1. Introduction
The volume of land-based ice, including Antarctica, Greenland, and ice caps and mountain glaciers, is roughly equivalent to a 70-m global sea level rise (Church et al. 2001). The potential loss of land-based ice can significantly affect global sea level in the next few centuries. Observational evidence suggests that the land-based ice is losing mass, at least since the 1990s, if not earlier [the loss of ice caps and mountain glaciers is even earlier (e.g., Lemke et al. 2007)], and this might be related to greenhouse gas–induced global warming (Peltier and Tushingham 1989; Rignot and Kanagaratnam 2006; Kaser et al. 2006; Meier et al. 2007; Lemke et al. 2007; Pritchard and Vaughan 2007; Cazenave et al. 2009; Peltier 2009; Pritchard et al. 2009; Velicogna 2009; van den Broeke et al. 2009; Rignot et al. 2011; Jacob et al. 2012). The estimated mass loss of mountain glaciers and ice caps (GIC) is about 402 ± 95 gigatons per year (Gt yr−1) in 2006, equivalent to a global mean sea level rise of 1.1 ± 0.24 mm yr−1 (Kaser et al. 2006; Meier et al. 2007), but a more recent estimation gives a much smaller number of 148 ± 30 Gt yr−1 for recent years (Jacob et al. 2012). The estimated mass loss from the Greenland Ice Sheet in recent years has been accelerating at an astonishing rate since the mid-1990s and has been estimated to be doubling in less than 10 yr (Rignot and Kanagaratnam 2006), possibly reaching a historical high of about 300 Gt yr−1 (Velicogna 2009; Rignot et al. 2011). Moreover, the Antarctic Ice Sheet, previously believed to be stable, also shows a significantly increased net mass loss of 165 ± 72 Gt yr−1 (Jacob et al. 2012) based on GRACE satellite observations, but it could be up to 250 Gt yr−1 (Velicogna 2009; van den Broeke et al. 2009; Rignot et al. 2011). In addition, ice sheet movement observations suggest that the thinning of the ice sheet at the edges due to ice dynamics is also accelerating for both the Greenland and Antarctic Ice Sheets in recent years (Pritchard et al. 2009). This mass loss along the edges of the ice sheets, especially on the ice shelf, would expose more grounded ice to warm oceanic currents and thus could further accelerate the mass loss of the ice sheets (Yin et al. 2011). The contribution of land-based ice loss to global sea level change has become dominant since approximately the late 1990s (Bindoff et al. 2007). For the increase of sea level rise from 1972 to 2008, it has been estimated that the largest contributions come from ocean thermal expansion (0.8 mm yr−1) and the melting of glaciers and ice caps (0.7 mm yr−1), with melt from Greenland and Antarctica contributing about 0.4 mm yr−1 (Church et al. 2011). On the other hand, Levitus et al. (2009) presented a much smaller contribution of the ocean thermal expansion to global sea level rise. For the most recent years, studies suggest that the contribution from land-based ice loss to global sea level rise is increasing and may have become the dominant force for global sea level rise (e.g., Peltier 2009; Cazenave et al. 2009; Jacob et al. 2012). Given the complexity of the estimation of the observed sea level rise, different authors have given slightly (sometimes significantly) different values for the contributions from each of the different components, especially the contributions from land-based ice loss (Rignot and Kanagaratnam 2006; Kaser et al. 2006; Stouffer et al. 2007; Meier et al. 2007; Lemke et al. 2007; Pritchard and Vaughan 2007; Cazenave 2006; Stammer 2008; Peltier 2009; Pritchard et al. 2009; Velicogna 2009; van den Broeke et al. 2009; Cazenave et al. 2009; Rignot et al. 2011; Jacob et al. 2012). Nevertheless, all of these estimates have pointed in the same direction, namely, that land-based ice loss has played a more and more dominant role in sea level rise in the past 30 yr. This implies that it is very likely that, in the twenty-first century, sea level rise associated with elevated atmospheric CO2 concentrations may be driven mainly from the runoff of meltwater from land-based ice into the ocean. This meltwater could provide another important forcing of the climate system in addition to greenhouse gases and aerosols. Here, our focus is on understanding what would be the impact of this potential loss of land-based ice on aspects other than global mean sea level, such as the global surface climate and the Atlantic meridional overturning circulation (AMOC).
The loss of land-based ice could contribute significant amounts of fresh water into the North Atlantic, affecting the strength of the AMOC (e.g., Manabe and Stouffer 1988; Rahmstorf 1996; Stouffer et al. 2006, 2007). The AMOC [or the thermohaline circulation (THC)] is a global scale ocean circulation. It pulls warm and salty upper-ocean water into the subpolar North Atlantic, where this water loses its heat to the atmosphere, becomes dense, sinks, flows southward, and upwells elsewhere in the World Ocean (Dickson and Brown 1994; Stocker and Broecker 1994). This circulation is important for the global climate because it transports significant amounts of heat into the subpolar North Atlantic region, resulting in a mild climate downstream in Europe, and also causes global surface temperature variations. Multiple modeling studies suggest that a dramatic weakening of this circulation can cool the North Atlantic and surrounding regions (e.g., Keigwin and Jones 1994; Sakai and Peltier 1997; Stouffer et al. 2006; Hu et al. 2012a), providing a physical mechanism to explain the abrupt climate change events revealed by paleoproxy records, such as Dansgaard–Oeschger events or Heinrich events (Dansgaard et al. 1993; Ditlevsen et al. 2005; Heinrich 1988; Hemming 2004). Multiple modeling studies also indicate that the AMOC could affect the climate response to increased greenhouse gas forcing in the future (e.g., Manabe et al. 1991; Delworth et al. 1993; Timmermann et al. 1998; Schmittner and Stocker 1999; Hu et al. 2004b; Gregory et al. 2005; Schmittner et al. 2005; Timmermann et al. 2005a,b; Dahl et al. 2005; Zhang and Delworth 2005; Levermann et al. 2005; Broccoli et al. 2006).
