1. Introduction
The global-mean surface temperature has risen by about 0.7°C since the second half of the nineteenth century, which is mainly caused by increased greenhouse gases (GHGs) (Meehl et al. 2007). Both the observations and modeling results have shown an accelerating global warming (Ji et al. 2014; Karl et al. 2015; S. J. Smith et al. 2015; Chavaillaz et al. 2016) and a declining Arctic sea ice (Holland and Bitz 2003; Graversen et al. 2008; Serreze et al. 2009; Screen and Simmonds 2010). Sea ice melting provides freshwater to the ocean, which can weaken the deep-water formation in the North Atlantic and thus the Atlantic meridional overturning circulation (AMOC) (Jahn and Holland 2013; Drijfhout 2015; Liu et al. 2017; Thornalley et al. 2018). A slowdown of the AMOC can result in cooling over the Northern Hemisphere (NH) and warming in the Southern Hemisphere (SH) (Zhang and Delworth 2005; Cheng et al. 2007; Yang et al. 2013; Sevellec et al. 2017; Caesar et al. 2018). The competition between global warming and cooling associated with AMOC weakening has been of great interest to the climate research community (Brown et al. 1997; Kim and An 2013; Drijfhout et al. 2012; Drijfhout 2015).
The evolution of global surface temperatures in the past 150 years is the product of two factors: the direct radiative flux change caused by increased GHG and the global hydrological change caused by temperature rising that would in turn feed back to global warming. For simplicity and clarity, we define the former as the “heating effect” and the latter as the “hosing effect” of GHGs. The heating effect can directly change the surface temperature and atmosphere water vapor through altering the surface energy budget and the Clausius–Clapeyron relation (Held and Soden 2006; Andrews et al. 2009). Water vapor itself is one of the main GHGs, providing a positive feedback to amplify global warming (Zelinka and Hartmann 2012). The hosing effect, in this work, refers to ocean freshwater flux change, which can directly change ocean salinity (Durack and Wijffels 2010) and thus upper-ocean buoyancy, affecting mainly the thermohaline circulation and thus the oceanic heat transport (OHT) (Swingedouw et al. 2007, 2009; Yang et al. 2013, 2017).
The role of ocean freshwater flux change in a warming climate has been a critical concern in recent climate research. Gregory et al. (2005) analyzed the results of 11 coupled models and showed that freshwater flux change would have either positive or negative influence on the AMOC under a quadrupling of CO2. Drijfhout (2015) found that the weakened AMOC caused by sea ice melting would lead to global cooling. In addition to AMOC change, the role of freshwater in the change of sea surface temperature (SST) in a warming climate has also been investigated extensively. Williams et al. (2007) suggested that an amplified hydrological cycle may lead to a cooling over the global ocean as a result of enhanced downward diffusive heat flux. The cooling effect of freshwater change in global warming has also been studied in Drijfhout (2015). Zhang and Wu (2012) furthered the modeling study of Williams et al. (2007) by restoring the freshwater flux to the model climatology in a double-CO2 experiment; they found that the change in freshwater flux in global warming tends to enhance rather than reduce global warming, because the change in atmospheric moisture content in high latitudes can trap CO2-induced warming. The role of freshwater flux in a warming climate remains quite controversial. Furthermore, sea ice melting is important to the change of ocean circulation (Aagaard and Carmack 1989; Driesschaert et al. 2007), which is not included in Williams et al. (2007) or Zhang and Wu (2012). How would the AMOC change in response to sea ice melting under global warming? What is the role of the AMOC in SST change under global warming? If the freshwater flux change could indeed mitigate the SST warming, then Earth’s climate system would not warm as fast.
In this study, the heating and hosing effects are separated in a coupled climate model through two groups of global warming experiments. It is shown that the freshwater change has a significant cooling effect that can mitigate the global surface warming by as much as ~30%. Two significant regional cooling centers appear: one in the subpolar Atlantic and one in the Southern Ocean; both are triggered by sea ice melting but are sustained by different mechanisms. The subpolar Atlantic cooling is maintained by the weakened AMOC in the NH, while the Southern Ocean surface cooling is maintained by the enhanced northward Ekman flow related to strengthened westerly wind (Ferreira et al. 2015; Kostov et al. 2017). In these two regions, the effect of freshwater flux change dominates over that of radiation flux change, controlling the SST change in the warming climate.
This work is also a part of our series on Bjerknes compensation (BJC) in Earth’s climate system (Yang and Dai 2015; Dai et al. 2017; Yang et al. 2016; Zhao et al. 2016; Yang et al. 2017, 2018). The BJC rate is defined as the ratio of anomalous atmosphere heat transport (AHT) to OHT, which has been found to depend on local climate feedback (Liu et al. 2016; Yang et al. 2016). The theoretical study of Yang et al. (2016) showed that the BJC would fail under global warming because of a violation of global energy conservation, but it remains uncertain whether the BJC would be achieved in a more complex coupled model. In this work we show that the BJC in a warming climate can still occur, because the hosing effect can cause out-of-phase changes in ocean and atmosphere meridional overturning circulations. The AHT change can compensate the OHT change by about 80% in the extratropics during the transient stage of global warming. The mechanism of BJC is explained by large-scale circulation changes, which is consistent with existing studies (Yang and Dai 2015; Dai et al. 2017; Yang et al. 2016). In general, the global hydrological change can play a stabilizing role in a warming climate, containing the global overall temperature rising and maintaining Earth’s energy balance.
This paper is organized as follows. An introduction to models and experiments is given in section 2. Transient and equilibrium responses are analyzed in section 3. Freshwater balance and the mechanisms for SST change are presented in section 4. A summary and a discussion are given in section 5.
2. Model and experiments
The model used in this study is the Community Earth System Model (CESM) of the National Center for Atmospheric Research (NCAR), which was used in our previous studies (e.g., Dai et al. 2017). CESM is a fully coupled global climate model that provides state-of-the-art simulations of Earth’s past, present, and future climate states (http://www.cesm.ucar.edu/). CESM (version 1.0) consists of five components and one coupler: The Community Atmosphere Model, version 5 (CAM5; Park et al. 2014); the Community Land Model, version 4 (CLM4; Lawrence et al. 2012); the Community Ice Code, version 4 (CICE4; Hunke and Lipscomb 2008); the Parallel Ocean Program, version 2 (POP2; Smith et al. 2010); the Community Ice Sheet Model (Glimmer-CISM); and the CESM coupler, version 7 (CPL7). CESM1.0 is widely used and validated by researchers in the community.
The CESM used in this study has the model grid of T31_gx3v7. The atmospheric model CAM5 has 26 vertical levels, with the finite volume nominally 3.75° × 3.75° in the horizontal. It is essentially a new atmospheric model with more realistic formulations of radiation, boundary layer, and aerosols (Meehl et al. 2013; Neale et al. 2013). The features of the model formulation can be found in Neale et al. (2010) and Park et al. (2014). The CLM4 has the same horizontal resolution as the CAM5. The ocean model POP2 uses the grid gx3v7, which includes 60 vertical levels and a uniform 3.6° spacing in the zonal direction. In the meridional direction, the POP2 grid is nonuniformly spaced: It is 0.6° near the equator, gradually increases to the maximum 3.4° at 35°N/S, and then decreases poleward. Details of the model physics can be found in Danabasoglu et al. (2012). The ice model CICE4 has the same horizontal grid as the POP2. No flux adjustments are used in CESM1.0.
To explore the individual roles of heating and hosing effects in global warming, two groups of experiments are carried out. Each group consists of two parallel runs. The first group includes a 3500-yr control run (CTRL) and a 2000-yr doubled-CO2 experiment (2CO2). The CTRL starts from the rest, with standard configurations and forced by the preindustrial CO2 concentration of 285 ppm (http://www.cesm.ucar.edu/experiments/cesm1.0/). The model climate in CTRL reaches quasi equilibrium after 1000 years of integration (Yang et al. 2015). Experiment 2CO2 starts from year 1501 of CTRL, with the CO2 concentration increasing by 1% yr−1 for 70 years and held constant at the doubled level from year 71 to year 2000. The second group is configured the same way as the first group, except the freshwater flux at the ocean surface is prescribed using the mean seasonal cycle of CTRL. The freshwater flux consists of precipitation, evaporation, river runoff, and sea/land ice. The ice includes sea ice formation or melting in high latitudes, continental ice influx caused by land model snow capping, and salt flux caused by the salinity gradient between ice and water. The prescribed freshwater flux is applied only to the ocean model. The two parallel experiments, FixFW-CTRL and FixFW-2CO2, also start from year 1501 of CTRL and are integrated for 2000 years. The total global warming effect is obtained by subtracting that of CTRL from that of 2CO2. The heating effect in global warming is obtained by subtracting that of FixFW-CTRL from that of FixFW-2CO2. The hosing effect in global warming is then obtained by subtracting the heating effect from the total global warming effect. It is worth noting that the mean seasonal cycles in CTRL and FixFW-CTRL are nearly identical (figure not shown), and there is no long-term climate drift in the mean state.
