 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
Global Planetary Change, Vol. 30, pp. 257-270, 2001
Accepted 26 September 2000
Keywords: ocean circulation, climate change, global warming, paleoclimate
1 .Introduction
One of the important new findings in oceanography is the fact that the ocean, including its abyss, is warming by a rate of a half degree C or more per century (Levitus et al., 2000). This ocean warming during last several decades is linked to global climate change. Historical observations and paleoclimate data reveal significant climate variability on decadal and shorter to millennial time scales. Many factors contribute to this variability, but significant aspects of the observations can be related to ocean-atmosphere interaction. In particular, a number of model studies identify the North Atlantic and its associated deepwater formation as a focal point of decade-to-century variations (Maier-Reimer et al., 1991; Manabe and Stouffer, 1988; 1995; 1997). These model studies, combined with observations, provide considerable evidence that changes in freshwater influx into the Northern Atlantic Ocean can substantially modify the deep ocean circulation, in turn, dramatically influencing the climate state and driving rapid climate change. Much effort, beginning from pioneer work by Bryan (1986), have been directed to understanding of sensitivity of the high-latitudinal oceans to freshwater fluxes (e.g., Manabe and Stouffer (1995); Marotzke and Willebrand (1991); Rahmstorf (1995a); Weaver et al. (1993)). These studies have substantially advanced our understanding of the dynamics of the North Atlantic Deep Water (NADW) circulation. Importantly, this sensitivity to surface fluxes has also become a major factor in explaining variability during the glacial cycles of the Pleistocene (Broecker, 1998; Fichefet et al., 1994; Ganopolski et al., 1998; Rahmstorf, 1995a; Seidov and Haupt, 1997; Seidov et al., 1996; Stocker and Wright, 1991; Stocker and Wright, 1996; Weaver and Hughes, 1994). In addition, they have sparked considerable interest in whether human-induced global warming may be amplified or may produce climate “surprises” through the impact of meltwater events on the deepwater circulation (Broecker, 1994). The deepwater circulation associated with NADW production has become known as the thermohaline conveyor or “salinity conveyor belt” (Broecker, 1991).
The success of the above-mentioned modeling effort in identifying a key mechanism of climate variability is noteworthy, but our understanding of the impacts of changes in freshwater fluxes to the ocean is far from complete. NADW production has been almost exclusively the focus of model study when, in fact, the thermohaline circulation is operated by two major deepwater sources, the North Atlantic and the Southern Ocean surrounding Antarctica (primarily the Weddell and Ross Seas). The role of the southern source of deepwater, namely the Antarctic Bottom Water (AABW) in modulating variability has received much less attention, limiting the development of a complete understanding of decadal to millennial time-scale climate change. An issue is whether the NADW is the ultimate driver of the conveyor, and whether substantial additional variability is generated by freshwater impacts in the Southern Ocean.
Several lines of evidence support a significant Southern Ocean role. First, many examples of climate intermittency during the glacial cycles of Pleistocene remain poorly understood, even though they seem to correlate with major deglaciations (Bond et al., 1992; Bond and Lotti, 1995; Broecker, 1994). Second, recent studies reveal a bi-polar nature of the glacial cycles of Pleistocene, including an essential element of southern variability (Blunier et al., 1998; Broecker, 1994; Stocker, 1998). For example, Vidal et al. (1999) argue that the southern Atlantic leads in the occurrence of several Heinrich Events. Broecker et al. (1999) suggest that deepwater production in the Southern Ocean may have reduced from 15 Sv (1 Sv = 106 m3/s) to 5 Sv over last 800 years. They also argue that the Little Ice Age (about only 500 years ago) was caused by far stronger deep ocean ventilation in the Southern Ocean. One of the reasons for the speed up of the southern ventilation could be an increase in Atlantic Ocean salinity (Broecker et al., 1999), whereas the slowdown could be caused by reduced surface salinity, associated warming, and sea ice or ice sheet melting in the Southern Ocean after the Little Ice Age (Broecker, 2000).
The source and character of the meltwater in the Southern Hemisphere is substantially different from the Northern Hemisphere. Sea surface salinity controls both the AABW and Antarctic Intermediate Water northward incursions (England, 1992; Stocker et al., 1992). Basically, sea surface salinity in the Southern Ocean can be either increased due to brine rejection when sea ice is formed, or decreased due to sea ice melting. Both early studies (Toggweiler and Samuels, 1980), as well as recent model simulations (Goosse and Fichefet, 1999), confirm the importance of the brine rejection during the formation of sea ice on the salinity, and therefore the role of sea ice formation in the southern circulation regime. In addition, the Antarctic ice sheet also plays an important role in governing freshwater fluxes into the Southern Ocean. Substantial concern has also been expressed about the stability of the West Antarctic Ice Sheet (WAIS) (see review ref. Oppenheimer (1998)). Uncertainties about Antarctic ice sheet mass balance and its contribution to global sealevel rise is a major issue of debate (Vaughan et al., 1999) even without large-scale collapse of the WAIS. Hence, the potential for changes in freshwater fluxes or salinity variations to influence the Southern Ocean is clearly evident. Moreover, Schmittner and Stocker (1999) give another reason for diluting of sea surface during global warming, invoked by an increased equator-to-pole freshwater transport in a warmer atmosphere. The increased poleward moisture flux results in increased precipitation in the high latitudes, which in concert with cryosphere meltdown, could enhance the meltwater impact on the ocean circulation.