For the future climate, modeling studies of land-based ice loss to date are mostly focused on the Greenland Ice Sheet since runoff water from this ice sheet would directly flow into the regions where deep convection occurs in the North Atlantic. In the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report, it was shown that models that did not include the potential land-based ice melt flux all produced a weakened AMOC in the next two centuries, mainly associated with the warming that induced a more stably stratified ocean (e.g., Dixon et al. 1999; Mikolajewicz and Voss 2000; Thorpe et al. 2001; Kamenkovich et al. 2003; Hu et al. 2004b; Gregory et al. 2005; Schmittner et al. 2005; Meehl et al. 2007). Recent modeling studies suggest that the inclusion of idealized meltwater from the Greenland Ice Sheet in model simulations of future climate could slow the AMOC further, but only if the meltwater flux is strong enough (Hu et al. 2009, 2011). Other modeling studies show mixed results: some suggest a dramatic weakening of the AMOC in a future warmer climate (Fichefet et al. 2003; Swingedouw et al. 2006, 2007), and some show a negligible effect (e.g., Huybrechts et al. 2002; Ridley et al. 2005; Jungclaus et al. 2006; Gerdes et al. 2006; Vizcaino et al. 2008). In comparison with these studies, the results from Hu et al. (2009, 2011) are more realistic because of an improved experimental setup, that is, different rates of Greenland Ice Sheet melting are proposed on the basis of observations, and the meltwater is only added to the ocean during the summer half of the year. As suggested by Hu et al. (2011), by adding the meltwater in summer only, the response of the AMOC to the meltwater is weaker relative to the simulation when the same annual amount of meltwater was added in all seasons. This is because, if the meltwater is added in summer only, part of the meltwater will be transported out of the deep convection region by surface ocean currents, thus reducing the meltwater’s impact on winter deep convection in the subpolar North Atlantic.
On the basis of the research of Hu et al. (2009, 2011), we expand the scope of those studies by including not only the effects of Greenland Ice Sheet loss on the AMOC, surface climate, and sea level in the next two centuries, but also by including the loss of glaciers, ice caps, and the West Antarctic Ice Sheet. Our study here is unique in comparison with many previous studies (e.g., Huybrechts et al. 2002; Fichefet et al. 2003; Ridley et al. 2005; Jungclaus et al. 2006; Gerdes et al. 2006; Swingedouw et al. 2006, 2007; Stouffer et al. 2006, 2007) in that our simulations are not only run under the IPCC A1B scenario for projected changes of greenhouse gases over the next two centuries with idealized melting scenarios based on observations for the Greenland and West Antarctic Ice Sheets, ice caps, and mountain glaciers, but they are also run with the ice sheet melt flux added into the ocean only during boreal/austral summer months. This latter approach significantly differs from many previous studies of this kind, where the meltwater runoff was applied during the whole year (e.g., Huybrechts et al. 2002; Fichefet et al. 2003; Ridley et al. 2005; Jungclaus et al. 2006; Gerdes et al. 2006; Swingedouw et al. 2006, 2007; Stouffer et al. 2006, 2007). As shown in more detail in the next section, the runoff from the land-based ice increases linearly with time in our simulations, which also differs from previous studies where the runoff was kept constant (e.g., Stouffer et al. 2006, 2007).
The focus of this study is on the response of the AMOC to this prescribed freshwater runoff and the potential influence of this runoff on global and regional surface temperature and precipitation. Since the resulting change of the steric and dynamic sea level caused by a further weakening of the AMOC in Community Climate System Model version 3 (CCSM3) has been discussed in more detail in previous studies by Hu et al. (2009, 2011), we omit it from this paper. The rest of the paper is organized as follows: section 2 outlines the model and experiment design, section 3 describes the AMOC changes, section 4 relates the changes of the surface climate to the AMOC changes, and section 5 gives the conclusions.
2. Model and experiments
The coupled climate model used in this study is the National Center for Atmospheric Research (NCAR) CCSM3, which was developed by NCAR scientists in collaboration with U.S. Department of Energy research laboratories and university scientists (Collins et al. 2006). The atmospheric component in CCSM3 is the Community Atmospheric Model version 3 (CAM3), which uses spectral dynamics at T42 resolution (grid points roughly every 280 km) and 26 vertical hybrid levels. The ocean model is a version of the Parallel Ocean Program (POP) developed at Los Alamos National Laboratory with 1° horizontal resolution and enhanced meridional resolution (½°) in the equatorial tropics and the North Atlantic with 40 vertical levels. The sea ice model is the Community Sea Ice Model version 5 (CSIM5) with elastic–viscous–plastic dynamics, a subgrid-scale thickness distribution, and energy-conserving thermodynamics. The land model is the Community Land Model version 3 (CLM3).
