The greater role of Southern Ocean warming compared to Arctic Ocean warming in shifting future tropical rainfall patterns

The greater role of Southern Ocean warming compared to Arctic Ocean warming in shifting future tropical rainfall patterns | Nature Communications

Abstract

The recent rapid decline in Antarctic sea ice highlights the need to understand whether rising Southern Ocean temperatures have an influence on global climate. While Arctic warming has been extensively studied, the importance of Southern Ocean warming is emerging only now. Here, using multi-model simulations, we show that over 1.5 °C of Southern Ocean surface warming can offset the projected northward shift in tropical zonal-mean precipitation by the mid-21st century, driven by stronger northern high-latitude warming under rising greenhouse gas concentrations. SST nudging experiments suggest that a 1.0 °C warming in the Southern Ocean could impact tropical precipitation as significantly as a 1.5 °C warming in the Arctic. Regionally, Southern Ocean warming increases rainfall in northeastern Brazil while heightening drought risks in the Sahel. These effects are comparable to, or slightly more pronounced than, those caused by a weakening Atlantic Meridional Overturning Circulation and the associated development of the North Atlantic warming hole. Thus, Southern Ocean warming may play a more crucial role than Arctic warming in shaping tropical climate patterns in the coming decades.

Similar content being viewed by others

Introduction

Polar amplification, the phenomenon where polar regions experience more rapid surface warming compared to the global average, has been evident in the Arctic. Since satellite observations began in 1978-79, the summer Arctic sea ice extent (SIE) has decreased by approximately 50%, with a declining rate of about 13% per decade in September 1. Arctic surface temperatures have warmed 2–4 times more than the global average 2, 3, 4, 5. In contrast, the Southern Ocean has warmed more slowly, and Antarctic SIE showed a slight increase over the past 40 years until around 2014 (refs. 6, 7, 8), although this was followed by rapid declines in 2015-16 and 2023-24 (refs. 9, 10). This observational trend in Antarctic SIE has been a highly controversial issue, particularly because no single coupled climate model has successfully reproduced it. While this trend could fall within natural variability (e.g., ref. 11), several studies suggest it highlights deficiencies in climate models’ ability to simulate interactions between the Southern Ocean and Antarctic sea ice 12, 13.

The pronounced warming difference between the Arctic and Antarctic has been cited as evidence that future polar amplification is expected to be more pronounced in the Arctic 14, 15, 16, 17. The net equatorward transport of oceanic heat over the southern high latitudes is likely to mitigate Southern Ocean warming and slow the loss of Antarctic sea ice 18, 19. Nevertheless, Antarctic SIE has been decreasing since 2016, reaching an exceptionally low levels during June-July-August (JJA) of 2023, unprecedented in the satellite record 9, 10. The recent decline in Antarctic SIE aligns with multi-model projections of a decline in Antarctic sea ice in the coming decades due to increasing greenhouse gas concentrations 13.

The relative warming of the Southern Ocean compared to the Arctic Ocean could have far-reaching global impacts, particularly on regional tropical precipitation patterns. Climate models emphasize the role of hemispheric temperature gradients in shaping future ITCZ positioning 20, 21, highlighting the need to investigate the individual contributions of Arctic and Southern Ocean warming. While the influence of rapid Arctic Ocean warming on tropical precipitation has been extensively explored 22, 23, 24, the impact of Southern Ocean warming is only now gaining attention 25.

Recent modeling studies indicate that the loss of Antarctic sea ice may contribute more to the rise in tropical surface temperatures than the loss of Arctic sea ice 26, 27. This may be partly explained by the sensitivity of the tropical southeastern Pacific SSTs to Southern Ocean surface warming, particularly through cloud-radiative feedback 28, 29. A CMIP6 multi-model statistical analysis suggests that future Antarctic sea ice loss could account for approximately 45% of the variation in future cross-equatorial SST gradient changes by the late 21st century, when normalized by the global warming effect 30. These recent findings challenge the conventional belief that projected reductions in aerosol emissions 31 and continuing Arctic warming will cause a northward shift in the Intertropical Convergence Zone (ITCZ) in the future 32, 33, 34. Despite recent advances in understanding how Southern Ocean warming – particularly due to the decline in Antarctic sea ice – affects tropical climates, the relative significance of Southern Ocean versus Arctic warming remains uncertain.

In this study, we utilize CMIP6 multi-model analysis in conjunction with SST nudging simulations to investigate the differential impacts of future surface warming in the Arctic and Southern Ocean on tropical climate projections by the mid-21st century. The focus on the mid-21st century (2046–2065) aims to provide practical insights into near-term climate changes. Our findings suggest that Southern Ocean warming may exert a more substantial influence on remote tropical climate systems, particularly in altering precipitation patterns and atmospheric circulation in the coming decades.

Results

Inter-model diversity in future Arctic and Southern Ocean warming

Figure 1 a and b show historical and future projections of Arctic and Southern Ocean surface temperatures under the Shared Socioeconomic Pathways 2-4.5 scenario (SSP2-4.5). This study focuses on the SSP2-4.5 medium-emission scenario, representing a ‘middle-of-the-road’ pathway 35. Comparable analyses with the SSP5-8.5 scenario are provided in Supplementary Fig. 1. In this context, the ‘Arctic Ocean’ refers to the pan-Arctic Ocean, defined as latitudes north of 60°N. Future changes in Arctic Ocean SST, denoted as ΔArctic-SST, exhibit a continuing upward trend from 1980 through the late 21st century (Fig. 1a). From 1980 to 2100, the multi-model mean Arctic SST is projected to increase approximately 2 °C, with inter-model ΔArctic-SST ranging from 0.4 °C (CESM2) to 5 °C (CMCC-CM2-SR5) under SSP2-4.5. The suppressed pan-Arctic warming of CESM2 is attributed to significant Atlantic meridional overturning circulation (AMOC) weakening 36. The observed Arctic SST anomaly (red line in Fig. 1a) closely aligns with the multi-model mean (thick black line in Fig.

