1. Introduction
The Atlantic meridional overturning circulation (AMOC) plays an essential role in the redistribution of heat globally (e.g., Gregory et al. 2005; Stouffer et al. 2006; Hu et al. 2010, 2011, 2012a, 2013a,b). Observation-based studies suggest that AMOC transports significant heat into the North Atlantic from elsewhere in the world oceans (e.g., Ganachaud and Wunsch 2000; Talley et al. 2003; Trenberth and Fasullo 2017; Trenberth et al. 2019), resulting in a relatively mild climate in western Europe. Earlier work on the dependence of the AMOC’s stability on the sea surface salinity change in the subpolar North Atlantic was done by Bryan (1986) using an ocean standalone model, and the pioneering work of Manabe and Stouffer (1988) was the first to use a coupled atmosphere–ocean model to demonstrate the bistability of the AMOC. Using a box model, Rahmstorf (1996) was the first to point out that if AMOC transports freshwater southward in the Atlantic basin, AMOC bistability is possible. Following these earlier studies, more work has been done by various authors using models with different complexities to show that the collapse (or reactivation) of the AMOC can induce significant global-scale climate changes, such as a significant cooling (or warming) in the subpolar North Atlantic and its surrounding regions [e.g., Delworth et al. 1993; Timmermann et al. 1998; Vellinga and Wood 2002; Stocker 2002; Timmermann et al. 2005a,b; Dahl et al. 2005; Zhang and Delworth 2005; Levermann et al. 2005; Schmittner et al. 2005; Broccoli et al. 2006; Stouffer et al. 2006; Hu et al. 2008, 2010, 2011, 2013a, 2015; see Weijer et al. (2019) for a review of the AMOC stability mechanisms]. On the other hand, paleo-proxy records indicate a frequent occurrence of abrupt climate change events during the last glacial period, such as Heinrich events (Heinrich 1988; Hemming 2004), Dansgaard–Oeschger events (Dansgaard et al. 1993; Ditlevsen et al. 2005), and the Younger Dryas (e.g., Rasmussen et al. 2006). The most plausible cause of these abrupt climate change events is proposed to be the bistability of AMOC (Broecker 1997).
Earlier studies using standalone ocean models or Earth system models of intermediate complexity (EMICs) have shown that a weakening (or strengthening) of the AMOC may induce a strengthening (or weakening) of the Pacific deep meridional overturning circulation (PMOC), and vice versa (e.g., Seidov and Haupt 2003, 2005; Saenko et al. 2004). Therefore, the AMOC and PMOC can form a seesaw-like variability between North Atlantic and North Pacific. Later, Stouffer et al. (2007) showed a seesaw-like variability of the AMOC and the Antarctic Bottom Water formation by adding additional freshwater into the Southern Ocean mimicking the melting water runoff from the Antarctic ice sheet. They showed that a stronger North Atlantic Deep Water formation is related to a weaker Antarctic Bottom Water formation, and vice versa. By analyzing the paleo record, Barker et al. (2009) showed an interhemispheric Atlantic seesaw response in the last deglacial period, such that a warming of Greenland corresponds to a cooling of the Antarctic, and they linked this response to the adjustment of the ocean circulation related to AMOC, a result agreeing well with Stouffer et al. (2007).
By combining newly available paleo-observations and EMIC model simulations, Okazaki et al. (2010) demonstrated that the seesaw-like AMOC and PMOC relationship is possible under paleoclimate conditions, such that a weakening of AMOC triggers a setup and strengthening of PMOC, and vice versa for a strengthening of AMOC. This seesaw-like AMOC and PMOC relationship is also simulated during Heinrich event 1 in a transient simulation from the Last Glacial Maximum to the present day using a low-resolution fully coupled model (TraCE21ka; Liu and Hu 2015). On the other hand, Hu et al. (2012b) indicated that this seesaw-like AMOC and PMOC relationship might be true only under paleoclimate conditions with a closed Bering Strait. Under modern-day conditions, the open Bering Strait can transport the freshwater added into the subpolar North Atlantic via the Arctic Ocean into the subpolar North Pacific by reversing the direction of the Bering Strait throughflow (from a northward transport to a southward transport), inducing a more stabilized surface oceanic stratification that prevents the deep convection from occurring. However, with a closed Bering Strait, the reduced precipitation associated with the cooler climate induced by a collapsed AMOC initiates the deep convection in the subpolar North Pacific, and the transport of the salty subtropical water into the subpolar North Pacific strengthens the deep convection and forms the PMOC (Hu et al. 2012a,b).
Here, by analyzing a set of specially designed sensitivity experiments using the Community Earth System Model version 1 (CESM1; Hurrell et al. 2013), we will explore the stability of the AMOC and, by extension, its influence on global-scale ocean circulation when AMOC weakens significantly or collapses under modern conditions, and also will further explore whether AMOC and PMOC form a natural pair of seesaw-like variability under modern day conditions. In these experiments, freshwater forcing is added to or subtracted from a targeted region in order to force a significant change of the AMOC and the associated changes in the global ocean. The rest of the paper is organized as follows: section 2 gives the model details and experimental design; section 3 describes our results and is followed by discussion and conclusions in section 4.
2. Model and experiments
The model used here is the CESM1 with nominal 1° horizontal resolution for all model components which is developed at the National Center for Atmospheric Research (NCAR) in collaboration with scientists at laboratories of the U.S. Department of Energy (DOE) and universities (Hurrell et al. 2013). The atmospheric component of CESM1 is the Community Atmospheric Model version 6 (CAM6; Neale et al. 2010), the ocean component is the Parallel Ocean Program version 2 (POP2; Danabasoglu et al. 2012), the land component is the Community Land Model version 4 (CLM4; Lawrence et al. 2011), and the sea ice component is the Community Sea Ice Code version 4 (CICE4; Hunke and Lipscomb 2008). The twentieth-century climate simulated by CESM1 agrees reasonably well with observations in many aspects (Meehl et al. 2013; Si and Hu 2017; Hu et al. 2018).
To study the stability of AMOC and the associated changes in global oceanic circulation, we conducted five idealized sensitivity experiments, each of which is branched at the same time point of the preindustrial control simulation of CESM1 (see Table 1 for a summary of all sensitivity experiments). In the first two experiments, freshwater is added into the subpolar North Atlantic between 50° and 70°N (orange area in Fig. 1) at a rate of 0.2 and 0.4 Sv (1 Sv ≡ 106 m3 s−1), respectively (equivalent to an additional freshwater input of 2.29 and 4.59 mm day−1 in this region; for comparison, the control run precipitation rate in this region is 2.80 mm day−1). To keep the global mean salinity unchanged, we compensate for the freshwater added to the North Atlantic by uniformly removing an equivalent amount of freshwater from the global ocean (exclusive of the North Atlantic). This freshwater compensation is accomplished by applying a negative precipitation rate of roughly −0.05 and −0.10 mm day−1 uniformly over the global ocean; for comparison, the mean precipitation rate in the control run in regions outside of the subpolar North Atlantic is 3.36 mm day−1 (cyan and light yellow areas in Fig. 1). In essence, we remove freshwater from the global ocean and add it to the subpolar North Atlantic. Note that this is equivalent to removing the same amount of salt flux from the subpolar North Atlantic and uniformly adding it elsewhere in the World Ocean. Hereafter these two experiments are called 0.2 Sv GLOBx and 0.4 Sv GLOBx, respectively, where GLOBx indicates the freshwater added to the North Atlantic is compensated by a negative freshwater flux over the global ocean, exclusive of the North Atlantic.
Freshwater hosing and compensation regions. The orange region is where freshwater forcing is added to the subpolar North Atlantic. The light yellow region in the subpolar North Pacific is where the freshwater compensation is applied in the NPAC experiments, and the cyan plus light yellow regions indicate where the freshwater is compensated for in the GLOBx experiments. The light yellow region also represents where additional salt is added in the PACSALT experiment, which was compensated as freshwater elsewhere. The red and blue lines represent the cross section at 20°N in the Pacific and Atlantic, respectively, the yellow line represents the cross section of 20°S across all ocean basins, the green line is for the cross section of Drake Passage, and the black line is for the cross section at the southern tip of Africa.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Freshwater hosing and compensation regions. The orange region is where freshwater forcing is added to the subpolar North Atlantic. The light yellow region in the subpolar North Pacific is where the freshwater compensation is applied in the NPAC experiments, and the cyan plus light yellow regions indicate where the freshwater is compensated for in the GLOBx experiments. The light yellow region also represents where additional salt is added in the PACSALT experiment, which was compensated as freshwater elsewhere. The red and blue lines represent the cross section at 20°N in the Pacific and Atlantic, respectively, the yellow line represents the cross section of 20°S across all ocean basins, the green line is for the cross section of Drake Passage, and the black line is for the cross section at the southern tip of Africa.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Freshwater hosing and compensation regions. The orange region is where freshwater forcing is added to the subpolar North Atlantic. The light yellow region in the subpolar North Pacific is where the freshwater compensation is applied in the NPAC experiments, and the cyan plus light yellow regions indicate where the freshwater is compensated for in the GLOBx experiments. The light yellow region also represents where additional salt is added in the PACSALT experiment, which was compensated as freshwater elsewhere. The red and blue lines represent the cross section at 20°N in the Pacific and Atlantic, respectively, the yellow line represents the cross section of 20°S across all ocean basins, the green line is for the cross section of Drake Passage, and the black line is for the cross section at the southern tip of Africa.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Experiments (FW is freshwater; NA is North Atlantic; NP is North Pacific).
