years minus positive to negative NANA-SPV IPV years). b Composites of the geopotential height at 50 hPa in Dec16 Feb10 (the top panel), the geopotential height at 500 hPa in Dec16 Feb10 (contours) and in Jan11 Feb28 (shadings), the wind at 850 hPa in Jan11 Feb28 (arrows, the middle panel), and the SAT in Jan11 Feb28 (the bottom panel) according to NANA-SPV IPV transition during 1979/80 to 2023/24. The cyan arrows represent the propagation of the Plumb wave at the 500 hPa in Nov1 Dec15 and Dec16 Feb10. The contours and shadings in the top and middle panels and the white dots in the bottom panels indicate that the results are significant above the 95% confidence level. Fig. 3. Physical mechanisms of the stratospheric polar vortex morphology transition over North America-North Atlantic impacting the phase reversal of “warm Arctic-cold Eurasia”. Open in a new tab Composites of the geopotential height (shadings) and the vertical and zonal components of anomalous Plumb wave activity flux (vectors) averaged over 60°–70°N as a function of longitude and pressure in ( a ) Nov1 Dec15 and ( b ) Dec16 Feb10 according to the transition of stratospheric polar vortex morphology over North America-North Atlantic (NANA-SPV IPV ) during 1979/80 to 2023/24 (negative to positive NANA-SPV IPV years minus positive to negative NANA-SPV IPV years). The magnitude of the Plumb flux is scaled by (1000/ p ) 1/2 . c Composites of the daily variations of the pressure-temporal geopotential height (shadings) and zonal wind (contours) along 0–120°W, 60–70°N, and the daily variations of NANA-SPV IPV , Ural blocking (UB) and “warm Arctic-cold Eurasia” pattern (WACE T ) index according to NANA-SPV IPV transition during 1979/80 to 2023/24 (negative to positive NANA-SPV IPV years minus positive to negative NANA-SPV IPV years). The composited variations after 45-day low-pass filtering are represented by the smooth lines. The thicker lines indicate that the results are significant above the 95% confidence level. The white dots indicate that the results are significant above the 95% confidence level. In the troposphere, the positive geopotential height anomaly over North America-North Atlantic and the negative anomaly over Western Europe form a Rossby wave propagating eastward (Supplementary Fig. 11a ). This is conducive to a notable positive PV anomaly at 315 K and anomalously increased static stability in the Ural mountains region (Supplementary Fig. 12 ), thereby suppressing the Ural blocking activities 33 (Fig. 2a ). The weakened Ural high can last until mid-January and form the sustained southerly winds in the mid-high latitudes of Eurasia (Supplementary Fig. 11c ), which hinder the southward transport of cold air and gather the cold air in the Barents-Kara Seas, thus resulting in a significant CAWE anomaly during early winter (Figs. 2a and 3c ). When the NANA-SPV IPV transits from negative to positive, the SPV appears to be stretched towards North America-North Atlantic in Dec16 Feb10. The entire Arctic region is covered by an enhanced SPV, with the center of anomalies biased towards North America-North Atlantic (Fig. 2b ). Over North America-North Atlantic, the anomalous downward propagation of the Plumb vertical flux is observed (Fig. 3b and Supplementary Fig. 10b ). The negative geopotential height anomalies and intensified zonal winds propagate strongly and persistently from the stratosphere to the troposphere (Fig. 3c ), contributing to the formation of a positive NAO phase (Fig. 2b ). The Rossby wave train triggered by the stratospheric atmosphere anomaly is generated in the mid-latitudes of the North Atlantic (Supplementary Fig. 11b, d ), serving as a wave source 34 . The wave activity flux in the troposphere propagates eastward and causes the abnormal high pressure over Western Europe and the mid-high latitudes of Eurasia. The persistence of this Rossby wave makes the Ural high enhanced and sustained in late winter (Fig. 2b ). In addition, the eddy vorticity forcing forms a positive synoptic eddy feedback (see “Methods”), and further enhances the tropospheric response 35 . Under the influence of enhanced SPV, the eddy vorticity forcing in the North Atlantic is conducive to the maintenance of a positive NAO phase. The divergent eddy vorticity fluxes over Western Europe and the Ural region generate an anticyclonic forcing (Supplementary Fig. 13 ), which strengthens and maintains the Ural blocking in late winter. Driven by the intensified Ural high, a pronounced WACE phase occurs in late winter that is opposite to that in early winter (Figs. 2b and 3c ). In the real case observed in 1983/84, a typical NANA-SPV IPV transition from negative to positive occurred, leading to a WACE T reversal with a lag of about 25 days (Supplementary Fig. 14 ). An abnormal positive pressure center at 50 hPa was located over North America-North Atlantic in Nov1 Dec15 of 1983/84, but there was a negative pressure center over Eurasia, representing a contraction of the SPV area over North America-North Atlantic (Supplementary Fig. 14a ). While in Dec16 Feb10, the SPV expanded sharply towards North America-North Atlantic, with a remarkable negative geopotential height anomaly at 50 hPa (Supplementary Fig. 14b ). Affected by the downward influence of the stratospheric anomalies and the propagation of tropospheric Rossby waves (Supplementary Fig. 15 ), the Ural high intensity underwent a significant variation from weakened in early winter to enhanced in late winter, thus causing a phase reversal from CAWE to WACE (Supplementary Fig. 14c ). The stratospheric anomaly in Dec16 Feb10 was stronger than that in the other sub-period, thereby causing a more pronounced response of Ural high and WACE T in late winter. East Asia was affected by the significant WACE phase and experienced a severe cold wave in