Timeseries of SSTs (°C) in the (a) Arctic Ocean (north of 60°N), (b) Southern Ocean (south of 50°S), and (c) cross-equatorial SST gradient (°C) between the northern tropics (0-25°N) and the southern tropics (0-25°S). (d) Same as (c), but for cross-equatorial rainfall (mm day-1) gradient. The dashed gray lines depict the 42 individual CMIP6 models, the solid blue lines indicate the ±95% interval from CMIP6, the solid black line represents the multi-model mean from CMIP6, and the solid red and green lines indicate observations ERSST v5 and GPCP, respectively. All data are presented as deviations from 1980 and are smoothed using a 5-year running average. Composites of SST changes (°C) in (e) warm Arctic Ocean models, (f) warm Southern Ocean models, and (g) zonal mean of SST changes (°C) from warm Arctic (solid red lines) and warm Southern Ocean (solid blue lines) models. (h, i, j) Same as (e, f, g), but for rainfall changes (mm day-1). Composite patterns are calculated by subtracting the present-day climate (1995-2014 average) under the historical simulation from the future climate (2046-2065 average) under the SSP2-4.5 scenario. For the spatial composite fields, statistically significant values (p < 0.05) are indicated with stippling.

Compared to Arctic sea surface warming, future Southern Ocean warming (ΔSouthern-SST) is relatively moderate (Fig. 1b). The multi-model mean Southern Ocean SST (south of 50°S) is projected to increase by 1.1 °C from 1980 to 2100, though there is considerable variation among models. In several climate models, such as MIROC6 and CanESM5-CanOE, Southern Ocean SSTs show minimal increase by the end of the 21st century, while models like CIESM and CESM2-WACCM project increases of up to 2 °C. The observed Southern Ocean SST anomaly (red line in Fig. 37, 38.

In general, CMIP6-simulated Arctic Ocean surface warming is approximately twice as large as Southern Ocean surface warming throughout the 21st century (compare Fig. 1 a and b). This difference is projected to increase the cross-equatorial SST gradient (Fig. 1c). The cross-equatorial SST gradient (hereafter referred to as ‘Cross-SST’) is defined as the SST gradient between the northern tropics (0-25°N) and the southern tropics (0-25°S) (see Methods for details). A positive cross-equatorial SST gradient (Cross-SST > 0) indicates that SSTs in the northern tropical oceans are relatively higher than those in the southern tropical oceans. Figure 2a,b).

a, b Inter-model relationship between changes in Arctic Ocean SST (ΔArctic-SST, °C per °C) and Southern Ocean SST (ΔSouthern-SST, °C per °C) with changes in the cross-equatorial SST gradient (ΔCross-SST, °C per °C). The cross-equatorial SST gradient is defined as the difference between the northern tropics (0-25°N) and southern tropics (0-25°S). Inter-model correlation coefficients (r, unitless) and p-values are shown in each figure, with dashed black lines indicating the regression slopes. c, d Same as (a, b), but for the cross-equatorial rainfall gradient (ΔCross-Precip, mm day-1 per °C). e, f Timeseries of the cross-equatorial SST gradient (°C) (e) and rainfall gradient (mm day-1) (f) for the warm Arctic Ocean group (red) and warm Southern Ocean group (blue) under the SSP2-4.5 scenario. Shaded regions denote ±1 standard deviation across models. Future changes are calculated as the difference between the 2046-2065 average under the SSP2-4.5 scenario and the 1995-2014 average in the historical simulation.

Consistent with previous studies, future increases in ΔCross-SST are accompanied by increased tropical rainfall in the Northern Hemisphere. Figure 1d shows the multi-model evolution of the cross-equatorial precipitation change (ΔCross-Precip), defined as the precipitation gradient between the northern tropics (0–25°N) and the southern tropics (0–25°S). While the multi-model mean ΔCross-Precip shows a positive trend from 1980 to the mid-21st century, the substantial interannual variability of precipitation somewhat obscures the signal.

Figure 1 e and f show composite SST anomaly maps depicting differences between the mid- 21st century (2046-2065 average) and the late 20th century (1995-2014 average). The top four models simulating the strongest Arctic warming and Southern Ocean warming are selected for the warm Arctic Ocean (Fig. 1e) and warm Southern Ocean (Fig. 1f) composites, respectively. While both the warm Arctic and warm Southern Ocean composites show mild El Niño-like SST warming in the eastern equatorial Pacific, this pattern is more pronounced in the warm Southern Ocean composite, consistent with ref. 39. Zonal-mean SST anomalies indicate a positive cross-equatorial SST gradient in the strong Arctic Ocean warming composite (red line in Fig. 1g) and a negative gradient in the strong Southern Ocean warming composite (blue line in Fig.

The warm Southern Ocean composite also reveals SST cooling in the North Atlantic Ocean (Fig. 1f). This feature, often referred to as the ‘North Atlantic warming hole’ (NAWH), is frequently linked to a weakening of the Atlantic Meridional Overturning Circulation (AMOC), although other factors, such as teleconnections from the Indian Ocean 40, 41 and wind-driven changes in the subpolar gyre 42, may also contribute to its formation. While NAWH-like low SST anomalies are present in the warm Southern Ocean composite, these cold SST anomalies are not statistically significant (statistically significant SST anomalies are stippled in Fig. 1f). Furthermore, the inter-model relationship between the magnitude of Southern Ocean warming and AMOC weakening is relatively weak (r = 0.22).

Figure 1 h, i show composite precipitation anomalies. In the strong warm Arctic composite, precipitation increases over South Asia and off-equatorial North Africa but decreases over the tropical South Indian Ocean and off-equatorial South Pacific Ocean (Fig. 1h). In the strong Southern Ocean composite, precipitation increases over the off-equatorial South Pacific and South Atlantic Oceans (Fig. 1i). The zonal-mean tropical precipitation anomaly shows that equatorial precipitation increases in both the warm Arctic and warm Southern Ocean composites due to global warming. A close examination reveals that precipitation in the off-equatorial Northern Hemisphere (5–8°N) increases by around 0.5 mm day-1, indicating an intensification of the ITCZ in the warm Arctic Ocean composite (red line in Fig. 1j). Recent studies indicate that the meridional width of the ITCZ narrows and strengthens in response to global warming 27. Meanwhile, zonal-mean precipitation in the equatorial region (5°S–5°N) increases by around 0.4 mm day-1 in the warm Southern Ocean composite (blue line in Fig. 1j), suggesting a slight southward shift of the ITCZ.

Future increases in polar SSTs and cross-equatorial precipitation changes

To what extent does inter-model diversity in Arctic and Southern Ocean SST warming correlate with cross-equatorial SST and precipitation gradients? Because future polar ocean warming is closely tied to the degree of global warming 5, 43, the increase in global-mean surface air temperature by the mid-21st century is well correlated with both Arctic and Southern Ocean SST increases (Supplementary Fig. 2). To exclude the effects of global warming, Arctic and Southern Ocean surface temperature changes are normalized by the degree of global-mean temperature changes (see Methods for details). After normalization, Arctic and Southern Ocean SST changes are expressed per °C of global warming.