In this study, nine simulations (summarized in Table 1) are carried out using CCSM3, and all simulations use the IPCC Special Report on Emissions Scenarios (SRES) A1B scenario as the climatic forcing. This is a midrange greenhouse gas emission scenario among the SRES scenarios. The CO2 concentrations in this scenario range from 368.5 ppm at year 1999 to 688.5 ppm at year 2099 and then are kept constant at year 2099 levels afterward. The first experiment is the standard A1B simulation without the inclusion of the land-based ice melting (hereafter A1Bexp). All experiments included in this study are branched from a single realization of the twentieth-century all-forcing simulation at year 2000 and are run from the years 2000 to 2199.
List of experiments, the initial rate of meltwater, the increment, the maximum rate of meltwater, and the corresponding global mean sea level rise.
In the first two sensitivity experiments, the meltwater from the Greenland Ice Sheet is assumed to only flow into the seas surrounding the southern half of Greenland (Fig. 1a, red shaded area) during the summer season from April to October. This meltwater flux is added uniformly in these seas (58°–72°N and 12°–60°W) with an initial annual mean melting rate of 0.01 Sv (1 Sv ≡ 106 m3 s−1 or 1 mm yr−1 global sea level equivalent). This melting rate is close to or a bit larger than the observed current rate of Greenland Ice Sheet mass loss (Lemke et al. 2007; Velicogna 2009; Jacob et al. 2012). This initial rate of melting increases by 1% or 3% yr−1, compounded until 2099, and then is kept at year 2099 values until 2199 (hereafter referred to as Gr 1% and Gr 3%, respectively).
Oceanic regions where the land ice meltwater is put into the ocean. (a) The red (blue) area indicates the region where Greenland (West Antarctic) meltwater is uniformly added into the model. (b) Regions where the meltwater from glaciers and ice caps is added into the ocean. As the color changes from blue to red, the runoff volume changes from small to large. Because of the significant variations of the river runoff among different rivers, we could not assign a reasonable scale here; instead, we just show the relative importance of each river in our simulation.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
Next, we simulate the effect of the loss of the ice caps and mountain glaciers. Since the meltwater from the ice caps and mountain glaciers would go into the global river system, this meltwater is parameterized as river runoff into the ocean, as shown in Fig. 1b. The colors shown in Fig. 1b represent the relative amount of runoff in each river, changing from minimum (blue) to maximum (red). In our simulations, we initially increase global river runoff by 0.01 Sv at year 2000. This value then increases 1% or 3% yr−1 until 2099, when it becomes constant at the year 2099 value to the year 2199 (hereafter referred to as Roff 1% and Roff 3%, respectively).
For the melting of the West Antarctic Ice Sheet, we spread the meltwater uniformly south of 65°S, between 160°E and 20°W (light blue shading in Fig. 1a). The initial rate of melting is 0.002 Sv at year 2000, approximately equivalent to the observed rate of Antarctic Ice Sheet mass loss in the early twenty-first century inferred from Gravity Recovery and Climate Experiment (GRACE) satellite measurements (Velicogna 2009), but a bit smaller than the most recent observed estimation of the West Antarctic Ice Sheet mass loss (e.g., Jacob et al. 2012). This initial rate of melting is increased by 3% yr−1 until year 2099, and it is kept constant at the year 2099 value to year 2199 (Ant 3%).
After considering the mass loss from ice sheets, ice caps, and mountain glaciers individually, three addition experiments are carried out to assess the combined effect of the land ice loss on global climate. First, we combine the loss of the Greenland and West Antarctic Ice Sheets together. The combined rate of mass loss is 0.012 Sv, and this rate is assumed to increase 3% yr−1 until 2099, when it is kept constant at the year 2099 level until 2199 (GrAn 3% hereafter). Next, we consider all land-based ice loss in two simulations with a combined initial rate of melting flux 0.022 Sv (0.01 Sv for the Greenland Ice Sheet, 0.01 Sv for the glaciers and mountain ice caps, and 0.002 Sv for the West Antarctic Ice Sheet). This melting flux increases 1% or 3% yr−1 to year 2099, when it is kept at year 2099 levels until 2199 (GrAnRf 1% and GrAnRf 3%, respectively).
The rates of the land-based ice loss and resulting global mean sea level rise due to these land-based ice losses in our experiments are shown in Fig. 2 and summarized in Table 1. The maximum rate of the ice mass loss is 0.027 Sv for Gr 1% and Roff 1%, 0.192 Sv for Gr 3% and Roff 3%, 0.038 Sv for Ant 3%, 0.230 Sv for GrAn 3%, 0.059 Sv for GrAnRf 1%, and 0.422 Sv for GrAnRf 3%. This maximum rate of ice mass loss lasts for 100 yr, from 2100 to 2199, in our simulations. The potential eustatic global mean sea level rise due to the idealized land ice mass loss is 14.9 cm for the Gr 1% and Roff 1% by 2099 and reaches 38.5 cm by 2199. It is 53.8 cm for the Gr 3% and Roff 3% by 2099 and 221.3 cm by 2199. For the Ant 3% and GrAn 3% cases, the eustatic sea level rise is 10.8 cm and 64.5 cm by 2099 and 44.3 cm and 265.5 cm by 2199. For the GrAnRf 1% and GrAnRf 3% cases, the eustatic sea level rise is 32.9 cm and 118.3 cm by 2099 and 84.6 cm and 487.2 cm by 2199.
