Arctic sea ice loss warmed the temperate East Asian winter in the mid-Holocene

The study addresses the debated relationship between rapid Arctic sea ice decline and recent occurrences of colder winters in midlatitude Eurasia, known as the warm Arctic–cold continents pattern. Observational analyses and climate simulations have linked Eurasian winter cooling to Arctic sea-ice loss via planetary waves, storms, and jet stream changes, but the link is uncertain due to limited records, internal variability, intermittency of teleconnections, influences from lower-latitude climate change, anthropogenic greenhouse forcing, and lagged ocean–atmosphere state effects. To overcome these limitations, the authors examine the mid-Holocene (8.2–4.2 ka), a warm period with stronger summer insolation and prominent Arctic amplification but minimal human influence, as a natural experiment. They propose that reconstructing Arctic sea ice changes and the East Asian winter monsoon (EAWM)—a key indicator of East and Central Asian winter temperature linked to the Siberian High and East Asian Trough—on millennial timescales can clarify the high–midlatitude winter climate connection. The focus is on the eastern Arctic, especially the East Siberian Sea where seasonal sea ice has decreased strongly in autumn and influences basin-scale sea ice, and on the temperate East China Sea, which lies along the core pathway of the EAWM.

Prior work identified the warm Arctic–cold continents pattern over recent decades and offered mechanisms linking Arctic sea ice loss to midlatitude winter cooling via atmospheric circulation changes (planetary waves, storm tracks, jet stream shifts). However, divergent conclusions exist due to internal variability and observational limits. Proxy reconstructions indicate the mid-Holocene was warmer globally, with pronounced Arctic amplification and likely intensified sea-ice decline. Theory and models suggest that reduced equator-to-pole temperature gradients in warm periods increase climate stability and reduce variance, potentially limiting internal variability impacts. The EAWM strength correlates positively with the Siberian High and East Asian Trough and inversely with Central Asian winter temperature, making it a proxy for winter climate. Previous transient CESM and loess-based studies argued that since the mid-Holocene the EAWM intensified with decreasing annual mean temperatures and increasing Arctic sea ice; other modeling suggested that Arctic summer heat absorbed due to sea ice loss can persist into winter and compensate for weaker winter insolation, potentially warming as far south as ~50°N. Contrasting findings also exist wherein Arctic sea-ice loss weakens the stratospheric polar vortex and cools midlatitudes, highlighting the complexity and nonlinearity of stratosphere–troposphere coupling and regional dependence on where sea ice is lost (e.g., Barents–Kara vs. Chukchi–Bering).

- Field sampling and cores: Two sediment cores were collected. Core LV77-36-1 (155.66°E, 74.10°N; 36 m water depth; 376 cm; mud) from the East Siberian Sea shelf was retrieved in September 2016 (R/V Akademik Lavrentiev). Core ECMZ (122.17°E, 28.5°N; 40 m water depth; upper mud length 1546 cm; silt and clay) from the East China Sea shelf was collected in September 2015 (vessel Kan 407). Event-forced sand layers at 1450–1350 cm in ECMZ were excluded from climate interpretation. - Chronologies: Age models were built using AMS 14C dates on 12 bivalve shells and 10 mixed benthic foraminifera (63–250 µm). Calibrations used Marine20 via Calib 8.2 with local reservoir corrections (ΔR: −121 ± 76 yr for East China Sea; −95 ± 61 yr for East Siberian/Laptev Seas). Interpolated ages and 95% CIs were computed with Clam 2.2 in R. Surface sediments were dated using 210Pb and 137Cs, and 0 cm was set to collection year. Records span ~7.5 ka to present. - IRD proxy for Arctic sea ice: From LV77-36-1, 77 samples at ~100-yr spacing were processed. Samples were dried (48 h, 60°C), wet-sieved at 63 µm, then dry-sieved at 125 and 250 µm. Terrigenous individual minerals and rock debris in 125–250 µm were counted as IRD under a ZEISS Stemi 2000 stereoscope; authigenic minerals (e.g., pyrite, gypsum) were excluded. >250 µm grains were excluded to minimize iceberg influence. IRD counts were normalized by bulk dry weight (counts g−1). - Zr/Rb ratio proxy for EAWM: From ECMZ, 76 bulk samples at ~100-yr spacing were digested (HNO3+HF) and analyzed by iCAP RQ ICP-MS. Replicate analytical errors were <5%. Zr/Rb reflects hydrodynamic sorting during long-distance winter transport by the EAWM-driven alongshore current; higher Zr/Rb indicates stronger current/EAWM and lower ratios indicate weaker EAWM. Regional tidal changes were minimal from mid- to late Holocene, and trends differ from precipitation, sediment supply, river mouth grain size, and typhoon activity, supporting the current-strength interpretation. - Synthesis of external proxies: IRD results were compared with dinocyst/biomarker records (Laptev, Kara, Barents, Chukchi/Fram) and IP25/quartz indicators to infer eastern Arctic seasonal sea-ice evolution; land-based loess grain size and dust MAR records and biomarker proxies in East Asian marginal seas were compiled to reconstruct EAWM and winter temperature changes. - Climate model simulations: CESM1.0 (CAM5, CLM4, POP2, CICE4) sensitivity experiments were performed to isolate the effect of eastern Arctic sea-ice decline via surface albedo perturbations. A 2000-year preindustrial control (CTRL) and a 500-year perturbed run (0.8A) were conducted; in 0.8A the ocean-ice surface albedo was set to 0.8× its original value globally over ice-covered ocean to mimic lower sea-ice concentration/extent mainly in the eastern Arctic. CO2 and orbital parameters were fixed at preindustrial values. Analyses used the last 200 years of 0.8A after quasi-equilibrium. Student’s t tests assessed significance (>95% confidence). CESM outputs were compared to proxy-based sea-ice and pollen-based eastern Arctic winter temperature anomalies for consistency. Additional CMIP6-PMIP4 context and a CESM transient simulation were referenced to support interpretations.