Figure 2 a, b illustrate the multi-model correlation coefficients between future Arctic and Southern Ocean warmings and the normalized future cross-equatorial SST gradient. The correlation between ΔArctic-SST and ΔCross-SST is low (r = 0.14; Fig. 2a), while the correlation between ΔSouthern-SST and ΔCross-SST is strong and negative (r = −0.67; Fig. 2b). This indicates that future Southern Ocean warming can significantly reduce the cross-equatorial SST gradient by enhancing southern tropical (0–25°S) SSTs more than northern tropical (0–25°N) SSTs.

Changes in the interhemispheric SST gradient are associated with shifts in atmospheric circulation patterns, influencing rainfall distribution across the equatorial region 33, 34. Similar to the cross-equatorial SST gradient, models with stronger Arctic warming tend to exhibit an increased cross-equatorial rainfall gradient (Fig. 2c). The multi-model correlation coefficient between ΔArctic-SST and ΔCross-Precip is low (r = 0.15; Fig. 2d). ΔSouthern-SST exhibits a relatively high negative correlation with ΔCross-Precip (r = −0.55; Fig.

When are these distinct cross-equatorial SST and precipitation gradients changes likely to occur? To address this question, we adopt a storyline approach 44, 45. Climate model groups simulating strong Arctic Ocean warming and strong Southern Ocean warming by the late 21st century show that notable differences in ΔCross-SST and ΔCross-Precip between these two groups begin to emerge around 2040 (Fig. 2 e, f). In the warm Arctic Ocean models (CMCC-CM2-SR5, CMCC-ESM2, MIROC6, and CanESM5 models), ΔCross-SST and ΔCross-Precip exhibit distinct northward migration starting around 2030 and continuously amplify by the late 21st century (red lines in Fig. 2 e, f). In contrast, in the warm Southern Ocean models (CESM2, CESM2-WACCM, CIESM, and CNRM-CM6-1-HR models), ΔCross-SST and ΔCross-Precip remain relatively unchanged until around 2070 and then shift slightly southward by 2080.

These findings suggest that significant Southern Ocean warming will lead to minimal changes in cross-equatorial precipitation (ΔCross-Precip ≈ 0) by the mid-21st century. This challenges the prevailing belief that differential heating between the hemispheres under global warming will likely shift zonal-mean tropical precipitation northward 23, 24, 25.

Tropical Atlantic rainfall changes

Due to the strong meridional rainfall gradient across the climatological ITCZ, its displacement can lead to droughts and floods in neighboring countries 46. In the tropical Atlantic, the movement of the rainfall band induces a dipole precipitation pattern between the West African monsoon region and northeastern Brazil (Fig. 3). Here, we focus on the Sahel and northeastern Brazil, two tropical regions highly vulnerable to high latitude climate changes 22, 47, 48.

Composites of rainfall changes (% per °C) in (a) warm Arctic Ocean models, (b) warm Southern Ocean models. The blue box denotes northeastern Brazil and the red box indicates the Sahel region. c Inter-model relationship between changes in Arctic Ocean SST (°C per °C) and Sahel rainfall (% per °C). d Same as (c), but for Southern Ocean SST (°C per °C). e Same as (c), but for northeastern Brazil rainfall (% per °C). f Same as (d), but for northeastern Brazil rainfall (% per °C). Inter-model correlation coefficients (r, unitless) and p-values are shown in each figure, with dashed black lines indicating the regression slopes. Here, future changes are calculated as the difference between the future climate (2046-2065 average) in the SSP2-4.5 scenario and the present-day climate (1995-2014 average) in the historical simulation. Normalization is based on the corresponding changes in global mean surface air temperature. For the spatial composite fields, statistically significant values (p < 0.05) are indicated with stippling.

Figure 3 a, b show the composite rainfall anomalies for the warm Arctic model group and the warm Southern Ocean group, respectively. Future precipitation changes are normalized by the global-mean surface temperature increases, resulting in units of fractional precipitation changes per °C of global warming. Consistent with previous studies 22, 48, anomalously high SSTs in the Arctic are associated with a noticeable increase in annual-mean precipitation over the Sahel region (Fig. 3a). Conversely, in the warm Southern Ocean model group, precipitation over the Sahel decreases by approximately 10% per 1.0 °C of global warming (Fig. 3b).

The multi-model correlation between ΔArctic-SST and precipitation changes in the Sahel region is 0.57 (Fig. 3c), indicating a strong connection between the degree of future Arctic warming and Sahel precipitation. In contrast, the correlation between Southern Ocean SST changes (ΔSouthern-SST) and precipitation changes in the Sahel is weakly negative, with a correlation coefficient of −0.25 (Fig. 3d). In the off-equatorial South Atlantic region, rainfall changes in northeastern Brazil (central-eastern Amazon) are more strongly correlated with Southern Ocean warming than with Arctic Ocean warming (Fig. 3e, f). The multi-model correlation between ΔSouthern-SST and precipitation changes in northeastern Brazil is 0.49 (Fig. 3f), accounting for 24% of the inter-model variances. In contrast, the connection between ΔArctic-SST and precipitation changes in northeastern Brazil is weak, with a correlation coefficient of -0.30 (Fig. 3e), which is marginally significant (p = 0.05). These multi-model correlations suggest that Southern Ocean warming can increase the chance of precipitation increases over northeastern Brazil.

Would other global warming scenarios produce similar impacts on regional precipitation changes? To explore this, the same analysis is applied to other scenarios – SSP1-2.6, SSP3-7.0, and SSP5-8.5 – with the results summarized in Supplementary Fig. 3. Across all scenarios, the warm Arctic Ocean group generally exhibits increased Sahel rainfall, whereas the warm Southern Ocean group shows a reduction in Sahel rainfall. This pattern is mostly pronounced under SSP1-2.6 and gradually weakens toward SSP5-8.5. In SSP1-2.6, the correlation between ΔSouthern-SST and Sahel precipitation change is statistically significant, with a correlation coefficient of −0.41. The sensitivity to scenario suggests that future Southern Ocean surface warming is more effective at inducing Sahel drought under low-emission scenarios, whereas under medium- or high-emission scenarios, the response is more driven by global warming.

