1. Introduction
In the last few years (after 2020), concentrations of atmospheric greenhouse gases, in particular carbon dioxide and methane, have increased at unprecedented rates (NOAA 2024a). Global warming continues as a result, even as strong natural variability also occurs. Global surface atmospheric temperatures have also continued to increase, as one manifestation, with 2024 being the warmest year on record. The magnitude of these increases is not well explained by climate model simulations (Schmidt 2024). Top-of-atmosphere (TOA) radiation imbalances also show record high increases since 2020 (Loeb et al. 2024). This paper addresses how these trends have come about and are changing along with the atmosphere and ocean circulations, with profound consequences for ocean heat content (OHC) (Cheng et al. 2022a).
As more than 90% of Earth’s energy imbalance (EEI) in the past half-century has accumulated in the ocean, increasing ocean temperatures and OHC reflect the record high values year after year (Johnson and Lyman 2020; Cheng et al. 2022a, 2024b) (Fig. 1). Changes in OHC also impact air–sea and ice–sea interactions and thus exert a considerable influence over the other components of the climate system, providing critical feedback through energy, water, and carbon cycles (Wu et al. 2012; von Schuckmann et al. 2016; Cheng et al. 2022a; Trenberth 2022).
(left) Global mean OHC (Cheng et al. 2024a) for 0–2000 m relative to a base period 1981–2010 (ZJ). The 95% confidence intervals are shown (sampling and instrumental uncertainties). (right) Trend from 2000 to 2023 in OHC for 0–2000 m (W m−2). The stippled areas show places where the trend is not significant at the 5% level.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
The focus of this paper is from 2000 through 2023, as 2000 is when reliable TOA radiation data became available. Accordingly, the OHC for the 0–2000-m depth is shown not only for the global mean but also as spatial trends over the 2000–23 period (Fig. 1); see methods in section 2. The global values from 1980 show increased confidence after 2005 or so, when Argo data became available globally (Cheng et al. 2017, 2024b). The spatial patterns of trends are of considerable interest because, although the ocean is warming nearly everywhere, by far the greatest increases are in the midlatitudes: in western boundary currents east of Japan in the Kuroshio Extension region of the Pacific and in the Gulf Stream extension in the Atlantic, and nearly everywhere from 35° to 50°S in the Southern Hemisphere (SH). Wu et al. (2012) earlier noted that the warming rate in subtropical western boundary currents in all ocean basins far exceeds the globally averaged surface ocean warming rate. Of particular interest is why the midlatitudes are warming the most.
In between the TOA radiation and the ocean is the atmosphere with its volatile dynamics moving energy around. By combining the TOA radiation with the divergence of the vertically integrated atmospheric energy transports, an estimate can be made of the net surface energy exchanges, the surface fluxes, which are dominated by sensible and latent energy turbulent fluxes plus surface radiation. The flux can be combined with observed changes in OHC to produce estimates of the ocean heat divergence as a residual (Trenberth et al. 2019). Often, this has been a fraught exercise owing to accumulation of errors, but by carefully ensuring a balance at both the northern and southern extremes of the oceans, zonal mean meridional heat transports have been determined from the divergences, both for the mean annual cycle and as time series since 2000 (Trenberth et al. 2019; Trenberth and Zhang 2019). In this way, a check can be made of the physical consistency of changes and trends. Accordingly, it is expected that changes over time in OHC are closely linked to the surface energy fluxes but with the ocean heat transports modulating those relationships through the redistribution of heat. The latter includes the wind-driven Ekman transports in the upper layers of the ocean.
Analyses of OHC changes both globally and regionally in each ocean basin (Cheng et al. 2022b, 2024b) split into northern, tropical, and southern regions, reveal part of what is occurring. But it turns out that the patterns of increases in OHC are remarkably simple when viewed from a zonal mean perspective. Although interannual variability, primarily associated with El Niño–Southern Oscillation (ENSO), adds complexity in the tropics, the extratropics are less variable and exhibit extraordinarily steady upward trends in both hemispheres near 40° latitude over the past two decades. The biggest changes are in the SH, which is where most of the ocean lies. This raises major questions about why.
Much speculation has occurred about whether presumed reductions in atmospheric aerosols, associated with the cleanup of ship fuels, have led to fewer clouds since 2020 and increases in absorbed solar radiation (Schmidt 2024; Loeb et al. 2024; Quaas et al. 2022; Hansen et al. 2023; Hodnebrog et al. 2024), but these effects are mostly in the North Atlantic. In the North Pacific, the abatement of industrial aerosol emissions by China from 2010 to 2020 has been shown to play a role in surface warming (Wang et al. 2024). Both are in the Northern Hemisphere (NH), where aerosol-related cloud changes may have amplified sea surface temperature (SST) changes but are not a suitable explanation for OHC trends overall. Moreover, Morgenstern (2024) convincingly demonstrates that aerosol effects in models are likely overestimated.
In this paper, we document the nature of the changes going on, along with some preliminary analysis of related fields to throw light on the character and causes of the changes.
2. Methods and data
a. Datasets
We use the latest updated gridded version of OHC called IAPv4 (Cheng et al. 2024b), which is at 1° × 1° horizontal resolution and monthly temporal resolution from the near surface to 2000-m depth. This dataset includes the latest sets of corrections and adjustments for inhomogeneities in the data streams from bottle casts, mechanical bathythermographs (MBTs), expendable bathythermographs (XBTs), and autonomous pinniped data, as well as Argo. All data were extensively quality controlled (Cheng et al. 2017, 2024b), and time series of global and regional results are presented (Cheng et al. 2024a) (Fig. 1, and Fig. 1 in the online supplemental material). The SST product used is the IAPv4 analysis at 1-m depth. There are only small differences with other in situ products (Cheng et al. 2024b); see supplemental Fig. 2 from Ishii et al. (2017), Zuo et al. (2018), and Lyman and Johnson (2023).