In the second two experiments, the added freshwater forcing is the same, but the compensation region is changed from the global ocean to the subpolar North Pacific only (40°–60°N; light yellow area in Fig. 1). The negative precipitation rate in the North Pacific is equivalent to −1.15 and −2.30 mm day−1, respectively, while the mean precipitation in this region in the control run is 3.46 mm day−1. Hereafter these two experiments are called 0.2 Sv NPAC and 0.4 Sv NPAC, where NPAC indicates that the freshwater added to the subpolar North Atlantic is only compensated for in the subpolar North Pacific. The freshwater forcing is applied for 500 years in all four experiments; then the forcing is switched off and the experiments are integrated for another 300 years to watch the recovery of the AMOC.
The purpose of the first two experiments is to test whether a 0.2 or 0.4 Sv additional freshwater input into the subpolar North Atlantic would induce a collapsed AMOC. The choice for the rates of this additional freshwater input is based on previous studies showing that the AMOC will not collapse for a freshwater input at the 0.1 Sv level but will for a freshwater input at the 1 Sv level (e.g., Stouffer et al. 2006; Hu et al. 2011; Jackson et al. 2015). The second two experiments in combination with the first two experiments are used to test under what conditions PMOC will form and whether the setup of PMOC depends on where the freshwater input in the subpolar North Atlantic is compensated.
The last experiment is to test whether formation of deep convection in the North Pacific would collapse the AMOC to form a seesaw-like circulation change between the Pacific and Atlantic as suggested by a few previous studies (Saenko et al. 2004; Seidov and Haupt 2003, 2005; Okazaki et al. 2010; Hu et al. 2012b). To do this, a hypothetical negative freshwater flux of 0.4 Sv (or −2.30 mm day−1; the precipitation rate is 3.46 mm day−1 in the control run) is taken from the subpolar North Pacific (light yellow area in Fig. 1) and added uniformly to the global ocean (~0.10 mm day−1; cyan and orange areas in Fig. 1) for 250 years. Then this freshwater deficit is turned off and the experiment is run for another 100 years to see whether the changes in Pacific deep convection are stable without the freshwater deficit. Hereafter this experiment is called PACSALT, indicating that we are effectively adding salt to the subpolar North Pacific by applying a negative freshwater flux.
It is worth noting that in the above experimental design, the exact source of the added freshwater flux or salt flux and the exact processes that redistribute the compensation flux are not specifically explained. Based on previous studies, the atmospheric processes have played an important role in redistributing the global moisture via excessive evaporation over precipitation in the subtropics and the excessive precipitation over evaporation in the subpolar and tropical regions (e.g., Seidov and Haupt 2005; Haupt and Seidov 2007; Marsh et al. 2007; Singh et al. 2016). The interbasin moisture transport from the subtropical Atlantic into the Pacific via the easterly wind may have helped to maintain the sea surface salinity contrast between subpolar North Pacific and subpolar North Atlantic and, in turn, helped to maintain an active AMOC (Reagan et al. 2018) and a suppressed PMOC (Seidov and Haupt 2005; Haupt and Seidov 2007). For our experiments, with the additional freshwater flux into the subpolar North Atlantic, the source of this freshwater forcing is proposed to be the melting of the ice sheet and other land-based ice, and the compensation of this freshwater flux in other ocean basins is assumed to be increased evaporation. The salt flux added in the subpolar North Pacific in the PACSALT experiment is purely hypothetical, but can be considered as reduced atmospheric moisture transport from the tropical and subtropical oceans to the subpolar North Pacific as indicated in previous studies under different climate backgrounds in order to trigger an active PMOC (Burls et al. 2017; Wen and Yang 2020).
3. Results
a. Changes in AMOC and PMOC indices and the meridional streamfunctions
The top-left panel in Fig. 2 shows the time evolution of the AMOC, defined as the maximum of the Atlantic meridional streamfunction below 500 m depth between 80°N and 34°S, in six experiments. The mean AMOC strength in the preindustrial control run is 25.7 Sv (20 Sv at 26.5°N), which is comparable to observed estimates [e.g., 17.5 Sv at 26.5°N in Smeed et al. (2019)]. In all experiments where freshwater is added into the subpolar North Atlantic, AMOC collapses after 20–180 years of hosing (0.2 Sv GLOBx, black line; 0.4 Sv GLOBx, red line; 0.2 Sv NPAC, blue line; and 0.4 Sv NPAC, green line). The stronger the freshwater forcing, the more rapid the collapse of the AMOC. Importantly, the location of the freshwater compensation also affects the timing of the AMOC collapse: when freshwater compensation is applied globally (GLOBx), the AMOC collapses more quickly than when it is applied only in the North Pacific (NPAC). AMOC recovery initiates almost immediately after the added freshwater flux is turned off in the 0.2 Sv experiments, but shows a 100-yr delay for the 0.4 Sv experiments. The AMOC overshoots its control run strength by almost 100% within 100 years of its recovery, before falling back to its prehosing strength.
Time evolution of the AMOC and PMOC, and the transports of key straits: Bering Strait, Indonesian Throughflow, Drake Passage, and the Agulhas Current. Here positive numbers represent eastward (northward) flow and negative numbers represent westward (southward) flow.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Time evolution of the AMOC and PMOC, and the transports of key straits: Bering Strait, Indonesian Throughflow, Drake Passage, and the Agulhas Current. Here positive numbers represent eastward (northward) flow and negative numbers represent westward (southward) flow.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Time evolution of the AMOC and PMOC, and the transports of key straits: Bering Strait, Indonesian Throughflow, Drake Passage, and the Agulhas Current. Here positive numbers represent eastward (northward) flow and negative numbers represent westward (southward) flow.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
This overshoot is mainly related to salt advection. As the AMOC weakens and collapses, the ocean stratification changes in the subpolar North Atlantic with a much fresher and colder surface layer overlying a warmer and slight fresher subsurface layer (figure not shown). When the additional freshwater forcing is switched off, surface cooling becomes the dominant forcing, which leads to the initial recovery of deep convection and AMOC (Hu et al. 2012a). As the AMOC starts to recover, the transport of saltier subtropical water into the subpolar North Atlantic further weakens stratification in the slightly fresher and warmer subsurface water, leading to a significantly strengthened deep convection and the AMOC (Liu et al. 2009).
Next, we test the conditions that initiate and sustain PMOC with an open Bering Strait, and attempt to answer the more fundamental question of why there is no deep convection in the subpolar North Pacific under modern day conditions. As shown in the top-right panel (Fig. 2), PMOC does not automatically occur when AMOC weakens and collapses (the PMOC index is also defined as the maximum of the Pacific meridional streamfunction below 500 m depth). In these experiments, we found that PMOC initiation was a function of how we compensated for the added freshwater in the subpolar North Atlantic, or more physically, this depends on the changes of the surface water salinity (density) in the subpolar North Pacific since the precondition for the deep convection to occur is a dense surface layer that generates instability in the water column, as the denser surface water sinks, replacing the less dense layers below. In our GLOBx simulations (where the added freshwater is compensated globally) there is no PMOC formation, regardless of the strength of the freshwater forcing. In other words, sea surface salinity in the North Pacific is not high enough to initiate overturning (Figs. 3a–c). In fact, the sea surface salinity is actually lower in many parts of the subpolar North Pacific than that in the control run, which makes the sinking of the surface water impossible (Fig. 3). However, in the two simulations where the freshwater compensation is only in the subpolar North Pacific (NPAC), PMOC indeed sets up and maintains a strength over 20 Sv in the 0.4 Sv NPAC experiment (slightly over 10 Sv in the 0.2 Sv NPAC experiment; Table 2 and top-right panel in Fig. 2). In these two experiments, the sea surface salinity is about 1–5 salinity units higher than that of the control run (Figs. 3d,e), and this much higher salinity makes the surface water dense enough to sink to depth, triggering the initial deep convection and sustaining the strength of the PMOC. Moreover, after the initial setup of the PMOC, additional saline subtropical water is transported into the subpolar North Pacific, leading to a strengthening trend of the PMOC (Fig. 2, top right). Therefore, we conclude that a collapse of the AMOC does not induce the strengthening of the PMOC under modern climate conditions without a hypothetical addition of salt into the subpolar North Pacific (or the occurrence of severely reduced precipitation in association with reduced moisture transport from the tropics; e.g., Burls et al. 2017; Wen and Yang 2020).