Another key factor contributing to the southward shift of the tropical Atlantic ITCZ is the projected weakening of the AMOC and formation of the NAWH 48, 49. Supplementary Fig. 4 shows that the multi-model correlations between Southern Ocean warming and AMOC weakening, as well as with the NAWH, by the mid-21st century are weak (r = 0.22 and r = −0.18, respectively). The Atlantic sector rainfall anomalies in CMIP6 models with significant AMOC weakening by the mid-21st century show reduced rainfall over the off-equatorial North Atlantic Ocean and northern Africa, with this pattern slightly more amplified in the strong NAWH composite. This precipitation anomaly pattern associated with the strong NAWH composite resembles that caused by strong Southern Ocean warming. However, while the patterns are similar, Atlantic sector-mean latitudinal precipitation anomalies indicate that cross-equatorial precipitation anomalies are slightly more pronounced in the warm Southern Ocean model group than in the strong NAWH model group (Supplementary Fig.

SST nudging simulations: testing the effects of polar ocean warming on tropical climate

Composite analyses of CMIP6 models highlight the potential influence of Arctic and Southern Ocean warming on future tropical climate patterns. To verify these effects and to better understand the underlying physical processes, we conducted a series of sensitivity experiments using Community Earth System Model version 2 (CESM2). In these experiments, SSTs in the Arctic and Southern Ocean were restored (nudged) to specified values. We selected CESM2 because it successfully simulates an enhanced zonal SST gradient across the equatorial Pacific and a strengthening of Pacific trade winds from 1980 to 2010, in response to human-induced stratospheric ozone depletion and/or aerosol changes 50.

Figure 4 a and b illustrate the global SST responses to warming in the Arctic (north of 57°N) and Southern Ocean (south of 52°S), respectively. Since SSTs increase gradually over 100 years in CMIP6 SSP scenario simulations, we incrementally increased SSTs in the polar oceans over 100 years (see Methods). The global SST responses show that warming in the Arctic and Southern Ocean effectively increases temperature in the off-equatorial Northern and Southern Hemisphere, respectively (Fig. 4c). These findings suggest that the interhemispheric SST gradient contrasts observed in the CMIP6 composites of the warm Arctic Ocean and warm Southern Ocean model groups (Fig. 1 e, f) can be partially driven by Arctic and Southern Ocean surface warming, respectively.

SST changes (°C) in (a) warm Arctic Ocean simulation, (b) warm Southern Ocean simulation, (c) zonal-mean SST (°C) changes in warm Arctic (red) and Southern Ocean simulations (blue). d, e, f Same as (a, b, c), but for rainfall changes (mm day-1). The analysis period is from years 41 to 60 of the 100-years simulation. Statistically significant values (p < 0.05) are indicated with stippling.

A close examination of Fig. 4 a, b reveals that the impact of Southern Ocean warming on global SSTs is more pronounced than that of Arctic Ocean warming. Specifically, Southern Ocean warming not only affects SSTs in the Southern Hemisphere but also induces a positive Pacific Decadal Oscillation (PDO)-like SST pattern in the North Pacific Ocean (Fig. 4b). This is consistent with recent modeling studies suggesting that Southern Ocean surface warming, particularly due to the loss of Antarctic sea ice, can deepen the Aleutian Low and drive positive PDO-like SST anomalies in the North Pacific 26. Conversely, Southern Ocean cooling is shown to drive Aleutian low weakening and North Pacific warming 25.

Moreover, the Southern Ocean warming experiment shows an El Niño-like SST pattern and an associated Atlantic Niño (Fig. 4b), which are not clearly simulated by the Arctic Ocean warming experiment (Fig. 4a). Instead, the tropical SST response to Arctic Ocean warming (Fig. 5). This outcome contrasts somewhat with the CMIP6 ‘warm Arctic SST’ composite (Fig. 1e), which shows a weak El Niño-like SST warming pattern. This difference likely arises because the SST nudging experiment isolates the effects of Arctic Ocean warming alone, whereas the CMIP6 warm Arctic Ocean composite includes both Arctic Ocean warming and the effects of greenhouse gas-induced global warming. Additionally, differences may also stem from both internal variability, model diversity, and the timescale of perturbations 51.

In response to the cross-equatorial SST gradient, the zonal-mean cross-equatorial precipitation changes accordingly (Fig. 4d–f). The Southern Ocean warming experiment shows an increase in precipitation in the off-equatorial Southern Hemisphere and a corresponding decrease in the off-equatorial Northern Hemisphere (Fig. 4e). This pattern aligns with the findings from recent modeling studies 26, 27, which investigated the effects of reduced Antarctic sea ice cover. The zonal-mean precipitation anomalies show that Southern Ocean warming increases precipitation by up to 150 mm yr−¹ in the 5°S–10°S range (blue line in Fig. 4f), while a similar reduction in annual precipitation occurs in the off-equatorial Northern Hemisphere. Similarly, in the Arctic Ocean warming experiment, the opposite pattern emerges, with a comparable magnitude of annual-mean precipitation anomalies (red line in Fig. 1h–j), as both show maximum rainfall anomalies in the off-equatorial Southern Hemisphere for Southern Ocean warming and in the Northern Hemisphere for Arctic warming. Over the tropical Atlantic region, the Arctic Ocean warming can effectively decrease precipitation over northeastern Brazil by about 400 mm yr-1 (Fig. 4d). Although the CMIP6 multi-model correlation between Arctic Ocean warming and precipitation over northeastern Brazil is weak (Fig. 3e), our idealized SST nudging simulations suggest that future Arctic warming could potentially increase drought risk in northeastern Brazil.

More detailed teleconnection processes explaining how Southern Ocean warming at high latitudes (50°-80°S) increases SSTs in lower latitudes are illustrated in Fig. 5. In response to Southern Ocean warming, sea level pressure (SLP) decreases across broad areas of the southern mid-latitudes and subtropics (10°-60°S; Fig. 5b). These low SLP anomalies induce anomalous cyclonic circulation in the southern subtropics, along with cross-equatorial flows anomalies over the eastern equatorial Pacific and Atlantic Oceans (vectors in Fig. 5 a, b), indicating a weakening of the prevailing wind patterns, particularly the trade winds.

a Climatological sea level pressure (SLP, hPa; shaded) and surface winds (m s-1, vector). Green boxes indicate the western coasts of Australia, South America, and South Africa. b Changes in sea level pressure (ΔSLP; shaded) and surface wind vectors (Δwind). Changes in (c) downward shortwave radiation (Δ\({{SW}}^{\downarrow }\)), (d) downward longwave radiation (Δ\({{LW}}^{\downarrow }\)), (e) upward shortwave radiation (\(\Delta {{SW}}^{\uparrow }\)), (f) upward longwave radiation (Δ\({{LW}}^{\uparrow }\)). Changes in (g) surface turbulent heat flux (ΔSHF + ΔLHF) and (h) net surface heat flux (ΔNSHF). For (c-h), positive values indicate downward heat fluxes and statistically significant values (p < 0.05) are indicated with stippling.