The TOA radiation is from Clouds and Earth’s Radiant Energy System (CERES) (Loeb et al. 2024) as the Energy Balanced and Filled (EBAF4.2), limiting the analysis to post 2000. Several atmospheric reanalyses have been used to make computations: 1) the European Centre for Medium-Range Weather Forecasts (ECMWF) ERA5, 2) the Modern-Era Retrospective Analysis for Research and Applications (MERRA-2), and 3) the most recent Japanese 55-year Reanalysis (JRA-55) (Mayer et al. 2024). Independent data are used for the reanalyses to test the robustness of the results where possible, as was also done by Loeb et al. (2022). Our evaluations of the reanalyses (Trenberth et al. 2011, 2016; Trenberth and Fasullo 2018; Mayer et al. 2021, 2022, 2024) using regional energy budgets and evaluation of physical constraints reveal notable differences, especially in the tropics. However, our evaluations have demonstrated that by far the best for our purposes is ERA5 (Loeb et al. 2022; Mayer et al. 2024). An ensemble of these does not make sense given the known biases; however, uncertainties in the results are reflected in the discussion.
Atmospheric reanalyses from ERA5 (Hersbach et al. 2020) are therefore used to produce mass corrections and compute all components of the energy and energy transports within the atmosphere (Mayer et al. 2024). Following Trenberth et al. (2019), these are vertically integrated and then combined with TOA radiation to compute Fs, the net surface energy flux. The inferred surface fluxes are superior to those obtained directly from ERA5 forecasts (Mayer et al. 2024). ERA5-based divergences are conservatively regridded to the 1° grid of IAPv4 and EBAF4.2. The reanalyses are also used to explore changes and trends in atmospheric variables including winds, temperatures, and moisture, and the atmospheric moisture budget is used to compute evaporation minus precipitation as a residual.
Surface wind stress from ERA5 is used to compute the curl of the wind stress that relates to the Ekman (frictional)-driven circulation in the ocean and changes in ocean currents. Two products are available. Belmonte Rivas and Stoffelen (2019) evaluated surface winds from ERA5 and found some modest biases. Mayer et al. (2025) found that ERA5 short-term (12-hourly) forecasts have weakened 10-m wind (and hence surface stress) trends compared to the analyses, which suggests that the model loses the observed trend rather quickly, and analyses are thus more appropriate for trends. Hence, we used the hourly analyzed values here, although it turns out there is almost no difference between the two products after our processing.
b. Computations
The ocean energy convergence is computed as the sum of ocean heat content tendency (computed from IAPv4-centered monthly differences from the surface to 2000 m) and net Fs. The mean annual cycle is computed from monthly averages from 2001 through 2023. A small global residual computed from the global ocean divergence (typically less than 1 W m−2) is distributed evenly over the global ocean (Trenberth et al. 2019). No temporal smoothing is applied. Although some quantities have been computed as residuals, in general, the results are somewhat reproducible using different reanalyses except in the deep tropics where strong rainfall and convection prevail.
Trends in this study have mostly been estimated by applying a locally weighted scatterplot smoothing (LOWESS) to the time series (25-yr window, equal to an effective 15-yr smoothing), and then, the OHC difference between the first and the end year is used to calculate the rate (Cheng et al. 2022b). This approach provides an effective method to quantify the local trend by minimizing the impact of year-to-year variability and influence of start/end points. A Monte Carlo simulation is used to generate 1000 surrogates, smoothed with LOWESS, and if the trend is larger than twice the standard deviation, the signal is deemed significant for a 95% confidence interval (or 1.65 times for 90%). For Fig. 11 (shown later), the statistical significance is based on the standard deviation of the annual mean winds.
3. Changes in patterns of OHC
Global OHC time series for 0–2000 m after 1980 (Fig. 1) and its partition into layers 0–700 and 700–2000 m for 1955–2023 (supplemental Fig. 1) show a robust ocean heating, with a linear warming rate of 6.4 ± 0.3 ZJ yr−1 (0–2000 m), 4.4 ± 0.2 ZJ yr−1 (0–700 m), and 2.0 ± 0.1 ZJ yr−1 (700–2000 m) (Cheng et al. 2022a). Warming is strongest near the surface (supplemental Fig. 1).
The striking increases near 40° latitude in both hemispheres (Fig. 1) are robustly revealed when zonal averages are taken (Fig. 2) relative to a 2000–04 base period to reveal the variability and trends. Here, gridpoint values (J m−2) have been integrated around a latitude circle over a 1° latitude band and thereby build in the convergence of meridians at higher latitudes, so that the units are zettajoules per degree of latitude. The strongest increases in OHC are evident after about 2010, but what is especially striking is the pattern that emerges. Sizable increases are evident from 10°N to 20°S, populated by large interannual variability. Similar magnitude trends are evident in the NH from 35° to 45°N, but the strongest trends are from 35° to 50°S. In between, in the subtropics of both hemispheres and at higher latitudes, the trends are much smaller. At 20°N, the trends are negligible, and trends are quite small at 30°S. Warming trends south of 50°S are lower as the amount of ocean diminishes, and the northern ocean volume is smallest at 60°N. Similar patterns are evident in other OHC datasets (supplemental Fig. 2).
Zonal mean OHC for 0–2000 m relative to a 2000–04 baseline for 1° wide strips (ZJ per degree latitude).
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
The global and zonal mean values broken down into the contributions from each basin (Fig. 3) show that all three oceans contribute to strong warming in the SH. The oceans are strongly linked south of about 35°S, south of Africa, such as through the Antarctic Circumpolar Current (ACC).
Zonal mean OHC anomalies relative to 2000–04 for Pacific, Atlantic, and Indian Oceans for 0–2000 m in zettajoules per degree latitude. Given in the bottom panel is the NOAA ONI time series. Some vertical dashed lines are added to help align features.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
In the Pacific (Fig. 3), strong cross-equatorial dipoles appear in OHC especially after 2014, all associated with ENSO (bottom panel Fig. 3). Strong El Niño warming [Oceanic Niño Index (ONI) increasing] from 5°N to 10°S centered on the equator for a year or so is opposed by strong cooling from 10° to 20°N and then abruptly reverses in sign as El Niño fades or the next La Niña kicks in (Cheng et al. 2019). The ENSO-related changes are associated with huge changes in the thermocline and ocean dynamics.
