Thermal coupling of the Indo-Pacific warm pool and Southern Ocean over the past 30,000 years

The study addresses how the Indo-Pacific warm pool (IPWP) and its coupling to the Southern Ocean influenced climate during the last deglaciation. Traditional views link deglacial warming primarily to Northern Hemisphere insolation and ice-sheet retreat, but growing evidence suggests Southern Hemisphere sea surface temperature (SST) and IPWP changes preceded Northern Hemisphere and global ice volume variations, implying a southern driver of global climate change through orbital forcing. The tropical Pacific–Southern Ocean teleconnections, mediated by ocean–atmosphere processes and ENSO, exert strong control on global climate. However, the magnitude and pattern of glacial-to-deglacial SST changes in the IPWP and eastern equatorial Pacific (EEP) are heterogeneous, leading to conflicting reconstructions of ENSO-like states during the last glacial period. This study aims to reconstruct spatiotemporal SST and subsurface temperature (subT) evolution, define WPWP extent changes, quantify zonal gradients indicative of ENSO-like variability, and investigate the thermal linkage between the IPWP and Southern Ocean over the past 30 kyr.

Previous work shows divergent LGM SST cooling magnitudes: 2–5 °C in the IPWP based on Mg/Ca and alkenone proxies, and 0.3–3.5 °C in the EEP cold tongue. Some ENSO reconstructions used sites distant from the WPWP core or nearshore locations influenced by upwelling, freshwater runoff, and coastal currents, or relied on meridional gradients debated as ENSO indicators versus seasonality. Consequently, studies reached opposing conclusions, inferring either dominant La Niña-like or El Niño-like conditions during the last glacial. Modeling and data indicate Southern Ocean processes modulate tropical Pacific variability via atmospheric teleconnections and oceanic pathways (subduction of Subtropical Mode Water, Subantarctic Mode Water, and Antarctic Intermediate Water) that ventilate the equatorial thermocline, a key control on ENSO. The Southern Ocean has dominated historical ocean heat uptake, with preferential heat storage where surface waters are subducted. Precessional forcing is known to modulate ENSO-like variability by altering stratification, upwelling, and thermocline tilt, suggesting potential orbital pacing of tropical variability.

• Compiled 340 published SST records (planktonic foraminiferal Mg/Ca and Uk'37 alkenone indices) and 7 published subsurface temperature records from the tropical and subtropical Pacific and eastern Indian Ocean spanning the last 30 kyr. Defined regional groups (e.g., ECS, SCS, EIO, WWP/EWP, CEP/EEP, SWP, SCP, SEP) using cluster analysis (per Supplementary information) and mapped against modern annual mean SST (World Ocean Atlas 2018) to contextualize the WPWP and EEP cold tongue.
• Generated a new western Pacific warm pool (WPWP) ocean heat content (OHC, 0–300 m) record from core KX22-4 using Mg/Ca from five planktonic foraminiferal species. Assessed mixed-layer structure and barrier layer signals using species-specific Mg/Ca–SST and residual seawater δ18O (δ18Osw-iv) for Globigerinoides ruber and Trilobatus sacculifer.
• Quantified ENSO-like changes via zonal annual SST anomaly (SSTA) stacks and the annual zonal SST difference (ASST) between the central WPWP and EEP. SSTA stacks computed with a sliding rectangular window of 1 kyr using STATNARY 1.2. Compared reconstructed ASST to modern annual mean ASST from WOA2018.
• Reconstructed spatiotemporal WPWP extent (top 1 °C) and boundary shifts between the last glacial and early Holocene using Mg/Ca-derived SST distributions (with comparisons to mixed proxies). Assessed latitudinal extent, west boundary position, and meridional displacement of WPWP boundaries.
• Evaluated zonal and meridional heat transfer proxies via SST gradients among key regions: EWP–ECS, EWP–EIO, Equatorial Pacific–South Pacific. Compared with planetary radiative imbalance, orbital precession, and seasonal insolation at 30°S.
• Used the transient TRACE simulation to evaluate differential glacial cooling between regions, timing of WPWP warming, WPWP area expansion, and deglacial temperature anomalies and heat pathways (including vertical sections illustrating a hypothesized oceanic tunnel connecting Southern Ocean subduction zones to the equatorial Pacific).