(a) The time-evolving annual mean rate of meltwater from different sources into the ocean and (b) the associated global mean sea level rise due to these meltwaters. Note that, because the initial freshwater runoff rate and the rate of increment later on are the same for Gr 1% (Gr 3%) and Roff 1% (Roff 3%), the dark blue (red) line is overwritten by the sky blue (orange) line.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
The rate of land-based ice loss increase used here is lower than that suggested by the most recent observations (e.g., Velicogna 2009; Rignot et al. 2011; Jacob et al. 2012). These observations indicate a doubling of the Greenland and Antarctic ice loss in less than 10 yr. Since it is still unclear whether this rate of increased ice loss from these ice sheets would be maintained in the future, we use more moderate rates of ice loss in our studies, which constitutes an increase of ice loss by either 1% yr−1 or 3% yr−1. If a higher rate of ice mass loss occurred in the future, it would not change our results qualitatively based on the studies of Hu et al. (2009, 2011).
3. Influence of the land-based ice loss on the AMOC and the associated MHT
a. Changes of the AMOC
The time evolution of the AMOC index, defined as the maximum of the Atlantic meridional overturning streamfuction below 500-m depth, is shown in Fig. 3a. The mean AMOC is about 19.4 Sv (Table 2, Fig. 4a) in the late twentieth century, similar to that suggested from observations (e.g., Ganachaud and Wunsch 2000). In the standard A1Bexp simulation, the AMOC (Fig. 3a, black line) weakens to about 15 Sv by the end of the twenty-first century, then stabilizes in the second half of the twenty-second century at about 14 Sv (Table 2). In comparison to A1Bexp, the AMOC in all sensitivity experiments is not significantly different from that in the A1Bexp in the early half of the twenty-first century (Fig. 3a). After the mid-twenty-first century, the AMOC in Gr 1% (dark blue line), Roff 1% (sky blue line), and Ant 3% (purple line) varies in a similar way to A1Bexp. By the end of the twenty-second century (Fig. 5, Table 2), the AMOC is about 0.7 Sv weaker in the Gr 1% and Roff 1% and 1.1 Sv weaker in Roff 3% simulations but about 0.4 Sv stronger in the Ant 3% simulation relative to A1Bexp. This suggests that a low rate of land-based ice loss from Greenland, glaciers and ice caps, or the West Antarctic will not affect the AMOC much in the next two centuries.
Changes of the AMOC and Atlantic MHT: (a) time-evolving AMOC index, (b) time-evolving Atlantic MHT at 24°N, (c) the annual mean Atlantic MHT averaged over 2080–99 for the sensitivity simulations and averaged over 1980–99 (20C, dashed line), and (d) the annual mean Atlantic MHT averaged over 2180–99 for the sensitivity simulations and averaged over 1980–99 (20C, dashed line).
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
The 20-yr mean AMOC index and Atlantic MHT at 24°N for each simulation in the late twenty-first and twenty-second centuries.
The 20-yr mean Atlantic meridional streamfuction (AMSF) in (a) the late twentieth century (1980–99) and (b) the late twenty-second century (2180–99). Also shown are the AMSF anomalies in the late twenty-second century (2180–99): (c) the A1Bexp AMSF anomaly relative to the mean of the last 20 yr of the twentieth century and the AMSF anomaly in (d) Gr 3%, (e) GrAn 3%, and (f) GrAnRf 3% simulations relative to the same 20-yr AMSF mean of the A1Bexp. Contour interval is 1 Sv.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
AMSF anomaly in the late twenty-second century (2180–99): the AMSF anomaly in (a) Gr 1%, (b) Ant 3%, (c) Roff 3%, and (d) GrAnRf 1% simulations relative to the same 20-yr AMSF mean of the A1Bexp. Contour interval is 0.02 Sv.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
Saenko et al. (2003), using an earth system model of intermediate complexity, suggested that the AMOC strength in a steady state is controlled by southward moisture transport from the subtropics to the subpolar southern oceans. An increased southward moisture transport (freshening) into the southern oceans would induce a strengthening of the AMOC by increasing the density contrast between the subpolar North Atlantic and the subpolar southern oceans. Hu et al. (2004a) indicated that, under transient greenhouse gas forcing, an increase of the southward moisture transport and a freshening of the southern oceans will not induce an enhanced AMOC because of the larger density reduction in the subpolar North Atlantic due to the warming from the greenhouse gas forcing. Stouffer et al. (2007), who added 1-Sv freshwater forcing in the region south of 60°S, found there is no significant impact on the strength of the AMOC from this freshwater forcing in the southern oceans. However, this does affect the sea surface temperature and salinity. In our Ant 3% simulation, the increased discharge of meltwater from the West Antarctic Ice Sheet (up to 0.038 Sv during the twenty-second century) leads to a cooler and fresher southern ocean in comparison to those in the A1Bexp (Figs. 6a,b). This result agrees with Stouffer et al. (2007). On the other hand, the North Atlantic becomes saltier and warmer. Overall, the salinity increase leads to an increase in surface water density in the Ant 3% simulation relative to the A1Bexp (Fig. 6c), resulting in slightly stronger convection and a somewhat more active AMOC (0.4 Sv stronger; Figs. 4b and 5b and Table 2).