For the convenience of further analyses, three stages are defined based on the evolution of AMOC (Fig. 2d). Stage I is from year 200 to year 500, representing an earlier “quasi equilibrium” stage of climate change, during which the ocean and atmosphere circulations, as well as the meridional heat transports, each reach a quasi-equilibrium state (Figs. 2d,g). Note that stage I is not a stage of an equilibrium response, since the net radiation flux at the top of the atmosphere (TOA) is still unbalanced (Fig. 1a). Stage II is from year 700 to year 1000, representing the recovery period of the global overturning circulation (Fig. 2d). Stage III is from year 1700 to year 2000, representing the final equilibrium stage of global warming, during which the net radiation flux at the TOA regains its balance (Fig. 1a) and Earth’s climate shifts to another regime. Climate changes in these stages are obtained by subtracting corresponding states in CTRL. Note that we have defined year 1501 of CTRL as the starting year of all experiments.
(a) Temporal evolutions of globally integrated net radiative flux (black), net downward SW (blue), and outgoing LW (red) at TOA (PW; positive for downward anomaly) under 2CO2 forcing. The red line at the top denotes the CO2 forcing. (b),(c) As in (a), but showing the hosing effect and heating effect, respectively. Each curve is smoothed with a 20-yr running mean. (d) Radiative flux change at the TOA in stage I of global warming, with net radiative flux (black), SW (blue), and LW (red). Stage I spans the years 200–500. (e),(f) As in (d), but showing the hosing effect and heating effect, respectively.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
Temporal evolutions of (a)–(c) SST (thick solid curves) and SAT (thin dashed curves) averaged over the globe (black), NH (red), and SH (blue) (°C); (d)–(f) percentage changes of the AMOC (black), the Indo-Pacific STC (blue), and HC (red; %); and (g)–(i) AHT (red), global OHT (blue), Atlantic OHT (green), and Indo-Pacific OHT (light blue) averaged over 30°–70°N (PW). The AMOC index is defined as the maximum of the streamfunction in the range of 0°–10°C isotherms over 20°–70°N in the Atlantic. The Indo-Pacific STC is similarly defined, but in the range of 20°–30°C isotherms over 0°–30°N. The HC index is defined as the maximum streamfunction between 200 and 1000 hPa over 0°–30°N. All indexes are normalized by their time-mean values in CTRL, which are 18, 36, and 92 Sv, respectively (1 Sv = 106 m3 s−1 for the ocean; 1 Sv = 109 kg s−1 for the atmosphere). Each curve is smoothed with a 20-yr running mean. Stage I spans the years 200–500 and represents an earlier quasi-equilibrium stage of global warming, based on the AMOC evolution. Stage II spans the years 800–1100 and represents the recovery stage of the AMOC. Stage III spans the years 1700–2000 and represents the equilibrium stage of global warming for (a),(d),(g) 2CO2 forcing; (b),(e),(h) hosing effect; and (c),(f),(i) heating effect.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
3. Climate changes caused by hosing and heating effects
a. Energy balance at the TOA
The net radiation flux at the TOA in 2CO2 does not reach equilibrium until year 1000 (Fig. 1a). There is a net energy gain by Earth’s system during the transient period of global warming. During the first 70 years and several decades after the doubling of CO2, Earth gains energy mainly as a result of the enhancing GHG effect; that is, the outgoing longwave (LW) radiation is weakened (red curve in Fig. 1a; positive for downward anomaly and negative for upward anomaly). Afterward the incoming shortwave (SW) radiation increases gradually (blue curve in Fig. 1a) because of decreasing polar albedo and global overall albedo (figure not shown), while the outgoing LW also increases gradually because of the rising surface temperature. Earth gains energy mainly from the enhanced incoming SW about 200 years after the doubling of CO2. The SW and LW radiations at the TOA finally reach a new balance in about 1000 years. Earth’s climate then enters a new equilibrium (Yang et al. 2018; Rind et al. 2018). Based on the evolution of the net TOA radiation flux, the transient stage of climate change under global warming can be defined as from year 1 to year 1000, and the equilibrium stage is defined as from year 1000 to year 2000 (Fig. 1a).
The hosing and heating perturbations have opposite effects on the SW and LW radiations at the TOA (Figs. 1b,c). Under global warming, the hosing effect in fact causes global cooling (Fig. 2b), which is accompanied by declined outgoing LW (red curve in Fig. 1b) and reduced downward SW (blue curve) because of more low-cloud reflection (figure not shown). The coupled response of SST cooling and more low-cloud reflection under the hosing effect is consistent with previous studies (e.g., Zhang et al. 2010; Trossman et al. 2016). The pure heating effect, however, increases both the downward SW (blue curve in Fig. 1c) and outgoing LW (red curve) at the TOA, because the surface warming leads to more outgoing LW from the surface, and smaller planetary albedo. The latter is related to fewer low clouds and less sea ice coverage.
During the first 100 years of global warming, the TOA energy imbalance is mainly attributed to the heating effect (Fig. 1c); thereafter, it is mainly attributed to the hosing effect (Fig. 1b). On a global scale, the TOA energy balance can be more easily fulfilled under the heating effect (Fig. 1c) than under the hosing effect (Fig. 1b). The radiation fluxes respond quickly to the heating effect and approach equilibrium by about 400 years (Fig. 1c). The TOA energy imbalance can be roughly thought as the ocean heat uptake (OHU), whose evolution determines the time scale of the climate system. Under the heating effect, the ocean becomes more stable because of enhanced stratification, since the upper ocean warms more quickly (figure not shown). The OHU occurs mainly in the upper 500 m (Fig. 6c). The equilibrium response time scale under the heating effect is thus only several hundred years. The adjustment of OHU to the TOA energy imbalance takes place on fairly a short time scale. This is consistent with observations and CMIP5 model results (D. M. Smith et al. 2015). Under the hosing effect, the ocean evolution is closely related to the thermohaline circulation and cryosphere in the high latitudes, which have millennium time scales. The hosing effect in global warming slows down the adjustment of the TOA radiation fluxes, significantly delaying the achievement of an equilibrium state (Fig. 1b).
Both the hosing and heating effects can result in compensation changes in the SW and LW at the TOA. For the global average, the compensation under the heating effect appears to be better than that under the hosing effect. The energy imbalance is about 0.1 PW (1 PW = 1015 W) for the former (Fig. 1c) and 0.2 PW for the latter (Fig. 1b) in stage I of global warming. However, at a regional scale, the compensation between SW and LW is nearly perfect at most latitudes under the hosing effect (Fig. 1e), leaving a small change in the regional net radiation flux, while under the heating effect, the regional energy imbalance at the TOA is stronger (Fig. 1f), which is particularly clear in the extratropics. Furthermore, the changes in TOA radiation fluxes under the heating effect are nearly symmetric about the equator (Fig. 1f), while those under the hosing effect are antisymmetric (Fig. 1e). The former is consistent with the symmetric temperature change in the two hemispheres under uniform GHG, while the latter is consistent with the bipolar seesaw in SST, which in turn is closely related to the ocean dynamics of thermohaline circulation (see more details in section 4b).
b. Temporal changes in surface temperature, overturning circulations, and MHT
The global-mean surface temperature is increased by 2°–3°C under the 2CO2 forcing in our model, consistent with projections in the IPCC Fifth Assessment Report (AR5) (IPCC 2014). The temporal changes of global surface temperature, atmosphere, and ocean meridional overturning circulations, as well as meridional heat transports (MHTs), are shown in Fig. 2. The surface air temperature (SAT) (thin curves in Fig. 2a) is 1°C warmer than the SST (thick curves) in the experiment. The SH exhibits a stronger warming than the NH, particularly during the transient stage of global warming. This asymmetric warming can be thought of as a result of a change in ocean thermohaline circulation (Fig. 2d). The global surface temperature under the hosing effect experiences a significant cooling during the transient stage (stage I) and then recovers to the state of CTRL during the equilibrium stage of global warming (stage III) (Fig. 2b). The cooling is stronger in the NH than in the SH (Fig. 2b). Under the heating effect, the evolution of global surface temperature exhibits a monotonic increase (Fig. 2c). The equilibrium responses of SST and SAT are the same as those under the 2CO2 forcing. This cooling effect of the freshwater on global surface temperature (Fig. 2b) can offset as much as ~30% of global warming (Fig. 2c), mitigating the global temperature rise significantly (Fig. 2a).