From about 10° to 33°N, there is no overall warming in the Pacific, and the same is true around 28°S. The strongest warming trends are evident at 10°–20°S and 40°–50°S, and about 40°N. Near 40°N and 45°S, the zonal mean warming is mostly unaffected by the large interannual variability in the equatorial region.
In the Indian Ocean (Fig. 3), warming is evident around 10° and 45°S and is missing entirely at 30°S. On the equator and farther north in the Indian Ocean, the Indonesian islands, Africa, and India play strong local roles, such as via the Indonesian Throughflow (ITF) (Trenberth and Zhang 2019; Mayer et al. 2018). The ITF is associated with a net volume flow around Australia accompanied by a heat transport, that varies with ENSO. We note that while ENSO variability is pronounced in the Pacific and Indian Oceans regionally, it is less evident in the extratropical zonal means.
In the Atlantic, warming is evident south of 45°N, although weaker in the subtropics near 20°N and 20°S, and heating is larger south of 35°S. The Atlantic contributions are relentless and especially striking, as noted for the southern oceans by Hague et al. (2024), who explored the processes involved with a model. They find that the South Atlantic’s upward trends in OHC are primarily due to zonal asymmetries in changing surface fluxes.
The patterns of heating vary considerably with depth. Indeed, the SSTs (Fig. 4) show considerably stronger increases in the NH after 2020, especially in the Pacific, perhaps highlighting the role of aerosols, noted above. However, such warming has been evident from 1900 to 2008 (Wu et al. 2012). Unlike OHC, SSTs are not weighted by mass or areal extent. The contrast with the SH arises, for instance, through stronger winds, mixing, and a deeper mixed layer in the SH, especially in the summer half year (Pan et al. 2023), as well as the much greater extent of the ocean in the SH.
Zonal mean SST anomalies (°C) relative to 2000–04 for global, Atlantic, Pacific, and Indian Oceans.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
The contributions to the OHC from 0 to 100 m (Fig. 5) also show stronger warming in the NH relative to the SH, particularly after about 2020. For 0–300-m depths (Fig. 5), the NH is slightly less dominant, an indication that heat is effectively subducted into the deep mixed layers formed north of the ACC and over greater volumes in the SH.
Zonal mean accumulated OHC anomalies in zettajoules per degree latitude relative to 2000–04 for the global, Atlantic, Pacific, and Indian Oceans for (top) 0–100- and (bottom) 0–300-m depths.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
The TOA net radiation from CERES (Loeb et al. 2024) has shown sharp positive trends since 2000. It has the strongest contributions from increases in absorbed solar radiation offset somewhat by increases in outgoing longwave radiation, but is quite noisy since it measures the radiation in watts per square meter. When this is integrated in time to assess the accumulated effects of the heating, while setting the mean values to zero for 2000–04, then the result allows the possible contributions to increases in OHC to be assessed (Fig. 6a). Tests that vary this base period make only minor differences (discussed below). The integration effectively acts as a low-pass filter. The result reveals increases in downward radiation at all latitudes and the greatest increases near the equator (Fig. 6a), in contrast to the meridional structure of OHC changes (Fig. 2).
Accumulated energy fluxes relative to the base period of 2000–04 for (a) CERES TOA net radiation down, (b) vertically integrated atmospheric energy divergence TEDIV, (c) inferred zonal mean net surface energy flux Fs for the ocean, and (d) the vertically integrated time integrated ocean heat divergence OEDIV inferred as a residual. The values are all zonal means as integrated values in zettajoules per degree latitude relative to the 2000–04 period.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
By computing all energy transports throughout the atmosphere and including all the tendencies in the state variables (Trenberth et al. 2019) using atmospheric reanalyses from ERA5 (Mayer et al. 2024), a vertical integral can be made to give the total atmospheric contributions, called divergence of the total energy transports (TEDIV) (Fig. 6b). TEDIV includes the vertically integrated divergence of transports of atmospheric moist static energy plus kinetic energy. This product requires information on all variables every hour and careful attention to first balancing the total atmospheric mass (Mayer et al. 2021). Again, it has been time integrated to give a result in terms of the energy changes. Now, a strongly striped meridional structure emerges, with especially strong anomalous convergence of energy into 15°–25°S and divergence from 30° to 42°S. Then, there is an increasingly strong convergence of atmospheric energy from 43° to 70°S.
When TEDIV is combined with the TOA radiation, the net surface energy flux Fs results as a residual (Fig. 6c). Here, we use results only for the ocean. Over land, which has limited heat capacity, results are expected to be, and indeed are, very small but with small-scale spatial noise near mountains, as detailed in Trenberth and Fasullo (2018). Note that Fs is positive downward (the opposite sign of that from Trenberth et al. 2019, but consistent with Mayer et al. 2024).
Now, a distinctive meridional structure more like that seen in OHC (Fig. 2) emerges. Long-term mean errors in Fs are expected to be less than about 5 W m−2 on basin scales (Mayer et al. 2022, 2024). Note that as 1 ZJ per degree over 5 years is about 13 W m−2 in the midlatitudes of the Southern Hemisphere, the values in Fig. 6c are substantial. For the last 5 years available, the anomalous Fs out of the ocean is strongest from 30° to 40°S and strongest into the ocean from about 12° to 27°S and from 45° to 70°S. However, these maxima are displaced from the main OHC changes, with the implication that this energy is influenced by ocean currents. This is illustrated by combining the net surface energy flux Fs with the changes in OHC, the result being the inferred vertically integrated ocean heat transport divergence (OEDIV) (Fig. 6d).