These factors have initiated a growing modeling effort designed to investigate the climatic role of the Southern Hemisphere (Goodman, 1998; Hirschi et al., 1999; Scott et al., 1999; Stoessel et al., 1998; Wang et al., 1999a; 1999b). These studies describe the potential importance of feedbacks between northern and southern sources of deepwater, suggest that freshwater forcing in the Southern Hemisphere may influence the NADW, and examine the importance of brine release rates on AABW formation. The results are of great interest, but as yet are insufficient since most of these studies operate with simplified models, the thermal response of the deep ocean is only marginally discussed, and there are no systematic attempts to quantify the implications of deglaciation in the Northern versus Southern Hemispheres in the real world ocean. A recent study by Goosse and Fichefet (1999) is an exception. Here, a more realistic study of southern versus northern freshwater impacts indicates that suppressing the AABW production may result in noticeable (up to an average 1ºC) deep ocean warming. Although this study is a significant advance, several important issues remain. First, Goosse and Fichefet (1999) considered only brine rejection reductions associated with sea ice formation, which reduced the sea surface salinity by 1 psu. Any variation in ice sheet melt, or calving of the ice sheets, or variation of sea ice melting impact was not considered. In addition, Goosse and Fichefet utilize a flux formulation in governing the salinity anomalies. Normally, relaxation boundary condition specifications are considered to be inferior to flux formulations. However, in this case it is more difficult to specify a specific salinity change to match observations using the flux formulation. Therefore, the Goosse and Fichefet experiments reflect a rather conservative, and less specific, estimate of a meltwater episode, even without additionally considering decay of the WAIS, which is thought to have happened sporadically during last million years (MacAyeal, 1992). In addition, Goosse and Fichefets’ (1999) ground-breaking study still does not fully examine the linkages between NADW and AABW, nor it’s implications for climate change and for subsequent sealevel rise.
The study described here examines the character of deepwater formation as a function of systematic changes in freshwater fluxes in both the Northern and Southern Hemispheres. It demonstrates the degree to which the thermohaline circulation is driven by both deepwater sources, and therefore, the importance of high latitude freshwater fluxes in general in governing past and future climates. The implications of these results are further investigated by examining the potential for changes in poleward heat transport in response to the changes in deepwater circulation and the potential for changes in sealevel in response to changes in the thermohaline structure of the ocean.
2. Setup of Numerical Experiments
In order to compare the competitive roles of the Northern and Southern Hemisphere deepwater sources, we have designed a series of simplified meltwater events, in which all ocean model parameters are held constant except salinity. The objective is simply to demonstrate the bi-polar nature of the earth’s climate sensitivity to the ocean-cryosphere interactions and to demonstrate the potential importance of both Northern and Southern Hemisphere meltwater events in governing climate variability. Although the meltwater events are highly idealized, they are derived from realistic Pleistocene episodes. Both the locations in the North Atlantic and the amplitudes of the freshwater signals conform to what is known from ocean proxies (Duplessy et al., 1996; Labeyrie et al., 1986; Sarnthein et al., 1994; 1995).
The numerical experiments are straightforward. They are completed using GFDL MOM version 2, a well documented and extensively utilized ocean circulation model (Pacanowski, 1996). The upper layer temperature and salinity are restored to the present-day annual mean sea surface temperature (SST) and sea surface salinity (SSS) (Levitus and Boyer, 1994; Levitus et al., 1994). The Hellerman-Rosenstein (Hellerman and Rosenstein, 1983) annual mean wind stress is used in all simulations. A low-salinity signal with the amplitude comparable to that found in paleoceanographic proxies is superimposed on the present-day SSS, retaining present-day SST and wind stress. A rather coarse resolution of 6ºx4º with 12 levels is employed to perform large number of runs reaching near equilibrium states. We consider this resolution as appropriate for a pilot comparison between several low-salinity regimes in which the focal point is the large-scale thermohaline circulation. Moreover, this resolution is comparable with other coarse resolution studies (Manabe and Stouffer, 1995; Rahmstorf, 1995b; Seidov and Haupt, 1999; Seidov and Maslin, 1999; Toggweiler et al., 1989) addressing similar problems, and proven to be sufficient for studying the response of ocean meridional overturning to freshwater signals. For example, (Seidov and Haupt, 1999) demonstrate that water transports, convection depths, and inter-basin water exchanges are reasonably well-simulated in a study with a similar spatial resolution using the same boundary conditions.
Thirteen experiments have been completed in our study (Table 1): Exp. 1 is the control run (CR) with present-day surface climatology; Exp. 2-5 have low-salinity signals in the high-latitudinal North Atlantic; Exp. 6-8 have a low-salinity signal in the Southern Ocean, Exp. 9-10 have low-salinity signals in both these two regions, and Exp. 11-13 are the runs with the low-salinity signal confined to the Weddell Sea only. The low-salinity signal is applied as a negative salinity anomaly with different amplitudes, from 0.5 to 3 psu in the North Atlantic (Exp. 2-5), from 0.2 to1 psu in the Southern Ocean (Exp. 6-8), and vary in the combined cases (Exp. 9-10). The low salinity in the Weddell Sea is lower than in CR, with maximum differences from 0.5 to 3 psu (Exp. 11-13). The low-salinity signal in the North Atlantic is applied as a band of low-salinity anomalies between the northern boundary of the basin at 80ºN to 60ºN with smooth merging the band with the present-day SSS at 50ºN. In the Southern Ocean the same kind of a low salinity band was superimposed on the present-day climatology between the Antarctica coast and 60ºS merging to the unchanged SSS at 50ºS. In the SO experiments the anomalies are circumglobal. In the WED experiments the low salinity signal also extends to 50ºS but is confined to the Weddell Sea location only. All runs are 2000 model years long, with 5-fold acceleration in the deep layers (which means that the deep ocean is effectively run for 10,000 years). In all numerical experiments complete steady states are reached.