Under the 2CO2 forcing, the AMOC is weakened by nearly 60%, while the Indo-Pacific subtropical cell (STC) and Hadley cell (HC) are enhanced by about 5% during stage I of global warming (Fig. 2d). The curves in Figs. 2d–f are normalized by their corresponding mean values in CTRL, so the relative changes can be clearly demonstrated. The weakening of the AMOC under global warming has been investigated in many studies (e.g., Gregory et al. 2005; Meehl et al. 2012; Drijfhout 2015). It is noticed that the hosing effect and heating effect influence the AMOC at different stages of global warming. During the first 100 years, the AMOC weakening is mainly attributed to the direct heating effect of GHG and polar amplification (Fig. 2f), consistent with Gregory et al. (2005), when the freshwater flux change has not taken effect yet. The surface heating effect directly reduces the surface density and weakens the deep-water formation in the North Atlantic during the first several decades. Thereafter, the hosing effect gradually becomes the dominant controlling factor of AMOC evolution (Fig. 2e), which shows a nearly identical evolution to that in Fig. 2d. Note that the AMOC will eventually recover to the state in CTRL under both the hosing and heating effects (Figs. 2e,f). Under the hosing effect, the AMOC recovery can be attributed to enhanced evaporation in the subpolar Atlantic. Under the heating effect, the AMOC recovery can be attributed to the negative feedback between AMOC and SST in the North Atlantic. More details will be presented later (Fig. 4)
The evolutions of STC and HC under the hosing effect are opposite of those under the heating effect (Figs. 2e,f). Under the heating effect, the STC and HC are weakened slightly, which persists for the whole period of global warming (Fig. 2f). Under the hosing effect, they are strengthened during the transient stage of global warming and then recover to the state in CTRL during the equilibrium stage of global warming (Fig. 2e). The strength of the HC is, at the zeroth order, determined by the meridional SST gradient, which is weakened (strengthened) under the heating (hosing) effect (details shown in Figs. 3c,d) (Yang et al. 2017). The linear relationship between the meridional atmosphere mass transport (i.e., the HC) and the meridional SST gradient was assumed earlier in Budyko (1969) and is thereafter widely used in energy balance models (EBMs). This linear relationship states the simplest coupled atmosphere–ocean dynamics in the meridional direction. The HC drives the wind-driven STC, and the two are strongly coupled. It is seen again that, on a different time scale, the heating and hosing effects play different roles under the 2CO2 forcing. The heating effect forces quick changes in the global circulation, while the hosing effect controls the slow evolution of the global climate.
Temporal evolutions of (a) the global-mean SST (°C) in 2CO2 (black) and that due to the heating effect (red) and hosing effect (blue). (b)–(d) Zonal-mean SST (8°C) in 2CO2 and that due to the hosing and heating effects, respectively.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
The evolution of MHT follows closely that of the meridional overturning circulation (Figs. 2g–i). The changes in AHT and OHT under the 2CO2 forcing show clearly the BJC (Fig. 2g), which is nearly perfect under the hosing effect during the transient stage of global warming (Fig. 2h). Because of the weakened AMOC, the poleward OHT in the NH declines by about 0.2 PW (green curve in Fig. 2g). The poleward AHT is enhanced by more than 0.1 PW as a result of the strengthened HC in the tropics and eddy activity in the extratropics, compensating nearly 60% of the OHT decrease. It is clearly seen that it is the hosing effect that results in the BJC (Fig. 2h). Under the heating effect, the BJC is weak and occurs only in the fast transient stage of global warming (Fig. 2i). The freshwater change actually maintains the overall energy balance of Earth’s climate system. As expected, this occurs at the slow transient stage of global warming, when the freshwater flux change takes control of the evolution of thermohaline circulation. This is also consistent with previous studies using a simple coupled box model (Yang et al. 2016, 2018).
For the convenience of later analyses, let us define the fast transient stage of global warming as from year 1 to year 100, during which the heating effect on the AMOC is important and the hosing effect has not yet taken effect. The slow transient stage is defined as from year 100 to year 1000, during which the hosing effect controls the evolution of the AMOC. This is not a precise definition but is widely used in global warming studies (e.g., Held et al. 2010).
c. Temporal–latitude evolutions of SST, SSS, and SSD
Figure 3a shows that in the transient stage, nearly 30% of global warming caused by the heating effect can be offset by the hosing effect. The global SST change in stage I shows a net 1.3°C warming in 2CO2 with respect to CTRL, resulting from a 1.8°C warming caused by the heating effect and a 0.5°C cooling caused by the hosing effect. Neglecting complicated dynamics and thermodynamics of the hydrological cycle in the climate system, we can simply understand the overall cooling magnitude by thinking about the rule of lapse rate. For the dry air, the lapse rate 
The temporal–latitude evolution of SST is shown in Figs. 3b–d. Polar amplification and the “warming hole” coexist in the extratropics. The heating effect causes nearly symmetric changes in SST in two hemispheres (Fig. 3d), regardless of the asymmetric ocean–continent contrast. The polar amplification of SST is clear under the heating effect (Bekryaev et al. 2010). The SST in the high latitudes also responds more quickly to external forcing than the SST in the lower latitudes, with the former leading the latter by about several decades (Fig. 3d); this lag is closely related to the rapid sea ice albedo positive feedback in the high latitudes (Jahn and Holland 2013). The hosing effect causes a strong (weak) cooling in the NH (SH) high latitudes, which expands gradually southward in the NH (Fig. 3c), in concert with the weakening of AMOC. Thus, the so-called warming hole appears in the subpolar region in the transient stage of global warming (Fig. 3b) (Drijfhout 2015; Sévellec et al. 2017). It is the hosing effect that results in interhemispheric asymmetric changes in the global surface climate (Fig. 3b). The warming hole will disappear in about 800 years. The heating effect will eventually dominate over the hosing effect in determining the surface ocean temperature under continuous CO2 forcing.
The hosing effect is also seen in the ocean salinity change. The global-mean sea surface salinity (SSS) decreases during the transient stage of global warming and then recovers to the level of CTRL during the equilibrium stage (Fig. 4a), agreeing well with the SST evolution under the hosing effect (Fig. 3a). The sea surface freshening is clear in the extratropics (Fig. 4b). Strong freshening occurs in the extratropical NH as a result of the sea ice melting in the subpolar oceans and in the Arctic Ocean, as well as the weakened poleward salinity advection. Weak freshening occurs in the Southern Ocean off the Antarctic coast (Fig. 4b), which is related to the sea ice melting there (Fig. 4c). Near the equator, the surface ocean becomes more saline because of enhanced evaporation (Fig. 4c). It is the extratropical surface freshening (Fig. 4b) that results in the global surface cooling (Fig. 3c). In the heating experiments, the surface freshwater flux is prescribed using the mean annual cycle of CTRL, so the SSS is hardly changed (Fig. 5i). The combined changes in SST and SSS determine the change in sea surface density (SSD). Under the 2CO2 forcing, the global-mean SSD decreases mainly because of the heating effect (black and red curves in Fig. 4d). Being fresher in the extratropics and more saline in the tropics results in a net, though negligible, increase in global-mean SSD (blue curve in Fig. 4d). However, the pattern of SSD change, or the weakening of the meridional SSD gradient (Fig. 4e), caused by the hosing effect is significant (Fig. 4f).
Temporal evolutions of SSS (psu), surface freshwater (mm day−1), and SSD (kg m−3). (a) Global-mean changes in SSS (black), ocean surface freshwater flux (blue), and sea ice melting (red; mm day−1) in 2CO2. (b),(c) Zonal-mean SSS and net surface freshwater flux changes in 2CO2, respectively. Positive freshwater flux into the ocean can result in surface freshening, i.e., negative SSS. (d) Global-mean SSD changes in 2CO2 (black) and that due to the heating effect (red) and hosing effect (blue). (e),(f) Zonal-mean SSD changes in 2CO2 and that due to the hosing effect, respectively.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
(left) Quasi-equilibrium changes in (a) SAT (°C), (d) SST (°C), (g) SSS (psu), and (j) SSD (kg m−3) averaged over stage I of global warming under 2CO2 forcing. (middle),(right) As in (left), but showing the hosing effect and heating effect, respectively. (e) Gray dashed rectangle boxes in cover the regions where strong cooling occurs, which are 30°–70°N, 80°W–20°E in the North Atlantic; 30°–60°N, 130°E–120°W in the North Pacific; and 50°–80°S, 150°E–120°W in the Southern Ocean. These boxes are used in Fig. 10 for term balance analyses.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
The SSD change during the transient stage of global warming suggests a significant weakening of deep-water formation in the subpolar NH. This could result in the shutdown of the AMOC. The dynamic consequence of freshwater change during this period is critical to the AMOC. During the equilibrium stage of global warming, the SSD is still decreased (Fig. 4e), mainly because of the dominant role of strong surface warming (Fig. 3d). However, since the SSD change in the NH is nearly uniform, the change in the meridional SSD gradient is negligible (Fig. 4e). In other words, compared to that in stage I, the SSD in the extratropical NH is increased, and the deep-water formation over there recovers, which actually results in the recovery of the AMOC. An important implication is that the global thermohaline circulation may not change too much under global warming in the long run. The AMOC shutdown in stage I and its recovery in stage II are due to freshwater change, rather than pure radiative heating.
d. Horizontal and vertical patterns of ocean buoyancy changes
To better illustrate the hosing and heating effects, both horizontal and vertical patterns of ocean buoyancy changes are examined here. Knowing where the primary changes occur is important for understanding regional dynamics and mechanisms. In this subsection, we will focus on the changes in stage I.