(a) Climatological mean sea surface salinity in the control run and (b)–(f) the anomalies in the sensitivity runs. The left label bar is for the control run with a contour interval of 1 psu, and the right label bar is for the other panels with a contour interval of 0.5 psu. The mean for control is an 800-yr average; except for the PACSALT experiment, the means for the sensitivity experiments are averaged from years 201 to 500 (or a 300-yr mean) when AMOC is in a collapsed state (see top-left panel of Fig. 2). For the PACSALT experiment, the mean is from years 101 to 250 (or a 150-yr mean) when PMOC is healthy (see top right-panel in Fig. 2). This same average is used for Figs. 5–12.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Climatological mean sea surface salinity in the control run and (b)–(f) the anomalies in the sensitivity runs. The left label bar is for the control run with a contour interval of 1 psu, and the right label bar is for the other panels with a contour interval of 0.5 psu. The mean for control is an 800-yr average; except for the PACSALT experiment, the means for the sensitivity experiments are averaged from years 201 to 500 (or a 300-yr mean) when AMOC is in a collapsed state (see top-left panel of Fig. 2). For the PACSALT experiment, the mean is from years 101 to 250 (or a 150-yr mean) when PMOC is healthy (see top right-panel in Fig. 2). This same average is used for Figs. 5–12.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Climatological mean sea surface salinity in the control run and (b)–(f) the anomalies in the sensitivity runs. The left label bar is for the control run with a contour interval of 1 psu, and the right label bar is for the other panels with a contour interval of 0.5 psu. The mean for control is an 800-yr average; except for the PACSALT experiment, the means for the sensitivity experiments are averaged from years 201 to 500 (or a 300-yr mean) when AMOC is in a collapsed state (see top-left panel of Fig. 2). For the PACSALT experiment, the mean is from years 101 to 250 (or a 150-yr mean) when PMOC is healthy (see top right-panel in Fig. 2). This same average is used for Figs. 5–12.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The mean strength of the various transports (Sv). The mean for control is an 800-yr average; except for the PACSALT experiment, the means for the sensitivity experiments are averaged from years 201 to 500 (or a 300-yr mean) when AMOC is in a collapsed state (see top-left panel of Fig. 2). For the PACSALT experiment, the mean is from years 101 to 250 (or a 150-yr mean) when PMOC is healthy (see top-right panel in Fig. 2). The unit is Sv (Sv ≡ 106 m3 s−1). For the control run, the mean value and the standard deviations (std) are given to determine the statistical significance for the changes in the sensitivity experiments. Here we used 2 std to represent 95% significance. The bold numbers represent changes that are not statistically significant.
To better illustrate this point, we conduct an additional experiment in which additional hypothetical salt is added into the subpolar North Pacific at a rate equivalent to a negative 0.4 Sv freshwater forcing (PACSALT). This salt flux is compensated elsewhere in the World Ocean as a freshwater flux. As shown in the top two panels in Fig. 2 (orange line), PMOC reaches a strength of 15 Sv and AMOC weakens by about 6 Sv after ~70 years, then stabilizes afterward (also see Table 2). The weakening of AMOC is related to the freshwater compensation as a small amount of additional freshwater (~0.10 mm day−1 or 0.0015 Sv in subpolar North Atlantic between 50° and 70°N) that is added into the subpolar North Atlantic in the form of slightly increased precipitation. Nevertheless, the strengthening of PMOC in this experiment does induce a weakening of AMOC. When the negative freshwater flux in the North Pacific is turned off, PMOC quickly collapses and AMOC recovers (top panels in Fig. 2). The implication for this result is twofold: 1) AMOC and PMOC could form a seesaw-like variability if the PMOC could form without additional forcing to weaken the AMOC, and 2) without artificially increasing salinity in the subpolar North Pacific, a stable PMOC cannot exist. Further, the initiation of PMOC does not appear to depend on the strength of AMOC under modern-day conditions with an open Bering Strait, a result consistent with previous studies (Hu et al. 2012a,b). Moreover, the higher rate of precipitation in the subpolar North Pacific relative to the subpolar North Atlantic prevents an active PMOC from existing without additional forcing (such as artificially removing freshwater from the subpolar North Pacific). Therefore, we conclude that under modern-day conditions, the higher surface freshwater input (mainly from precipitation) in the subpolar North Pacific relative to the subpolar North Atlantic in both model simulations and observations makes the maintenance of an active PMOC impossible, and thus the mean modern-day climate favors an active AMOC and a collapsed PMOC.
A recent study using the low-resolution version of CESM1 (horizontal resolution for atmospheric and land component is nominal 3.75° × 3.75°, and for ocean and sea ice component is roughly 3°) shows that by removing the Tibetan Plateau, PMOC is established in response to the changes of atmospheric circulation, leading to enhanced Ekman convergence north of 40°N in the North Pacific and reduced moisture convergence in the western tropical Pacific (Wen and Yang 2020). Resulting from these changes, more salty water is transported into the subpolar North Pacific along with enhanced Ekman downwelling, leading to the establishment of PMOC. At the same time, more moisture stays in the Atlantic basin which leads to a collapse of the AMOC, forming a seesaw-like change between the Atlantic and Pacific basins. This result further confirmed our conclusion that under a modern-day geographic configuration, the seesaw-like change of AMOC and PMOC is very unlikely.
After the discussion of the AMOC and PMOC indices, we evaluate the meridional streamfunction in the Atlantic and Pacific basins next by comparing the mean states of the streamfunction in the control run to the changes induced by the collapse of AMOC in the sensitivity experiments (Figs. 4–6). The climatology for the control run is shown as an 800-yr average (Figs. 4–12). With the exception of the PACSALT experiment, the climatology with a collapsed AMOC for the sensitivity experiments is averaged over years 201–500 (or a 300-yr mean) based on the top-left panel of Fig. 2. For the PACSALT experiment, the climatology with an active PMOC is an average from years 101 to 250 (or a 150-yr mean) based on the top-right panel in Fig. 2. The anomalies shown in these figures are the difference between the climatology in the sensitivity experiments and that in the control run.
The (left) Atlantic and (right) Pacific meridional streamfunction in the control run and the anomalies in the 0.2 and 0.4 Sv NPAC experiments. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The (left) Atlantic and (right) Pacific meridional streamfunction in the control run and the anomalies in the 0.2 and 0.4 Sv NPAC experiments. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The (left) Atlantic and (right) Pacific meridional streamfunction in the control run and the anomalies in the 0.2 and 0.4 Sv NPAC experiments. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The upper 600 m (left) Atlantic and (right) Pacific meridional streamfunction representing the subtropical cells or the shallow meridional overturning circulation in the control run and the 0.2 and 0.4 Sv NPAC experiments. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The upper 600 m (left) Atlantic and (right) Pacific meridional streamfunction representing the subtropical cells or the shallow meridional overturning circulation in the control run and the 0.2 and 0.4 Sv NPAC experiments. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The upper 600 m (left) Atlantic and (right) Pacific meridional streamfunction representing the subtropical cells or the shallow meridional overturning circulation in the control run and the 0.2 and 0.4 Sv NPAC experiments. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The (left) Atlantic and (right) Pacific meridional streamfunction in the PACSALT experiment. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The (left) Atlantic and (right) Pacific meridional streamfunction in the PACSALT experiment. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The (left) Atlantic and (right) Pacific meridional streamfunction in the PACSALT experiment. Shading is for the mean and contours for the anomaly. Positive anomalies are solid contours and negative anomalies are dashed contours. The contour interval is 2 Sv.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
The dominant feature of the Atlantic meridional streamfunction in the control run is the double-cell structure with the upper cell representing the overturning circulation associated with the North Atlantic Deep Water (NADW) formation and the bottom cell representing the overturning circulation associated with the Antarctic Bottom Water (AABW) formation (Fig. 5a). The upper cell penetrates to over 3500 m in depth south of 15°N and to the bottom of the ocean north of 15°N, agreeing with observed estimates very well [e.g., Lozier 2012; see Frajka-Williams et al. (2019) for a review]. In the upper 100 m, there are weak and shallow subtropical cells (STCs) located north and south of the equator extending to about 30° north and south (Figs. 5a and 6a). In the Pacific, there is no obvious deep convection and the overturning circulation is dominated by the much stronger STCs in the control run (Figs. 5b and 6b). The STCs in the Pacific penetrate as deep as 1000 m in the Southern Hemisphere, but no more than 500 m in the Northern Hemisphere. This asymmetry in STC strength between the Northern and Southern Hemispheres is consistent with previous observational and modeling studies (e.g., Kuntz and Schrag 2018; Zeller et al. 2019).
When the AMOC collapses due to the additional freshwater forcing in the subpolar North Atlantic, the overturning associated with NADW ceases (Figs. 5c,e; figure not shown for 0.2 and 0.4 Sv GLOBx experiments, which are nearly identical to Figs. 5c,e, respectively), agreeing with many previous studies (e.g., Stouffer et al. 2006; Hu et al. 2008, 2011). With this collapsed AMOC, the Atlantic STC changes are smaller without active PMOC, but larger with active PMOC, suggesting that PMOC may have exerted an influence on the Atlantic STCs when the AMOC collapses (Figs. 6c,e; Table 2). Overall, the STC in the South Atlantic strengthens more in the NPAC experiments (up to about 50%), but is nearly unchanged in the GLOBx experiments, and this southern STC deepens by more than 100 m in all experiments. The northern STC is either unchanged or weakened a bit with collapsed AMOC (Figs. 6c,e; Table 2). In the Pacific, the STC in the South Pacific weakens more in the GLOBx experiments than in the NPAC experiments and the STC in the North Pacific strengthens more in the GLOBx experiments than in the NPAC experiments, suggesting that an active AMOC or PMOC strengthens the southern STC, but weakens the northern STC (0.2 and 0.4 Sv GLOBx experiments). This asymmetric change of the STCs between the Northern and Southern Hemispheres could contribute to the shift of the intertropical convergence zone (ITCZ) in response to the collapse of AMOC, which needs to be studied further. With an active PMOC, a similar overturning structure to that in the Atlantic with an active AMOC appears in the Pacific with the maximum located roughly at the same latitude (~40°N; Figs. 5d,f; 0.2 and 0.4 Sv Pacific compensation experiments) as in the Atlantic. At the same time, the STCs in the Pacific weaken and shoal to a depth less than 500 m (Figs. 6d,f), indicating that an active deep overturning circulation (e.g., AMOC or PMOC) tends to weaken shallow subtropical overturning in both Atlantic and Pacific basins.