Remote ocean warmings are primarily driven by increased downward shortwave (Δ\({{SW}}^{\downarrow }\); Fig. 5c) and downward longwave (Δ\({{LW}}^{\downarrow }\); Fig. 5d) radiation. Δ\({{SW}}^{\downarrow }\) anomalies are positive over the Southern mid-latitudes and southeastern Pacific cold tongue region, exceeding 10 W m−2, due to cloud feedbacks that reduce shortwave reflection (Supplementary Fig. 6), consistent with a previous study 28. In contrast, upward shortwave radiation anomalies (Δ\({{SW}}^{\uparrow }\); Fig. 5e) show relatively small changes, highlighting the effectiveness of net shortwave radiation (Δ\({{SW}}^{\downarrow }+\Delta {{SW}}^{\uparrow }\)) in driving remote ocean surface warming. Meanwhile, downward longwave radiation anomalies (Δ\({{LW}}^{\downarrow }\); Fig. 5d) contribute to warming the central tropical Pacific and tropical Atlantic Oceans but have little influence on the eastern tropical Pacific cold tongue. Upward longwave radiation anomalies (Δ\({{LW}}^{\uparrow }\); Fig. 5f) generally mirror Δ\({{LW}}^{\downarrow }\), and with a negative sign, implying heat release from the ocean to the atmosphere. The combined changes in sensible heat flux (ΔSHF) and latent heat flux (ΔLHF) (Fig. 5g) are also generally negative, indicating negative feedback to sea surface warming. As a result, the combination of enhanced oceanic heat loss to the atmosphere (negative anomalies in ΔSHF + ΔLHF and Δ\({{LW}}^{\uparrow }\)) and increased downward radiations (Δ\({{SW}}^{\downarrow }\,\)+ Δ\({{LW}}^{\downarrow }\)) leads to a net surface heat flux anomaly (ΔNSHF; Fig. 5h). This downward-directed flux reinforces regional SST warming over the southern subtropics, the tropical Indo-Pacific near the Maritime Continent, and the eastern tropical Pacific cold tongue.

To further quantify the effects of Arctic and Southern Ocean warming on the tropical climate, additional idealized SST nudging simulations were conducted. In these simulations, SST anomalies were gradually increased from 0 °C to 4 °C in both the Arctic and Southern Oceans (see Methods). Figure 6 a, b illustrate the sensitivities of ΔCross-SST to Arctic and Southern Ocean warmings, respectively. The regression slope of ΔCross-SST to Southern Ocean warming is twice as steep as that to Arctic Ocean warming. Specifically, for each 1 °C increase in Arctic Ocean SST, ΔCross-SST rises by approximately 0.09 °C, and ΔCross-Precip increases by about 0.11 mm day−1 (Fig. 6a, c). In contrast, for each 1 °C increase in Southern Ocean SST, ΔCross-SST decreases by approximately 0.18 °C, and ΔCross-Precip declines by about 0.17 mm day−1 (Fig. 6b, d).

a Relationship between changes in Arctic Ocean SST (°C) and cross-equatorial SST gradient (°C). b Same as (a), but for Southern Ocean SST (°C) changes. c Same as (a), but for the cross-equatorial rainfall gradient (mm day-1). d Same as (b), but for the cross-equatorial rainfall gradient (mm day-1). Regression coefficients are also denoted in each figure, with black dashed line indicating regression slope. Note that the y-axis interval is the same between (a) and (b), and between (c) and (d).

The main findings from these CESM2-based SST nudging simulations align well with the CMIP6 multi-model relationship (compare Figs. 2, 6). In CMIP6, the sensitivity of ΔCross-Precip to Southern Ocean warming is approximately 50% higher than to Arctic Ocean warming. Specifically, ΔCross-Precip increases by 0.11 mm day-1 per 1 °C increase in Arctic Ocean SST (Fig. 2c), while it decreases by −0.17 mm day-1 per 1 °C increase in Southern Ocean SST (Fig. 2d). This agreement highlights the robustness of the simulated relationship between polar SST changes and tropical precipitation responses across different modeling frameworks. If the rapid loss of Antarctic sea ice persists, the relatively moderate surface warming of the Southern Ocean – compared to the Arctic – could still have a substantial impact on tropical climates due to its higher effectiveness per degree of warming.

Discussion

This study underscores the significant influence of Southern Ocean warming on future tropical climate patterns, suggesting it may exert a greater impact than Arctic warming. Our multi-model analysis and SST nudging experiments suggest that Southern Ocean warming, even at 60-70% of the level of Arctic Ocean warming, could counteract the anticipated northward shift of zonal-mean precipitation in the tropics caused by Arctic Ocean warming and greenhouse gas-induced global warming. Notably, our idealized SST nudging experiments indicate that a 1.0 °C rise in Southern Ocean temperatures might influence tropical precipitation patterns as much as a 1.5 °C rise in Arctic temperatures. The regional implications of Southern Ocean warming are particularly striking. Our findings indicate that Southern Ocean warming could lead to an increase in rainfall in northeastern Brazil while simultaneously exacerbating drought conditions in the Sahel region.

A limitation of our study is the inability to isolate the effect of anthropogenic sulfate aerosols, which could significantly influence temperature changes in the Northern Hemisphere. Aerosols were a major factor contributing to suppressed surface warming in the Northern Hemisphere during the late 20th century 52, 53. As aerosol forcing is expected to diminish in future climate change scenarios 54, climate models that are more sensitive to aerosols in the historical era may project stronger Northern Hemisphere warming 55, 56. Another limitation is that our idealized model simulations are based on a single climate model, CESM2, which exhibits a larger southward shift of the ITCZ in response to global warming compared to the CMIP6 multi-model average. Further studies using a broader range of climate models are necessary to validate these findings and evaluate potential variations in model responses.