Figure 5 shows horizontal patterns of surface buoyancy changes averaged over stage I of global warming. The SAT rises by as much as 5°C under the 2CO2 forcing. The strong warming hole seen in Fig. 3b occurs in the subpolar North Atlantic. There is also a weak warming hole occurring in the Ross Sea of the Southern Ocean (Figs. 5a,d). Observations also show a weak warming in the Southern Ocean (Armour et al. 2016) (weak warming implies that a weak cooling—i.e., a warming hole—can possibly appear in the future), which is believed to be caused by enhanced circumpolar upwelling and equatorward Ekman transport. It is straightforward that the heating effect results in global surface warming everywhere (Figs. 5c,f). The polar amplification is strong and nearly uniform at the subpolar latitudes, which is largely due to the sea ice positive feedback (Jahn and Holland 2013). The SAT in the polar regions can be ~3°C warmer than that in the tropics (Fig. 5c). For comparison, except some sporadic warmings in the SH (Figs. 5b,e), the hosing effect results in global surface cooling. There is an interhemispheric asymmetry in surface cooling, with a stronger cooling in the NH (Fig. 5b). The strong cooling in the subpolar North Atlantic causes the warming hole in global warming, reducing the polar amplification in the NH significantly (Fig. 5a). The bipolar seesaw of SST is clear in the Atlantic (Fig. 5e), which is consistent with previous coupled model results forced by freshwater hosing in the North Atlantic (e.g., Manabe and Stouffer 1995; Zhang and Delworth 2005; Cheng et al. 2007; Yang et al. 2013).
As discussed in Fig. 4, the hosing effect is seen in the SSS change, particularly in the North Atlantic (Figs. 5g,h). Strong freshening in the North Atlantic can be attributed to the sea ice melting in the subpolar and Arctic Oceans, as well as the weakened poleward salinity advection. The North Atlantic freshening is responsible for the global surface cooling, because it determines the SSD change there (Figs. 5j,k), weakening the deep-water formation and slowing down the AMOC, which reduces the northward OHT. In the heating experiments, SSS is hardly changed (Fig. 5i), and the decrease in SSD as a result of surface warming (Fig. 5l) in the North Atlantic is much weaker than that resulting from the hosing effect (Fig. 5k). Here, we would like to emphasize again that in terms of global average, the SSD decrease is mainly due to the heating effect (Fig. 4d); however, on a regional scale, the SSD change in the North Atlantic is determined by freshwater balance, particularly during the transient period of global warming. This can be clearly seen in Figs. 4b–f and in Figs. 5j,k. These regional changes in the North Atlantic have a profound global impact.
Both the surface warming and freshening can penetrate downward to 2000 m in stage I of global warming (Figs. 6a,d). The strongest warming occurs in the subsurface, instead of the ocean surface, because on the one hand, the upper-ocean cooling resulting from the hosing effect (Fig. 6b) offsets the upper-ocean warming resulting from the heating effect (Fig. 6c), and on the other hand, there is the baroclinic response of ocean thermal structure to the surface freshwater influx (Fig. 6b). The latter has been investigated comprehensively (e.g., Manabe and Stouffer 1995; Zhang and Delworth 2005; Stouffer et al. 2007). The pure heating effect causes nearly no salinity change in the ocean (Fig. 6f). In the extratropics, the upper-ocean salinity changes resulting from the hosing effect are nearly identical to those under the 2CO2 forcing. The upper-ocean freshening (Figs. 6d,e) and subsurface warming (Figs. 6a,b) jointly cause a significant decline in ocean density in the North Atlantic (Figs. 6g,h), weakening the AMOC significantly. The heating effect mainly controls the ocean density change in the tropics (Figs. 6g,i). Its effect on the density in the high latitudes is negligible.
(left) Depth–latitude section of quasi-equilibrium changes in (a) temperature (°C), (d) salinity (psu), and (g) potential density (kg m−3) averaged over stage I of global warming in 2CO2. (middle), (right) As in (left), but for showing the hosing and heating effects, respectively.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
It is straightforward that the ocean warming will strengthen and penetrate deeper with time (figures for stages II and III not shown). The surface cooling will gradually disappear, and the intermediate water (1000–3000 m) will gradually warm up. The surface freshwater influx in the northern subpolar ocean will eventually stop in stage II (Fig. 4c), because of enhanced surface evaporation and the lack of sea ice melting. The sea ice south of Iceland is completely gone as a result of continuous warming. The upper-ocean salinity in the subpolar ocean will increase continuously from stage II to stage III, which increases ocean density, resulting in recovery of deep-water formation there and thus recovery of the AMOC (Figs. 2d,g).
e. Changes in ocean overturning circulation and Hadley cell
To better understand the dynamics of these ocean changes, the changes in the meridional overturning circulations are shown in Fig. 7. Here, the quasi-equilibrium changes averaged over stage I of global warming are examined. As revealed by the AMOC index in Figs. 2d–f, the pattern of AMOC exhibits a significant weakening of its downward mass transport in the subpolar North Atlantic (Fig. 7a), which is mainly due to the hosing effect (Fig. 7b) (Zhang and Wu 2012). It is only slightly weakened under the heating effect (Fig. 7c). The northern branch of the Hadley cell is enhanced significantly (Fig. 7d) in response to the enhanced meridional SST gradient in the NH, which is mainly related to the hosing effect (Fig. 7e). The Hadley cell declines slightly under the heating effect (Fig. 7f). The wind-driven Indo-Pacific STC changes accordingly in response to the Hadley cell change; that is, it is enhanced under the hosing effect and declines under the heating effect (Figs. 7b,c). In stages II and III, the Hadley cell, the STC, and the AMOC almost return to the states in CTRL (figures not shown).
(a) Global-mean ocean meridional overturning circulation (Sv) in CTRL (gray contours) and its changes (shading) in stage I of 2CO2 experiment. (b),(c) The AMOC changes due to the hosing and heating effects, respectively. (d)–(f) As in (a)–(c), but for the HC. Color scale is the same for all plots.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
4. Freshwater balance and mechanism for SST change
Since this study is mainly about the cooling effect of freshwater change in global warming, we will further examine the pattern of surface freshwater change in this section. We will also investigate the mechanisms of SST changes by diagnosing the temperature equation.
a. Freshwater flux
Under the 2CO2 forcing, the net freshwater gain by the ocean and the ocean salinity change as a whole are negligible, as shown in Fig. 4a. It is calculated that in the 2000 years of global warming, the accumulated net freshwater gained by the ocean is about 0.27 × 1015 m3, which is about 2.5% of the global total liquid freshwater (~1016 m3). The global-mean SSS change is −0.03 psu and therefore negligible. However, the components of the freshwater budget, as well as of the regional freshwater balance, change remarkably (Fig. 8), affecting the surface buoyancy and thus circulations significantly. For example, the ocean loses about −6.13 × 1015 m3 of freshwater through precipitation and evaporation processes (red curve in Fig. 8a), of which about 4.23 × 1015 m3 of freshwater returns to the ocean via river runoff (cyan curve). Sea ice melting also provides about 1.9 × 1015 m3 of freshwater to the ocean (blue curve). These freshwater gains and losses are nearly balanced (black curve). However, the global hydrological cycle does change significantly under global warming, which can be seen in the horizontal patterns of freshwater balance (Fig. 8b). The surface freshwater flux change averaged in stage I of global warming can be simply described as “rich get richer”; that is, the wet areas in the deep tropics and midlatitudes become wetter because of more precipitation, and the dry regions in the subtropics become drier because of stronger evaporation (roughly within 10°–40°N/S) (Fig. 8b). This pattern is similar to that in previous studies and suggests an intensification of the hydrological cycle in response to increased GHGs (Allen and Ingram 2002; Held and Soden 2006; Wentz et al. 2007; Feng and Zhang 2015).
(a) Temporal evolutions in global-mean freshwater balance (Sv), showing precipitation minus evaporation (PmE; red), river runoff (light blue), sea ice melting (dark blue), and the sum of the three (PmE, river runoff, and sea ice melting; black). Positive values mean ocean gains freshwater. (b) Change in net ocean surface freshwater flux (mm day−1) in stage I of 2CO2 forcing. Blue (red) represents positive (negative) value, which means the ocean gains (loses) freshwater. The solid, dashed, and dotted black curves show the sea ice margin in CTRL, 2CO2, and FixFW-2CO2, respectively. The sea ice margin is defined as the location of 15% sea ice fraction. Freshwater flux equatorward (poleward) of the sea ice margin is contributed by PmE (sea ice melting).