Regarding the strength of the low-salinity signal, even 1 psu is a rather moderate
estimate of the possible dilution of sea surface water during a meltwater event.
For instance, Goosse and Fichefet (1999) argue that even the reduction of brine
rejection can alone cause a 1 psu decrease in sea surface salinity. Duplessy
et al. (1996) show the low salinity band around glacial Antarctica with anomalies
up to -1.8 psu. Labeyrie et al. (1986) argue that the periphery of the Antarctic
ice sheet was eroded during some of the glacial cycles of Pleistocene, and that
only 10,000 km3 of meltwater could have reduced the sea surface d18O
by 1‰ (which translates to approximately 2 psu in sea surface salinity (Duplessy
et al., 1996)). Anderson and Andrews (1999) revisited the problem of the late
Quaternary Antarctic meltdown, and argue that significant deglaciation of the
Weddell Sea continental shelf could have taken place prior to the last glaciation.
Birchfield and Broecker (1990) point out that a relatively small freshwater
flux converted to a low-salinity signal will hamper the conveyor operation.
For instance, they show that a freshwater flux of 0.1 to 0.3 Sv in the North
Atlantic can cause 0.3 to almost 1 psu reduction of salinity in 1000 years.
The 0.3 Sv flux during a thousand years would convert to 10,000 km3
a year, a value that is only about 4 times greater than the present-day annual
meltwater production in Antarctica of about 2,500 km3/year (e.g.,
Vaughan et al. (1999)).
As restoring surface boundary conditions are used (e.g., Bryan (1987)), we have
calculated apparent freshwater volumes that would be needed to dilute surface
layer to achieve a respective salinity change in the imposition domain. These
volumes can be thought of as virtual freshwater volumes that would have to be
added to a thin surface layer, where the SSS is specified, to dilute the water
in this layer to a prescribed SSS reduction. (In the runs with increased SSS,
this would be the freshwater to be removed to achieve the respective increase
in SSS caused by brine rejection). Following Manabe and Stouffer (1995), we
estimate the rates by which these virtual freshwater fluxes would have to be
added within 10 years (Table 1). The total
amount of freshwater added at those rates is fairly realistic. For instance,
to dilute a 10 m layer of water with salinity of 35 psu by 1 psu in 10 years
in one of the Southern Ocean experiments (Exp. 8 in Table
1), a freshwater (or equivalent sea ice) layer of about only 0.3 m thickness
would be needed.
3. Results
The results of the experiments in Table 1 are divided into an analysis of the sensitivity of the thermohaline circulation to low-salinity perturbations followed by an analysis of the differences in sealevel associated with the thermal response of the ocean in response to the salinity anomalies.
3.1. The Nature of the Thermohaline Circulation
The changes in the character of the thermohaline circulation and in the nature of the meridional oceanic heat flux are described as a function of low-salinity perturbations in the North Atlantic alone, the Southern Ocean alone, combined North Atlantic and Southern Ocean events, and perturbations to the Weddell Sea only. For brevity, only the results of Exp. 1, 4, 8 and 10 are illustrated.
3.1.1. North Atlantic Events
The results of the experiments simulating a low-salinity impact in the North Atlantic conform to what is already well known from previous work (Broecker, 1998; Fichefet et al., 1994; Maier-Reimer et al., 1991; Manabe and Stouffer, 1988; 1997; Mikolajewicz, 1996; Rahmstorf, 1994; 1995a; Seidov and Haupt, 1997; Seidov et al., 1996; Stocker and Wright, 1991; 1996; Weaver and Hughes, 1994). The conveyor is weaker and shallower. Figure 1 shows the overturning stream function. The left panel depicts the control run (Figure 1a), whereas the right panel represents the scenario with -2 psu anomalies in the North Atlantic (Figure 1b). Temperature differences between this low-salinity scenario and the control case (not shown) indicate cooling in high latitudes of the Atlantic Ocean. This occurs because the reduced NADW production led to a shallow conveyor, and cooler and fresher water than today in these latitudes characterizes the deep ocean water. Additionally, surface ocean may have more time to loose heat to the atmosphere because the overturning slowed. The reduced NADW outflow, coincident with reduced replacement water crossing the equator in the North Atlantic, has an evident imprint in the oceanic heat transport. Meridional oceanic heat transport in the Atlantic Ocean from Exp. 1 and 4 is shown in Figure 2. Northward cross-equatorial heat transport, which is a characteristic of the present-day climate, is dramatically reduced in the scenario with a strong freshwater impact. This result implies the possibility of a cold episode following a meltwater event. For instance, if the present-day global warming were potent enough to induce a low-salinity episode in the North Atlantic caused by iceberg and Arctic sea ice melting, the result could be a tendency to colder conditions in the Northern Hemisphere. However, even an excessive northern low-salinity signal of -2 psu did not cause a complete termination of the conveyor.