January and February of 1984 36 . The NANA-SPV IPV transition is approximately 25 days ahead of the WACE T phase reversal (Fig. 1c ), which may be related to the time required for the downward transmission of the stratospheric signal and the excitation and propagation of Rossby waves 18 . This time interval is coordinated with a maximum suppression of Ural blocking observed around day 24 following the SPV weakening 33 . The downward influences of the SPV morphology transition indicate the modulations from the stratospheric variability on the phase variations of the SAT pattern in the Arctic-Eurasian region, deepening the understanding of the widely concerned cold-warm transition events from a fresh perspective of stratosphere–troposphere coupling 37 . Simulations of the key processes and linkages To further determine the robustness of the linkages between the variation of SPV morphology over North America-North Atlantic and the WACE T phase reversal, and verify the proposed physical processes, we perform the analysis using historical simulations of the Coupled Model Intercomparison Project phase 6 (CMIP6) from 1950 to 2014 and the Community Earth System Model large ensemble (CESM-LM) experiments from 1920 to 2005 (see “Methods”). The simulations of 21 CMIP6 models and 39 CESM-LE members, which have daily data and are available for download, are used for analysis. Since the PV variable is not directly available in simulations, we use the 50 hPa geopotential height to identify the years with the SPV morphology transition 38 . The conclusions obtained from the reanalysis using the PV at 530 K and the geopotential height at 50 hPa are consistent (see “Methods” for detailed explanations). Specifically, the case is selected as the SPV morphology transition over North America-North Atlantic from contraction to stretching based on these two conditions: (1) the area-weighted mean of geopotential height anomalies at 50 hPa over North America-North Atlantic transits from positive in Nov1 Dec15 to negative in Dec16 Feb10; (2) the geopotential height anomalies in Nov1 Dec15 present a zonal asymmetry over North America-North Atlantic and Eurasia, which is defined as the simulated “Con-to-Stre NANA-SPV” case (Supplementary Fig. 16a, b ). Conversely, the case is defined as the simulated “Stre-to-Con NANA-SPV” to represent the SPV morphology transition over North America-North Atlantic from stretching to contraction (Supplementary Fig. 16c, d ). In the CMIP6 and CESM-LE simulations, the SPV morphology transitions over North America-North Atlantic between Nov1 Dec15 and Dec16 Feb10 occur in 27% and 28% of the years (Fig. 4a ), respectively, which are lower than the observed frequency of NANA-SPV IPV transition (40%). Fig. 4. Verifications of the physical linkages in simulations. Open in a new tab a Multi-member mean and the probability distribution of the simulated difference of “warm Arctic-cold Eurasia” pattern between early and late winter (Diff-WACE T ) in the simulated transition of stratospheric polar vortex morphology over North America-North Atlantic from contraction to stretching (Con-to-Stre NANA-SPV) years (red) and from stretching to contraction (Stre-to-Con NANA-SPV) years (blue) based on CMIP6 simulations and CESM-LE simulations. The center line, box limits and whiskers of the box-plots elements are defined as mean, upper and lower quartiles, maximum and minimum, respectively. The hollow triangles represent the reanalysis results. b The simulated “warm Arctic-cold Eurasia” pattern (WACE T ) in early and late winter in the simulated Con-to-Stre NANA-SPV years (red) and Stre-to-Con NANA-SPV years (blue) based on the high-top and low-top CMIP6 models. The center line, box limits and whiskers of the box-plots elements are defined as mean, upper and lower quartiles, maximum and minimum, respectively. c , d Composites of the pressure-temporal geopotential height along 0–120°W, 60–70°N, and the daily variation of the geopotential height anomalies at 500 hPa over the North America-North Atlantic (NANA Z500 ), Ural blocking (UB) and WACE T according to the simulated stratospheric polar vortex morphology transition (Con-to-Stre NANA-SPV years minus Stre-to-Con NANA-SPV years) that selected from the high-top and low-top CMIP6 models. The white dots and the thicker lines indicate that the results are significant above the 95% confidence level. The WACE T reversal represents a significant WACE T difference between early and late winter (the latter minus the former, defined as Diff-WACE T ). We first verify whether the SPV morphology transition over North America-North Atlantic can lead to the corresponding Diff-WACE T through the model simulations. In the 200 cases of simulated Con-to-Stre NANA-SPV, the mean Diff-WACE T is positive (Fig. 4a ), and the distribution of SAT difference over Arctic-Eurasia exhibits a notable WACE anomaly (Supplementary Fig. 17a ). Among these 200 cases, 72% show a positive Diff-WACE T , indicating that the transition of SPV morphology over North America-North Atlantic from contraction to stretching is conducive to the reversal towards the WACE phase. In the 176 cases of simulated Stre-to-Con NANA-SPV, the mean SAT difference in all such years shows a significant CAWE anomaly (Supplementary Fig. 17b ). The negative Diff-WACE T can be captured in 71% of these years, indicating a tendency and a higher probability to reverse to CAWE phase (Fig. 4a ). The two transition types based on CESM-LE simulations, which have a larger sample size, also strongly verify that the SPV morphology transition over North America-North Atlantic facilitates the reversal between WACE and CAWE phase from early to late winter (Supplementary Fig. 17c, d ). However, the responses of Diff-WACE T associated with the SPV morphology transitions over North