In conclusion, using a storyline approach, we find that Southern Ocean warming, strongly tied to Antarctic sea ice loss, is likely to play a pivotal role in shaping tropical SST patterns and altering regional hydrological cycles in the coming decades. A concurrent weakening of the AMOC and the development of the NAWH, which may constrain Northern high-latitude warming 36, 57, represents one plausible storyline. This scenario could trigger an abrupt southward shift of the ITCZ and significant regional precipitation changes.

Methods

CMIP6 models

We analyzed data from 42 coupled general circulation models within CMIP6 (refs. 58, 59). Details such as model names and ensemble members are listed in Supplementary Table 1. To represent the present-day climate, we used the 1995-2014 mean from historical simulations. For future climate projections, we used the 2046–2065 mean under the SSP2-4.5 and SSP5-8.5 scenarios, with differences between these periods reflecting future changes due to global warming.

As outlined by ref. 60, we analyze the general trends emerging from the average normalized patterns to assess how models respond to the same level of warming. These normalized patterns are calculated by comparing future climate (2046-2065 average) to the present-day climate (1995–2014 average) and then divided by the corresponding change in global mean surface air temperature. This process is conducted for each model and scenario at each grid point after re-gridding to a common 1° × 1° grid.

CESM2 experiments

The Community Earth System Model version 2 (CESM2) is used to identify the impact of Arctic and Southern Ocean SSTs on cross-equatorial SST gradients and tropical rainfall variations. CESM2 consists of coupled atmosphere (CAM6; ref. 61), ocean (POP2; ref. 62), and sea ice (CICE5; ref. 63) components. The horizontal resolution of CAM6 is 2.5° in longitude and 1.9° in latitude. POP2 and CICE5 share the same horizontal resolution, which is a nominal 1° with a uniform resolution of 1.125° in the zonal direction. The meridional horizontal resolution varies with the finest resolution of 0.27° at the equator.

SST nudging simulations (1): Arctic vs. Antarctic incremental SST increases (Fig. 4)

We conducted idealized SST nudging simulations to isolate the climate responses to warming in the Arctic and Southern Ocean SSTs. This involved restoring the CMIP6 model-simulated SSTs in the Arctic Ocean north of 57°N and the Southern Ocean south of 52°S, using a 5-day relaxation time scale. The regions between 57°N to 60°N in the Arctic Ocean and 50°S to 52°S in the Southern Ocean were set as buffer zones, where the nudging effects gradually diminished.

For the control simulations, SSTs in the Arctic (north of 57°N) and Southern Ocean (south of 52°S) were restored to the CMIP6 model-simulated climatological mean SSTs (referred to as Clim_AO and Clim_SO, respectively). The climatological mean is defined as the average from 1995 to 2014 across 42 CMIP6 climate models. Both control simulations were conducted for 100 years each.

For the warm Arctic Ocean simulation (Forc_AO), we used the SST average from 2080 to 2099 from four models that project the strongest Arctic Ocean warming under the CMIP6 SSP2-4.5 scenario: CMCC-CM2-SR5, CMCC-ESM2, MIROC6, and CanESM5. The average Arctic SST change between future and present-day climates in these models is approximately 3 °C. Accordingly, Arctic SST was nudged by an increase of 0.03 °C per year, resulting in a total increase of 3 °C over 100 years. To isolate the climate impact of the warm Arctic Ocean, the control simulation for the Arctic Ocean (Clim_AO) was subtracted from the warm Arctic Ocean simulation (Forc_AO minus Clim_AO).

Similarly, for the warm Southern Ocean simulation (Forc_SO), we used the SST average from 2080 to 2099 from four different climate models that project the strongest Southern Ocean warming under the CMIP6 SSP2-4.5 scenario: CESM2, CESM2-WACCM, CIESM, and CNRM-CM6-1-HR. The average Southern Ocean SST change in these models is about 1.5 °C. Accordingly, Southern Ocean SST was nudged by an increase of 0.015 °C per year, resulting in a total increase of 1.5 °C over 100 years. To isolate the climate impact of Southern Ocean warming, the control simulation for the Southern Ocean (Clim_SO) was subtracted from the warm Southern Ocean simulation (Forc_SO minus Clim_SO).

To identify the impacts of Arctic and Southern Ocean SST warming by the mid-21st century, we analyzed the average from years 41 to 60 of each 100-year simulation. A summary of these experiments is provided in Table S2.

SST nudging simulations (2): varying polar SSTs from 0–4 °C for sensitivity testing (Fig. 6)

To quantify how cross-equatorial SST and rainfall gradients respond to varying degrees of SST increases in the Arctic and Southern Oceans, we performed additional idealized SST nudging simulations. Using the same regions as in SST Nudging Simulation (1), we incrementally increased SSTs in both the Arctic and Southern Oceans by amounts ranging from 0 °C to 4 °C. Each simulation was initialized with the specified SST increase and run for 10 years to observe the transient responses to surface warming in these polar regions. We conducted a total of 10 experiments, with 5 focusing on Arctic SST warming and 5 on Southern Ocean SST warming. A summary of these experiments can be found in Table S3.

Cross-equatorial SST and rainfall gradient

We quantify the cross-equatorial SST and rainfall gradient by calculating the difference between the SST and precipitation in the northern tropics (0-25°N) and those in the southern tropics (0-25°S). A positive cross-equatorial SST and rainfall gradient indicates that the northern tropics are warmer and wetter than the southern tropics, while a negative cross-equatorial SST and rainfall gradient indicates that the northern tropics are colder and drier than the southern tropics.

Data availability

CMIP6 model data were obtained from the Earth System Grid Federation ( https://esgf-node.llnl.gov/projects/esgf-llnl/). The observed SST dataset (ERSSTv5) is provided by NOAA/OAR/ESRL PSD, Boulder, CO, USA ( https://www.esrl.noaa.gov/psd/data/gridded/tables/sst.html). The observed precipitation dataset (Global Precipitation Climatology Project, GPCP) Monthly Analysis Product is available from NOAA PSL, Boulder, Colorado, USA ( https://psl.noaa.gov). Climate model outputs from CESM2 SST nudging simulations conducted in this study are available at Dryad: https://doi.org/10.5061/dryad.73n5tb35v.

Code availability

The CESM source code and the standard setup files can be obtained from https://escomp.github.io/CESM/versions/cesm2.1/html/introduction.html. The data analysis was conducted using the software CDO and NCL. The processed source code for data analysis can be downloaded from https://doi.org/10.5061/dryad.73n5tb35v.