Note that, although the signal is strong, it is mostly confined to the north of Iceland, with the southern site of the North Atlantic deep convection not affected strongly. This confirms the results of Seidov and Maslin (1999) who argue that the Dansgaard-Oeschger events, when the meltwater events were confined largely to the Nordic Seas, were not a sufficient force to terminate the conveyor. Hence, a scenario with only the Nordic Seas and the northern part of the Labrador Sea affected by a freshwater signal would result in limited climatic consequences.
3.1.2. Southern Ocean Event
In contrast to the predictable and understandable results of the northern low-salinity impact, the results of the Southern Ocean surface freshening are less intuitive. Two aspects are particularly noteworthy. First, the circulation changes driven by the low-salinity signal were much stronger, and second, they led to a very strong warming of the deep ocean. Figure 3 shows the meridional overturning in the Atlantic Ocean in the Southern Ocean experiment with SSS anomaly of -1 psu (Exp. 8). The overturning is only illustrated for the Atlantic Ocean because the bi-polar climate seesaw phenomenon is believed to be inherent mostly to this ocean, and because the NADW is still the major player (even when the Southern Ocean, rather than the North Atlantic, is freshened at the surface). Temperature differences between the southern low-salinity scenario and the control case are shown in two sections, in the Atlantic Ocean and in the western Pacific Ocean (Figure 4). It has been recognized a while ago that the increase of NADW production can cause cooling of the upper waters of the Southern Atlantic as the poleward heat flux increases (Crowley, 1992). However, we emphasize that the deep ocean thermal trend in the southern meltwater impact scenario can be of opposite sign to those in the upper layers. As NADW is warmer than AABW, there is a substantial warming of the deep ocean in this scenario, in contrast to the North Atlantic meltwater scenario. The warming takes place over the entire deep ocean and its maximum shifts to the southern edges. This deep-sea warming is caused not only by a substantially increased (by 40-60%) NADW production, but also largely because the meridional overturning takes over the entire deep ocean, pushing away the lessened AABW. In the North Atlantic scenario, the meltwater impact on the conveyor caused thermal effects only in the deep Atlantic Ocean, whereas in the Southern Ocean, the meltwater scenario impact is global. The increased NADW outflow in the deep layers leads to increased compensating northward surface water flow. This flow carries more warm and salty subtropical water to convection sites, which might further increase NADW production until the atmosphere warms up to reduce the cooling of the sea surface and subsequently reduce deep convection. The positive feedback of NADW production and northward heat transport can be viewed as a first link toward high-latitudinal warming in the Northern Hemisphere caused by meltwater events in the Southern Ocean. Heat transport in the Southern Ocean meltwater is shown in Figure 5 (the present-day transport is also shown, as in Figure 2, to facilitate comparison). This interpretation is, however, incomplete because in a stand-alone ocean model the negative feedbacks inherent to a coupled ocean-atmosphere system are absent.
3.1.3. North Atlantic versus Southern Ocean
Although in Exp. 10 the amplitude in the North Atlantic (-3 psu) perturbation was three times the Southern Ocean (-1 psu) event, the deep-water regime is qualitatively similar to the experiment with a Southern Ocean-only perturbation. There is somewhat less deep-ocean warming, but it remains global and substantial. Basically, the results of the runs with perturbations to two sources, in the North Atlantic and Southern Ocean, demonstrates a more powerful response to a meltwater event in the Southern Ocean than for those in the North Atlantic. However, much of this power stems from increased NADW production, adding to the evidence of the importance of the North Atlantic region. Most importantly, the problem of deep-ocean teleconnections is now seen from a very different angle.
3.1.4 Weddell Sea scenarios
Surprisingly, the Weddell Sea scenario did not give as noticeable a warming as was found in the whole Southern Ocean scenario. Even with the amplitude of the freshwater signal in the Weddell Sea of -3 psu, the impact was far less than in the Southern Ocean scenario with only -0.2 psu. These model simulations imply that AABW formed around Antarctica may be more important for the conveyor dynamics than the major portion originating in the Weddell Sea. It is not clear, however, whether this would be the case in a coupled ocean-atmosphere model, with the low-salinity signal spreading from the Weddell Sea circumglobally. In this case, the results might conform more to the SO cases, rather than to Exp. 11-13.
Table 2 shows NADW production rates, the depths to which the convection reaches at the NADW convection sites, the meridional overturning at the critical latitude of 30ºS, and the outflow of NADW and inflow of AABW. The balance of these two flows determines the state and intensity of the conveyor.
Table 3 (supplemental to Figures 2 and 5) presents the northward heat flux across several critical latitudes in the North Atlantic. As can be seen, the state of the conveyor is best represented by the cross-equatorial heat transport. In cases with significant southern meltwater impacts, the Northern Hemisphere is warmed by the ocean currents. However, only in the strongest northern meltwater episode (Exp. 5), southward cross-equatorial heat transport in the Atlantic may occur. Hence, during the northern meltwater episodes, the Southern Hemisphere might get heated by the overturning.