America-North Atlantic simulated by CMIP6 and CESM-LE are both weaker than the reanalysis results (Fig. 4a ). However, there is a large spread of simulated Diff-WACE T in Con-to-Stre NANA-SPV and Stre-to-Con NANA-SPV years in different CMIP6 models, and not all models can capture the physical linkages between the two well. The height of the model top level is revealed to be one of the key factors restricting the model simulation ability of the stratosphere–troposphere interactions 39 . We categorize the models into high-top and low-top groups based on whether their model top level reaches 0.1 hPa, thereby comparing their simulation capabilities for the SPV morphology transition and its influence on WACE phase reversal. The high-top and low-top model groups consist of 9 and 12 models, respectively, with the ranges of top level being 4.5 × 10 −6 –0.1 and 0.2–10 hPa, respectively (Supplementary Table 1 ). Both the high-top and low-top models can simulate the variations of the geopotential height anomalies at 50 hPa during Nov1 Dec15 and Dec16 Feb10 associated with the SPV morphology transition over North America-North Atlantic (Fig. 4c, d ). In the years when the SPV morphology undergoes transition, the accuracy rates for the sign of the simulated Diff-WACE T in the high-top and low-top models are 76 and 69%. The accuracy rate refers to the proportion of cases in which a positive Diff-WACE T is simulated in Con-to-Stre NANA-SPV cases, and a negative Diff-WACE T is simulated in Stre-to-Con NANA-SPV cases. A higher accuracy rate of high-top models indicates that they have a better simulation ability for the linkage between the SPV morphology transition and WACE T phase reversal. Specifically, both the high-top and low-top models can simulate the WACE T response in early winter, but significant differences exist in simulations of the WACE T in late winter (Fig. 4b ). The high-top models can more accurately reproduce the WACE T phase reversal, and the opposite WACE T between early and late winter can be clearly distinguished. Especially for the type where the SPV changes from stretching to contraction, the high-top models effectively simulate the SAT pattern reversing from WACE to CAWE, while the low-top models cannot capture the phase reversal in late winter. The discrepancy in simulating the WACE T response between the high-top and low-top models may stem from the differences in their simulation of the stratospheric anomaly intensity in Dec16 Feb10. Both the high-top and low-top models are capable of simulating the weakened Ural high in early winter, which is associated with the downward influence of the stratospheric zonal asymmetry (Supplementary Figs. 18 and 19 ). The positive geopotential height anomalies over North America-North Atlantic and negative anomalies over Western Eurasia at 50 hPa in Nov1 Dec15 lead to the CAWE phase in early winter (Fig. 4c, d ). However, when the stratospheric anomaly undergoes a phase reversal, the high-top models simulate a stronger intensity of the geopotential height anomaly at 50 hPa over North America-North Atlantic in Dec16 Feb10 compared to that simulated by low-top models (Fig. 4c ), with an approximate enhancement of 18%. The high-top models can better capture the downward propagation intensity of stratospheric anomalies. The response of the geopotential height anomaly over the North America-North Atlantic region (defined as NANA Z500 ) is more significant (Supplementary Fig. 18d ), which is more conducive to the formation and propagation of Rossby waves in the troposphere, ultimately enhancing the Ural High and promoting a reversal toward the WACE phase in late winter (Fig. 4c and Supplementary Fig. 18f ). In contrast, the low-top models simulate weaker stratospheric signals and their corresponding downward propagation over North America-North Atlantic in Dec16 Feb10, and the response in the troposphere extends to the western European coast (Fig. 4d and Supplementary Fig. 19d ), which is unfavorable for the establishment of the Ural high in late winter and the formation of the WACE phase (Supplementary Fig. 19f ). The high top-level models, with comprehensive stratospheric representations, well simulate and verify the physical linkages through which the SPV morphology transition modulates the WACE T reversal between early and late winter through cross-continental process. Models with a higher top level are more conducive to simulating the stratosphere–troposphere coupling and its impacts on near-surface extreme climate, further providing key evidence that model design and construction should be extended to higher levels. Discussions In this study, we mainly reveal that the transition of SPV morphology over North America-North Atlantic plays a prominent role in modulating the phase reversal between WACE and CAWE (Fig. 5 ). The downward influence of stratospheric precursor and the dispersion of Rossby wave establish a bridge for SPV variations over North America-North Atlantic to impact the climate variability in Arctic-Eurasia. The high-top CMIP6 models effectively simulate these physical linkages and mechanisms, especially well reproducing the Ural high reversal in late winter compared to the low-top models. However, the verification using the CMIP6 and CESM-LE experiments is a qualitative simulation, and sensitivity experiments are needed to directly control the intensity and distribution of stratospheric anomalies. Such experiments may further clarify why stratospheric changes over North America-North Atlantic exhibit a stronger correlation with WACE T reversal. The triggering mechanism for such rapid NANA-SPV IPV phase reversal may be related to the changes in intensity of upward planetary waves entering the stratosphere and the eddy heat transport to the pole 40 . These changes can be traced back to