In the PACSALT experiment (Fig. 6), a clear and robust deep overturning circulation appears in both Atlantic and Pacific basins with a stronger response in the former. In comparison to the control run (Figs. 5a,b), the AMOC is weaker and shallower in this experiment (Fig. 7a) and PMOC evolves from being absent in the control run to having decent strength in this experiment (Fig. 7b). For the STCs, they become slightly stronger and deeper in the Atlantic (Fig. 7c), but weaker and shallower in the Pacific (Fig. 7d), a change consistent with the conclusion shown in our previous paragraph.
(a) Zonal velocity at the cross section of Drake Passage in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 2 cm s−1 for the control run and 0.5 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Zonal velocity at the cross section of Drake Passage in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 2 cm s−1 for the control run and 0.5 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Zonal velocity at the cross section of Drake Passage in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 2 cm s−1 for the control run and 0.5 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
In summary, the collapse of AMOC leads to a stronger STC in the South Atlantic with a weaker one in the North Atlantic, but a weaker STC in South Pacific with a stronger one in the North Pacific. With an active PMOC, these changes become smaller in the Pacific, but larger in the Atlantic, suggesting that AMOC can exert an influence on the Pacific STCs and PMOC can also influence the Atlantic STCs. When AMOC is active, the oceanic poleward meridional heat transport in the Pacific is higher in the South Pacific, but lower in the North Pacific; however, with a collapsed AMOC, this meridional heat transport is symmetric relative to the equator in the Pacific basin (Hu et al. 2013b). With an active PMOC and a collapsed AMOC, the Pacific STCs shoal and weaken in the South Pacific, but deepen and strengthen in the North Pacific, indicating that deep overturning circulation can modulate the strength of shallow subtropical overturning circulation (STC).
b. Changes of transports in key straits in connection to overturning circulations
As AMOC is considered to be a global-scale ocean circulation (e.g., Broecker 1991; Rahmstorf 2002), significant changes in AMOC can induce significant global circulation change. These changes include but are not limited to the transports in key ocean strait. Here we focus on transports in four key places: the Bering Strait, Indonesian Throughflow, Drake Passage transport, and the Agulhas Current because the transports in these places are an important part of the global-scale ocean overturning circulation as indicated by many previous studies (e.g., Broecker 1991; Rahmstorf 2002; Hu et al. 2012a,b).
A series of earlier studies (Shaffer and Bendtsen 1994; Reason and Power 1994; de Boer and Nof 2004a,b; Hu and Meehl 2005; Hu et al. 2007, 2008, 2009, 2011, 2012a,b) have shown a strong relationship between AMOC and Bering Strait throughflow, demonstrating that a stronger AMOC leads to a stronger Bering Strait throughflow. Changes to the Bering Strait throughflow shown in this study (middle-left panel of Fig. 2) agree with these previous assessments. The mean strength of the Bering Strait throughflow is 0.91 Sv in the CESM1 control run, agreeing reasonably well with observations (~0.8 Sv; Aagaard and Carmack 1989; Woodgate and Aagaard 2005). As AMOC weakens in all experiments, Bering Strait throughflow weakens dramatically to below 0.2 Sv in the GLOBx experiments, even reversing direction in the NPAC experiments (−0.18 Sv in 0.2 Sv NPAC; −0.64 Sv in 0.4 Sv NPAC; blue and green lines, respectively in the middle-left panel of Fig. 2; Table 2), consistent with previous studies (e.g., Hu et al. 2007, 2008, 2012a). Moreover, the setup of an active PMOC induces a strong reversal at the Bering Strait by inverting the driving force of this throughflow, the sea level difference between the North Pacific and North Atlantic (Shaffer and Bendtsen 1994), from higher sea level in the former to higher sea level in the latter (Hu et al. 2008). As indicated by previous studies (e.g., Shaffer and Bendtsen 1994; Hu et al. 2008), Bering Strait throughflow is dynamically driven by the sea level difference between North Pacific and North Atlantic. The lower density in the North Pacific is primarily caused by lower salinity leads to a higher sea level and higher density in the North Atlantic in association with higher salinity that leads to a low sea level in both the model control simulation and observations (e.g., the mean sea level in the subpolar North Pacific is about 0.7 m higher than that in subpolar North Atlantic). As the sea level becomes lower in the subpolar North Pacific with an active PMOC, more freshwater is diverted into the Pacific, slowing the collapse of AMOC in these experiments (0.2 and 0.4 Sv NPAC) relative to those without an active PMOC (0.2 and 0.4 Sv GLOBx) as noted in the previous section.
The Indonesian Throughflow is an important oceanic current connecting the Pacific and Indian Oceans and plays an important role in modulating the regional climate in the western Pacific and Indian Oceans (Schneider 1998; Song et al. 2007; Feng et al. 2018). As indicated by Broecker (1991) and Rahmstorf (2002), this throughflow is also part of the global-scale overturning circulation. The mean transport of this throughflow in the CESM1 control run is 11.8 Sv, which agrees with observations reasonably well [~15 Sv; see Feng et al. (2018) for details and the references therein]. As AMOC weakens, the Indonesian Throughflow is reduced by 25% (3 Sv) when PMOC is absent; the reduction is much larger (50%–75%, or 6.3–9.2 Sv) when PMOC is active (middle-right panel of Fig. 2; Table 2). This result suggests that at least 25% of the total Indonesian Throughflow transport could be attributed to AMOC. When the recovering AMOC overshoots its control run strength, the Indonesian Throughflow also overshoots its control run value by roughly 10%–20%. Furthermore, PMOC can also affect the strength of Indonesian Throughflow by altering surface currents, such as setting up a pattern of northward cross-equator flow in the western Pacific (figure not shown).
Drake Passage is a key strait between the southern tip of South America and the South Shetland Islands of Antarctica. Observations indicate that the transport through this passage ranges from 134 to 173 Sv (e.g., Whitworth 1983; Whitworth and Peterson 1985; Cunningham et al. 2003; Koenig et al. 2014; Chidichimo et al. 2014; Donohue et al. 2016) and is commonly referenced as 134 Sv. This transport is used to represent the strength of the major current in the southern oceans, the Antarctic Circumpolar Current (ACC). In the CESM1 control simulation, the strength of the Drake Passage transport is 154.5 Sv, which is on the high side of the commonly referenced observed value (although in the middle of the observed value range), but similar to a recent high-resolution model estimate (157.3 Sv) reported by Xu et al. (2020), who argue that this might be the true ACC transport. Further decomposition of the Drake Passage transport shows a westward flowing water transport of 3 Sv and a total eastward flow of 157.5 Sv, which is nearly identical to the Xu et al. (2020) estimate. As AMOC weakens, the Drake Passage transport increases up to 35 Sv (~23%, with a mean of ~33 Sv) in the two GLOBx simulations (without an active PMOC), but is only 15–25 Sv in the two simulations with the setup of PMOC (bottom-left panel of Fig. 2 with a mean of ~14 and ~23 Sv for the 0.2 Sv and 0.4 Sv NPAC, respectively; Table 2), suggesting that active AMOC and PMOC both impede the strength of the Drake Passage transport (or ACC strength). This effect can be clearly seen in the PACSALT experiment. The setup of PMOC and the weakening of AMOC in this experiment result in a nearly unchanged Drake Passage transport (Table 2) due to the competing effects (i.e., a weaker AMOC induces a stronger Drake Passage transport, but a stronger PMOC induces a weaker Drake Passage transport).
The Agulhas Current is a key current transporting Indian Ocean water into the Atlantic basin and is located along the southeast coast to the southern tip of Africa. It is also an important branch of the AMOC upper limb (e.g., Broecker 1991; Rahmstorf 2002). Observations indicate a westward transport of this current of roughly 84 Sv (Beal et al. 2015) with an earlier estimation of 70 Sv (Bryden et al. 2005). This transport in the CESM1 control run (measured as the westward flow from the southern tip of Africa to the northern edge of the eastward flowing ACC) is 82.4 Sv, which is nearly equivalent to the more recent estimate of Beal et al. (2015). With a collapsed AMOC, the Agulhas Current weakens by 10–14 Sv (or 12%–17%) with a slightly higher reduction in the simulations without an active PMOC (bottom-right panel of Fig. 2; Table 2). This weakening is less for the PACSALT experiment (6 Sv) due to a much smaller change in AMOC (a reduction of 6.5 Sv in PACSALT versus a reduction of 21–23 Sv in the other four experiments; Table 2). Therefore, the Agulhas Current is possibly similar to the Gulf Stream, which is a combined wind-driven western boundary current and thermohaline-driven overturning circulation.