3.2. Implications for Changes in Sealevel
The above results help elucidate the differential role of the two polar regions in explaining Pleistocene climate variability and they are also important for understanding potential future climate change. Notably, the implications are not limited to changes in oceanic circulation. For example, sealevel rise caused by melting of major ice sheets is a central issue of global warming forecasts (see references in Houghton (1997); Karl (1993); Warrick et al. (1993)). However, there is also an indirect sealevel effect of meltwater events caused by thermal restructuring of the world ocean. As the deep ocean warms up, the sea elevation will change because of the thermal expansion of sea water. Historic hydrographic data suggest that thermal expansion of the ocean can contribute tens of centimeters to the observed sealevel rise over the last century (Godfrey and Love, 1992). Some simulations (e.g., Church et al. (1991)) indicate that the thermal expansion of the ocean associated with a global warming of 3ºC temperature rise by the year 2050 results in up to 30 cm sealevel rise. On the other hand, cooling of large segments of the world ocean would compensate for the land ice sheet melting and reduce the sealevel rise caused by such melting. The issue of how changes in oceanic circulation will impact sealevel is largely unexplored.
The difference of the sea elevation relative to the sea floor was calculated for each of the sensitivity experiments in Table 1. These are the differences of steric sealevels relative to the ocean bottom calculated using density anomalies (see, e.g., Godfrey and Love (1992); Levitus (1982)). Steric heights in Levitus (1982)) are computed relative to 2000 m; since we are interested only in differences between the sensitivity runs and the control run (i.e., in relative, rather than absolute sealevels), the relative depth can be extended to the sea bottom to avoid uncertainties of an arbitrary chosen “level of no motion” (e.g. Sarkisyan (1977)).
The NA case with the anomaly of -2 psu minus the control run (Exp. 4 and Exp. 1), and the SO case with the anomaly of -1 psu minus the control run (Exp. 8 and Exp. 1) are illustrated in Figures 6a and Figure 6b. These differences in sea surface height are due to the differences in the 3-D density field caused by different T and S distribution in the world ocean. Significant sealevel rise (up to 2-3 m in the SO case) is evident. Notably, in many sensitive coastal areas the sea lever rise can be over 1 m. Hence, it is possible that a meltwater episode, especially in the Southern Ocean (with a substantial global deep ocean warming and salinity redistribution), could impact island nations and coastal regions through a noticeable sea-level change. Importantly, this sea-level rise could occur without significant melting of the ice sheets, including WAIS, which is considered the most vulnerable to climate change. If some melting of WAIS, or any other ice sheet were to happen, the effect would lead to even more dramatic changes than those shown in Figure 6.
The thermal response of the deep ocean and related sealevel change was remarkably fast. In an integration of the model for 1000 years with the low-salinity signal superimposed on the steady state of the control run, the first 70 % of total sealevel change and warming of deep ocean was reached within first 400 years, with a much slower increase followed.
4. Discussion and conclusions
A number of model studies identify the North Atlantic and its associated deepwater formation as a focal point of decade-to-century variations (Maier-Reimer et al., 1991; Manabe and Stouffer, 1988). As a result, the sensitivity of the ocean circulation to changes in North Atlantic surface fluxes has become a major factor in explaining variability during the glacial cycles of the Pleistocene and a major factor in contemplating the nature of future human-induced global warming. The role of the southern source of AABW in modulating variability has received much less attention, limiting the development of a complete understanding of decadal to millennial time-scale climate change.
The source and character of the salinity perturbations in the Southern Hemisphere is substantially different from the Northern Hemisphere, involving brine rejection in the formation of sea ice, freshwater fluxes from ablation and calving of the Antarctic Ice Sheet and the stability of the WAIS. Hence, the potential for changes in freshwater fluxes or salinity variations to influence the Southern Ocean is clearly evident.
The study described here examines the character of deepwater formation as a function of systematic changes in freshwater fluxes in both the Northern and Southern Hemispheres. It demonstrates the degree to which the thermohaline circulation is driven by both deepwater sources, and therefore, reveals the importance of high latitude freshwater fluxes in general in governing past and future climates. The results demonstrate that changes in surface salinity of the Southern Ocean can significantly alter deepwater structure and temperatures. In addition, the experiments demonstrate that the high latitude sea surface need not be very warm, or very salty, or both to produce significant deep ocean warming. In principle, an imbalance of meltwater impacts in high latitudes may lead to warming up of the deep ocean even if the ocean in high latitudes stays relatively cool in one of the hemispheres. In this case, the warming is in response to increased overturning in one of the hemispheres, whereas another hemispheric source of deepwater might become stagnated. In other words, a strong meltwater impact in one of the two hemispheres may lead to a strengthening of the thermohaline conveyor driven by the source in another hemisphere. This, in turn, may lead to an increased surface poleward compensating flow in the active hemisphere. Further, this warming can have a substantial sealevel impact even without ice sheet melting.
As regards the deep ocean warming, our study picks up where previous studies of the role of the southern deepwater source stopped. Most of those studies idealized simulations and they did not focus on either the deep ocean thermal response, or on the thermally induced sealevel rise. The most realistic study to date is that of Goosse and Fichefet (1999). Although the Goose and Fichefet study advances our understanding of the AABW formation, it does not reveal the hidden connection between the NADW and AABW in governing the deep ocean, nor does it consider the range of plausible freshwater input changes to the Southern Ocean. Importantly, in the study presented here, we have found that the warming of deep-ocean is caused by NADW intensification that accompanies the reduction of AABW, rather than by reduction of the southern source itself. The study presented here emphasizes the competitive nature of the northern and southern sources and indicates the role of the southern source as a strong modifier of NADW. Further, this study calls attention to the important implications of changes in the thermohaline circulation, both in terms of heat transport and sealevel rise.