tropospheric perturbations, such as blockings and the westerly jet 17 , 41 , as well as to the climate forcing factors, such as the reduction of Arctic sea ice 42 , the increase of Siberian snow cover 43 and the El Niño-southern oscillation 44 . Fig. 5. The timeline and physical process of the stratospheric polar vortex morphology transition over North America-North Atlantic modulating the phase reversal of “warm Arctic-cold Eurasia”. Open in a new tab The contracted stratospheric polar vortex (SPV) over North America-North Atlantic with zonal asymmetric anomalies in Nov1 Dec15 causes the suppressed Ural high in early winter through the vertical wave coupling and transmission of the Rossby wave, leading to the cold Arctic-warm Eurasia phase. While in Dec16 Feb10, the SPV stretches toward North America-North Atlantic and enhances across the Arctic, causing the sustained Ural blocking in late winter through downward-propagation and synoptic eddy feedback, resulting in the warm Arctic-cold Eurasia phase opposite to that in early winter. The red and blue arrows indicate abnormal anticyclones and cyclones. The dotted arrow indicates the evolution of stratospheric anomalies. The blue and yellow thick lines represent the geopotential height at 500 hPa. The clock indicates the period from November to March, and the red and gray shadows in the clock indicate the duration of the two stages of the SPV morphology and “warm Arctic-cold Eurasia” pattern. The SPV morphology transition leads the phase reversal of the “warm Arctic-cold Eurasia” pattern by about 25 days. Previous studies revealed that the synergistic effect of the preceding tropical-subtropical SST anomalies in the North Atlantic and the Indian Ocean can also cause the WACE T to reverse 6 . Based on a relatively simple and linear method of constructing a regression model (see “Methods”), the SPV variability and tropical-subtropical SST signals together explain 59% of the WACE T reversal intensity. These two factors are independent of each other, with the correlation coefficient at 0.26 (insignificant). The SPV and tropical-subtropical SST anomalies contribute 39% and 20%, respectively, indicating an effective driving force from the stratosphere. The revelation of stratospheric precursors is particularly important for the explanations of Arctic-Eurasia climate variability in late winter, compensating for the insufficiency of trigger factors and predictable sources within the troposphere 45 . The significant simulation differences between the high and low top-level models highlight the complexities of the potential mechanisms of stratosphere–troposphere coupling. Accurately simulating the interactions between the stratosphere and troposphere remains a considerable challenge 46 . The high-top models have high skill in simulating the downward influence of stratospheric precursor and its role in climate variability in the Arctic-Eurasia, which may offer an effective solution and future development direction for improving the simulation of the stratosphere–troposphere interaction. The pronounced variations in Arctic-Eurasia climate anomalies between early and late winter pose great challenges and obstacles for seasonal prediction 47 . The significant contribution of SPV variations to the WACE pattern provides a potential prediction implication for the climate variability in the Arctic and mid-high latitudes of Eurasia 48 . To truly predict the subseasonal to seasonal variations of the WACE pattern, it is necessary to comprehensively consider multiple signals, including the stratosphere, Arctic change and tropical SST. More importantly, the Arctic-Eurasia climate variability serves as a driver and predictable source for extreme climate and environmental disasters in the mid-low latitudes 49 , 50 . The persistent large-amplitude atmospheric circulations accompanied by the WACE pattern exacerbate the extremity of the climate anomalies 9 , 49 . The impacts of the stratospheric atmosphere on the Arctic-Eurasia climate variability can also have a chain reaction on the climate extremes in the mid-low latitudes. Explorations of changes in the stratosphere and their impacts enhance the understanding of the crucial role of stratosphere–troposphere coupling in extreme climate. Methods Data treatment and statistical methods In this study, we use the ERA5 reanalysis data for analysis 51 . The linear trend and its long-term (1979/80–2023/24) mean of all the daily data are removed, which indicates that the seasonal cycle is also removed. The method of detrending is to first calculate the slope and intercept of the least squares linear trend line of the original variables during 1979/80–2023/24, and obtain the linear trend term of year(i) as slope × year(i) + intercept, and then remove the linear trend term from the original variables to obtain the detrended results. In this study, the statistical methods, such as correlation coefficient, composite analysis, trend analysis, significance tests, and moving t-test are used 52 . The type of correlation coefficient used in this study is Pearson correlation, which measures the linear relationship between two random variables. The linear trends during the different sub-periods are calculated by the least squares method. A two-sided Student’s t test is used to test the statistical significance of the composite analysis. The significance of the correlation and the slope rates of the linear trends during the different sub-periods can be tested by using the Student’s t test. The 95% confidence level is denoted by p < 0.05. The moving t-test can examine whether the difference between the means of two sample groups is statistically significant, in order to detect the abrupt change point. In this study, a sample length of 45 days is adopted, which effectively captures the phase