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
In the polar and subpolar latitudes, sea ice melting (or formation) dominates the surface freshwater flux change (Fig. 8b). Sea ice melting in the polar ocean of the two hemispheres is enhanced as expected under the 2CO2 forcing. Consequently, the sea ice margin retreats poleward, which is clearer in the Southern Ocean, denoted by the solid and dashed black curves in Fig. 8b. However, sea ice formation also occurs in the subpolar ocean, specifically in the Greenland–Iceland–Nordic (GIN) seas and in the Ross Sea off the Antarctic, denoted by the negative freshwater flux. These sea ice formations are the consequence of strong surface cooling in these regions (Fig. 5e), which, in turn, comes from the hosing effect. In fact, under the 2CO2 forcing, the sea ice margin in the GIN seas expands southward slightly (dashed black curve for 2CO2 and black curve for CTRL in Fig. 8b). Here, we would like to emphasize three points. First, in the CESM used in this work, the deep-water formation in the Labrador Sea is most critical to the thermohaline circulation, so the sea ice melting there can weaken the AMOC significantly. Note that the land ice melting could also weaken the AMOC, but it is not considered in the CESM used in this work. Second, without the offset of sea ice melting to the global warming—that is, under only the heating effect—the shrinkage of sea ice coverage would be much more serious, as illustrated by the dotted black curve in Fig. 8b. Third, the negative freshwater flux in the subpolar ocean indicates two things: There is sea ice formation over the sea ice region, such like in the GIN seas and Ross Sea; or, there is no sea ice melting at all after global warming, compared to the sea ice in CTRL, such as in the Southern Ocean near 60°S, since there is sea ice melting (i.e., positive freshwater flux to the ocean) in CTRL in these regions.
As mentioned before, the hosing effect of the freshwater change in the North Atlantic results in the weakening and recovery of the AMOC in the slow transient period of global warming. However, during the fast transient stage of global warming, it is the heating effect that results in the weakening of the AMOC. Considering its critical role in ocean buoyancy and AMOC, the evolution of freshwater flux in the subpolar North Atlantic is further examined in Fig. 9a. Over the critical region where deep-water formation occurs, the change in sea ice melting determines the net freshwater flux (red and gray curves in Fig. 9a). This further determines the SSS and thus the SSD during the slow transient period of global warming (blue and black curves in Fig. 9a). Diagnostic analysis of the SSS equation over the Labrador Sea confirms the critical role of sea ice melting (figure not shown). During the fast transient stage of global warming (Fig. 9b), the heating effect lowers the SSD in the North Atlantic and thus results in the weakening of the AMOC. In Fig. 9b, the heating effect is represented by the SST warming (red curve); the SSD decrease caused by SST warming (dashed black curve) is roughly equal to the total SSD decrease (solid black curve), while the SSD decrease caused by freshening (dotted black curve) only becomes significant 170 years later.
(a) Temporal changes of surface freshwater flux (mm day−1), SSS (psu), and SSD (kg m−3) averaged between 40° and 60°N of the North Atlantic, showing SSD (black), SSS (blue), net freshwater flux (gray), and sea ice melting (red). (b) As in (a), but for the first 200 years of global warming, showing the SST change (°C) due to the heating effect (red) and SSS change due to the hosing effect (blue). Black solid, dashed, and dotted curves are for SSD changes under 2CO2 forcing, and that due to the heating and hosing effects, respectively. Each curve is smoothed with an 11-yr running mean.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
During the fast transient stage of global warming, the AMOC is weakened by nearly 30% (Fig. 2f). This is consistent with previous model results (Gregory et al. 2005) and the model results from phase 5 of the Coupled Model Intercomparison Project (CMIP5) (Cheng et al. 2013), in which the AMOC declines gradually in all models by 10%–50% over 140 years in response to a quadrupling of CO2, caused more by changes in surface heat flux than by changes in surface freshwater flux. In this work we emphasize that at a longer time scale beyond 200 years, the freshwater flux change in the North Atlantic plays a dominant role in the AMOC, as a result of sea ice melting, as well as the weakened northward saline-water advection. The positive feedback between AMOC strength and SSS in the North Atlantic could even shut down the AMOC.
b. Mechanism for SST change
For the North Atlantic, the surface warming is caused by enhanced surface heat flux, while the surface cooling—that is, the appearance of a warming hole—is due to weakened warm-water advection from the south and enhanced cold-water advection from the polar ocean. Term balance analysis reveals these processes explicitly (Figs. 10a–d). Figure 10a shows the net surface heat flux (red curve) is the main reason for the surface warming during the transient stage of global warming, while temperature diffusion, mainly vertical diffusion (horizontal diffusion is negligible and not plotted), is a cooling factor, balancing the surface warming. Figure 10b shows that the enhanced net surface heat flux is attributed to more downward SW and less sensible heat loss from the ocean surface. The solid and dashed black curves in Fig. 10b show the sea ice margins in CTRL and FixFW-2CO2, respectively. The region between these two curves indicates sea ice loss, which reduces the surface albedo significantly and results in more SW absorbed. The region south of the solid black curve warms up mainly because of less sensible heat loss, which in turn is due to reduced air–sea temperature difference, since GHG increase air temperature directly. Figure 10c shows horizontal temperature advection (blue curve) is the main process for the surface cooling during the slow transient stage of global warming, which largely offsets the warming effect of surface heat flux, resulting in the appearance of a warming hole in the North Atlantic. The vertical temperature advection is trivial and not plotted. Figure 10d shows changes in SST, surface current, and sea surface height (SSH) averaged in stage I of global warming. It shows clearly that on the one hand, the northward surface current—that is, the surface branch of the AMOC—is weakened and that on the other hand, the southward surface current from the Arctic Ocean is enhanced. This southward geostrophic current is driven by the sea surface pressure gradient (white contours). The anomalous high in the polar ocean is a result of the freshwater piling up, caused by sea ice melting (Saenko et al. 2017). These surface current changes bring less warm water from the south and more cold water from the north, jointly causing the strong warming hole in the North Atlantic in global warming, consistent with previous studies (Meehl et al. 2012; Drijfhout 2015; Sévellec et al. 2017).
(left) Temporal evolutions of terms in SST equation (°C month−1) and (right) the horizontal pattern of the most critical term for SST change averaged over stage I of global warming. (a),(c) Terms averaged in the North Atlantic box (defined in Fig. 5e), as a result of the heating effect and hosing effect, respectively. The net surface heat flux (red), horizontal advection (blue), vertical diffusion and convection (gray), and ∂T/∂t (black). (b) Changes in SW radiation and sensible heat flux (W m−2) in the North Atlantic due to the heating effect, in which the solid and dashed black curves denote sea ice margin in CTRL and FixFW-2CO2, respectively. (d) Changes in SST (color shading; 8°C), surface current (vector; cm s−1), and SSH (white contour; cm) in the North Atlantic due to the hosing effect. (e) As in (c), but for terms averaged over the Southern Ocean box (defined in Fig. 5e). (f) As in (d), but for changes in the Southern Ocean, with white contours represent Ekman pumping velocity (cm day−1; positive value for upward velocity).
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
For the Southern Ocean, the surface warming is also caused by more incoming SW and less sensible heat loss (figure not shown), as in the North Atlantic. The warming hole near the Ross Sea (Figs. 5a,d) is also attributed to changes in ocean dynamics, that is, enhanced northward Ekman transport as well as Ekman pumping. Figure 10e shows that advection (blue curve) is the only cooling factor for the ocean near the Ross Sea in the transient stage of global warming, which is balanced by the warming factors of surface heat flux (red curve) and vertical diffusion (gray curve). Figure 10f explicitly shows the enhanced northward Ekman flow, as well as the enhanced Ekman pumping (white contours) between 60° and 70°S. The latter is in good agreement with the surface cooling center. The mechanism of Southern Ocean cooling caused by the hosing effect is consistent with that given by Ma and Wu (2011) and Armour et al. (2016), in which they showed that a freshening of the Antarctic Ocean can induce a significant local cooling coupled with an intensification of the westerly winds. Our hosing experiment further reveals that the delayed warming should be attributed to the freshwater change caused by sea ice melting along the Antarctic continent.
For the North Pacific, the SST change is well correlated with but lags that in the North Atlantic (Fig. 11a). In both the heating and hosing experiments, the peak correlation (~0.85) occurs when the North Pacific SST lags the North Atlantic SST by about 3–5 years. This suggests a passive role of climate change in the North Pacific, which has a possible dynamic connection to the North Atlantic SST. Under global warming, the slower and weaker response in the North Pacific than that in the North Atlantic is quite counterintuitive, given the fact that the North Pacific is dominated by wind-driven circulation, while the North Atlantic is dominated by thermohaline circulation. The former should have a shorter time scale than the latter. We will not delve into this argument in this study.