Anomalies of (a) sea surface salinity, (b) temperature, and (c) density in the Ant 3% simulation relative to the A1Bexp at the end of the twenty-second century. Contour intervals are 0.01 psu for salinity, 0.01°C for temperature, and 0.01 kg m−3 for density.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
To further explore the AMOC’s response to increased southern ocean freshwater input, we plot zonal mean salinity, temperature, and density anomalies relative to A1Bexp in the Atlantic sector for the Ant 3% simulation (Fig. 7). In comparison to the A1Bexp, the upper ocean becomes fresher and cooler south of 30°N by the end of the twenty-second century with a density reduction, and in regions north of 30°N, the upper ocean becomes saltier and warmer with a density increase. This indicates that the salinity changes are the dominant factor in controlling the upper-ocean density changes in this simulation. However, in the deeper ocean, the temperature change seems to play a more dominant role. Although the exact mechanism is still not clear, earlier studies suggest that the meridional steric height gradient is linked to AMOC strength (e.g., Hughes and Weaver 1994; Thorpe et al. 2001; Hu et al. 2004a). Steric height is defined as vertically integrated density. Figure 7c shows an increased water column density at 60°N and a decreased density at 30°S in the Atlantic. This leads to an increase of the meridional steric height gradient and a stronger AMOC. A further calculation of the steric height gradient between 30°S and 60°N of the Atlantic shows an increase of 0.05 cm degree−1 latitude in the Ant 3% case than in the A1Bexp, averaged over the last 20 yr of the twenty-second century. Although this change is small, it is consistent with a slightly strengthened AMOC in Ant 3% simulation relative to the A1Bexp. Thus, we propose here that, in the Ant 3% simulation, as the freshwater forcing from the West Antarctic Ice Sheet increases, this freshwater is diverged northward (Stouffer et al. 2007), leading to a reduction of the upper-ocean density at 30°S. As this density reduction at 30°S becomes large enough in comparison with the density reduction at 60°N induced by warming associated with elevated greenhouse gas concentrations, the meridional steric height gradient increases, leading to a strengthening of the AMOC and an increased northward salt and heat transport. This results in a salinity and temperature increase in the subpolar North Atlantic. This suggests that if the runoff from the West Antarctic Ice Sheet is much larger than what we have used here, this runoff could potentially strengthen the AMOC more significantly, agreeing with Saenko et al. (2003).
Zonally averaged (a) salinity, (b) temperature, and (c) density anomalies in the Atlantic sector of the Ant 3% simulation relative to A1Bexp at the end of the twenty-second century. Contour intervals are 0.01 psu for salinity, 0.04°C for temperature, and 0.005 kg m−3 for density.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
With a potentially higher rate of land-based ice loss, the AMOC does weaken further in comparison with the A1Bexp in Gr 3% (red line), GrAn 3% (green line), and GrAnRf 3% (dark green line; Figs. 3a, 4 and Table 2). In the GrAnRf 1% (orange line) simulation, the AMOC weakens more (about 1.5 Sv) in the twenty-second century relative to A1Bexp. In the Roff 3% (brown line) simulation, the AMOC differs significantly from that in the A1Bexp, only in the late twenty-second century (about 2.1 Sv weaker). Overall, the most significant changes of AMOC from the A1Bexp only occur in those experiments where the ice mass loss increases at 3% yr−1 and the Greenland Ice Sheet loss is also included, for example, 4.1 Sv weaker in the Gr 3% case, 4.7 Sv in the GrAn 3% case, and 5.6 Sv in the GrAnRf 3% case (Table 2). This suggests that a high rate of Greenland ice loss may potentially play the most crucial role for future changes of the AMOC relative to land-based ice loss from other sources. This implies that only the freshwater discharged directly into the North Atlantic could potentially affect the strength of the AMOC over the next two centuries if there is not significant ice loss from the West Antarctic Ice Sheet. On the other hand, our results also suggest that a low rate of land-based ice loss would not significantly alter the AMOC, agreeing with Hu et al. (2009, 2011).
b. Changes of the MHT
The mean Atlantic meridional heat transport (MHT) at 24°N in the late twentieth century is about 1.05 PW (PW ≡ 1015 W), similar to the observed estimates (Ganachaud and Wunsch 2000). In the A1Bexp, the MHT at 24°N reduces to about 0.87 PW by the end of the twenty-first century and becomes flat afterward (Fig. 3b, Table 2). In the cases with a further slowdown of the AMOC, the MHT reduces even more at 24°N by 0.23 PW for Gr 3%, 0.26 PW for GrAn 3%, and 0.3 PW for the GrAnRf 3% simulations compared to that in the A1Bexp by the end of the twenty-second century (Fig. 3b, Table 2). This agrees with previous studies that suggest that the meridional heat transport in the Atlantic is quasi-linearly related to the strength of the AMOC (Smith et al. 2000; Hu 2001). Meridionally, this reduction of the MHT is obvious at almost all latitudes in the Atlantic basin, as shown in Figs. 3c and 3d, except at the latitudes north of 60°N. The increase of the northward MHT north of 60°N in our simulations is caused by enhanced deep convection in the Nordic Seas (Hu et al. 2004b). The potential impact of the reduced MHT on the regional climate around the North Atlantic will be discussed in the next section.
c. Upper-ocean freshwater budget of the North Atlantic
A freshwater budget analysis indicates that for the cases with a further AMOC weakening (GrAnRf 3%, GrAn 3%, and Gr 3%), the surface freshwater input into the North Atlantic between 40° and 80°N is higher than those cases with less changes of the AMOC relative to the standard A1Bexp (Fig. 8a). The difference of the sea ice melt flux into this subpolar North Atlantic region among the different cases is small during the first 80 yr of the twenty-first century, and then this flux becomes stabilized for the simulations with a further-weakened AMOC (Fig. 8b). However, it continuously declines in the other simulations, indicating a reduced sea ice export from the Arctic into the subpolar North Atlantic caused by the reduction of Arctic sea ice cover (Figs. 8e,f).