variation and shift on the subseasonal scale. The Liang–Kleeman (L–K) information flow analysis is applied to verify the causal relationship between the NANA-SPV IPV transition and the WACE T reversal. The causality is measured by the time rate of information flowing from one series to the other. In this study, a causal analysis is conducted on the 30-day time series of the NANA-SPV IPV phase transition process (December 1st to December 30th) and the WACE T reversal process (December 26th to January 25th). These 30 days represent the 15 days before and 15 days after the reversal point of NANA-SPV IPV and WACE T . According to theory, the maximum likelihood estimation form of the information flow from time series X 2 to time series X 1 under the assumption of the linear model is as follows: T 2 → 1 = C 11 C 12 C 2 , d 1 − C 12 2 C 1 , d 1 C 11 2 C 22 − C 11 C 12 2 1 where C i j is the sample covariance between X i and X j , and C i , d j is the covariance between X i and X j , n + 1 − X j , n / ∂ t , with ∂t representing the time interval. According to the L–K theory, causation implies correlation. However, the converse is that correlation does not imply causation. When T 2 → 1 > 0 , X 2 is a cause of X 1 ; and when T 2 → 1 = 0 , X 2 is not a cause of X 1 53 . Definitions of the WACE T phase reversal and NANA-SPV IPV transition The WACE T phase reversal needs to satisfy two conditions: (1) the WACE T in early winter and late winter is reversed after the removal of the trend and climate mean of 1979/80–2023/24; and (2) the sum of the standard deviation in early winter and late winter after the standardization of daily WACE T in December 1st to February 28th is <1.8. The second condition picks out the persistent WACE or CAWE stage in half-winter. The threshold of 1.8 is selected based on a summary of year-by-year analyses of the daily WACE T index that met the first condition, excluding years that primarily exhibited fluctuating variations. Minor adjustments to this threshold do not affect the feature of WACE T phase reversal. In this study, the early winter and late winter are divided by January 10th, meaning early winter spans from December 1st to January 10th, and late winter spans from January 11th to February 28th. To verify the reliability of the WACE T reversal years selected by this method, we perturb the division dates for a few days before and after January 10th and find the results consistent. Using the above similar methods to define the NANA-SPV IPV transition, two conditions also need to be met: (1) the NANA-SPV IPV in periods Nov1 Dec15 and Dec16 Feb10 is reversed after the removal of the trend and climate mean of 1979/80–2023/24; and (2) the sum of the standard deviation in periods Nov1 Dec15 and Dec16 Feb10 after the standardization of daily NANA-SPV IPV in November 1st to February 10th is <1.6. Similarly, the threshold of 1.6 is selected based on a summary of year-by-year analyses of the daily NANA-SPV IPV index that met the first condition, excluding years that primarily exhibited fluctuating variations. Minor adjustments to this threshold do not affect the conclusions of this study. In addition, when perturbing the division date (December 15th), the selected NANA-SPV IPV transition years remain consistent. By applying the empirical orthogonal function (EOF) analysis over the latitude-time (day) domain to the daily geopotential height anomaly at 50 hPa zonally averaged from 0° to 120°W 54 , the spatiotemporal evolution of the NANA-SPV IPV transition is further explained. The combined manifestation of EOF1 and EOF3 can effectively capture the variation of geopotential height anomaly associated with the NANA-SPV IPV transition (Supplementary Fig. 20 ). The presentation of the EOF pattern is robust, which has undergone sensitivity tests using the additional period 1950/51–1978/79. The specific manifestation is that the stratospheric anomaly over North America-North Atlantic reverses at around mid-December, and the anomaly is stronger during the Dec16 Feb10 period. The sum of these two mode’s time series in the NANA-SPV IPV reversal years is generally extreme (Supplementary Fig. 20g ), further indicating the importance of studying such events. Plumb wave activity flux We used the 3D Plumb wave activity flux to examine zonal, meridional, and vertical wave propagation of quasi-stationary waves 55 . ( F λ , F φ , F z ) = p cos φ v ′ 2 − 1 f a cos φ ∂ v ′ Φ ′ ∂ λ − u ′ v ′ + 1 f a cos φ ∂ u ′ Φ ′ ∂ λ f ∂ T ~ / ∂ z + κ T ~ / H v ′ T ′ − 1 f a cos φ ∂ T ′ Φ ′ ∂ λ 2 where λ is longitude, φ is latitude, z is height, and p is pressure. u is the zonal wind, v is the meridional wind, T is the temperature, and Φ is the geopotential height. f is the Coriolis parameter. a is Earth’s radius. κ is the specific gas constant of dry air divided by the specific heat of dry air. T̃ denotes the domain average of temperature. H is the log-pressure scale height. Primes denote the deviations from zonal means. Eddy vorticity forcing The stream function tendency induced by eddy-vorticity fluxes, namely eddy-vorticity forcing (EVF), can measure the dynamic synoptic eddy feedback onto the low-frequency flow 56 . The EVF can be expressed as ∂ ψ a ∂ t se = − Δ − 1 ∇ V ′ ζ ′ a 3 where V ′ ( u ′ , v ′ ) and ζ' denote the 2–8-day bandpass filtered zonal wind, meridional wind, and vorticity, respectively. [ ] a denotes the anomaly during the Dec16 Feb10 period. ∇ ⋅ ( ) and ∆ −1 ( ) are the horizontal divergence and the Laplacian inversion operators, respectively. ( ) se indicates the tendency induced by the synoptic eddy. The positive (negative) EVF can contribute to a positive (negative) tendency of the stream function, which expresses an anticyclonic (cyclonic) vorticity tendency. Calculation methods for the contribution of WACE T reversal intensity The intensity