In summary, the collapse of AMOC induces a weakened transport at the Bering Strait, Indonesian Throughflow, and Agulhas Current, but strengthened transport for Drake Passage. The active PMOC further weakens the transport at the Bering Strait and Indonesian Throughflow, and also weakens the transport at Drake Passage, but its influence on the Agulhas Current is small.
c. Transport changes at selected cross sections in association with AMOC/PMOC
A recent review paper by Bower et al. (2019) acknowledges that while AMOC is generally considered to be a three-dimensional global scale circulation, it is frequently simplified as a one-dimensional index. In section 3a, we discuss AMOC in one and two dimensions, while section 3b hints at the three-dimensionality of AMOC. In this section, we expand on our discussion of AMOC as a three-dimensional variable by analyzing the ocean transports at a few select cross sections to explore how changes in AMOC strength influence global-scale ocean circulation: 1) the Drake Passage (green vertical line in Fig. 1), 2) the southern tip of Africa (black vertical line in Fig. 1), 3) 20°S across all ocean basins (yellow horizontal line in Fig. 1), and 4) 20°N in the Pacific and Atlantic (red and blue horizontal lines, respectively, in Fig. 1) in Figs. 7 and 10–13.
For the Drake Passage cross section, the major core of the ACC (Drake Passage transport) is located at the northern half of this passage with a maximum velocity of more than 30 cm s−1 with a much weaker minor core at about 63.5°S (maximum velocity about 6 cm s−1; Fig. 7a). As AMOC collapses, the major core of the ACC weakens in the upper few hundred meters of depth, but strengthens at depth with a maximum strengthening of 3–5 cm s−1. At the same time, the minor core (located farther south) is nearly doubles its strength (centered around 63.5°S; Figs. 7b–e). Comparing simulations with and without an active PMOC, the weakening of the upper portion of the ACC is greater (by more than 5 cm s−1; Figs. 7b,c vs Figs. 7d,e) and the strengthening of the lower portion of the ACC is less pronounced for the major ACC core (by a couple of centimeters per second), but the strengthening of the minor core located farther south is stronger (by a few centimeters per second) for the simulations with an active PMOC than those without an active PMOC. As shown in Figs. 8 and 9 and Chen et al. (2019), the wind stress anomaly in the Southern Ocean strengthens roughly south of 55°S, and slightly weakens north of 55°S leading to the changes of the Drake Passage flow pattern. In the PACSALT experiment with active AMOC and PMOC, we see a dipole pattern of ACC velocity changes for the major core of ACC with a weakening in the upper and northern portion and a strengthening farther south (Fig. 7f), which is also associated with the much weaker changes of the westerlies in the southern oceans roughly south of 55°S (Figs. 8f and 9) relative to the other four experiments.
(a) Wind stress and surface temperature climatology for control run and (b)–(f) the anomalies for the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 5 K for the control run and 0.5 K for the positive anomalies, but 2.5 K for the negative anomalies. The reference vector for the control run is 0.1 N m−2, but 0.01 N m−2 for the wind stress anomalies in the sensitivity experiments.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Wind stress and surface temperature climatology for control run and (b)–(f) the anomalies for the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 5 K for the control run and 0.5 K for the positive anomalies, but 2.5 K for the negative anomalies. The reference vector for the control run is 0.1 N m−2, but 0.01 N m−2 for the wind stress anomalies in the sensitivity experiments.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Wind stress and surface temperature climatology for control run and (b)–(f) the anomalies for the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 5 K for the control run and 0.5 K for the positive anomalies, but 2.5 K for the negative anomalies. The reference vector for the control run is 0.1 N m−2, but 0.01 N m−2 for the wind stress anomalies in the sensitivity experiments.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Zonal-mean zonal wind stress in the control run (brown line) and the anomalies in the sensitivity experiments. For the control run, it is an 800-yr mean, and for the sensitivity runs, it is the mean of years 201–500 for the GLOBx and NPAC experiments when AMOC is collapsed, and it is the mean of years 101–250 for PACSALT experiment when both AMOC and PMOC are active.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Zonal-mean zonal wind stress in the control run (brown line) and the anomalies in the sensitivity experiments. For the control run, it is an 800-yr mean, and for the sensitivity runs, it is the mean of years 201–500 for the GLOBx and NPAC experiments when AMOC is collapsed, and it is the mean of years 101–250 for PACSALT experiment when both AMOC and PMOC are active.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Zonal-mean zonal wind stress in the control run (brown line) and the anomalies in the sensitivity experiments. For the control run, it is an 800-yr mean, and for the sensitivity runs, it is the mean of years 201–500 for the GLOBx and NPAC experiments when AMOC is collapsed, and it is the mean of years 101–250 for PACSALT experiment when both AMOC and PMOC are active.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
In the cross section near the tip of South Africa, there are three sections: a westward flow at the north side (representing the Agulhas Current; north of 38.5°S) with a transport of 82.4 Sv (Table 2), an eastward flow farther south (representing ACC and the return flow of the Agulhas Current; roughly between 38.5° and 55°S) with a mean transport of 270 Sv, and a westward flow even farther south with a mean transport of 32.1 Sv for the control run (Fig. 10a). Therefore, the net eastward flow due to the ACC is 155.5 Sv. Recalling the 154.5 Sv transport from the Drake Passage and 0.9 Sv from Bering Strait (Table 2), the total volume transport is balanced in the Atlantic–Arctic basin. As AMOC collapses in the sensitivity experiments, the changes of the flow pattern are more complicated in this cross section than those at Drake Passage (Figs. 10b–e) due to the more complicated wind patterns relative to the control run (Fig. 9a), such as a stronger westerly north of 45°S, and a weaker westerly between 45° and 70°S, and a stronger easterly south of 70°S. In this cross section, the ACC major core around 51°S weakens at nearly all depths and the ACC strengthens farther south and in the middepth north of 47°S. Overall, the Agulhas Current weakens and shifts offshore with a strengthening of the total ACC transports and a weakening of the westward transport south of 55°S. These changes are weaker in the experiments with an active PMOC (Figs. 10d,e) than in those without (Figs. 10b,c). The pattern of flow changes in the PACSALT experiment (Fig. 10f) is very similar to that in other experiments (Figs. 10b–e), but with a much smaller amplitude. These changes are also related to changes in the Southern Ocean westerlies with a weakening in the north and strengthening south of 55°S (Figs. 10b–f).
(a) Zonal velocity at the cross section at the southern tip of Africa in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 2 cm s−1 for the control run and 0.5 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Zonal velocity at the cross section at the southern tip of Africa in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 2 cm s−1 for the control run and 0.5 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Zonal velocity at the cross section at the southern tip of Africa in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 2 cm s−1 for the control run and 0.5 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
At the cross section of 20°S, the western boundary currents are very clear in the control run, such as the southward currents east of Africa and Australia (Fig. 11a) in the control run. In the Atlantic basin, there is a northward flow in the upper 1000 m east of the South American continent in association with the upper limb of the AMOC, and a deep western boundary current flowing southward below 1000 m depth, representing the lower limb of the AMOC. In the ocean interior, the flow in the upper ocean is generally northward in all ocean basins, reflecting the return flow of the subtropical gyres, and there are no strong deep currents. With a collapsed AMOC, all of the western boundary currents, deep or shallow, become weaker, especially in the Atlantic basin, which has a total absence of the deep western boundary current (Figs. 11b–e). Interestingly, the southward shallow western boundary current east of Australia reverses direction from south to northward, strengthening and deepening to a depth of more than 1000 m. These reversals may be related to changes in Indonesian throughflow due to the collapse of AMOC and need further study in the future. With an active PMOC, there is a stronger deep western boundary current slightly east of the international date line with a depth greater than 1000 m, which may represent the deep return flow of the PMOC. In the PACSALT experiment, the changes in all three ocean basins are smaller, with the exception that the southward flow east of Australia becomes deeper and stronger at depth (Fig. 11f), indicating that the weakening of the northward flow east of Australia in the experiments with an active PMOC and collapsed AMOC is due to the setup of PMOC (Figs. 11d,e vs Figs. 11b,c).
(a) Meridional velocity at the cross section of 20°S in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Meridional velocity at the cross section of 20°S in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Meridional velocity at the cross section of 20°S in the control run and (b)–(f) the anomalies in the sensitivity experiments. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Next, we discuss the changes to meridional flow farther north, at 20°N in the Atlantic and Pacific basins (Figs. 12 and 13). In the control run, the flow pattern in the upper 1000 m in the North Atlantic is mainly the northward flowing western boundary current (Gulf Stream) and the southward flowing interior flow representing the return flow of the subtropical gyre (Fig. 12a). Below 1000 m depth, the deep returning flow of the lower limb of the AMOC is located at east of 120°W with two centers. With a collapsed AMOC, the changes to the above mentioned flow pattern are nearly identical in the four experiments with additional freshwater added into the subpolar North Atlantic (Figs. 12b–e), such as a weakened northward flowing boundary current (associated with Gulf Stream) off the Cuban coast and the weakened southward flowing deep return flow at a depth below 1000 m (between 70° and 60°W). With an active PMOC and a weakened AMOC in the PACSALT experiment, changes to the flow pattern are also the same as in the other experiments but with much weaker amplitude (Fig. 12f), suggesting that the establishment of PMOC does not have a large effect on the flow pattern in the Atlantic. In the Pacific, the upper ocean flow pattern is very similar to that in the Atlantic with a strong northward flowing western boundary current representing the Kuroshio and a weak southward flowing interior and there are no obvious strong deep currents in the control run (Fig. 13a). With a collapsed AMOC and the absence of PMOC, the surface part of the Kuroshio weakens and the deep part strengthens with a slightly strengthened interior flow (Figs. 13b,c). The weakening of the surface part of the Kuroshio may be related to the weakening of the Bering Strait throughflow in these experiments. With an active PMOC, the Kuroshio strengthens in the upper 1000 m depth, and a deep southward flowing boundary current appears at a depth below 1000 m (Figs. 13d,e). A similar change in flow pattern is also shown in the PACSALT experiment with an active PMOC (Fig. 13f). These changes in the Pacific are consistent with the establishment of the PMOC, which transports more upper ocean water into the subpolar region where this water sinks to depth and induces a deep southward return flow similar to that in the Atlantic when AMOC is active.