(a) Correlations between SST changes averaged over the North Atlantic (red curves) (40°–70°N) and the North Pacific (blue curves) (30°–60°N), with bright colors for the hosing effect and faded colors for the heating effect. Note that the x axis uses the logarithmic scale. The peak correlation (~0.85) occurs when the North Atlantic SST leads the North Pacific SST by about 5 years. (b),(d) As in Figs. 10a and 10c, respectively, but for terms averaged over the North Pacific box (defined in Fig. 5e). (c),(e) Changes in SST (color shading, °C), surface current (vector; cm s−1), and SSH (white contour; cm) in the North Pacific due to the heating effect and hosing effect, respectively.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
The North Pacific warming, in general, is due to enhanced surface heat flux and northward advection (Fig. 11b). In contrast to that in the North Atlantic (Fig. 10a), the surface heat flux change is responsible only for the North Pacific warming during the fast transient stage. It is the enhanced northward warm-water advection that maintains the North Pacific surface warming in the later stage (blue curve in Fig. 11b). Figure 11c shows clearly the enhanced Kuroshio Extension, the North Pacific Current along 35°–40°N, and the subpolar gyre under the heating effect. The SSH change (white contours) in the North Pacific is also consistent with the changes in these geostrophic flows, which enhance the northward warm-water advection. The changes in SSH and currents are consistent with the findings in Sakamoto et al. (2005) and Li et al. (2017), in which an anomalous positive SSH in the North Pacific and intensified Kuroshio were identified in near-term global warming in the CMIP5 models. The mechanism of North Pacific surface cooling caused by the hosing effect (Fig. 11d) is similar to that in the North Atlantic, except that the cold water comes from the north via the Bering Strait (Fig. 11e). The freshwater piling up (white contours) as a result of sea ice melting in the polar ocean drives significant southward cold-water advection, which delays and weakens the North Pacific warming under global warming.
5. Summary and discussion
Understanding the change in the hydrological cycle in a warming climate is crucial to fully recognizing the consequences of global warming. As a first step, we separate the hosing effect and the heating effect in a unified coupled climate model through sensitivity experiments in this study. Special attention is paid to the hosing effect on global temperature and circulations in a warming climate. Our model results show that freshwater has a significant cooling effect in a warming climate that can potentially mitigate the global surface warming by as much as ~30%. Moreover, the compensation changes in AHT and OHT (i.e., the Bjerknes compensation) under global warming are also controlled by freshwater. We therefore conclude that in terms of global temperature and Earth’s energy balance, the freshwater change can play a stabilizing role in a warming climate.
The roles of the hosing effect and the heating effect in global surface temperature and meridional circulations in different stages of global warming are investigated in detail in this work. The relevant processes are summarized in Fig. 12. In the fast transient stage of global warming (in about 100 years), the heating effect dominates, causing surface warming and triggering the weakening of the AMOC. This is further confirmed by the ensemble experiments, showing clearly that the surface ocean is warming quickly under heating effect, which is also slightly offset by the hosing effect (Fig. 13a), and that the weakening of the AMOC is exclusively attributed to the heating effect (red curves in Fig. 13b). In the slow transient stage of global warming, the hosing effect dominates. The sea ice melting in the subpolar North Atlantic provides a large amount of freshwater, which suppresses the deep-water formation in the North Atlantic and results in significant weakening of the AMOC. This in turn causes significant cooling in the North Atlantic, that is, the warming hole. The negative feedback between the AMOC and SST in the North Atlantic slows the rising of the SST. In the equilibrium stage of global warming, the heating effect regains its control on surface temperature, but the surface density change is still mainly controlled by the hosing effect. Because of the lack of sea ice melting south of the GIN seas and enhanced evaporation in the North Atlantic, the SSS increases and the deep-water formation recovers. The AMOC and Hadley cell, as well as the OHT and AHT, all recover to the state of CTRL, although the global-mean temperature reaches a new level.
Schematic diagram summarizing the main processes in global warming. The upward (downward) arrows represent increase (decrease). NA: North Atlantic; AMOC: Atlantic meridional overturning circulation; VT (VS): SST (SSS) meridional advection; OHT (AHT): ocean (atmosphere) heat transport; EMP: evaporation minus precipitation. Ty represents the northward SST gradient. Here, “recover” means the state recovers to that of CTRL.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
Temporal evolutions of (a) SST changes in the North Atlantic (defined in Fig. 5e) and (b) percentage changes of the AMOC in the first 200 years, showing 2CO2 forcing (black), hosing effect (blue), and heating effect (red). Light color curves are for five ensemble members. Thick color curves are for the ensemble mean.
Citation: Journal of Climate 31, 23; 10.1175/JCLI-D-18-0297.1
By separating the heating effect and hosing effect in a warming climate, we quantified the critical role of the freshwater cycle in the long-term climate change. The role of freshwater flux has also been highlighted in the paleoclimate research. Observational evidence suggests that Earth’s climate has experienced remarkable changes since the Last Glacial Maximum (LGM) (Broecker 1998; Clark et al. 2012). The surface climate in the NH experienced significant cooling during the Older Dryas because of substantial melting of sea ice and thus the shutdown of the AMOC (Clark et al. 2012). The cooling event was followed by an abrupt warming that occurred at the onset of the Bølling–Allerød (McManus et al. 2004), accompanied by a rapid recovery of the AMOC (Liu et al. 2009). The climate evolution since the LGM can be well simulated in the NCAR coupled model, provided proper boundary conditions, such as orbital parameters, atmospheric GHG, and freshwater boundary condition (Liu et al. 2009). The freshwater condition, which includes the sea level records, the ice sheet boundaries, and meltwater discharges, is shown to be the most critical condition for achieving reasonable simulation of the climate evolution since the LGM (Liu et al. 2009). The AMOC shutdown during the Older Dryas and its recovery in the Bølling–Allerød can be reproduced in the NCAR model under freshwater forcing alone (Liu et al. 2009). This implies that great care should be paid to the hydrological condition when carrying out future climate simulations.
How the AMOC would change in reality under the current GHG changing situation remains a great concern. In a “model climate” the AMOC evolution is easy to recognize and understand. The results in this work show that even under the double-CO2 forcing, the AMOC will eventually recover, given a long-enough time. However, most model results from the CMIP5 show a common feature that the AMOC will be weakened notably but not catastrophically in the near future (Gregory et al. 2005). The same occurs in our model. This may be caused by the negative feedback between the AMOC and SST in the North Atlantic. Currently, it appears that meltwater discharge from sea ice and continental ice has not seriously impacted the AMOC intensity yet, although a number of observations have shown Arctic sea ice melting is faster and stronger than expected (Stroeve et al. 2007; http://nsidc.org/arcticseaicenews/), which has provided substantial freshwater to the polar and subpolar oceans. Given the expectation that the long-term declining trend of sea ice is very likely to continue (IPCC 2014), people should keep a wary eye on the AMOC change in the future. Once the positive feedback between the AMOC and SSS in the North Atlantic steps in, the catastrophic consequences may become possible.
This work is the first step toward quantifying the individual contributions of the heating and hosing effects to an evolving climate. Only the surface temperature and large-scale circulations are examined in this paper. Many other aspects have not been considered in the present study. For example, how do the hosing effect and the heating effect influence the wind-driven circulations and subduction in the midlatitudes, the intermediate water formation in the subpolar Antarctic, and the atmospheric monsoon system? How are climate variabilities, such as El Niño–Southern Oscillation, the Pacific decadal oscillation, and the Atlantic multidecadal oscillation, modulated by freshwater change and radiation forcing? It should be recognized that climate models have many limitations and that a climate shift in the model experiments may also exist. Many questions remain about the roles of the hydrological cycle in global change.
Acknowledgments
This work is jointly supported by the National Key Research and Development Program of China (2016YFA0601802) and the National Natural Science Foundation of China (41725021, 91737204, and 41376007). All experiments were performed on the supercomputer at LaCOAS at Peking University.
REFERENCES
Aagaard, K., and E. C. Carmack, 1989: The role of sea ice and other fresh water in the Arctic circulation. J. Geophys. Res., 94, 14 485–14 498, https://doi.org/10.1029/JC094iC10p14485.
Allen, M. R., and W. J. Ingram, 2002: Constraints on future changes in climate and the hydrologic cycle. Nature, 419, 224–232, https://doi.org/10.1038/nature01092.
Andrews, T., P. M. Forster, and J. M. Gregory, 2009: A surface energy perspective on climate change. J. Climate, 22, 2557–2570, https://doi.org/10.1175/2008JCLI2759.1.
Armour, K. C., J. Marshall, J. R. Scott, A. Donohoe, and E. R. Newsom, 2016: Southern Ocean warming delay by circumpolar upwelling and equatorial transport. Nat. Geosci., 9, 549–554, https://doi.org/10.1038/ngeo2731.