of the WACE T reversal is denoted as the WACE T difference between early and late winter (defined as Diff-WACE T ). The NANA-SPV IPV difference between Nov1 Dec15 and Dec16 Feb10 serves as the intensity of the stratospheric anomalies. The SST anomalies in the tropical Atlantic and Indian oceans have significant primary impacts on the WACE/CAWE pattern in early and late winter, respectively 6 . Their synergistic effect effectively modulates the phase reversal between the WACE and CAWE. Therefore, the sum of the standardized time series of these two SST factors is used to represent the tropical-subtropical SST anomaly signals. After standardizing the time series of tropical-subtropical SST signals and stratospheric signals, a linear regression is performed against Diff-WACE T . The contribution of stratospheric anomalies to the WACE T reversal intensity in a given year is calculated by multiplying the regression coefficient of the stratospheric anomalies by their standardized anomaly value in that year 57 . The average across all WACE T reversal years represents the contribution of stratospheric anomalies. The same method is applied to obtain the contribution of tropical-subtropical SST signals. CMIP6 and CESM-LE simulations To verify the strong linkage between NANA-SPV A transition and WACE T reversal, the CMIP6 58 historical simulations and the Community Earth System Model large ensemble (CESM-LE) 59 simulations are used to complement the statistical analyses. Daily SAT and geopotential height on pressure levels from the historical simulations of 21 available CMIP6 models from 1950 to 2014 and from 39 members of the Community Earth System Model from 1920 to 2005 are employed in this study. Similar to the process of observations, the linear trends during 1950–2014 in the CMIP6 simulations and during 1920–2005 in the CESM-LE simulations are removed to obtain the daily anomaly. Since the PV variable is not directly available in simulations, we use the 50 hPa geopotential height to identify the years with the SPV morphology transition 38 . In the WACE T reversal year, both the PV at 530 K and the geopotential height at 50 hPa over North America-North Atlantic undergo significant phase reversals (Supplementary Fig. 21 ). The daily variations of these two indices are basically the same, only with opposite signs. The reversal time points of these two indices are basically consistent. When the NANA-SPV IPV transits, the geopotential height anomaly also shows a phase reversal over North America-North Atlantic between Nov1 Dec15 and Dec16 Feb10, with a zonal asymmetric pattern in Nov1 Dec15 (Supplementary Fig. 2a ). Among the 18 years of observed NANA-SPV IPV transition, 17 years exhibit these characteristics. Therefore, using these definitions to identify the cases of SPV morphology transition over North America-North Atlantic in the simulation is relatively reasonable and reliable. In CMIP6 and CESM-LE simulations, the cases are selected as the SPV morphology transition over North America-North Atlantic from contraction to stretching based on these two conditions: (1) the area-weighted mean of geopotential height anomalies at 50 hPa over North America-North Atlantic transits from positive in Nov1 Dec15 to negative in Dec16 Feb10; (2) the geopotential height anomalies in Nov1 Dec15 present a zonal asymmetry over North America-North Atlantic and Eurasia. Conversely, the cases represent the SPV morphology transition over North America-North Atlantic from stretching to contraction. In addition, we explore the simulation capabilities of the CMIP6 models with different top height. The models are divided into high-top and low-top groups based on whether the model top level reaches 0.1 hPa. The high-top and low-top model groups consist of 9 and 12 models, respectively, with the ranges of top level being 4.5 × 10 −6 –0.1 and 0.2–10 hPa, respectively. Supplementary information Supplementary Information (4.6MB, pdf) Transparent Peer Review File (5.6MB, pdf) Acknowledgements Funding: National Natural Science Foundation of China 42394125 to Z.Y., 42505058 to Y.Z., and the Postdoctoral Fellowship Program of CPSF GZB20250077 to Y.Z. Author contributions Y.Z., Z.Y. conceived and designed the study; Y.Z., Z.Y., and S.H. performed the analyses; Y.Z., Z.Y., wrote the draft paper; W.T., S.H., and P.H. helped improve the paper. All authors discussed the results and contributed to writing the paper. Peer review Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available. Data availability Daily meteorological data on surface and pressure levels from ERA5, including surface air temperature, potential vorticity, geopotential height, zonal wind, meridional winds are available at https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels?tab=overview and https://cds.climate.copernicus.eu/datasets/reanalysis-era5-pressure-levels?tab=overview . Daily surface air temperature and geopotential height of the CMIP6 historical simulation are available at https://esgf-metagrid.cloud.dkrz.de/search . Daily surface air temperature and geopotential height at 50 hPa of the CESM-LE historical simulation are available at https://gdex.ucar.edu/datasets/d651027/dataaccess/# . The data underlying each figure of this study are available in the Zenodo repository 10.5281/zenodo.18623601. Code availability The computer codes for analyzing data and drawing plots are developed in NCAR Command Language (available at https://www.ncl.ucar.edu/ ). The computer codes used in this study are available in the Zenodo repository 10.5281/zenodo.18623601. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Yijia Zhang, Zhicong Yin. Supplementary information The online version contains Supplementary material available at 10.1038/s41467-026-70100-3. References 1. Cohen, J. et al. Recent Arctic amplification and extreme mid-latitude weather. Nat. Geosci. 