(a) Meridional velocity at the cross section of 20°N in the control run and (b)–(f) the anomalies in the sensitivity experiments in the Atlantic. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Meridional velocity at the cross section of 20°N in the control run and (b)–(f) the anomalies in the sensitivity experiments in the Atlantic. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Meridional velocity at the cross section of 20°N in the control run and (b)–(f) the anomalies in the sensitivity experiments in the Atlantic. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Meridional velocity at the cross section of 20°N in the control run and (b)–(f) the anomalies in the sensitivity experiments in the Pacific. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Meridional velocity at the cross section of 20°N in the control run and (b)–(f) the anomalies in the sensitivity experiments in the Pacific. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
(a) Meridional velocity at the cross section of 20°N in the control run and (b)–(f) the anomalies in the sensitivity experiments in the Pacific. The left label bar is for (a) and the right label bar is for (b)–(f). The contour interval is 0.5 cm s−1 for the control run and 0.25 cm s−1 for the anomalies.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
d. Changes of the surface climate
As shown in Fig. 8, the basic structure of surface climate simulated in the control run is similar to the observations (e.g., Meehl et al. 2013; Hurrell et al. 2013, and the references therein). In response to the collapse of AMOC, there is a cooling in the Northern Hemisphere and a warming in the Southern Hemisphere with the most significant cooling in the subpolar North Atlantic and the Nordic seas (Figs. 8b–e), agreeing with many previous studies (e.g., Stouffer et al. 2006; Hu et al. 2011; Jackson et al. 2015). This pattern of surface climate change is a direct result of the collapsed AMOC, which leads to a reduced northward meridional heat transport in the Atlantic basin (e.g., Rahmstorf 2002; Stouffer et al. 2006; Hu et al. 2010, 2011, 2012a, 2015). With the establishment of PMOC, there is warming in the subpolar and eastern subtropical North Pacific; this warming signal is stronger when PMOC is stronger (Figs. 8d,e), a feature that has been shown in Hu et al. (2012a,b) with a closed Bering Strait. Note that in experiments with a closed Bering Strait, the collapse of AMOC will naturally induce the establishment of PMOC and does not require additional salt flux (Hu et al. 2012a,b). In the PACSALT experiment (Fig. 8f), there is an opposite change in surface climate relative to the experiments with a collapsed AMOC and an inactive PMOC (Figs. 8b,c). Although weakened AMOC in this experiment reduces the northward meridional heat transport in the Atlantic basin, which leads to a potential cooling of the North Atlantic, the increased northward meridional heat transport by the establishment of PMOC in the Pacific basin is significantly larger than the reduction in the Atlantic, resulting in a net increase of the northward meridional heat transport and a warming in almost the entire Northern Hemisphere (except the subpolar North Atlantic, where the small cooling is caused by the weakened AMOC) and a cooling in the Southern Hemisphere.
Changes in AMOC/PMOC induce changes not only in surface temperature, but also in surface wind patterns (and changes in ocean circulation as discussed in prior sections). As mentioned earlier, a collapse of AMOC leads to a strengthening of the westerlies south of 55°S and a weakening between 55° and 35°S, and a consequent southward shift of the ACC (Figs. 8b–e). This implies that not only the Southern Ocean winds can affect the AMOC mean state (e.g., Buizert and Schmittner 2015), but also the mean state changes in AMOC can also affect the Southern Ocean winds. In the Pacific basin, the easterly winds strengthen between 40° and 30°S, and between the equator and 30°N (except the northwest portion of the Pacific), but weaken between the equator and 30°S, and the westerlies strengthen between 30° and 60°N. Similar changes occur in the Atlantic basin. In the Indian Ocean basin, the annual mean monsoon wind is dominated by the boreal summer monsoon, such as the southeasterly over the southern Indian Ocean and the southwesterly over the northern Indian Ocean (Fig. 8a). With a collapsed AMOC, there is a weakening of the monsoon wind (especially the summer monsoon) since the anomalous winds go against the boreal summer wind pattern over the Indian Ocean. This change is mainly associated with the reduced temperature contrast between India and Australia. These changes are generally weaker in simulations with an active PMOC (Figs. 8d,e vs Figs. 8b,c). The effect of PMOC on the wind pattern changes can be seen more clearly in the PACSALT experiment (Fig. 8f). With an active PMOC (and a weakened AMOC), the westerlies weaken in most parts of the Southern Ocean. The easterlies in the Northern Hemisphere weaken, but strengthen in the Southern Hemisphere in both Pacific and Atlantic basins, accompanied by a weakening of the westerlies in the northern midlatitudes. By increasing the temperature contrast between India and Australia, the monsoon circulation strengthens in the Indian Ocean (especially the summer monsoon).
4. Discussion and conclusions
We have studied the influence of a collapsed AMOC on global ocean circulation and surface climate and explored the hypothesis that AMOC and PMOC form a natural seesaw-like variability, such that a weakening (strengthening) of AMOC induces a strengthening (weakening) of PMOC, and vice versa. As shown in the schematic in Fig. 14, there are four possible states: 1) active AMOC with an inactive PMOC, which represents the oceanic mean state under modern conditions (Fig. 14a; control run); 2) inactive AMOC and PMOC, which could be a state under future warmer climate condition (Fig. 14b; GLOBx experiments); 3) active PMOC with an inactive AMOC (Fig. 14c; NPAC experiments), a condition that might be true during Heinrich events (e.g., Okazaki et al. 2010; Liu and Hu 2015); and 4) active AMOC and PMOC (Fig. 14d; PACSALT experiment), a condition that may be true during the Pliocene (Burls et al. 2017; Thomas et al. 2021).
Schematic plot for different AMOC and PMOC combinations: (a) the control run state with active AMOC but inactive PMOC, (b) the GLOBx experiments with inactive AMOC and PMOC, (c) the NPAC experiments with inactive AMOC but active PMOC, and (d) the PACSALT experiment with active AMOC and PMOC. The arrows represent the flow direction and the colors represent roughly the water mass temperature. The purple color represents the deep water and the red to blue color represents the upper-ocean water. Note that (a) is a reproduction of the Fig. 2 available at https://rapid.ac.uk/background.php (National Oceanography Center) using Python.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Schematic plot for different AMOC and PMOC combinations: (a) the control run state with active AMOC but inactive PMOC, (b) the GLOBx experiments with inactive AMOC and PMOC, (c) the NPAC experiments with inactive AMOC but active PMOC, and (d) the PACSALT experiment with active AMOC and PMOC. The arrows represent the flow direction and the colors represent roughly the water mass temperature. The purple color represents the deep water and the red to blue color represents the upper-ocean water. Note that (a) is a reproduction of the Fig. 2 available at https://rapid.ac.uk/background.php (National Oceanography Center) using Python.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Schematic plot for different AMOC and PMOC combinations: (a) the control run state with active AMOC but inactive PMOC, (b) the GLOBx experiments with inactive AMOC and PMOC, (c) the NPAC experiments with inactive AMOC but active PMOC, and (d) the PACSALT experiment with active AMOC and PMOC. The arrows represent the flow direction and the colors represent roughly the water mass temperature. The purple color represents the deep water and the red to blue color represents the upper-ocean water. Note that (a) is a reproduction of the Fig. 2 available at https://rapid.ac.uk/background.php (National Oceanography Center) using Python.
Citation: Journal of Climate 34, 18; 10.1175/JCLI-D-21-0119.1
Our results further indicate that significant changes in AMOC affect ocean circulation not only in the Atlantic basin, but also globally. For example, a collapsed AMOC weakens the Bering Strait and Indonesian throughflows and the Agulhas Current, but strengthens the ACC (measured as the Drake Passage transport). The collapse of AMOC would not automatically cause a strengthening of the PMOC unless additional salt flux is added into the subpolar North Pacific under modern conditions. This is due to two reasons: 1) a much higher precipitation (thus higher net freshwater input) in the subpolar North Pacific than that in the subpolar North Atlantic and 2) the weaker AMOC inducing a weakening (even reversing direction) of the Bering Strait throughflow, which leads to a reduced freshwater being transported into the Atlantic from the North Pacific via the Bering Strait (and thus more freshwater being kept in the subpolar North Pacific), leading to a much fresher surface water that prevents the deep convection to occur (Hu et al. 2012a,b).