Freshwater budget analysis: (a) changes of the surface freshwater input (positive represents oceanic freshwater gain), (b) surface freshwater input due to melting of sea ice, (c) meridional freshwater divergence in the North Atlantic between 40° and 80°N, and (d) the freshwater budget for the upper 1000 m of the North Atlantic (40°–80°N, positive represents oceanic freshwater divergence). Also shown are the time-evolving Northern Hemisphere (e) maximum and (f) minimum sea ice extent.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
The oceanic freshwater divergence, which is calculated relative to the global mean salinity, for the upper 1000 m between 40° and 80°N shows, in general, a weakening trend in almost all simulations due to the weakening of the AMOC, which reduces the ocean’s ability to diverge the freshwater anomaly there (Fig. 8c). In fact, the freshwater export from the Arctic into the North Atlantic increases in our simulation primarily because of the sea ice reduction in both extent and thickness. This sea ice reduction leads to a decrease of the sea ice export but an increase in liquid freshwater export from the Arctic into the North Atlantic (Hu et al. 2009, 2011). However, because of the significantly increased surface freshwater input, this divergence strengthens a bit in those simulations with a further-weakened AMOC.
Overall, the net freshwater budget for the upper 1000-m ocean layer shows that, in the subpolar North Atlantic region, the freshwater deficit (positive values in Fig. 8d) declines in all simulations. However, throughout the 200-yr simulation, this freshwater deficit never changed to a net freshwater gain in the A1Bexp. In other words, the slowing of the AMOC in the A1Bexp is not primarily induced by changes of the freshwater forcing in the upper 1000-m ocean layer but by the warming related to elevated greenhouse gases, as indicated by previous studies (e.g., Dixon et al. 1999; Mikolajewicz and Voss 2000; Thorpe et al. 2001; Kamenkovich et al. 2003; Hu et al. 2004b; Gregory et al. 2005; Schmittner et al. 2005; Hu et al. 2011). On the other hand, only in those simulations where the freshwater deficit turns into a net freshwater gain in the twenty-first or twenty-second century does the AMOC spin down further relative to the A1Bexp simulation.
To further analyze the changes of North Atlantic deep convection and the AMOC, we show the sea surface salinity (SSS) and the maximum March mixed layer depth in Figs. 10, 11 (shown later) . The simulated SSS mean state in the late twentieth century agrees reasonably well with observations (Fig. 9a). Certain biases of the SSS in CCSM3 relative to observations have been discussed by Large and Danabasoglu (2006). In the late twenty-second century, SSS increases in the subtropical regions and is reduced in the polar regions, in general, in A1Bexp (Figs. 9b,c). The elevated SSS in the subtropics is due primarily to intensified evaporation, and SSS reduction in the polar region is due to the reduction of sea ice cover (Figs. 8e,f), increased precipitation (Figs. 14b,d, shown in section 4.), and river runoff into the Arctic basin. All of these processes are associated with global warming induced by increased greenhouse gas forcing. It can be clearly seen that the sea ice meltwater is being transported out of the Arctic along the east coast of Greenland, then into the northern Labrador Sea and along the west coast of Greenland (Figs. 9b,c). Part of this freshwater is diverged into the subtropical gyre though the Gulf Stream extension. However, the salinity actually increases in the Labrador Sea in the A1Bexp (Fig. 9c) since the fresher sea ice meltwater is not able to be transported into the central Labrador Sea, and the local surface freshwater input is less than that in the late twentieth century. Corresponding to these salinity changes, deep convection in the North Atlantic also changes (Figs. 10a–c). Deep convection in the A1Bexp strengthens in the Labrador Sea but weakens in the Irminger Sea and the Nordic Sea regions. Thus, it is the reduction of deep convection in the Nordic (especially in the northern part of the Nordic Sea) and Irminger Seas that leads to the weakening of the AMOC in the A1Bexp. In contrast, deep convection strengthens significantly in the Labrador Sea in the A1Bexp (Fig. 10c).
Simulated 20-yr mean SSS for (a) the late twentieth century using observed historical climate forcing and (b) the late twenty-second century for A1Bexp. Also shown are (c) the SSS anomaly for A1Bexp in the late twenty-second century relative to the late twentieth century and the SSS anomaly of the (d) Gr 3%, (e) GrAn 3%, and (f) GrAnRf 3% simulations relative to A1Bexp in the late twenty-second century. Contour interval is 1 psu for (a) and (b), 0.3 psu for(c), and 0.2 psu for (d)–(f).
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
Simulated 20-yr mean March maximum mixed layer depth over the upper 300 m for (a) the late twentieth century using observed historical climate forcing and (b) the late twenty-second century for A1Bexp. Also shown are (c) the March maximum mixed layer depth anomaly for A1Bexp in the late twenty-second century relative to the late twentieth century and the March maximum mixed layer depth anomaly of the (d) Gr 3%, (e) GrAn 3%, and (f) GrAnRf 3% simulations relative to A1Bexp in the late twenty-second century. Contour interval is 100 m for (a)–(c) and 50 m for (d)–(f).
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
In the three 3% simulations with the Greenland Ice Sheet included, the North Atlantic becomes significantly fresher (up to −2 psu) compared to the A1Bexp by the end of the twenty-second century because of the additional freshwater input from the Greenland Ice Sheet, ice caps, and glaciers (Figs. 9d–f). This freshening of the surface water makes the surface water lighter and leads to a significant weakening of deep convection in the Labrador Sea in these three simulations (Figs. 10d–f). As a result, the AMOC weakens further in these three simulations relative to the A1Bexp.