7 , 627–637 (2014). [ Google Scholar ] 2. Mori, M. et al. A reconciled estimate of the influence of Arctic sea-ice loss on recent Eurasian cooling. Nat. Clim. Change 9 , 123–219 (2019). [ Google Scholar ] 3. Yin, M. et al. Amplified wintertime Arctic warming causes Eurasian cooling via nonlinear feedback of suppressed synoptic eddy activities. Sci. Adv. 11 , eadr6336 (2025). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Cohen, J. et al. Divergent consensuses on Arctic amplification influence on midlatitude severe winter weather. Nat. Clim. Change 10 , 20–29 (2020). [ Google Scholar ] 5. Gong, H. et al. Teleconnection from Arctic warming suppresses long-term warming in central Eurasia. Sci. Adv. 11 , eadq9461 (2025). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 6. Yin, Z. C. et al. Subseasonal variability and the “Arctic warming-Eurasia cooling” trend. Sci. Bull. 68 , 528–535 (2023). [ DOI ] [ PubMed ] [ Google Scholar ] 7. Blackport, R. & Screen, J. Weakened evidence for mid-latitude impacts of Arctic warming. Nat. Clim. Change 10 , 105–106 (2020). [ Google Scholar ] 8. Outten, S. D. & Esau, I. A link between Arctic sea ice and recent cooling trends over Eurasia. Clim. Change 110 , 1069–1075 (2012). [ Google Scholar ] 9. Yang, H. & Fan, K. The causes of intraseasonal alternating warm and cold variations over China in winter 2021/22. J. Clim. 37 , 5153–5170 (2024). [ Google Scholar ] 10. Yin, Z. C. et al. Why super sandstorm 2021 in North China? Natl. Sci. Rev. 9 , nwab165 (2022). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Kug, J. et al. Two distinct influences of Arctic warming on cold winters over North America and East Asia. Nat. Geosci. 8 , 759–762 (2015). [ Google Scholar ] 12. Mori, M. et al. Robust Arctic sea-ice influence to frequent Eurasian cold winters in the recent past. Nat. Geosci. 7 , 869–873 (2014). [ Google Scholar ] 13. Matsumura, S. & Kosaka, Y. Arctic–Eurasian climate linkage induced by tropical ocean variability. Nat. Commun. 10 , 3441 (2019). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Luo, B. et al. Decadal variability of winter warm Arctic-cold Eurasia dipole patterns modulated by Pacific decadal oscillation and Atlantic multidecadal oscillation. Earths Future 10 , e2021EF002351 (2022). [ Google Scholar ] 15. Mitchell, D. M. et al. The influence of stratospheric vortex displacements and splits on surface climate. J. Clim. 26 , 2668–2682 (2013). [ Google Scholar ] 16. He, S. et al. Eurasian cooling linked to the vertical distribution of Arctic warming. Geophy. Res. Lett. 47 , e2020GL087212 (2020). [ Google Scholar ] 17. Xu, X. et al. Atmospheric contributions to the reversal of surface temperature anomalies between early and late winter over Eurasia. Earths Future 10 , e2022EF002790 (2022). [ Google Scholar ] 18. Kidston, J. et al. Stratospheric influence on tropospheric jet streams, storm tracks and surface weather. Nat. Geosci. 8 , 433–440 (2015). [ Google Scholar ] 19. Shaw, T. A., Perlwitz, J. & Harnik, N. Downward wave coupling between the stratosphere and troposphere: the importance of meridional wave guiding and comparison with zonal-mean coupling. J. Clim. 23 , 6365–6381 (2010). [ Google Scholar ] 20. Baldwin, M. P. et al. Sudden stratospheric warmings. Rev. Geophys. 59 , e2020RG000708 (2021). [ Google Scholar ] 21. Ding, X. et al. Extreme stratospheric wave activity as harbingers of cold events over North America. Commun. Earth Environ. 4 , 187 (2023). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Davini, P., Cagnazzo, C. & Anstey, J. A. A blocking view of the stratosphere-troposphere coupling. J. Geophys. Res. Atmos. 119 , 11100–11115 (2014). [ Google Scholar ] 23. Chen, Q., Xu, L. & Cai, H. Impact of stratospheric sudden warming on East Asian winter monsoons. Adv. Meteorol. 2015 , 640912 (2015). [ Google Scholar ] 24. Kolstad, E. W., Breiteig, T. & Scaife, A. A. The association between stratospheric weak polar vortex events and cold air outbreaks in the Northern Hemisphere. Q. J. R. Meteorol. Soc. 136 , 886–893 (2010). [ Google Scholar ] 25. Kretschmer, M. et al. More-persistent weak stratospheric polar vortex states linked to cold extremes. Bull. Am. Meteorol. Soc. 99 , 49–60 (2018). [ Google Scholar ] 26. Zou, C. et al. Contrasting physical mechanisms linking stratospheric polar vortex stretching events to cold Eurasia between autumn and late winter. Clim. Dyn. 62 , 2399–2417 (2024). [ Google Scholar ] 27. Hu, D. et al. Longer duration of extreme cold events over East Asia affected by the stratospheric polar vortex extension. J. Geophys. Res. Atmos. 130 , e2024JD042299 (2025). [ Google Scholar ] 28. Domeisen, D. I. V. et al. The role of the stratosphere in subseasonal to seasonal prediction: 2. Predictability arising from stratosphere-troposphere coupling. J. Geophys. Res. Atmos. 125 , e2019JD030923 (2020). [ Google Scholar ] 29. Zhang, M., Yang, X. Y. & Huang, Y. Impacts of sudden stratospheric warming on extreme cold events in early 2021: an ensemble-based sensitivity analysis. Geophys. Res. Lett. 49 , e2021GL096840 (2022). [ Google Scholar ] 30. Shan, Q. & Fan, K. The transition of stratospheric polar vortex intensity: a case study of winter 1987/88. J. Geophys. Res. Atmos. 127 , e2022JD036511 (2022). [ Google Scholar ] 31. Zhang, C. et al. Impacts of stratospheric polar vortex shift on the East Asian trough. J. Climate 35 , 5605–5621 (2022). [ Google Scholar ] 32. Karami, H., Jacobi, C. & Kumar, A. The morphology of the stratospheric polar vortex under stratospheric aerosol intervention scenarios. Int. J. Climatol. 45 , e8838 (2025). [ Google Scholar ] 33. Qian, C. et al. Impacts of the Arctic stratospheric polar vortex weakening on Ural blocking in boreal winter. J. Geophys. Res. Atmos. 