- Proxy reconstructions indicate a synchronous, enhanced decline in seasonal sea ice across the eastern Arctic during the mid-Holocene relative to the late Holocene. In the East Siberian Sea, IRD (125–250 µm terrigenous grains g−1) increased since the mid-Holocene, consistent with increased East Siberia-sourced ice-rafted Fe grains at Chukchi/Fram and with dinocyst and biomarker evidence for stronger spring–summer sea-ice melt and episodic summer sea-ice-free conditions. - EAWM weakened in the mid-Holocene relative to the late Holocene. In the East China Sea, lower Zr/Rb ratios indicate weaker winter alongshore current and thus weaker EAWM in the mid-Holocene. Organic biomarkers in southern ECS corroborate reduced terrigenous biomass transport in the mid-Holocene versus enhanced transport in the late Holocene. Continental loess grain size and dust MAR records show reduced EAWM strength and variability in the mid-Holocene, consistent with a weaker Siberian High and warmer Central Asian winters. - Despite weaker winter insolation at midlatitudes during the mid-Holocene, proxies indicate relatively warmer winter conditions in East and Central Asia occurring concurrently with enhanced Arctic sea-ice decline, implying a linkage between sea-ice loss and East Asian winter warming on millennial scales. - CESM sensitivity experiments show that reduced sea-ice concentration (via lower ocean-ice albedo) increases summer shortwave absorption, stores heat in the upper ocean/shelf waters, and releases it in autumn–winter through enhanced vertical mixing and upward sensible heat flux due to thinner/less extensive ice. This raises Arctic winter surface air temperature. - The wintertime heat release suppresses meridional heat transport from low to high latitudes, trapping more heat at lower latitudes. A larger low–midlatitude temperature gradient strengthens the westerlies, damps quasi-stationary planetary waves, weakens the East Asian Trough and the Siberian High, and induces southerly anomalies over northern East Asia and eastern Siberia. Consequently, the EAWM weakens and fewer Arctic cold surges reach East Asia, yielding warmer winters. The combined proxy–model evidence supports a warm Arctic–warm continents pattern in the mid-Holocene.

The research demonstrates that under warmer-than-present boundary conditions of the mid-Holocene, enhanced eastern Arctic sea-ice loss transferred summer-absorbed heat to the winter atmosphere, modifying large-scale atmospheric circulation to yield weaker EAWM and warmer winters in East and Central Asia. This finding directly addresses the question of how Arctic sea-ice decline influences midlatitude winter climate on long timescales, showing a warming influence via seasonal heat storage/release and wave–mean flow interactions. The mechanism—strengthened westerlies and damped planetary wave activity leading to a weaker East Asian Trough and Siberian High—explains reduced southward cold-air intrusions. The results reconcile the Holocene proxy evidence with coupled model dynamics and contrast with some short-term observational inferences of cold Eurasian winters tied to sea-ice loss, highlighting the role of timescale and background climate state. While other studies emphasize stratospheric pathways that could cool midlatitudes, the coupled ocean–ice–atmosphere mechanism presented here yields a net warming response in East Asia during the mid-Holocene and underscores regional dependence and seasonality of sea-ice forcing.

By combining new marine proxies (IRD from the East Siberian Sea and Zr/Rb from the East China Sea) with a synthesis of Holocene sea-ice and EAWM records and CESM sensitivity simulations, the study shows that enhanced eastern Arctic sea-ice decline in the mid-Holocene coincided with weaker EAWM and warmer winters in East and Central Asia. The mechanism involves increased summer heat uptake due to reduced sea ice, seasonal heat storage in the upper ocean, and enhanced winter heat release, which suppresses meridional heat transport, strengthens westerlies, damps planetary waves, and weakens the East Asian Trough and Siberian High. These findings imply that intermittent cold winters observed recently do not necessarily offset the longer-term trend toward winter warming in East Asia as Arctic sea ice declines. Future research should further resolve regional contrasts in Arctic sea-ice forcing (e.g., Barents–Kara vs. Chukchi–Bering), improve constraints on winter temperature reconstructions beyond monsoon proxies, and explore the roles of stratosphere–troposphere coupling and varying background boundary conditions.

- Regional heterogeneity: Sea-ice loss in different Arctic basins can produce contrasting Northern Hemisphere circulation responses, limiting inference from single-core records; the study mitigates this by focusing on overall eastern Arctic sea-ice loss but residual uncertainty remains. - Proxy limitations: Direct midlatitude winter surface air temperature reconstructions are scarce; EAWM is used as a proxy for winter temperature. Early Holocene records are not used due to reworking from rapid sea-level rise. - Modeling simplifications: CESM sensitivity runs impose uniform ocean-ice albedo reductions and fix CO2 and orbital parameters at preindustrial levels, not the exact mid-Holocene forcing. Ocean circulation was not drastically altered by design; stratospheric processes are complex and may contribute additional pathways not fully resolved. - Chronological and analytical uncertainties: Radiocarbon reservoir corrections, age–depth model interpolation, and geochemical measurement errors introduce uncertainties, though efforts were made to quantify them.