Our results on the amplitude of deep-ocean warming may match projected temperature trends. As shown by Levitus et al. (2000), the global volume warming is 0.06ºC per 40 years. Our global warming rates vary between 4ºC to 7ºC per 1000 years which is about twice to three times higher than the observed present-day trend (Levitus et al., 2000), would it continue for a thousand years. This could be because we have no atmospheric feedbacks that could slow down the ocean warming induced by the southern meltdown (see below). Alternatively, the warming could have speed up as the AABW weakens with the addition of meltdown (hence it may turn out to be the major accelerator of the deep-ocean warming trend). Note that the ocean warming is not linear in time (see above), even if the low-salinity signal is kept constant. However, we indicate that our results are at least in a qualitative agreement with Levitus et al. (2000) analyses and with Broecker’s (2000) arguments about the AABW reduction as the global warming cause.
The above results are not without limitation. A stand-alone ocean model helps
to outline the nature of the problem and enables a wealth of sensitivity experiments.
However, this is largely a first-order analysis. As such, the more extreme scenarios
provide a higher level of sensitivity that clearly illustrate the potential
response of the ocean without atmospheric and cryospheric feedbacks. The response
of a coupled model may be different than of a stand-alone ocean model because
of the potential importance of feedbacks associated with the atmospheric response
to an altered poleward ocean heat transport, or the impact of wind stress changes
on global thermohaline overturning. Hence, our work is only a first step in
assessing the climate response to changes in freshwater inputs at high latitudes.
Despite this limitation, the result is clear. The Southern Ocean can overpower
the Northern Hemisphere and become a major climatic player in long-term climate
change. The NADW appears to be a universal driver of the conveyor, amplified
by southern meltwater episodes or reduced by the northern meltwater impacts,
but remaining largely the strong player in bi-polar thermohaline conveyor variability.
The two hemisphere sources, if reduced, lead to principally different consequences:
If the NADW is reduced, the deep ocean cools down, whereas reduction of AABW
leads to warmer abyssal waters. Therefore, a change of sea-ice or WAIS state
tending toward less saline surface waters in the Southern Ocean can cause unfavorable
sea-level changes, whereas the collapse of the northern source might cause cooling
of the northern climate. Both scenarios may pose a serious threat to climate-sensitive
environments.
Acknowledgments
We are grateful to anonymous reviewers for their useful comments. One of the
reviewers provided extended list of comments that were especially helpful for
improving the manuscript. This study is a partly supported by NSF (NSF project
#9975107).
References
Anderson, J.B. and Andrews, J.T., 1999. Radiocarbon constraints on ice sheet advance and retreat in Weddell Sea, Antarctica. Geology(27): 179-182. Birchfield, G.E. and Broecker, W.S., 1990. A salt oscillator in the glacial Atlantic? 2. A "scale analysis model. Paleoceanography, 5: 835-843. Blunier, T. et al., 1998. Asynchrony of Antarctic and Greenland climate change during the last glacial period. Nature, 394: 739-743. Bond, G. et al., 1992. Evidence for massive discharges of icebergs into the North Atlantic ocean during the last glacial period. Nature, 360: 245-249. Bond, G. and Lotti, R., 1995. Iceberg discharges into the North Atlantic on millennial time scales during the last glaciation. Science, 267: 1005-1010. Broecker, W., 1991. The great ocean conveyor. Oceanography, 1: 79-89. Broecker, W.S., 1994. Massive iceberg discharges as triggers for global climate change. Nature, 372: 421-424. Broecker, W.S., 1998. Paleocean circulation during the last deglaciation: A bipolar seesaw? Paleoceanography, 13: 119-121. Broecker, W.S., 2000. Was a change in thernohaline circulation responsible for the Little Ice Age? Proc. Nat. Acad. Sci., 97(4): 1339-1342. Broecker, W.S., Sutherland, S. and Peng, T.-H., 1999. A possible 20th century slowdown of Southern Ocean deep water formation. Science, 286: 1132-1135. Bryan, F., 1986. High-latitude salinity effects and interhemispheric thermohaline circulations. Science, 323: 301-304. Bryan, F., 1987. Parameter sensitivity of primitive equation ocean general circulation models. Journal of Physical Oceanography, 17: 970-985. Church, J.A., Godfrey, J.S., Jacket, D.R. and McDougall, T.R., 1991. A model of sealevel rise caused by ocean thermal expansion. Journal of Climate, 4: 438-456. Crowley, T.J., 1992. North Atlantic deep water cools the southern hemisphere. Paleoceanography, 7: 489-497. Duplessy, J.