From the zonal mean salinity and potential density fields (Figs. 11, 12), the North Atlantic becomes fresher in the upper 50 m north of 30°N, reflecting an increased export of liquid freshwater from the Arctic into the North Atlantic due to the loss of sea ice cover in the A1Bexp (Fig. 11a). The subsurface water actually becomes saltier associated with the weakened deep convection, which leads to less surface freshwater being mixed into the subsurface ocean. The zonal mean density field shows a uniform density decrease with the maxima at the surface (Fig. 12a). This further indicates that the weakening of deep convection is induced primarily by surface warming in the A1Bexp. With the added land-based ice runoff into the North Atlantic, the ocean becomes fresher in comparison to the A1Bexp, especially in the upper North Atlantic (Figs. 11b–d). This leads to a significant surface density reduction (Figs. 12b–d), which suppresses deep convection in the subpolar North Atlantic, leading to a further weakening of the AMOC in these three simulations.
Atlantic sector zonal mean salinity anomalies in the late twenty-second century (2180–99): (a) A1Bexp salinity anomaly relative to the mean of the last 20-yr period in the twentieth century and the salinity anomaly in (b) Gr 3%, (c) GrAn 3%, and (d) GrAnRf 3% simulations relative to the same 20-yr AMSF mean of the A1Bexp. Contour interval is 0.1 psu.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
As in Fig. 11, but for potential density. Contour interval is 0.03 kg m−3.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
d. Influence of the Bering Strait on the AMOC’s response to land-based ice loss
Another interesting point we will review here is the potential influence of the Bering Strait transport on the AMOC’s response to land ice discharge in the future. This influence has been addressed by a series of previous studies (De Boer and Nof 2004a,b; Hu and Meehl 2005; Hu et al. 2007, 2008, 2010, 2011, 2012a,b). The Bering Strait is a shallow and narrow strait connecting the Pacific and North Atlantic via the Arctic. In our simulation, about 0.9 Sv of fresher North Pacific water flows into the Arctic via the Bering Strait in the late twentieth century, similar to that of the observations (0.8 Sv; Woodgate and Aagaard 2005). A comparison of the Bering Strait mass and freshwater transport (Fig. 13) with the AMOC index (Fig. 3a) clearly shows that, as the AMOC weakens, the Bering Strait mass transport is also reduced and less freshwater is transported from the North Pacific into the Arctic and the North Atlantic. These changes at the Bering Strait when the AMOC weakens are because the Bering Strait transport is primarily controlled by the sea level difference between the North Pacific and Arctic (Shaffer and Bendtsen 1994). As the AMOC weakens, the sea level difference becomes smaller and the Bering Strait throughflow becomes weaker (Hu et al. 2008). This reduction of freshwater transport leads to reduced freshwater convergence in the subpolar North Atlantic deep convection regions. Therefore, the net effect of this reduced Bering Strait transport due to AMOC weakening is to destabilize the oceanic stratification in the subpolar North Atlantic by reducing the freshwater convergence there, thus tending to strengthen the AMOC or to work against the trend of AMOC weakening due to the warming from increased greenhouse gases. This suggests that, if the Bering Strait transport did not reduce, the AMOC would have weakened further in all our simulations.
Time-evolving Bering Strait (a) mass transport and (b) freshwater transport. Positive indicates northward transport.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
4. Effect of the land-based ice discharge on global temperature and precipitation
a. Global surface temperature changes
In general, as the global mean temperature rises because of greenhouse gas forcing, the global precipitation also increases (figure not shown). By the end of the twenty-second century, the global mean temperature in A1Bexp rises by 2.86°C. By including the idealized land-based ice mass loss in our simulations, the global mean temperature does not significantly differ from that in A1Bexp during most of the twenty-first century. By the late twenty-second century, the simulations with a further-weakened AMOC show a moderate reduction of global warming, with a maximum reduction of 0.41°C (or 13%) in the GrAnRf 3% simulation (Fig. 6b). Therefore, moderate land-based ice loss will not alter the global mean temperature much in the near future, consistent with Hu et al. (2009, 2011).
Regionally, the warming in A1Bexp is much higher in the northern high latitudes, with a maximum warming of more than 8°C (Figs. 14a,c). This is consistent with the model results from the IPCC Fourth Assessment Report (Meehl et al. 2007). Studies indicate that this amplified polar warming possibly receives contributions from the combined effect of the CO2 forcing, increased meridional latent heat flux, cloud longwave and surface albedo radiative feedbacks (including sea ice and snow albedo feedback processes), and atmospheric dynamical feedbacks (e.g., Schmittner and Stocker 1999; Cai 2006; Hegerl et al. 2007; Lu and Cai 2010). The polar amplified warming can be illustrated more clearly by removing the global mean warming from Figs. 14a and 14c, as displayed in Figs. 15a and 15c. In general, the warming on land is greater than that over the oceans, except in the Arctic region, which shows the amplified polar warming. Earlier studies using a surface energy budget analysis suggest that the lower water availability over land, which leads to less evaporative cooling over land than ocean, might be the major reason for more warming over land than over ocean with increasing greenhouse gases (Manabe et al. 1991; Sutton et al. 2007).