Bekryaev, R. V., I. V. Polyakov, and V. A. Alexeev, 2010: Role of polar amplification in long-term surface air temperature variations and modern Arctic warming. J. Climate, 23, 3888–3906, https://doi.org/10.1175/2010JCLI3297.1.
Broecker, W. S., 1998: Paleocean circulation during the Last Deglaciation: A bipolar seesaw? Paleoceanogr. Paleoclimatol., 13, 119–121, https://doi.org/10.1029/97PA03707.
Brown, M. G. L., E. K. Vizy, and K. H. Cook, 1997: Combined effects of global warming and an Atlantic meridional overturning circulation shutdown on West African and European climate. Object Mag., 7, 32–39.
Budyko, M. I., 1969: The effect of solar radiation variations on the climate of the Earth. Tellus, 21, 611–619, https://doi.org/10.3402/tellusa.v21i5.10109.
Caesar, L., S. Rahmstorf, A. Robinson, G. Feulner, and V. Saba, 2018: Observed fingerprint of a weakening Atlantic Ocean overturning circulation. Nature, 556, 191–196, https://doi.org/10.1038/s41586-018-0006-5.
Chavaillaz, Y., S. Joussaume, A. Dehecq, P. Braconnot, and R. Vautard, 2016: Investigating the pace of temperature change and its implications over the twenty-first century. Climatic Change, 137, 187–200, https://doi.org/10.1007/s10584-016-1659-4.
Cheng, W., C. M. Bitz, and J. C. H. Chang, 2007: Adjustment of the global climate to an abrupt slowdown of the Atlantic meridional overturning circulation. Ocean Circulation: Mechanisms and Impacts—Past and Future Changes of Meridional Overturning, Geophys. Monogr., Vol. 173, Amer. Geophys. Union, 295–313, https://doi.org/10.1029/173GM19.
Cheng, W., J. C. H. Chiang, and D. Zhang, 2013: Atlantic meridional overturning circulation (AMOC) in CMIP5 models: RCP and historical simulations. J. Climate, 26, 7187–7197, https://doi.org/10.1175/JCLI-D-12-00496.1.
Clark, P. U., and Coauthors, 2012: Global climate evolution during the last deglaciation. Proc. Natl. Acad. Sci. USA, 109, E1134–E1142, https://doi.org/10.1073/pnas.1116619109.
Dai, H., H. Yang, and J. Yin, 2017: Roles of energy conservation and climate feedback in Bjerknes compensation: A coupled modeling study. Climate Dyn., 49, 1513–1529, https://doi.org/10.1007/s00382-016-3386-y.
Danabasoglu, G., S. Bates, B. P. Briegleb, S. R. Jayne, M. Jochum, W. G. Large, S. Peacock, and S. G. Yeager, 2012: The 542 CCSM4 ocean component. J. Climate, 25, 1361–1389, https://doi.org/10.1175/JCLI-D-11-00091.1.
Driesschaert, E., and Coauthors, 2007: Modeling the influence of Greenland ice sheet melting on the Atlantic meridional overturning circulation during the next millennia. Geophys. Res. Lett., 34, L10707, https://doi.org/10.1029/2007GL029516.
Drijfhout, S., 2015: Competition between global warming and an abrupt collapse of the AMOC in Earth’s energy imbalance. Sci. Rep., 5, 14 877, https://doi.org/10.1038/srep14877.
Drijfhout, S., G. J. van Oldenborgh, and A. Cimatoribus, 2012: Is a decline of AMOC causing the warming hole above the North Atlantic in observed and modeled warming patterns? J. Climate, 25, 8373–8379, https://doi.org/10.1175/JCLI-D-12-00490.1.
Durack, P. J., and S. E. Wijffels, 2010: Fifty-year trends in global ocean salinities and their relationship to broad-scale warming. J. Climate, 23, 4342–4362, https://doi.org/10.1175/2010JCLI3377.1.
Feng, H., and M. Zhang, 2015: Global land moisture trends: Drier in dry and wetter in wet over land. Sci. Rep., 5, 18018, https://doi.org/10.1038/srep18018.
Ferreira, D., J. Marshall, C. M. Bitz, S. Solomon, and A. Plumb, 2015: Antarctic Ocean and sea ice response to ozone depletion: A two-time-scale problem. J. Climate, 28, 1206–1226, https://doi.org/10.1175/JCLI-D-14-00313.1.
Graversen, R. G., T. Mauritsen, M. Tjernström, E. Källén, and G. Svensson, 2008: Vertical structure of recent Arctic warming. Nature, 451, 53, https://doi.org/10.1038/nature06502.
Gregory, J. M., and Coauthors, 2005: A model intercomparison of changes in the Atlantic thermohaline circulation in response to increasing atmospheric CO2 concentration. Geophys. Res. Lett., 32, L12703, https://doi.org/10.1029/2005GL023209.
Held, I. M., and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. J. Climate, 19, 5686–5699, https://doi.org/10.1175/JCLI3990.1.
Held, I. M., M. Winton, K. Takahashi, T. Delworth, F. R. Zeng, and G. K. Vallis, 2010: Probing the fast and slow components of global warming by returning abruptly to preindustrial forcing. J. Climate, 23, 2418–2427, https://doi.org/10.1175/2009JCLI3466.1.
Holland, M. M., and C. M. Bitz, 2003: Polar amplification of climate change in coupled models. Climate Dyn., 21, 221–232, https://doi.org/10.1007/s00382-003-0332-6.
Hunke, E. C., and W. Lipscomb, 2008: CICE: The Los Alamos Sea Ice Model, documentation and software user’s manual, version 4.0. Los Alamos National Laboratory Tech. Rep. LA-CC-06-012, 73 pp.
IPCC, 2014: Climate Change 2014: Synthesis Report. Cambridge University Press, 151 pp., http://ipcc.ch/pdf/assessment-report/ar5/syr/SYR_AR5_FINAL_full_wcover.pdf.
Jahn, A., and M. M. Holland, 2013: Implications of Arctic sea-ice changes for North Atlantic deep convection and the meridional overturning circulation in CCSM4-CMIP5 simulations. Geophys. Res. Lett., 40, 1206–1211, https://doi.org/10.1002/grl.50183.
Ji, F. W. Z., J. Huang, and E. P. Chassignet, 2014: Evolution of land surface air temperature trend. Nat. Climate Change, 4, 462–466, https://doi.org/10.1038/nclimate2223.
Karl, T. R., and Coauthors, 2015: Possible artifacts of data biases in the recent global surface warming hiatus. Science, 348, 1469–1472, https://doi.org/10.1126/science.aaa5632.
Kim, H., and S. I. An, 2013: On the subarctic North Atlantic cooling due to global warming. Theor. Appl. Climatol., 114, 9–19, https://doi.org/10.1007/s00704-012-0805-9.
Kostov, Y., J. Marshall, U. Hausmann, K. C. Armour, D. Ferreira, and M. M. Holland, 2017: Fast and slow responses of Southern Ocean sea surface temperature to SAM in coupled climate models. Climate Dyn., 48, 1595–1609, https://doi.org/10.1007/s00382-016-3162-z.
Lawrence, D. M., K. W. Oleson, M. G. Flanner, C. G. Fletcher, P. J. Lawrence, S. Levis, S. C. Swenson, and G. B. Bonan, 2012: The CCSM4 land simulation, 1850–2005: Assessment of surface climate and new capabilities. J. Climate, 25, 2240–2260, https://doi.org/10.1175/JCLI-D-11-00103.1.
Li, R., Z. Jing, Z. Chen, and L. Wu, 2017: Response of the Kuroshio Extension path state to near-term global warming in CMIP5 experiments with MIROC4h. J. Geophys. Res. Oceans., 122, 2871–2883, https://doi.org/10.1002/2016JC012468.
Liu, W., S.-P. Xie, Z. Liu, and J. Zhu, 2017: Overlooked possibility of a collapsed Atlantic Meridional Overturning Circulation in warming climate. Sci. Adv., 3, e1601666, https://doi.org/10.1126/sciadv.1601666.
Liu, Z., and Coauthors, 2009: Transient simulation of last deglaciation with a new mechanism for Bølling-Allerød warming. Science, 325, 310–314, https://doi.org/10.1126/science.1171041.
Liu, Z., H. Yang, C. He, and Y. Zhao, 2016: A theory for Bjerknes compensation: The role of climate feedback. J. Climate, 29, 191–208, https://doi.org/10.1175/JCLI-D-15-0227.1.
Ma, H., and L. Wu, 2011: Global teleconnections in response to freshening over the Antarctic Ocean. J. Climate, 24, 1071–1088, https://doi.org/10.1175/2010JCLI3634.1.
Manabe, S., and R. J. Stouffer, 1995: Simulation of abrupt climate change induced by freshwater input to the North Atlantic Ocean. Nature, 378, 165–167, https://doi.org/10.1038/378165a0.