A narrow band of anomalous ocean heat divergence right along the equator is offset by anomalous convergence of heat in the ocean over a band just north of the equator and much broader bands from 2° to 12°S and strong convergence from 28° to 50°S. Anomalous ocean heat divergence occurs from 5° to 21°N and 13° to 27°S, accompanied by anomalous convergence of energy over most of 25°–60°N, with the strongest increases near 40°N. There is some offset of anomalous OHC from the surface flux Fs south of 45°S but from 30° to 40°S ocean transport is implied to make up for the increases in OHC and the flux of energy into the atmosphere (Fig. 2).
The climatological time-average MHT is updated through 2023 (Fig. 7). The mean annual cycle of MHT through 2016 and the long-term mean (Trenberth et al. 2019) show strong heat divergence away from 10°N to 10°S, with peak poleward ocean heat transports of order 2 PW near 20° in the subtropics of both hemispheres. In the Atlantic, average MHTs are northward at all latitudes (Fig. 7). In the Pacific, the MHTs are poleward in both hemispheres, and the Pacific and Indian Oceans dominate in the SH. These heat transports systematically contribute to the observed changes in OHC, especially through boundary currents in the NH, the Gulf Stream (Atlantic) and Kuroshio (Pacific) (Fig. 1).
Total MHT for the oceans, global, Atlantic plus Arctic, and Indo-Pacific from 2000 through 2023, including a small adjustment to ensure closure as in Trenberth et al. (2019).
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
The time series of zonal mean anomalous MHT through 2016 (Trenberth et al. 2019) have been extended through 2023 (Fig. 8) and are plotted here relative to the 2000–04 mean to emphasize the trends. In the NH, anomalous southward MHT peaks from 40° to 50°N while anomalous northward MHT from 10° to 35°N combines to add to the convergence near 40°N. The largest southward anomalies in MHT are in the Atlantic at all latitudes (not shown, but see Fig. 12 of Trenberth et al. 2019) owing to quite strong northward MHT from 2000 to 2004. The Atlantic is the one region where the 2000–04 base period matters somewhat and there the evidence indicates a nonsignificant weakened poleward heat transport by the Atlantic meridional overturning circulation (AMOC). Hence, in the NH in recent years, there is a net anomalous heat input from the atmosphere into the ocean (Fs > 0, Fig. 6c) together with heat convergence (Figs. 6d and 8); both drive the increase in OHC around 40°N.
Latitude–time series of zonal mean MHT (PW) relative to a base period of 2000–04. Arrows are indicated in a few places to show the direction of anomalous transport.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
In the last 5 years in the SH, the largest southward anomalies in MHT occur from the equator to 40°S and combine with modest northward anomalous MHT from 40° to 65°S to produce the huge convergence of heat near 40°S. Heat is also input into the ocean (Fs > 0) south of about 40°S, and there is strong anomalous ocean heat convergence from 30° to 40°S (Fig. 6d) via ocean currents. We note that the switch from anomalous southward to northward MHT (Fig. 8) occurs just south of Africa and Australia, indicating the important role of boundary currents and the ACC.
The OEDIV and MHT calculations are at the end of a chain of computations involving net TOA radiation, TEDIV, Fs, and OHC changes, and errors accumulate. It is difficult to assess error bars. Nevertheless, the distinctive divergence of ocean heat away from the subtropics into the midlatitudes is apparent in both hemispheres and offers a viable explanation of Fig. 2. The implications from these diagnostics are that both the atmospheric energy transports and ocean redistribution of heat have played major coupled roles in setting up the changes in OHC seen in Fig. 2.
Surface wind stress is a key driver of ocean currents. As a partial check on these results, the Ekman meridional heat transports (not shown) have been inferred using the surface westerly wind stress from ERA5 over the oceans; for anomalies in zonal wind stress, see Fig. 9 and also Hague et al. (2024). At the surface, the Ekman current is 45° to the right (NH) and to the left (SH) of the wind stress, but for the ocean column, there is a transport 90° to the right (NH) and to the left (SH). Moreover, climatological wind stress between about 30°N and 30°S is negative as easterly. Accordingly, with decreasing westerly wind stress from 5°N to 40°S but increasing wind stress from 40° to 65°S (Fig. 9), there is anomalous Ekman-induced convergence near 40°S. In the NH, the anomalous Ekman-induced convergence is near 50°N and with quite strong northward transport from 25° to 48°N.
Accumulated zonal wind stress from ERA5 relative to a base period of 2000–04 in 1016 N s per degree latitude. Anomalous zonal mean eastward (westward) wind stress (red) indicates southward (northward) Ekman transports in the NH but northward (southward) Ekman transports in the SH.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
The anomalous Ekman currents make a modest contribution through the advection of heat (not shown), but their primary contribution appears to be one of a mass transport that deepens the thermoclines near 40°N and 40°S. It therefore contributes to the OHC changes (Figs. 1 and 2) and moreover contributes to changes in sea level (Fig. 10). The linear trends in sea level for 2000–23 with the global mean removed (Fasullo and Nerem 2018) (Fig. 10), indicate the biggest changes from 30° to 50°S in the Pacific and Atlantic, and east of Japan in the North Pacific, consistent with OHC (Fig. 1).
Linear trends in sea level [sea surface height (SSH)], from satellite altimetry (Fasullo and Nerem 2018) (mm yr−1) from 2000 to 2023 but with the global mean removed. Spatial smoothing is applied at T63, or about 180-km resolution.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
The surface zonal wind stress over the oceans undergoes an annual cycle and is larger in winter than in summer of each hemisphere from about 20° to 50° latitude. However, when the trends are examined separately for November–April versus May–October, the patterns are very similar to Fig. 9, except the decrease at 40°N begins more abruptly about 2010 in northern winter (not shown).
An examination of changes in zonal mean winds over just the ocean (Fig. 11) shows a poleward shift in the overall westerlies and associated jet stream. The full mean field over the oceans is in supplemental Fig. 3. Figure 11 highlights the increases in the westerlies near 60°S that extend from the surface well into the stratosphere. Weak decreases also occur near 40°–45°S (strictly not significant, and hence stippled, but significant in the context of the movement). In the NH, a dipole structure suggests a more modest poleward shift, with less influence at the surface.