Our experiments also show that by hypothetically adding additional salt input in the subpolar North Pacific, PMOC can set up and this formation of PMOC does induce a weakening of AMOC if there is additional freshwater input into the subpolar North Atlantic, leading to a seesaw-like change between AMOC and PMOC (e.g., the AMOC and PMOC changes in PACSALT experiment). Further investigation using our experiments indicates that under modern-day climate conditions, the states of active PMOC or inactive AMOC are not stable states without the additional salt input in the subpolar North Pacific or additional freshwater flux into the subpolar North Atlantic because without these forcings, PMOC collapses and AMOC reactivates. This result implies that although AMOC and PMOC are capable of generating global-scale climate changes, these changes due to either AMOC or PMOC or both are not strong enough to fundamentally change the dynamics of atmospheric moisture transport, ocean freshwater redistribution, and local air–sea interactions. Once the perturbation is removed, the AMOC and PMOC will go back to their preferred state, and under modern-day conditions, these are the active AMOC and inactive PMOC states. Therefore AMOC and PMOC do not form a natural seesaw under modern climate and geographic conditions.
Acknowledgments
We thank Val Byfield and Meric Srokosz from the UK National Oceanography Centre for providing the editable PDF version of the AMOC schematic figure. Portions of this study were supported by the Regional and Global Model Analysis (RGMA) component of the Earth and Environmental System Modeling Program of the U.S. Department of Energy’s Office of Biological and Environmental Research (BER) via National Science Foundation IA 1844590. The National Center for Atmospheric Research is sponsored by National Science Foundation of the United States of America. Computing resources (ark:/85065/d7wd3xhc) were partially provided by the Climate Simulation Laboratory at NCAR’s Computational and Information Systems Laboratory, sponsored by the National Science Foundation and other agencies. This research also used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract DE-AC02-05CH11231.
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.
Barker, S., P. Diz, M. J. Vautravers, J. Pike, G. Knorr, I. R. Hall, and W. S. Broecker, 2009: Interhemispheric Atlantic seesaw response during the last deglaciation. Nature, 457, 1097–1102, https://doi.org/10.1038/nature07770.
Beal, L. M., S. Elipot, A. Houk, and G. M. Leber, 2015: Capturing the transport variability of a western boundary jet: Results from the Agulhas Current time-series experiment (ACT). J. Phys. Oceanogr., 45, 1302–1324, https://doi.org/10.1175/JPO-D-14-0119.1.
Bower, A., and Coauthors, 2019: Lagrangian views of the pathways of the Atlantic meridional overturning circulation. J. Geophys. Res. Oceans, 124, 5313–5335, https://doi.org/10.1029/2019JC015014.
Broccoli, A. J., K. A. Dahl, and R. J. Stouffer, 2006: Response of the ITCZ to Northern Hemisphere cooling. Geophys. Res. Lett., 33, L01702, https://doi.org/10.1029/2005GL024546.
Broecker, W. S., 1991: The great ocean conveyor. Oceanography, 4, 79–89, https://doi.org/10.5670/oceanog.1991.07.
Broecker, W. S., 1997: Thermohaline circulation, the Achilles heel of our climate system: Will man-made CO2 upset the current balance? Science, 278, 1582–1588, https://doi.org/10.1126/science.278.5343.1582.
Bryan, F., 1986: High-latitude salinity effects and interhemispheric thermohaline circulations. Nature, 323, 301–304, https://doi.org/10.1038/323301a0.
Bryden, H. L., L. M. Beal, and L. M. Duncan, 2005: Structure and transport of the Agulhas Current and its temporal variability. J. Oceanogr., 61, 479–492, https://doi.org/10.1007/s10872-005-0057-8.
Buizert, C., and A. Schmittner, 2015: Southern Ocean control of glacial AMOC stability and Dansgaard–Oeschger interstadial duration. Paleoceanography, 30, 1595–1612, https://doi.org/10.1002/2015PA002795.
Burls, N. J., A. V. Fedorov, D. M. Sigman, S. L. Jaccard, R. Tiedemann, and G. H. Haug, 2017: Active Pacific meridional overturning circulation (PMOC) during the warm Pliocene. Sci. Adv., 3, e1700156, https://doi.org/10.1126/sciadv.1700156.
Chen, C., W. Liu, and G. Wang, 2019: Understanding the uncertainty in the 21st century dynamic sea level projections: The role of the AMOC. Geophys. Res. Lett., 46, 210–217, https://doi.org/10.1029/2018GL080676.
Chidichimo, M. P., K. A. Donohue, R. D. Watts, and K. L. Tracey, 2014: Baroclinic transport time series of the Antarctic Circumpolar Current measured in Drake Passage. J. Phys. Oceanogr., 44, 1829–1853, https://doi.org/10.1175/JPO-D-13-071.1.
Cunningham, S. A., S. G. Alderson, B. A. King, and M. A. Brandon, 2003: Transport and variability of the Antarctic Circumpolar Current in Drake Passage. J. Geophys. Res. Oceans, 108, 8084, https://doi.org/10.1029/2001JC001147.
Dahl, K. A., A. J. Broccoli, and R. J. Stouffer, 2005: Assessing the role of North Atlantic freshwater forcing in millennial scale climate variability: A tropical Atlantic perspective. Climate Dyn., 24, 325–346, https://doi.org/10.1007/s00382-004-0499-5.
Danabasoglu, G., S. C. Bates, B. P. Briegleb, S. R. Jayne, M. Jochum, W. G. Large, S. Peacock, and S. G. Yeager, 2012: The CCSM4 ocean component. J. Climate, 25, 1361–1389, https://doi.org/10.1175/JCLI-D-11-00091.1.
Dansgaard, W., and Coauthors, 1993: Evidence for general instability of past climate from a 250-kyr ice-core record. Nature, 364, 218–220, https://doi.org/10.1038/364218a0.
de Boer, A. M., and D. Nof, 2004a: The exhaust valve of the North Atlantic. J. Climate, 17, 417–422, https://doi.org/10.1175/1520-0442(2004)017<0417:TEVOTN>2.0.CO;2.
de Boer, A. M., and D. Nof, 2004b: The Bering Strait’s grip on the Northern Hemisphere climate. Deep-Sea Res. I, 51, 1347–1366, https://doi.org/10.1016/j.dsr.2004.05.003.
Delworth, T. L., S. Manabe, and R. J. Stouffer, 1993: Interdecadal variations of the thermohaline circulation in a coupled ocean–atmosphere model. J. Climate, 6, 1993–2011, https://doi.org/10.1175/1520-0442(1993)006<1993:IVOTTC>2.0.CO;2.
Ditlevsen, P. D., M. S. Kristensen, and K. K. Andersen, 2005: The recurrence time of Dansgaard–Oeschger events and limits on the possible periodic component. J. Climate, 18, 2594–2603, https://doi.org/10.1175/JCLI3437.1.
Donohue, K. A., K. L. Tracey, D. R. Watts, M. P. Chidichimo, and T. K. Chereskin, 2016: Mean Antarctic Circumpolar Current transport measured in Drake Passage. Geophys. Res. Lett., 43, 11 760–11 767, https://doi.org/10.1002/2016GL070319.
Feng, M., N. Zhang, Q. Liu, and S. Wijffels, 2018: The Indonesian throughflow, its variability and centennial change. Geosci. Lett., 5, 3, https://doi.org/10.1186/s40562-018-0102-2.
Frajka-Williams, E., and Coauthors, 2019: Atlantic meridional overturning circulation: Observed transport and variability. Front. Mar. Sci., 6, 260, https://doi.org/10.3389/fmars.2019.00260.
Ganachaud, A., and C. Wunsch, 2000: Improved estimates of global ocean circulation, heat transport and mixing from hydrographic data. Nature, 408, 453–457, https://doi.org/10.1038/35044048.
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.
Haupt, B. J., and D. Seidov, 2007: Strengths and weaknesses of the global ocean conveyor: Inter-basin freshwater disparities as the major control. Prog. Oceanogr., 73, 358–369, https://doi.org/10.1016/j.pocean.2006.12.004.
Heinrich, H., 1988: Origin and consequences of cyclic ice rafting in the northeast Atlantic Ocean during the past 130,000 years. Quat. Res., 29, 142–152, https://doi.org/10.1016/0033-5894(88)90057-9.
Hemming, S. R., 2004: Heinrich events: Massive late Pleistocene detritus layers of the North Atlantic and their global climate imprint. Rev. Geophys., 42, RG1005, https://doi.org/10.1029/2003RG000128.
Hu, A., and G. A. Meehl, 2005: Bering Strait throughflow and the thermohaline circulation. Geophys. Res. Lett., 32, L24610, https://doi.org/10.1029/2005GL024424.
Hu, A., G. A. Meehl, and W. Han, 2007: Role of the Bering Strait in the thermohaline circulation and abrupt climate change. Geophys. Res. Lett., 34, L05704, https://doi.org/10.1029/2006GL028906.
Hu, A., B. L. Otto-Bliesner, G. A. Meehl, W. Han, C. Morrill, E. C. Brady, and B. Briegleb, 2008: Response of thermohaline circulation to freshwater forcing under present day and LGM conditions. J. Climate, 21, 2239–2258, https://doi.org/10.1175/2007JCLI1985.1.
Hu, A., G. A. Meehl, W. Han, and J. Yin, 2009: Transient response of the MOC and climate to potential melting of the Greenland Ice Sheet in the 21st century. Geophys. Res. Lett., 36, L10707, https://doi.org/10.1029/2009GL037998.
Hu, A., and Coauthors, 2010: Influence of Bering Strait flow and North Atlantic circulation on glacial sea level changes. Nat. Geosci., 3, 118–121, https://doi.org/10.1038/ngeo729.