129 , e2024JD041362 (2024). [ Google Scholar ] 34. Song, L., Wu, R. & Jiao, Y. Relative contributions of synoptic and intraseasonal variations to strong cold events over eastern China. Clim. Dyn. 50 , 4619–4634 (2018). [ Google Scholar ] 35. Domeisen, D. I. V., Sun, L. & Chen, G. The role of synoptic eddies in the tropospheric response to stratospheric variability. Geophys. Res. Lett. 40 , 4933–4937 (2013). [ Google Scholar ] 36. Kurihara, K. & Kawahara, M. Extremes of East Asian weather during the post ENSO years of 1983/84. J. Meteorol. Soc. Japan 64 , 493–503 (1986). [ Google Scholar ] 37. Kretschmer, M. et al. The different stratospheric influence on cold-extremes in Eurasia and North America. NPJ Clim. Atmos. Sci. 1 , 44 (2018). [ Google Scholar ] 38. Rao, J. et al. Mean state of the Northern Hemisphere stratospheric polar vortex in three generations of CMIP models. J. Clim. 35 , 4603–4625 (2022). [ Google Scholar ] 39. Domeisen, D. I. V. et al. The role of the stratosphere in subseasonal to seasonal prediction: 1. Predictability of the stratosphere. J. Geophys. Res. Atmos. 125 , e2019JD030920 (2019). [ Google Scholar ] 40. Polvani, L. M. & Waugh, D. W. Upward wave activity flux as a precursor to extreme stratospheric events and subsequent anomalous surface weather regimes. J. Clim. 17 , 3548–3554 (2004). [ Google Scholar ] 41. Castanheira, J. M. & Barriopedro, D. Dynamical connection between tropospheric blockings and stratospheric polar vortex. Geophys. Res. Lett. 37 , L13809 (2010). [ Google Scholar ] 42. Kim, B.-M. et al. Weakening of the stratospheric polar vortex by Arctic sea-ice loss. Nat. Commun. 5 , 4646 (2014). [ DOI ] [ PubMed ] [ Google Scholar ] 43. Cohen, J. et al. Linking Siberian snow cover to precursors of stratospheric variability. J. Clim. 27 , 5422–5432 (2014). [ Google Scholar ] 44. Domeisen, D. I. V., Garfinkel, C. I. & Butler, A. H. The teleconnection of El Niño Southern Oscillation to the stratosphere. Rev. Geophys. 57 , 5–47 (2019). [ Google Scholar ] 45. Fan, K., Yang, H. & Dai, H. Research progress on inter-monthly winter temperature variation in East Asia and climate prediction. Atmos. Ocean. Sci. Lett. 16 , 100372 (2023). [ Google Scholar ] 46. Tian, W. et al. Role of stratospheric processes in climate change: advances and challenges. Adv. Atmos. Sci. 40 , 1379–1400 (2023). [ Google Scholar ] 47. Meng, J. et al. Arctic weather variability and connectivity. Nat. Commun. 14 , 6574 (2023). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 48. Huang, J. et al. Stratospheric influence on the development of the 2018 late winter European cold air outbreak. J. Geophys. Res. Atmos. 127 , e2021JD035877 (2022). [ Google Scholar ] 49. Xu, T. et al. Hybrid seasonal prediction of meridional temperature gradient associated with “warm Arctic-cold Eurasia. Adv. Atmos. Sci. 40 , 1649–1661 (2023). [ Google Scholar ] 50. Yin, Z. et al. Traditional Meiyu-Baiu has been suspended by global warming. Natl. Sci. Rev. 11 , nwae166 (2024). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 51. Hersbach, H. et al. ERA5 hourly data on single levels from 1940 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS). 10.24381/cds.adbb2d47 (2023). 52. Wilks, D. S. Statistical Methods in the Atmospheric Sciences, 3rd edn (Academic Press, 2011). 53. Liang, X. S. Unraveling the cause-effect relation between time series. Phys. Rev. 90 , 052150 (2014). [ DOI ] [ PubMed ] [ Google Scholar ] 54. Ma, J. et al. Skillful seasonal predictions of continental East-Asian summer rainfall by integrating its spatio-temporal evolution. Nat. Commun. 16 , 273 (2025). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 55. Plumb, R. A. On the three-dimensional propagation of stationary waves. J. Atmos. Sci. 42 , 217–229 (1985). [ Google Scholar ] 56. Ren, H. et al. Dynamic synoptic eddy feedbacks contributing to maintenance and propagation of intraseasonal NAO. Geophys. Res. Lett. 49 , e2021GL096508 (2022). [ Google Scholar ] 57. Zhang, R. et al. A stratospheric precursor of East Asian summer droughts and floods. Nat. Commun. 15 , 247 (2024). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. Eyring, V. et al. Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev. 9 , 1937–1958 (2016). [ Google Scholar ] 59. Kay, J. E. et al. The Community Earth System Model (CESM) large ensemble project: a community resource for studying climate change in the presence of internal climate variability. Bull. Am. Meteorol. Soc. 96 , 1333–1349 (2015). [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary Information (4.6MB, pdf) Transparent Peer Review File (5.6MB, pdf) Data Availability Statement Daily meteorological data on surface and pressure levels from ERA5, including surface air temperature, potential vorticity, geopotential height, zonal wind, meridional winds are available at https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels?tab=overview and https://cds.climate.copernicus.eu/datasets/reanalysis-era5-pressure-levels?tab=overview . Daily surface air temperature and geopotential height of the CMIP6 historical simulation are available at https://esgf-metagrid.cloud.dkrz.de/search . Daily surface air temperature and geopotential height at 50 hPa of the CESM-LE historical simulation are available at https://gdex.ucar.edu/datasets/d651027/dataaccess/# . The data underlying each figure of this study are available in the Zenodo repository 10.5281/zenodo.18623601. The computer codes for analyzing data and drawing plots are developed in NCAR Command Language (available at https://www.ncl.ucar.edu/ ). The computer codes used in this study are available in the Zenodo repository 10.5281/zenodo.18623601. Articles from Nature Communications are provided here courtesy of Nature Publishing Group ACTIONS View on publisher site PDF (2.0 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top