1.
Wei, J., Hang, R. & Luo, J.-J. Prediction of pan-arctic sea ice using attention-based LSTM neural networks. Front. Mar. Sci. 9, 860403 (2022).
2.
Serreze, M., Barrett, A., Stroeve, J., Kindig, D. & Holland, M. The emergence of surface-based Arctic amplification. Cryosphere 3, 11–19 (2009).
3.
Pithan, F. & Mauritsen, T. Arctic amplification dominated by temperature feedbacks in contemporary climate models. Nat. Geosci. 7, 181–184 (2014).
4.
Dai, A., Luo, D., Song, M. & Liu, J. Arctic amplification is caused by sea-ice loss under increasing CO2. Nat. Commun. 10, 121 (2019).
5.
Rantanen, M. et al. The Arctic has warmed nearly four times faster than the globe since 1979. Commun. Earth Environ. 3, 168 (2022).
6.
Turner, J., Hosking, J. S., Phillips, T. & Marshall, G. J. Temporal and spatial evolution of the Antarctic sea ice prior to the September 2012 record maximum extent. Geophys. Res. Lett. 40, 5894–5898 (2013).
7.
Zunz, V., Goosse, H. & Massonnet, F. How does internal variability influence the ability of CMIP5 models to reproduce the recent trend in Southern Ocean sea ice extent? Cryosphere 7, 451–468 (2013).
8.
Comiso, J. C. et al. Positive trend in the Antarctic sea ice cover and associated changes in surface temperature. J. Clim. 30, 2251–2267 (2017).
9.
Purich, A. & Doddridge, E. W. Record low Antarctic sea ice coverage indicates a new sea ice state. Commun. Earth Environ. 4, 314 (2023).
10.
Kuhlbrodt, T., Swaminathan, R., Ceppi, P. & Wilder, T. A glimpse into the future: the 2023 ocean temperature and sea ice extremes in the context of longer-term climate change. Bull. Am. Meteorol. Soc. 105, E474–E485 (2024).
11.
Polvani, L. M. & Smith, K. L. Can natural variability explain observed Antarctic sea ice trends? New modeling evidence from CMIP5. Geophys. Res. Lett. 40, 3195–3199 (2013).
12.
Rosenblum, E. & Eisenman, I. Sea ice trends in climate models only accurate in runs with biased global warming. J. Clim. 30, 6265–6278 (2017).
13.
Roach, L. A. et al. Antarctic sea ice area in CMIP6. Geophys. Res. Lett. 47, e2019GL086729 (2020).
14.
Marshall, J. et al. The ocean’s role in the transient response of climate to abrupt greenhouse gas forcing. Clim. Dyn. 44, 2287–2299 (2015).
15.
Salzmann, M. The polar amplification asymmetry: Role of Antarctic surface height. Earth Syst. Dyn. 8, 323–336 (2017).
16.
Singh, H. A., Garuba, O. A. & Rasch, P. J. How asymmetries between Arctic and Antarctic climate sensitivity are modified by the ocean. Geophys. Res. Lett. 45, 13,031–013,040 (2018).
17.
Li, X. et al. Tropical teleconnection impacts on Antarctic climate changes. Nat. Rev. Earth Environ. 2, 680–698 (2021).
18.
Armour, K. C., Marshall, J., Scott, J. R., Donohoe, A. & Newsom, E. R. Southern Ocean warming delayed by circumpolar upwelling and equatorward transport. Nat. Geosci. 9, 549–554 (2016).
19.
Rackow, T. et al. Delayed Antarctic sea-ice decline in high-resolution climate change simulations. Nat. Commun. 13, 637 (2022).
20.
McFarlane, A. A. & Frierson, D. M. The role of ocean fluxes and radiative forcings in determining tropical rainfall shifts in RCP8. 5 simulations. Geophys. Res. Lett. 44, 8656–8664 (2017).
21.
Byrne, M. P., Pendergrass, A. G., Rapp, A. D. & Wodzicki, K. R. Response of the intertropical convergence zone to climate change: Location, width, and strength. Curr. Clim. Change Rep. 4, 355–370 (2018).
22.
Park, J.-Y., Bader, J. & Matei, D. Northern-hemispheric differential warming is the key to understanding the discrepancies in the projected Sahel rainfall. Nat. Commun. 6, 5985 (2015).
23.
Zhang, R. & Delworth, T. L. Simulated tropical response to a substantial weakening of the Atlantic thermohaline circulation. J. Clim. 18, 1853–1860 (2005).
24.
Sun, L., Alexander, M. & Deser, C. Evolution of the global coupled climate response to Arctic sea ice loss during 1990–2090 and its contribution to climate change. J. Clim. 31, 7823–7843 (2018).
25.
Kang, S. M. et al. Global impacts of recent Southern Ocean cooling. Proc. Natl. Acad. Sci. 120, e2300881120 (2023).
26.
Hay, S. & Kushner, P. J. The Relative Importance of Antarctic Sea Ice Loss within the Response to Greenhouse Warming. J. Clim. 37, 3323–3344 (2024).
27.
England, M. R., Polvani, L. M., Sun, L. & Deser, C. Tropical climate responses to projected Arctic and Antarctic sea-ice loss. Nat. Geosci. 13, 275–281 (2020).
28.
Kim, H., Kang, S. M., Kay, J. E. & Xie, S.-P. Subtropical clouds key to Southern Ocean teleconnections to the tropical Pacific. Proc. Natl. Acad. Sci. 119, e2200514119 (2022).
29.
Kang, S. M., Ceppi, P., Yu, Y. & Kang, I.-S. Recent global climate feedback controlled by Southern Ocean cooling. Nat. Geosci. 16, 775–780 (2023).
30.
Geng, Y.-F. et al. CMIP6 intermodel spread in interhemispheric asymmetry of tropical climate response to greenhouse warming: Extratropical ocean effects. J. Clim. 35, 4869–4882 (2022).
31.
Lamarque, J.-F. et al. Global and regional evolution of short-lived radiatively-active gases and aerosols in the Representative Concentration Pathways. Climatic Change 109, 191–212 (2011).
32.
Friedman, A. R., Hwang, Y.-T., Chiang, J. C. & Frierson, D. M. Interhemispheric temperature asymmetry over the twentieth century and in future projections. J. Clim. 26, 5419–5433 (2013).
33.
Tomas, R. A., Deser, C. & Sun, L. The role of ocean heat transport in the global climate response to projected Arctic sea ice loss. J. Clim. 29, 6841–6859 (2016).
34.
Green, B., Marshall, J. & Campin, J.-M. The ‘sticky’ITCZ: Ocean-moderated ITCZ shifts. Clim. Dyn. 53, 1–19 (2019).
35.