For the ocean only, zonal average winds are shown in gray shading and differences between 2012–23 and 2000–11 in color, where significant at the 5% level. Areas not significant are blanked out (to see the background), except for 20°–45°S where insignificant regions are stippled.
Citation: Journal of Climate 38, 9; 10.1175/JCLI-D-24-0609.1
A preliminary analysis of transient eddy storm-track activity (Trenberth 1991) at 850 hPa (not shown) reveals increasing poleward heat transports by transient eddies from 40° to 65°S and increasing poleward moisture transports from 35° to 60°S, consistent with Fig. 6b which shows stronger atmospheric energy convergence there and stronger divergence to the north. These quantities are often referred to as (υ′T′) and (υ′q′), where the prime is the departure from monthly means, υ is the northward velocity, T is the temperature, and q is the specific humidity. The NH does not show up well in these statistics because zonal averages involve land areas as well as the ocean. However, as changes in the mean flow, including the Hadley circulation, also matter, these aspects should be detailed by season.
4. Discussion
Here, we have used 2000–04 as a base period to examine subsequent trends. Natural variability, therefore, can slightly influence results, although a series of tests changing the dates indicate slight changes only in the Atlantic.
Changes in ocean heat transport affect OHC and thus regional thermosteric sea level change (Fig. 10), especially in the North Atlantic (Zanna et al. 2019; Moat et al. 2024). Indeed, Moat et al. (2024) conclude that “ocean heat transport convergence controls OHC tendency in most regions of the North Atlantic,” mainly through changes in geostrophic currents and anomalous advection across the time-mean temperature gradient, consistent with our results here. Earlier, Wu et al. (2012) noted that accelerated warming in boundary currents for 1900–2008 is associated with a synchronous poleward shift and/or intensification of global subtropical western boundary currents in conjunction with a systematic change in winds over both hemispheres, thereby indicating some changes well before the current period.
Linear trends from 2000 to February 2020 (Loeb et al. 2022) complement the results here, and by further exploring discrepancies between TOA radiation observed and that from ERA5 atmospheric reanalyses, they build confidence in our results. In particular, the regional trend pattern of divergence of total atmospheric energy transport over the ocean using ERA5 analyzed fields was similar to that inferred from the difference between TOA and surface fluxes from ERA5.
Previous studies have also noted trends for 2005–19 in the zonal mean temperatures in the upper 2000 m of the ocean (Shi et al. 2021) with the biggest increases from 45° to 50°S above about 1000-m depths accompanied by the strongest increases in zonal geostrophic currents between 48° and 58°S, while zonal currents weakened between 30° and 37°S. The main increases were from 2010 to 2020 (the end of their data period). Changes in the annual cycle of SST were documented (Shi et al. 2024) and attributed to human activities. Evidently, the Southern Ocean’s eastward flow intensified and narrowed during this period, and changes in zonal currents are largely associated with the meridional gradient of ocean buoyancy change rather than wind stress change (Shi et al. 2021); although, the evidence here suggests a strong role for the latter. Hague et al. (2024) note the vital role of surface fluxes for the southern oceans, with roughly equal contributions from changes in wind stress and surface heat flux, especially in the South Atlantic north of the ACC.
Other processes contributing to the uptake of heat arise from wind-driven upwelling of unmodified deep water which keeps the surface ocean cold, and the cold water absorbs anthropogenic heat which is then exported to the northern flank of the ACC by the background overturning circulation. Then, it is subducted on the north side of the ACC near 40°S (Armour et al. 2016; Gao et al. 2017; Huguenin et al. 2022). Furthermore, anomalous Fs is negative 30°–40°S, indicating that the ocean is taking up less heat. Hence, the anomalous MHT explains why the maximum OHC increase is around 40°S.
The implications from our work are that there are strong changes in the atmospheric circulation, jet streams, and storm tracks, that are reflected in surface heat fluxes and Ekman transports in the ocean. These can be influenced by the large natural variability in the atmosphere associated with weather and the coupled climate system. Indeed, systematic changes reveal the zonal mean jet stream moving poleward in both hemispheres over the ocean, and especially increasing the westerlies at 60°S (Fig. 11). Li and Ding (2024) recently found a global poleward shift in atmospheric rivers in winter in both hemispheres. An increase has also occurred in the poleward energy and moisture transports by transient eddies from 50° to 65°S. However, much more needs to be done to pin down all that is occurring in the coupled system and especially the atmosphere.
These kinds of changes are not unexpected from model results and some data studies. Much attention has been paid to an intensification and poleward shift of the westerlies over the last four decades in the SH (Swart and Fyfe 2012; Swart et al. 2015; Thomas et al. 2015; Thompson et al. 2011; Goyal et al. 2021; Chemke et al. 2022). Earlier trends associated with the Southern Annular Mode and stratospheric ozone loss (Fogt and Marshall 2020) are factors in Fig. 11, and positive SAM anomalies continue into mid-2023. The poleward shift in the southern jet stream from 1979 to 2019 has been linked to the known sensitivity of the circulation to tropical warming (Woollings et al. 2023).
OHC varies greatly in the deep tropics associated with ENSO (Fig. 1), yet ENSO signals are much less apparent at higher latitudes in the ocean in zonal means. On decadal time scales, the Pacific decadal oscillation (PDO) and closely related interdecadal Pacific oscillation (IPO) have typically modulated ENSO, and there is a strong correlation pattern in OHC variations associated with the PDO. Capotondi et al. (2023) comprehensively review tropical Pacific decadal variability and processes potentially involved in both the atmosphere and ocean, which also occur during ENSO. The simultaneous correlation of PDO and Niño-3.4 indices is 0.50 for 1991–2020, and statistically, ENSO leads PDO by 3–4 months through teleconnections via an atmospheric bridge, with El Niño (La Niña) associated with positive (negative) PDO index (NOAA 2024b). However, the PDO has unexpectedly been in a negative phase since January 2020, throughout the latest 2023–24 El Niño event, and negative values strengthened from May through October 2024 and continued negative through December 2024. The negative PDO has contributed to the high SST anomalies in the extratropical Pacific in both hemispheres (Qiu et al. 2023), and hence, changes in the IPO and natural variability may be underway.