-C. et al., 1996. High latitude deep water sources during the Last Glacial Maximum and the intensity of the global oceanic circulation. In: G. Wefer, W.H. Berger, G. Siedler and D. Webb (Editors), The South Atlantic. Springer, NY, pp. 445-460. England, M.H., 1992. On the formation of Atlantic intermediate and bottom water in ocean general circulation models. Journal of Physical Oceanography, 22: 918-926. Fichefet, T., Hovine, S. and Duplessy, J.-C., 1994. A model study of the Atlantic thermohaline circulation during the Last Glacial Maximum. Nature, 372: 252-255. Ganopolski, A., Rahmstorf, S., Petoukhov, V. and Claussen, M., 1998. Simulation of modern and glacial climates with a coupled global model of intermediate complexity. Nature, 391: 351-356. Godfrey, J.S. and Love, G., 1992. Assessment of sealevel rise, specific to the South Asian and Australian situations In: Seal Level Changes: Determination and Effects, Seal Level Changes: Determination and Effects. Geophys. Monograph 69. AGU, Washington DC, pp. 87-94. Goodman, P.J., 1998. The role of North Atlantic Deep Water formation in an OGCM's ventilation and thermohaline circulation. J. Phys. Oceanogr., 28: 1759-1785. Goosse, H. and Fichefet, T., 1999. Importance of ice-ocean interactions for the global ocean circulation: A model study. Journal of Geophysical Research, 104(C10): 23337-23355. Hellerman, S. and Rosenstein, M., 1983. Normal monthly wind stress over the world ocean with error estimates. Journal of Physical Oceanography, 13: 1093-1104. Hirschi, J., Sander, J. and Stocker, T.F., 1999. Intermittent convection, mixed boundary conditions and the stability of the thermohaline circulation. Climate Dynamics, 15(4): 277-291. Houghton, J., 1997. Global warming. Cambridge University Press, N.Y., 251 pp. Karl, T.R., 1993. A new perspective on global warming. Eos, Transactions, American Geophysical Union, 74: 25. Labeyrie, L.D. et al., 1986. Melting history of Antarctica during the past 60,000 years. Nature, 322: 701-706. Levitus, S., 1982. Climatological Atlas of the World Ocean. Levitus, S., Antonov, J.I., Boyer, T.P. and Stephens, C., 2000. Warming of the World Ocean. Science, 287: 2225-2229. Levitus, S. and Boyer, T.P., 1994. World Ocean Atlas, vol. 3, Salinity, 99 pp., Natl. Ocean and Atmos. Admin., Washington, D. C. Levitus, S., Burgett, R. and Boyer, T.P., 1994. World Ocean Atlas, vol. 4, Temperature, 117 pp., Natl. Ocean and Atmos. Admin., Washington, D. C. MacAyeal, D.R., 1992. Irregular oscillations of the West Antarctic ice sheet. Nature, 359: 29-32. Maier-Reimer, E., Mikolajewicz, U. and Hasselmann, K., 1991. On the sensitivity of the global ocean circulation to changes in the surface heat flux forcing. Manabe, S. and Stouffer, R., 1997. Coupled ocean-atmosphere model response to freshwater input: Comparison to Younger Dryas event. Paleoceanography, 12: 321-336. Manabe, S. and Stouffer, R.J., 1988. Two stable equilibria of a coupled ocean-atmosphere model. Journal of Climate, 1: 841-866. Manabe, S. and Stouffer, R.J., 1995. Simulation of abrupt change induced by freshwater input to the North Atlantic Ocean. Nature, 378: 165-167. Marotzke, J. and Willebrand, J., 1991. Multiple equlibria of the global thermohaline circulation. Journal of Physical Oceanography, 21: 1372-1385. Mikolajewicz, U., 1996. A meltwater induced collapse of the 'conveyor belt' thermohaline circulation and its influence on the distribution of D14C and d18O in the oceans. 189, Report No. 189, 25 pp., Max-Planck Institut für Meteorologie, Hamburg, Germany. Oppenheimer, M., 1998. Global warming and the stability of the west Antarctic ice sheet. Nature, 393: 325-332. Pacanowski, R.C., 1996. MOM 2. Documentation, User's Guide and Reference Manual, GFDL Ocean Technical Report #3.2 edited by R. C. Pacanowski, Geophysical Fluid Dynamics Laboratory/NOAA, Princeton, N.J. Rahmstorf, S., 1994. Rapid climate transitions in a coupled ocean-atmosphere model. Nature, 372: 82-85. Rahmstorf, S., 1995a. Bifurcations of the Atlantic thermohaline circulation in response to changes in the hydrological cycle. Nature, 378: 145-149. Rahmstorf, S., 1995b. Climate drift in an ocean model coupled to a simple, perfectly matched atmosphere. Climate Dynamics, 11: 447-458. Sarkisyan, A.S., 1977. The diagnostic calculations of large-scale oceanic circulation. In: E.D. Goldberg et al. (Editor), The Sea, Vol. 6. J. Wiley and Sons, New York, pp. 363-458. Sarnthein, M. et al., 1995. Variations in Atlantic Ocean paleoceanography, 50°-85°N: A time-slice record of the last 30,000 years. Paleoceanography, 10: 1063-1094. Sarnthein, M. et al., 1994. Changes in east Atlantic deepwater circulation over the last 30,000 years: Eight Time Slice Reconstructions. Paleoceanography, 9: 209-267. Schmittner, A. and Stocker, T.F., 1999. The stability of the thermohaline circulation in global warming experiments. Journal of Climate, 12: 1117-1133. Scott, J.R., Marotzke, J. and Stone, P.E., 1999. Interhemispheric thermohaline circulation in a coupled box model. Journal of Physical Oceanography, 29: 351-365. Seidov, D. and Haupt, B.J., 1997. Simulated ocean circulation and sediment transport in the North Atlantic during the Last Glacial Maximum and