McManus, J. F., R. Francois, J. M. Gherardi, L. D. Keigwin, and S. Brown-Leger, 2004: Collapse and rapid resumption of Atlantic meridional circulation linked to deglacial climate changes. Nature, 428, 834–837, https://doi.org/10.1038/nature02494.
Meehl, G. A., and Coauthors, 2007: Global climate projections. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 747–845.
Meehl, G. A., and Coauthors, 2012: Climate system response to external forcings and climate change projections in CCSM4. J. Climate, 25, 3661–3683, https://doi.org/10.1175/JCLI-D-11-00240.1.
Meehl, G. A., and Coauthors, 2013: Climate change projections in CESM1 (CAM5) compared to CCSM4. J. Climate, 26, 6287–6308, https://doi.org/10.1175/JCLI-D-12-00572.1.
Neale, R. B., and Coauthors, 2010: Description of the NCAR Community Atmosphere Model (CAM5.0). NCAR Tech. Note NCAR/TN-486+STR, 268 pp., www.cesm.ucar.edu/models/cesm1.1/cam/docs/description/cam5_desc.pdf.
Neale, R. B., J. Richter, S. Park, P. Lauritzen, S. Vavrus, P. Rasch, and M. Zhang, 2013: The mean climate of the Community Atmosphere Model (CAM4) in forced SST and fully coupled experiments. J. Climate, 26, 5150–5168, https://doi.org/10.1175/JCLI-D-12-00236.1.
Park, S., C. S. Bretherton, and P. J. Rasch, 2014: Integrating cloud processes in the Community Atmosphere Model, version 5. J. Climate, 27, 6821–6856, https://doi.org/10.1175/JCLI-D-14-00087.1.
Rind, D., G. A. Schmidt, J. Jonas, R. Miller, L. Nazarenko, M. Kelley, and J. Romanski, 2018: Multicentury instability of the Atlantic meridional circulation in rapid warming simulations with GISS ModelE2. J. Geophys. Res. Atmos., 123, 6331–6355, https://doi.org/10.1029/2017JD027149.
Saenko, O. A., D. Yang, and P. G. Myers, 2017: Response of the North Atlantic dynamic sea level and circulation to Greenland meltwater and climate change in an eddy-permitting ocean model. Climate Dyn., 49, 2895–2910, https://doi.org/10.1007/s00382-016-3495-7.
Sakamoto, T. T., H. Hasumi, M. Ishii, S. Emori, T. Suzuki, T. Nishimura, and A. Sumi, 2005: Responses of the Kuroshio and the Kuroshio Extension to global warming in a high-resolution climate model. Geophys. Res. Lett., 32, L14617, https://doi.org/10.1029/2005GL023384.
Screen, J. A., and I. Simmonds, 2010: Increasing fall-winter energy loss from the Arctic Ocean and its role in Arctic temperature amplification. Geophys. Res. Lett., 37, L16707, https://doi.org/10.1029/2010GL044136.
Serreze, M. C., A. P. Barrett, J. C. Stroeve, D. N. Kindig, and M. M. Holland, 2009: The emergence of surface-based Arctic amplification. Cryosphere, 3, 11–19, https://doi.org/10.5194/tc-3-11-2009.
Sévellec, F., A. V. Fedorov, and W. Liu, 2017: Arctic sea-ice decline weakens the Atlantic Meridional Overturning Circulation. Nat. Climate Change, 7, 604–610, https://doi.org/10.1038/nclimate3353.
Smith, D. M., and Coauthors, 2015: Earth’s energy imbalance since 1960 in observations and CMIP5 models. Geophys. Res. Lett., 42, 1205–1213, https://doi.org/10.1002/2014GL062669.
Smith, R. D., and Coauthors, 2010: The Parallel Ocean Program (POP) reference manual: Ocean component of the Community Climate System Model (CCSM) and Community Earth System Model (CESM). Los Alamos National Laboratory Tech. Rep. LAUR-10-01853, 140 pp.
Smith, S. J., J. Edmonds, C. A. Hartin, A. Mundra, and K. Calvin, 2015: Near-term acceleration in the rate of temperature change. Nat. Climate Change, 5, 333–336, https://doi.org/10.1038/nclimate2552.
Stouffer, R. J., D. Seidov, and B. J. Haupt, 2007: Climate response to external sources of freshwater: North Atlantic versus the Southern Ocean. J. Climate, 20, 436–448, https://doi.org/10.1175/JCLI4015.1.
Stroeve, J., M. M. Holland, W. Meier, T. Scambos, and M. Serreze, 2007: Arctic sea ice decline: Faster than forecast. Geophys. Res. Lett., 34, L09501, https://doi.org/10.1029/2007GL029703.
Swingedouw, D., P. Braconnot, P. Delecluse, E. Guilyardi, and O. Marti, 2007: The impact of global freshwater forcing on the thermohaline circulation: Adjustment of North Atlantic convection sites in a CGCM. Climate Dyn., 28, 291–305, https://doi.org/10.1007/s00382-006-0171-3.
Swingedouw, D., T. Fichefet, H. Goosse, and M. Loutre, 2009: Impact of transient freshwater releases in the Southern Ocean on the AMOC and climate. Climate Dyn., 33, 365–381, https://doi.org/10.1007/s00382-008-0496-1.
Thornalley, D. J. R., and Coauthors, 2018: Anomalously weak Labrador Sea convection and Atlantic overturning during the past 150 years. Nature, 556, 227–230, https://doi.org/10.1038/s41586-018-0007-4.
Trossman, D. S., J. B. Palter, T. M. Merlis, Y. Huang, and Y. Xia, 2016: Large-scale ocean circulation-cloud interactions reduce the pace of transient climate change. Geophys. Res. Lett., 43, 3935–3943, https://doi.org/10.1002/2016GL067931.
Wentz, F. J., L. Ricciardulli, K. Hilburn, and C. Mears, 2007: How much more rain will global warming bring? Science, 317, 233–235, https://doi.org/10.1126/science.1140746.
Williams, P. D., E. Guilyardi, R. Sutton, J. Gregory, and G. Madec, 2007: A new feedback on climate change from the hydrological cycle. Geophys. Res. Lett., 34, L08706, https://doi.org/10.1029/2007GL029275.
Yang, H., and H. Dai, 2015: Effect of wind forcing on the meridional heat transport in a coupled climate model: Equilibrium response. Climate Dyn., 45, 1451–1470, https://doi.org/10.1007/s00382-014-2393-0.
Yang, H., Y. Wang, and Z. Liu, 2013: A modelling study of the Bjerknes compensation in the meridional heat transport in a freshening ocean. Tellus, 65A, 18480, https://doi.org/10.3402/tellusa.v65i0.18480.
Yang, H., Q. Li, K. Wang, Y. Sun, and D. Sun, 2015: Decomposing the meridional heat transport in the climate system. Climate Dyn., 44, 2751–2768, https://doi.org/10.1007/s00382-014-2380-5.
Yang, H., Y. Zhao, and Z. Liu, 2016: Understanding Bjerknes compensation in atmosphere and ocean heat transports using a coupled box model. J. Climate, 29, 2145–2160, https://doi.org/10.1175/JCLI-D-15-0281.1.
Yang, H., Q. Wen, J. Yao, and Y. Wang, 2017: Bjerknes compensation in meridional heat transport under freshwater forcing and the role of climate feedback. J. Climate, 30, 5167–5185, https://doi.org/10.1175/JCLI-D-16-0824.1.
Yang, Q., Y. Zhao, Q. Wen, J. Yao, and H. Yang, 2018: Understanding Bjerknes compensation in meridional heat transports and the role of freshwater in a warming climate. J. Climate, 31, 4791–4806, https://doi.org/10.1175/JCLI-D-17-0587.1.
Zelinka, M. D., and D. L. Hartmann, 2012: Climate feedbacks and their implications for poleward energy flux changes in a warming climate. J. Climate, 25, 608–624, https://doi.org/10.1175/JCLI-D-11-00096.1.
Zhang, L., and L. Wu, 2012: Can oceanic freshwater flux amplify global warming? J. Climate, 25, 3417–3430, https://doi.org/10.1175/JCLI-D-11-00172.1.
Zhang, R., and T. L. Delworth, 2005: Simulated tropical response to a substantial weakening of the Atlantic thermohaline circulation. J. Climate, 18, 1853–1860, https://doi.org/10.1175/JCLI3460.1.
Zhang, R., S. M. Kang, and I. M. Held, 2010: Sensitivity of climate change induced by weakening of the Atlantic meridional overturning circulation to cloud feedback. J. Climate, 23, 378–389, https://doi.org/10.1175/2009JCLI3118.1.
Zhao, Y., H. Yang, and Z. Liu, 2016: Assessing Bjerknes compensation for climate variability and its timescale dependence. J. Climate, 29, 5501–5512, https://doi.org/10.1175/JCLI-D-15-0883.1.