Wind forcing caused by CO2 increases and ozone changes are likely the main drivers of a dipole in the structure of the annual cycle of the zonal SST pattern between 40° and 55°S (Shi et al. 2021, 2024), and here, we show that these effects extend beyond the annual cycle, that they analyzed, to the mean flow through 2023. The upper-layer zonal-velocity changes, which likely involve the changes in sea level and geostrophic currents in the ocean, are associated with the observed heat-content changes.
Very recently, Shaw et al. (2024) reviewed changes in atmospheric circulation in observations and models, and the main forcings involved. They discuss aspects of the poleward shift in the jet stream and associated storm tracks and possible mechanisms involved. However, they do not consider the role of the ocean. Much more analysis is desirable to fully determine what has happened throughout the atmosphere and attribute the changes to external forcings and natural variability, but these will be pursued elsewhere.
5. Conclusions
Heating in the climate system from 2000 to 2023 is most clearly manifested in zonal mean OHC for 0–2000-m depth. It occurs primarily in the top 300 m and is evident in SSTs. The SST changes emphasize surface warming in the NH, but the strongest energy increases are in the SH, where ocean area and volume are greater. In the NH, heating occurs at all latitudes in the Atlantic with some modulation and slightly reduced MHT from the south, but in the North Pacific, strong warming near 40°N is countered by cooling near 20°N. The zonal mean across all oceans is more robust than a focus on any particular ocean basin.
Estimates of TOA radiation, atmospheric energy transports, surface fluxes of energy, and redistribution of energy by surface winds and ocean currents reveal that the patterns of OHC warming are mostly caused by systematic changes in the atmospheric circulation, which alters ocean currents. The coupled atmospheric changes have resulted in a striking pattern of changes in the vertically integrated atmospheric energy divergence which is strongly reflected in surface wind stress and anomalous net surface heat fluxes into and out of the ocean.
In response to the wind changes, the ocean redistributes heat meridionally (Fig. 8), especially in western boundary currents in the NH (Fig. 1). Hence, the patterns are not directly related to TOA radiation imbalances but arise primarily from coupled atmosphere–ocean changes. In turn, those influence storms and cloudiness, and thus TOA radiation. Changes in atmospheric aerosols and associated clouds may have played a role in the North Pacific and North Atlantic, likely in amplifying SST anomalies, although, because land is warming a lot more than the oceans, advection of warmer air from continents over the northern oceans may also be in play.
In the NH, changes are associated with the western boundary currents, but the associated atmospheric changes require analysis of more than a zonal mean framework, as continents play a major role. Nonetheless, it is clear that the atmosphere and ocean currents are systematically redistributing heat from global warming, profoundly affecting local climates.
Acknowledgments.
The IAP/CAS analysis is supported by the National Natural Science Foundation of China (Grants 42122046 and 42076202). NSF NCAR is sponsored by the U.S. National Science Foundation. Fasullo was also supported by NASA Award 80NSSC22K0046 and by the Regional and Global Model Analysis (RGMA) component of the Earth and Environmental System Modeling Program of the U.S. Department of Energy. Mayer received funding from the Austrian Science Fund (FWF) P33177.
Data availability statement.
The code used in this paper includes data quality control and is available at http://www.ocean.iap.ac.cn/. IAPv4 global ocean temperature and OHC products are available at https://doi.org/10.12157/IOCAS.20240117.002 and https://doi.org/10.12157/IOCAS.20240117.001. CERES data were obtained from http://ceres.larc.nasa.gov/compare_products.php. ERA5 data are publicly available via the Copernicus Climate Change Service at https://cds.climate.copernicus.eu/ and processed data at https://cds.climate.copernicus.eu/cdsapp#!/dataset/derived-reanalysis-energy-moisture-budget?tab=overview. Ishii data are available at https://www.data.jma.go.jp/gmd/kaiyou/english/ohc/ohc_global_en.html.
REFERENCES
Armour, K. C., J. Marshall, J. R. Scott, A. Donohoe, and E. R. Newsom, 2016: Southern Ocean warming delayed by circumpolar upwelling and equatorward transport. Nat. Geosci., 9, 549–554, https://doi.org/10.1038/ngeo2731.
Belmonte Rivas, M., and A. Stoffelen, 2019: Characterizing ERA-Interim and ERA5 surface wind biases using ASCAT. Ocean Sci., 15, 831–852, https://doi.org/10.5194/os-15-831-2019.
Capotondi, A., and Coauthors, 2023: Mechanisms of tropical Pacific decadal variability. Nat. Rev. Earth Environ., 4, 754–769, https://doi.org/10.1038/s43017-023-00486-x.
Chemke, R., Y. Ming, and J. Yuval, 2022: The intensification of winter mid-latitude storm tracks in the Southern Hemisphere. Nat. Climate Change, 12, 553–557, https://doi.org/10.1038/s41558-022-01368-8.
Cheng, L., K. E. Trenberth, J. Fasullo, T. Boyer, J. Abraham, and J. Zhu, 2017: Improved estimates of ocean heat content from 1960 to 2015. Sci. Adv., 3, e1601545, https://doi.org/10.1126/sciadv.1601545.
Cheng, L., K. E. Trenberth, J. T. Fasullo, M. Mayer, M. Balmaseda, and J. Zhu, 2019: Evolution of ocean heat content related to ENSO. J. Climate, 32, 3529–3556, https://doi.org/10.1175/JCLI-D-18-0607.1.
Cheng, L., and Coauthors, 2022a: Past and future ocean warming. Nat. Rev. Earth Environ., 3, 776–794, https://doi.org/10.1038/s43017-022-00345-1.
Cheng, L., G. Foster, Z. Hausfather, K. E. Trenberth, and J. Abraham, 2022b: Improved quantification of the rate of ocean warming. J. Climate, 35, 4827–4840, https://doi.org/10.1175/JCLI-D-21-0895.1.