Riahi, K. et al. The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: An overview. Glob. Environ. Change 42, 153–168 (2017).
36.
Orihuela-Pinto, B., England, M. H. & Taschetto, A. S. Interbasin and interhemispheric impacts of a collapsed Atlantic Overturning Circulation. Nat. Clim. Change 12, 558–565 (2022).
37.
Meehl, G. A. et al. Sustained ocean changes contributed to sudden Antarctic sea ice retreat in late 2016. Nat. Commun. 10, 14 (2019).
38.
Wilson, E. A., Bonan, D. B., Thompson, A. F., Armstrong, N. & Riser, S. C. Mechanisms for abrupt summertime circumpolar surface warming in the Southern Ocean. J. Clim. 36, 7025–7039 (2023).
39.
Kang, S. M. et al. Walker circulation response to extratropical radiative forcing. Sci. Adv. 6, eabd3021 (2020).
40.
Keil, P. et al. Multiple drivers of the North Atlantic warming hole. Nat. Clim. Change 10, 667–671 (2020).
41.
Hu, S. & Fedorov, A. V. Indian Ocean warming as a driver of the North Atlantic warming hole. Nat. Commun. 11, 4785 (2020).
42.
Ghosh, R. et al. Two distinct phases of North Atlantic eastern subpolar gyre and warming hole evolution under global warming. J. Clim. 36, 1881–1894 (2023).
43.
Huang, J. et al. Recently amplified arctic warming has contributed to a continual global warming trend. Nat. Clim. Change 7, 875–879 (2017).
44.
Zappa, G. & Shepherd, T. G. Storylines of atmospheric circulation change for European regional climate impact assessment. J. Clim. 30, 6561–6577 (2017).
45.
Shepherd, T. G. et al. Storylines: an alternative approach to representing uncertainty in physical aspects of climate change. Climatic Change 151, 555–571 (2018).
46.
Liu, Y., Cai, W., Lin, X. & Li, Z. Increased extreme swings of Atlantic intertropical convergence zone in a warming climate. Nat. Clim. Change 12, 828–833 (2022).
47.
Liu, Y. & Chiang, J. Coordinated abrupt weakening of the Eurasian and North African monsoons in the 1960s and links to extratropical North Atlantic cooling. J. Clim. 25, 3532–3548 (2012).
48.
Chen, Z. et al. Observationally constrained projection of Afro-Asian monsoon precipitation. Nat. Commun. 13, 2552 (2022).
49.
Park, I.-H. & Yeh, S.-W. Projections of the North Atlantic warming hole can be constrained using ocean surface density as an emergent constraint. Commun. Earth Environ. 5, 98 (2024).
50.
Chung, E.-S. et al. Tropical eastern Pacific cooling trend reinforced by human activity. npj Clim. Atmos. Sci. 7, 170 (2024).
51.
Tseng, H.-Y. et al. Fast and slow responses of the tropical pacific to radiative forcing in northern high latitudes. J. Clim. 36, 5337–5349 (2023).
52.
Flynn, C. M. & Mauritsen, T. On the climate sensitivity and historical warming evolution in recent coupled model ensembles. Atmos. Chem. Phys. 20, 7829–7842 (2020).
53.
Gillett, N. P. et al. Constraining human contributions to observed warming since the pre-industrial period. Nat. Clim. Change 11, 207–212 (2021).
54.
Lund, M. T., Myhre, G. & Samset, B. H. Anthropogenic aerosol forcing under the Shared Socioeconomic Pathways. Atmos. Chem. Phys. 19, 13827–13839 (2019).
55.
Andreae, M. O., Jones, C. D. & Cox, P. M. Strong present-day aerosol cooling implies a hot future. Nature 435, 1187–1190 (2005).
56.
Watson-Parris, D. & Smith, C. J. Large uncertainty in future warming due to aerosol forcing. Nat. Clim. Change 12, 1111–1113 (2022).
57.
Pedro, J. B. et al. Beyond the bipolar seesaw: Toward a process understanding of interhemispheric coupling. Quat. Sci. Rev. 192, 27–46 (2018).
58.
Eyring, V. et al. Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geoscientific Model Dev. 9, 1937–1958 (2016).
59.
O’Neill, B. C. et al. The scenario model intercomparison project (ScenarioMIP) for CMIP6. Geoscientific Model Dev. 9, 3461–3482 (2016).
60.
Tebaldi, C. et al. Climate model projections from the scenario model intercomparison project (ScenarioMIP) of CMIP6. Earth Syst. Dyn. 12, 253–293 (2021).
61.
Danabasoglu, G. et al. The community earth system model version 2 (CESM2). J. Adv. Modeling Earth Syst. 12, e2019MS001916 (2020).
62.
Smith, R. et al. The parallel ocean program (POP) reference manual ocean component of the community climate system model (CCSM) and community earth system model (CESM). LAUR-01853 141, 1–140 (2010).
63.
Hunke, E. C., Lipscomb, W. H., Turner, A. K., Jeffery, N. & Elliott, S. Cice: the los alamos sea ice model documentation and software user’s manual version 4.1 la-cc-06-012. T-3 Fluid Dynamics Group,. Los Alamos Natl. Lab. 675, 500 (2010).

Acknowledgements

This work was supported by the National Research Foundation (NRF), Korea, under project BK21 FOUR. H.J was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (No. RS-2023-00241228). E.-S.C. was supported by Korea Polar Research Institute (KOPRI) grant funded by the Ministry of Oceans and Fisheries (KOPRI PE25030). The main calculations were performed by the supercomputing resource of the Korea Meteorological Administration (National Center for Meteorological Supercomputer).

Ethics declarations

Competing interests

The authors declare no competing interests.

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.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

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/.

About this article

Cite this article

Jeong, H., Park, HS., Kang, S.M. et al. The greater role of Southern Ocean warming compared to Arctic Ocean warming in shifting future tropical rainfall patterns. Nat Commun 16, 2790 (2025). https://doi.org/10.1038/s41467-025-57654-4

-
Received:
-
Accepted:
-
Published:
-
DOI: https://doi.org/10.1038/s41467-025-57654-4