• ENSO-like state and zonal gradients: • The glacial zonal annual SST difference (ASST) between WPWP and EEP was 0.67 ± 0.24 °C smaller than the modern annual mean, indicating a more El Niño-like mean state during the last glacial period. • TRACE simulation indicates 0.82 ± 0.27 °C less glacial cooling in the EEP than in the WPWP, consistent with a deeper EEP thermocline and reduced Walker circulation during the LGM.
• WPWP extent and magnitude of change: • WPWP experienced 2.41 ± 0.63 °C cooling during the glacial, migrated slightly equatorward, and its latitudinal extent was reduced by about 4° relative to the Holocene; the west boundary contracted eastward to ~150°E in the glacial. • In the early Holocene, WPWP meridional extent approximately spanned 19°N to 14°S with the west boundary near ~120°E, aligning with the modern 28 °C isotherm footprint. • TRACE indicates the WPWP first reached 28 °C at ~14 ka (Bølling–Allerød) and remained ≥28 °C after ~12 ka (post-Younger Dryas), coincident with precession minimum.
• Deglacial state: La Niña-like conditions strengthened as the WPWP warmed more intensely than the EEP, increasing the zonal gradient. WPWP area expanded substantially around ~12 ka when La Niña-like conditions and OHC peaked.
• Barrier layer mechanism: • Species-specific analyses from core KX22-4 show reduced SST contrast but markedly increased δ18Osw-iv contrast between G. ruber and T. sacculifer during deglaciation, implying T. sacculifer resided in saltier waters within a barrier layer. This supports enhanced surface heat retention, favoring La Niña-like conditions.
• Southern Ocean–tropical coupling and timing: • Subsurface temperature and OHC in the WPWP rose earlier than SST, with WPWP subT warming beginning by ~22 ka, in phase with South Pacific subT changes and orbital precession. • Earlier subT warming also occurred in the EEP cold tongue (fed by the Equatorial Undercurrent), and South Pacific regions, whereas East China Sea and eastern Indian Ocean showed delayed subT warming and increased vertical gradients. • Deglacial temperature reconstructions and Δ14C evidence support a Southern Ocean source for waters ventilating the equatorial Pacific thermocline. TRACE indicates WPWP upper waters absorbed heat from the south/lower waters and exported heat northward; the equator–South Pacific SST gradient decreased.
• Orbital pacing: • Decreases in the precession parameter during deglaciation correlate with rises in WPWP subT, OHC, and ENSO-like variability. Warmer austral winter insolation enhanced formation temperatures of overturning waters, subsequently transmitted to the western tropical Pacific.
• Carbon cycle implications: • Under La Niña-like conditions, enhanced Walker circulation and westward warm water buildup steepened thermocline tilt, shoaled the EEP thermocline, and facilitated CO2 outgassing from the EEP, contributing to deglacial atmospheric CO2 rise.
• Regional correlations during deglaciation (SST vs vertical ΔT): ECS R=0.91, EIO R=0.79, EEP R=0.38, EWP R=0.62, SWP R=0.77, SCP R=0.67; vertical gradient fits showed strong statistical significance (P < 0.0001 for most regions).

The findings resolve aspects of the debate on tropical–extratropical linkages by demonstrating that enhanced deglacial warming in the WPWP preceded or coincided with increases in subsurface temperature and OHC, implicating an oceanic pathway from the Southern Ocean to the equatorial Pacific thermocline. This supports a mechanism in which Southern Ocean heat uptake and subduction (SAMW/AAIW) ventilate the tropical thermocline, preconditioning the tropical Pacific for a La Niña-like state as deglaciation proceeds. The barrier layer within the WPWP further amplified surface heat retention, restricting entrainment of cooler thermocline waters and reinforcing zonal gradients and Walker circulation. Orbital precession modulated these processes by altering austral winter insolation and the properties of subducted waters, pacing ENSO-like variability. The enhanced zonal and meridional gradients during deglaciation invigorated heat transport to higher latitudes and promoted EEP CO2 outgassing via thermocline shoaling, linking ocean heat redistribution to the carbon cycle and global temperature rise. Collectively, the study highlights the central role of WPWP–Southern Ocean thermal coupling in orchestrating deglacial climate evolution and provides a framework to interpret future warming pathways under continued Southern Ocean heat uptake.

By integrating 340 SST and 7 subsurface temperature records with a new WPWP OHC reconstruction, the study shows that the last glacial mean state was more El Niño-like (reduced zonal ASST), whereas deglaciation featured intensified WPWP warming, expansion, and a shift to La Niña-like conditions. Subsurface warming and OHC increases in the WPWP preceded SST rise and were synchronized with South Pacific subT and precession changes, evidencing an oceanic tunnel linking the Southern Ocean to the tropical Pacific. A strengthened barrier layer in the WPWP facilitated surface heat buildup, enhanced zonal and meridional temperature gradients, invigorated heat export, and contributed to EEP CO2 outgassing via thermocline shoaling. These results underscore the importance of thermal coupling between the IPWP and Southern Ocean in deglacial climate dynamics and its relevance for anticipating future global warming. Future work should expand spatial coverage and temporal resolution of subsurface and OHC reconstructions, further quantify barrier layer variability, and employ coupled climate models to isolate the relative roles of precession forcing, Southern Ocean heat uptake, and internal variability on ENSO-like transitions.

• The synthesis relies on heterogeneous proxy datasets (Mg/Ca and Uk'37) with differing seasonality and calibration uncertainties; although the WPWP region shows minimal seasonal bias, higher latitude areas exhibit greater divergence, particularly where Uk'37 data are more prevalent (discussed in Supplementary material).
• Spatial coverage and site selection, especially near coasts or upwelling zones, can introduce local influences; the study mitigates this via clustering but residual heterogeneity remains.
• Subsurface constraints are limited to seven published subT records and one new OHC core, potentially undersampling regional variability.
• Some inferences depend on TRACE model simulations and reanalysis products, which carry model structural uncertainties and boundary condition assumptions.