Cheng, L., and Coauthors, 2024a: IAPv4 ocean temperature and ocean heat content gridded dataset. Earth Syst. Sci. Data, 16, 3517–3546, https://doi.org/10.5194/essd-16-3517-2024.
Cheng, L., and Coauthors, 2024b: New record ocean temperatures and related climate indicators in 2023. Adv. Atmos. Sci., 41, 1068–1082, https://doi.org/10.1007/s00376-024-3378-5.
Fasullo, J. T., and R. S. Nerem, 2018: Altimeter-era emergence of the patterns of forced sea-level rise in climate models and implications for the future. Proc. Natl. Acad. Sci. USA, 115, 12 944–12 949, https://doi.org/10.1073/pnas.1813233115.
Fogt, R. L., and G. J. Marshall, 2020: The Southern Annular Mode: Variability, trends, and climate impacts across the Southern Hemisphere. Wiley Interdiscip. Rev.: Climate Change, 11, e652, https://doi.org/10.1002/wcc.652.
Gao, L., S. R. Rintoul, and W. Yu, 2017: Recent wind-driven change in Subantarctic Mode Water and its impact on ocean heat storage. Nat. Climate Change, 8, 58–63, https://doi.org/10.1038/s41558-017-0022-8.
Goyal, R., A. Sen Gupta, M. Jucker, and M. H. England, 2021: Historical and projected changes in the Southern Hemisphere surface westerlies. Geophys. Res. Lett., 48, e2020GL090849, https://doi.org/10.1029/2020GL090849.
Hague, M., M. Münnich, and N. Gruber, 2024: Zonally asymmetric increase in Southern Ocean heat content. J. Climate, 37, 6585–6604, https://doi.org/10.1175/JCLI-D-23-0623.1.
Hansen, J. E., and Coauthors, 2023: Global warming in the pipeline. Oxford Open Climate Change, 3, kgad008, https://doi.org/10.1093/oxfclm/kgad008.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Hodnebrog, Ø., and Coauthors, 2024: Recent reductions in aerosol emissions have increased Earth’s energy imbalance. Commun. Earth Environ., 5, 166, https://doi.org/10.1038/s43247-024-01324-8.
Huguenin, M. F., R. M. Holmes, and M. H. England, 2022: Drivers and distribution of global ocean heat uptake over the last half century. Nat. Commun., 13, 4921, https://doi.org/10.1038/s41467-022-32540-5.
Ishii, M., Y. Fukuda, S. Hirahara, S. Yasui, T. Suzuki, and K. Sato, 2017: Accuracy of global upper ocean heat content estimation expected from present observational data sets. SOLA, 13, 163–167, https://doi.org/10.2151/sola.2017-030.
Johnson, G. C., and J. M. Lyman, 2020: Warming trends increasingly dominate global ocean. Nat. Climate Change, 10, 757–761, https://doi.org/10.1038/s41558-020-0822-0.
Li, Z., and Q. Ding, 2024: A global poleward shift of atmospheric rivers. Sci. Adv., 10, eadq0604, https://doi.org/10.1126/sciadv.adq0604.
Loeb, N. G., and Coauthors, 2022: Evaluating twenty-year trends in Earth’s energy flows from observations and reanalyses. J. Geophys. Res. Atmos., 127, e2022JD036686, https://doi.org/10.1029/2022JD036686.
Loeb, N. G., S.-H. Ham, R. P. Allan, T. J. Thorsen, B. Meyssignac, S. Kato, G. C. Johnson, and J. M. Lyman, 2024: Observational assessment of changes in Earth’s energy imbalance since 2000. Surv. Geophys., 45, 1757–1783, https://doi.org/10.1007/s10712-024-09838-8.
Lyman, J. M., and G. C. Johnson, 2023: Global high-resolution random forest regression maps of ocean heat content anomalies using in situ and satellite data. J. Atmos. Oceanic Technol., 40, 575–586, https://doi.org/10.1175/JTECH-D-22-0058.1.
Mayer, J., M. Mayer, and L. Haimberger, 2021: Consistency and homogeneity of atmospheric energy, moisture, and mass budgets in ERA5. J. Climate, 34, 3955–3974, https://doi.org/10.1175/JCLI-D-20-0676.1.
Mayer, J., M. Mayer, L. Haimberger, and C. Liu, 2022: Comparison of surface energy fluxes from global to local scale. J. Climate, 35, 4551–4569, https://doi.org/10.1175/JCLI-D-21-0598.1.
Mayer, M., M. A. Balmaseda, and L. Haimberger, 2018: Unprecedented 2015/2016 Indo-Pacific heat transfer speeds up tropical Pacific heat recharge. Geophys. Res. Lett., 45, 3274–3284, https://doi.org/10.1002/2018GL077106.
Mayer, M., and Coauthors, 2024: Assessment of atmospheric and surface energy budgets using observation-based data products. Surv. Geophys., 45, 1827–1854, https://doi.org/10.1007/s10712-024-09827-x.
Mayer, M., M. A. Balmaseda, F. Vitart, and S. Tietsche, 2025: Tropical Pacific trends in the ECMWF seasonal system and implications for predictions of the 2020–2022 triple-dip La Niña. J. Climate, in press.
Moat, B. I., and Coauthors, 2024: Ocean heat convergence and North Atlantic multidecadal heat content variability. J. Climate, 37, 4723–4742, https://doi.org/10.1175/JCLI-D-23-0370.1.
Morgenstern, O., 2024: Using historical temperature to constrain the climate sensitivity, the transient climate response, and aerosol-induced cooling. Atmos. Chem. Phys., 24, 8105–8123, https://doi.org/10.5194/acp-24-8105-2024.
NOAA, 2024a: Trends in Carbon Dioxide CO2. Accessed 2 May 2024, https://gml.noaa.gov/ccgg/trends/.
NOAA, 2024b: NCEP Global Ocean Data Assimilation System (GODAS). Accessed 22 July 2024, https://www.cpc.ncep.noaa.gov/products/GODAS.