Hu, A., G. A. Meehl, W. Han, and J. Yin, 2011: Effect of the potential melting of the Greenland Ice Sheet on the meridional overturning circulation and global climate in the future. Deep-Sea Res. II, 58, 1914–1926, https://doi.org/10.1016/j.dsr2.2010.10.069.
Hu, A., and Coauthors, 2012a: Role of the Bering Strait on the hysteresis of the ocean conveyor belt circulation and glacial climate stability. Proc. Natl. Acad. Sci. USA, 109, 6417–6422, https://doi.org/10.1073/pnas.1116014109.
Hu, A., G. A. Meehl, W. Han, A. Abe-Ouchi, C. Morrill, Y. Okazaki, and M. O. Chikamoto, 2012b: The Pacific–Atlantic seesaw and the Bering Strait. Geophys. Res. Lett., 39, L03702, https://doi.org/10.1029/2011GL050567.
Hu, A., G. A. Meehl, W. Han, J. Yin, B. Wu, and M. Kimoto, 2013a: Influence of continental ice retreat on future global climate. J. Climate, 26, 3087–3111, https://doi.org/10.1175/JCLI-D-12-00102.1.
Hu, A., G. A. Meehl, W. Han, J. Lu, and W. G. Strand, 2013b: Energy balance in a warm world without the ocean conveyor belt and sea ice. Geophys. Res. Lett., 40, 6242–6246, https://doi.org/10.1002/2013GL058123.
Hu, A., G. A. Meehl, W. Han, B. Otto-Bliesner, A. Abe-Ouchi, and N. Rosenbloom, 2015: Effects of the Bering Strait closure on AMOC and global climate under different background climates. Prog. Oceanogr., 132, 174–196, https://doi.org/10.1016/j.pocean.2014.02.004.
Hu, Z., A. Hu, and Y. Hu, 2018: Contributions of interdecadal Pacific oscillation and Atlantic multidecadal oscillation to global ocean heat content distribution. J. Climate, 31, 1227–1244, https://doi.org/10.1175/JCLI-D-17-0204.1.
Hunke, E., and W. Lipscomb, 2008: CICE: The Los Alamos Sea Ice Model user’s manual, version 4. Los Alamos National Laboratory Tech. Rep., LA-CC-06-012, 68 pp.
Hurrell, J. W., and Coauthors, 2013: The Community Earth System Model: A framework for collaborative research. Bull. Amer. Meteor. Soc., 94, 1339–1360, https://doi.org/10.1175/BAMS-D-12-00121.1.
Jackson, L. C., R. Kahana, T. Graham, M. A. Ringer, T. Woollings, J. V. Mecking, and R. A. Wood, 2015: Global and European climate impacts of a slowdown of the AMOC in a high resolution GCM. Climate Dyn., 45, 3299–3316, https://doi.org/10.1007/s00382-015-2540-2.
Koenig, Z., C. Provost, R. Ferrari, N. Sennechael, and M.-H. Rio, 2014: Volume transport of the Antarctic Circumpolar Current: Production and validation of a 20 year long time series obtained from in situ and satellite observations. J. Geophys. Res. Oceans, 119, 5407–5433, https://doi.org/10.1002/2014JC009966.
Kuntz, L. B., and D. P. Schrag, 2018: Hemispheric asymmetry in the ventilated thermocline of the tropical Pacific. J. Climate, 31, 1281–1288, https://doi.org/10.1175/JCLI-D-17-0686.1.
Lawrence, D. M., and Coauthors, 2011: Parameterization improvements and functional and structural advances in version 4 of the Community Land Model. J. Adv. Model. Earth Syst., 3, 27, https://doi.org/10.1029/2011MS000045.
Levermann, A., A. Griesel, M. Hofmann, M. Montoya, and S. Rahmstorf, 2005: Dynamic sea level changes following changes in the thermohaline circulation. Climate Dyn., 24, 347–354, https://doi.org/10.1007/s00382-004-0505-y.
Liu, W., and A. Hu, 2015: The role of the PMOC in modulating the deglacial shift of the ITCZ. Climate Dyn., 45, 3019–3034, https://doi.org/10.1007/s00382-015-2520-6.
Liu, Z., and Coauthors, 2009: Transient simulation of last deglaciation with a new mechanism for Bolling-Allerod warming. Science, 325, 310–314, https://doi.org/10.1126/science.1171041.
Lozier, M. S., 2012: Overturning in the North Atlantic. Annu. Rev. Mar. Sci., 4, 291–315, https://doi.org/10.1146/annurev-marine-120710-100740.
Manabe, S., and R. J. Stouffer, 1988: Two stable equilibria of a coupled ocean–atmosphere model. J. Climate, 1, 841–866, https://doi.org/10.1175/1520-0442(1988)001<0841:TSEOAC>2.0.CO;2.
Marsh, R., W. Hazeleger, A. Yool, and E. J. Rohling, 2007: Stability of the thermohaline circulation under millennial CO2 forcing and two alternative controls on Atlantic salinity. Geophys. Res. Lett., 34, L03605, https://doi.org/10.1029/2006GL027815.
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 (CAM 5.0). NCAR Tech. Note NCAR/TN-486+STR, 268 pp., www.cesm.ucar.edu/models/cesm1.1/cam/docs/description/cam5_desc.pdf.
Okazaki, Y., and Coauthors, 2010: Deepwater formation in the North Pacific during the last glacial termination. Science, 329, 200–204, https://doi.org/10.1126/science.1190612.
Rahmstorf, S., 1996: On the freshwater forcing and transport of the Atlantic thermohaline circulation. Climate Dyn., 12, 799–811, https://doi.org/10.1007/s003820050144.
Rahmstorf, S., 2002: Ocean circulation and climate during the past 120,000 years. Nature, 419, 207–214, https://doi.org/10.1038/nature01090.
Rasmussen, S. O., and Coauthors, 2006: A new Greenland ice core chronology for the last glacial termination. J. Geophys. Res., 111, D06102, https://doi.org/10.1029/2005JD006079.
Reagan, J., D. Seidov, and T. Boyer, 2018: Water vapor transfer and near-surface salinity contrasts in the North Atlantic Ocean. Sci. Rep., 8, 8830, https://doi.org/10.1038/s41598-018-27052-6.
Reason, C. J. C., and S. B. Power, 1994: The influence of the Bering Strait on the circulation in a coarse resolution global ocean model. Climate Dyn., 9, 363–369, https://doi.org/10.1007/BF00223448.
Saenko, O. A., A. Schmittner, and A. J. Weaver, 2004: The Atlantic–Pacific seesaw. J. Climate, 17, 2033–2038, https://doi.org/10.1175/1520-0442(2004)017<2033:TAS>2.0.CO;2.
Schmittner, A., M. Latif, and B. Schneider, 2005: Model projections of the North Atlantic thermohaline circulation for the 21st century assessed by observations. Geophys. Res. Lett., 32, L23710, https://doi.org/10.1029/2005GL024368.
Schneider, N., 1998: The Indonesian throughflow and the global climate system. J. Climate, 11, 676–689, https://doi.org/10.1175/1520-0442(1998)011<0676:TITATG>2.0.CO;2.
Seidov, D., and B. J. Haupt, 2003: Freshwater teleconnections and ocean thermohaline circulation. Geophys. Res. Lett., 30, 1329, https://doi.org/10.1029/2002GL016564.
Seidov, D., and B. J. Haupt, 2005: How to run a minimalist’s global ocean conveyor. Geophys. Res. Lett., 32, L07610, https://doi.org/10.1029/2005GL022559.
Shaffer, G., and J. Bendtsen, 1994: Role of the Bering Strait in controlling North Atlantic Ocean circulation and climate. Nature, 367, 354–357, https://doi.org/10.1038/367354a0.
Si, D., and A. Hu, 2017: Internally generated and externally forced multidecadal oceanic modes and their influence on the summer rainfall over East Asia. J. Climate, 30, 8299–8316, https://doi.org/10.1175/JCLI-D-17-0065.1.
Singh, H. K. A., A. Donohoe, C. M. Bitz, J. Nusbaumer, and D. C. Noone, 2016: Greater aerial moisture transport distances with warming amplify interbasin salinity contrasts. Geophys. Res. Lett., 43, 8677–8684, https://doi.org/10.1002/2016GL069796.
Smeed, D., B. Moat, D. Rayner, W. E. Johns, M. O. Baringer, D. Volkov, and E. Frajka-Williams, 2019: Atlantic meridional overturning circulation observed by the RAPID-MOCHA-WBTS (RAPID-Meridional Overturning Circulation and Heatflux Array-Western Boundary Time Series) array at 26°N from 2004 to 2018. British Oceanographic Data Centre, National Oceanography Centre, accessed 2020, https://doi.org/10.5285/8cd7e7bb-9a20-05d8-e053-6c86abc012c2.
Song, Q., G. A. Vecchi, and A. J. Rosati, 2007: The role of the Indonesian throughflow in the Indo–Pacific climate variability in the GFDL coupled climate model. J. Climate, 20, 2434–2451, https://doi.org/10.1175/JCLI4133.1.
Stocker, T. F., 2002: North–south connections. Science, 297, 1814–1815, https://doi.org/10.1126/science.1075870.
Stouffer, R. J., and Coauthors, 2006: Investigating the causes of the response of the thermohaline circulation to past and future climate changes. J. Climate, 19, 1365–1387, https://doi.org/10.1175/JCLI3689.1.