today. Paleoceanography, 12: 281-306. Seidov, D. and Haupt, B.J., 1999. Last glacial and meltwater interbasin water exchanges and sedimentation in the world ocean. Paleoceanography, 14: 760-769. Seidov, D. and Maslin, M., 1999. North Atlantic Deep Water circulation collapse during the Heinrich events. Geology, 27: 23-26. Seidov, D., Sarnthein, M., Stattegger, K., Prien, R. and Weinelt, M., 1996. North Atlantic ocean circulation during the Last Glacial Maximum and subsequent meltwater event: A numerical model. Journal of Geophysical Research, 101: 16305-16332. Stocker, T.F., 1998. The seesaw effect. Science, 282: 61-62. Stocker, T.F. and Wright, D.G., 1991. Rapid transitions of the ocean's deep circulation induced by changes in surface water fluxes. Nature, 351: 729-732. Stocker, T.F. and Wright, D.G., 1996. Rapid changes in ocean circulation and atmospheric radiocarbon. Paleoceanography, 11(6): 773-795. Stocker, T.F., Wright, D.G. and Broecker, W.S., 1992. The influence of high-latitude surface forcing on the global thermohaline circulation. Paleoceanography, 7: 529-541. Stoessel, A., Kim, S.-J. and Drijfhout, S.S., 1998. The impact of Southern Ocean sea ice in global ocean model. J. Phys. Oceanogr., 28: 1999-2018. Toggweiler, J.R., Dixon, K. and Bryan, K., 1989. Simulations of radiocarbon in a coarse-resolution world ocean circulation model, 1, Steady state prebomb distribution. Journal of Geophysical Research, 94: 8217-8242. Toggweiler, J.R. and Samuels, B., 1980. Effect of sea ice on the salinity of Antarctic Bottom Water. J. Phys. Oceanogr., 25: 1980-1997. Vaughan, D.G., Bamber, J.L., Giovinetto, M., Russel, J. and Cooper, A.P.R., 1999. Reassessment of net surface mass balance in Antarctica. Journal of Climate, 29(4): 933-946. Vidal, L. et al., 1999. Link between the North and South Atlantic during the Heinrich events of the last glacial period. Climate Dynamics, 15: 909-919. Wang, X., Stone, P. and Marotzke, J., 1999a. Global thermohaline circulation. Part I. Sensitivity to atmospheric moisture transport. J. Climate, 12: 71-82. Wang, X., Stone, P. and Marotzke, J., 1999b. Global thermohaline circulation. Part II. Sensitivity with interactive atmospheric transports. J. Climate, 12: 83-91. Warrick, R.A., Barrow, E.M. and Wigley, T.M.L. (Editors), 1993. Climate and Sea Level Change: Observations, Projections and Implications. Cambridge University Press, N.Y. Weaver, A.J. and Hughes, T.M.C., 1994. Rapid interglacial climate fluctuations driven by North Atlantic ocean circulation. Nature, 367: 447-450. Weaver, A.J., Marotzke, J., Cummins, P.F. and Sarachik, E.S., 1993. Stability and variability of the thermohaline circulation. Journal of Physical Oceanography, 23: 39-60.
| * | Corresponding author: | Tel.: +1-814-865-1921 |
| Fax: + 1-814-865-3191 | ||
| E-mail address: dseidov@essc.psu.edu | ||
Table 1. Amplitudes of sea surface salinity anomalies (in psu) and corresponding
effective freshwater fluxes in Sv (see text).
|
|
|
|
|
|
CR - Control Case (annual
mean present-day sea surface climatology); NA - North Atlantic; SO - Southern Ocean; WED - the Weddell Sea.EFWF - effective freshwater flux relative to CR (see text) Salinity anomalies are added to the present-day annual mean sea surface salinity in the bands between 60ºN and 80ºN in NA, and/or 50ºS and the coast of the Antarctica (SO). The modified salinity was merged using a cosine filter to the unchanged field within two latitudinal grid points (8º). In the SO the anomalies are circumglobal. The WED low salinity is confined to the Weddell Sea only. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Table 2. Meridional Overturning in the Atlantic Ocean (north of 30ºS)
in Sv (1 Sv =106 m3/s)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||
|
|
|
|||
Table 3. Northward heat transport (in PW; 1 PW = 1015 W) in
Atlantic Ocean. Positive numbers mean northward heat transport, negative - southward
heat transport.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||
|
|
|
|||
a)
b)
Figure 1. Meridional overturning in the Atlantic Ocean (in Sv; 1 Sv=106
m3/s) in the control run (Exp. 1; a) and in NA (Exp. 4) with -2 psu
surface salinity anomaly (b). Positive values indicate clockwise motion (as
seen from east to west here); Negative values indicate counter-clockwise motion
(shaded).
Figure 2. Northward heat flux in the Atlantic Ocean (in PW; 1 PW=1015
W) in control run (solid line) and in the NA scenario (dashed line) depicted
in Figure 1b.
Figure 3. As in Figure 1 in the
SO scenario (Exp. 8) with surface salinity anomaly of -1 psu in the Southern
Ocean.
a)
b)
Figure 4. Meridional sections of the temperature anomalies (TSO-1psu-TCR)
in the Atlantic (32°W; in ºC) (a) and Pacific (170°W) (b) Oceans
in the case SO-1 psu (Exp. 8). The areas with negative anomalies are shaded.
Figure 5. As in Figure 2 for SO-1psu
(Exp. 8).
b)
Figure 6. Differences of the sealevel elevations (in cm) relative to the bottom (a) between the NA-2 psu (Exp. 4) and CR (Exp. 1), and (b) between the SO-1 psu (Exp. 8) and CR (Exp. 1).