Pan, Y., and Coauthors, 2023: Annual cycle in upper-ocean heat content and the global energy budget. J. Climate, 36, 5003–5026, https://doi.org/10.1175/JCLI-D-22-0776.1.
Qiu, B., S. Chen, and E. Oka, 2023: Why did the 2017 Kuroshio large meander event become the longest in the past 70 years? Geophys. Res. Lett., 50, e2023GL103548, https://doi.org/10.1029/2023GL103548.
Quaas, J., and Coauthors, 2022: Robust evidence for reversal of the trend in aerosol effective climate forcing. Atmos. Chem. Phys., 22, 12 221–12 239, https://doi.org/10.5194/acp-22-12221-2022.
Schmidt, G., 2024: Climate models can’t explain 2023’s huge heat anomaly—We could be in uncharted territory. Nature, 627, 467, https://doi.org/10.1038/d41586-024-00816-z.
Shaw, T. A., and Coauthors, 2024: Emerging climate change signals in atmospheric circulation. AGU Adv., 5, e2024AV001297, https://doi.org/10.1029/2024AV001297.
Shi, J.-R., L. D. Talley, S.-P. Xie, Q. Peng, and W. Liu, 2021: Ocean warming and accelerating Southern Ocean zonal flow. Nat. Climate Change, 11, 1090–1097, https://doi.org/10.1038/s41558-021-01212-5.
Shi, J.-R., B. D. Santer, Y.-O. Kwon, and S. E. Wijffels, 2024: The emerging human influence on the seasonal cycle of sea surface temperature. Nat. Climate Change, 14, 364–372, https://doi.org/10.1038/s41558-024-01958-8.
Swart, N. C., and J. C. Fyfe, 2012: Observed and simulated changes in the Southern Hemisphere surface westerly wind-stress. Geophys. Res. Lett., 39, L16711, https://doi.org/10.1029/2012GL052810.
Swart, N. C., J. C. Fyfe, N. Gillette, and G. J. Marshall, 2015: Comparing trends in the southern annular mode and surface westerly jet. J. Climate, 28, 8840–8859, https://doi.org/10.1175/JCLI-D-15-0334.1.
Thomas, J. L., D. W. Waugh, and A. Gnanadesikan, 2015: Southern Hemisphere extratropical circulation: Recent trends and natural variability. Geophys. Res. Lett., 42, 5508–5515, https://doi.org/10.1002/2015GL064521.
Thompson, D. W. J., S. Solomon, P. J. Kushner, M. H. England, K. M. Grise, and D. J. Karoly, 2011: Signatures of the Antarctic ozone hole in Southern Hemisphere surface climate change. Nat. Geosci., 4, 741–749, https://doi.org/10.1038/ngeo1296.
Trenberth, K. E., 1991: Storm tracks in the Southern Hemisphere. J. Atmos. Sci., 48, 2159–2178, https://doi.org/10.1175/1520-0469(1991)048<2159:STITSH>2.0.CO;2.
Trenberth, K. E., 2022: The Changing Flow of Energy through the Climate System. Cambridge University Press, 319 pp., https://doi.org/10.1017/9781108979030.
Trenberth, K. E., and J. Fasullo, 2018: Applications of an updated atmospheric energetics formulation. J. Climate, 31, 6263–6279, https://doi.org/10.1175/JCLI-D-17-0838.1.
Trenberth, K. E., and Y. Zhang, 2019: Observed interhemispheric meridional heat transports and the role of the Indonesian Throughflow in the Pacific Ocean. J. Climate, 32, 8523–8536, https://doi.org/10.1175/JCLI-D-19-0465.1.
Trenberth, K. E., J. T. Fasullo, and J. Mackaro, 2011: Atmospheric moisture transports from ocean to land and global energy flows in reanalyses. J. Climate, 24, 4907–4924, https://doi.org/10.1175/2011JCLI4171.1.
Trenberth, K. E., J. T. Fasullo, K. von Schuckmann, and L. Cheng, 2016: Insights into Earth’s energy imbalance from multiple sources. J. Climate, 29, 7495–7505, https://doi.org/10.1175/JCLI-D-16-0339.1.
Trenberth, K. E., Y. Zhang, J. T. Fasullo, and L. Cheng, 2019: Observation-based estimates of global and basin ocean meridional heat transport time series. J. Climate, 32, 4567–4583, https://doi.org/10.1175/JCLI-D-18-0872.1.
von Schuckmann, K., and Coauthors, 2016: An imperative to monitor Earth’s energy imbalance. Nat. Climate Change, 6, 138–144, https://doi.org/10.1038/nclimate2876.
Wang, H., and Coauthors, 2024: Atmosphere teleconnections from abatement of China aerosol emissions exacerbate Northeast Pacific warm blob events. Proc. Natl. Acad. Sci. USA, 121, e2313797121, https://doi.org/10.1073/pnas.2313797121.
Woollings, T., M. Drouard, C. H. O’Reilly, D. M. H. Sexton, and C. McSweeney, 2023: Trends in the atmospheric jet streams are emerging in observations and could be linked to tropical warming. Commun. Earth Environ., 4, 125, https://doi.org/10.1038/s43247-023-00792-8.
Wu, L., and Coauthors, 2012: Enhanced warming over the global subtropical western boundary currents. Nat. Climate Change, 2, 161–166, https://doi.org/10.1038/nclimate1353.
Zanna, L., S. Khatiwala, J. M. Gregory, J. Ison, and P. Heimbach, 2019: Global reconstruction of historical ocean heat storage and transport. Proc. Natl. Acad. Sci. USA, 116, 1126–1131, https://doi.org/10.1073/pnas.1808838115.
Zuo, H., M. A. Balmaseda, K. Mogensen, and S. Tietsche, 2018: OCEAN5: The ECMWF ocean reanalysis system and its real-time analysis component. ECMWF Tech. Memo. 823, 46 pp., https://doi.org/10.21957/la2v0442.