The 20-yr mean surface (a),(c) temperature and (b),(d) precipitation anomalies: anomalies in the A1B simulation averaged over the last 20 years of the (a),(b) twenty-first and (c),(d) twenty-second centuries relative to the late twentieth century (1980–99). Contour interval is 0.8°C for surface temperature and 0.1 m yr−1 for precipitation.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
The 20-yr mean surface temperature in A1Bexp with (a),(c) the global mean warming removed and (b),(d) the zonal mean warming removed for (a),(b) the late twenty-first century and (c),(d) the late twenty-second century. Contour interval is 0.5°C.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
By removing the zonal mean warming (Figs. 15b,d) in A1Bexp, there is a different regional warming pattern comparing Figs. 14a and 14c. Although the land still warms more than ocean, the relative minimum of warming in the North Atlantic, especially the subpolar North Atlantic, becomes much more prominent. On the other hand, more warming can be seen in the South Atlantic. This pattern of temperature change is at least partly associated with the weakening of the AMOC, which transports less heat meridionally from the South Atlantic into the subpolar North Atlantic, inducing a cooling in the North Atlantic and a warming in the South Atlantic (e.g., Crowley 1992; Stouffer et al. 2006). This reduced warming signal in the subpolar North Atlantic further affects the temperature change downstream in Europe, where there is relatively less warming, amounting to nearly 1°C in most parts of Europe. Moreover, the weakened AMOC also reduces the magnitude of warming in Greenland, agreeing with a previous study (Hu et al. 2010). This reduced warming could lead to a slower ice loss from the Greenland Ice Sheet, thereby representing a negative feedback.
Compared to A1Bexp, by the end of the twenty-first century, the regional surface temperature changes in the three simulations (Gr 3%, GrAn 3%, and GrAnRf 3%) with a further slowdown of the AMOC show a pattern that resembles the negative phase of the interdecadal Pacific oscillation (IPO; based on Pacific basin-wide SSTs; Fig. 16) (Power et al. 1999; Arblaster et al. 2002; Meehl and Hu 2006), which is comparable in many ways to the Pacific decadal oscillation (PDO; based on North Pacific SSTs) (Mantua et al. 1997). This pattern is especially evident in the GrAnRf 3% simulation (Fig. 16). The negative phase of the IPO is characterized by a pattern of negative SST anomalies in the tropical Pacific that becomes wider from the west Pacific to the east Pacific. Negative SST anomalies also are evident along the west coast of the Americas toward both the Northern and Southern Hemispheres and extend to the west in the subpolar North Pacific. In the subtropical northwest and southwest Pacific, there are opposite-sign sea surface temperature anomalies (Meehl and Hu 2006). This negative phase of the IPO pattern has been linked to drought conditions in the southwest United States (e.g., Meehl and Hu 2006).
The 20-yr mean surface temperature anomalies in (a),(b) Gr 3%, (c),(d) GrAn 3%, and (e),(f) GrAnRf 3% relative to the same 20-yr mean of the A1Bexp simulation. The left-hand panels are the anomalies for the mean of 2080–99 and the right-hand panels are the anomalies for the mean of 2180–99. Contour interval is 0.5°C.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
By the end of the twenty-second century, less warming occurs, in general, everywhere in the Northern Hemisphere and the equatorial tropics in these three simulations because of the stronger slowdown of the AMOC. This agrees well with previous studies (e.g., Stouffer et al. 2006). Nevertheless, the IPO-like pattern still can be identified, except that the positive sea surface temperature anomaly in the subtropical northwest Pacific becomes weaker, with the positive anomalies in the subtropical southwest Pacific becoming much stronger in comparison to the late twenty-first century. This strengthened warming in the southwest Pacific is associated with reduced heat transport from the Pacific into the Atlantic basin because of the further weakening of the AMOC. The reduced warming in the northwest Pacific is related to two factors. First, a weaker AMOC transports less heat into the subpolar North Atlantic, resulting in a cooling there. This cooling signal is transported to the North Pacific by stimulating long waves in the atmosphere (Hu et al. 2004a). Second, a weaker AMOC induces a reduction of Bering Strait mass and freshwater transport, resulting in a freshwater gain in the North Pacific and stronger upper-ocean stratification. This stronger oceanic stratification weakens the wintertime convection, leading to a stronger winter cooling in the North Pacific region (Hu et al. 2007, 2008, 2011).
To further identify this IPO-like SST pattern associated with a slowdown of the AMOC when the land-based ice melt flux is considered for the future projection, an EOF analysis of the SST anomalies in the three 3% simulations with Greenland Ice Sheet melt included relative to the A1Bexp is done. The first EOF pattern of this SST anomaly in these three simulations (Gr 3%, GrAn 3%, and GrAnRf 3%) displays an IPO-like pattern. Over the next two centuries, this IPO-like pattern trends downward toward a negative phase (Fig. 17). Note that when the EOF analysis is done, the mean anomaly of the 200-yr time series in each simulation is removed. This is why the EOF pattern turns negative at the end of the twenty-first century in Fig. 17. These EOF patterns explain about 36%–43% of the SST anomaly variance. We calculate the second EOF of the SST anomalies for these three simulations, which also shows an IPO-like pattern, but the principal component (PC) time series does not have a trend (figure not shown; the variance explained by EOF2 varies from 22% to 30% in these three simulations). Therefore, we conclude that when the land-based ice loss is strong enough to induce a further slowdown of the AMOC relative to the same future projection simulation without considering the land-based ice loss runoff, this further weakening of the AMOC will induce a negative IPO-like SST anomaly.
First EOF of the SST anomalies in (a) Gr 3%, (c) GrAn 3%, and (e) GrAnRf 3% simulations relative to A1Bexp and (b),(d), and (f) the respective PC time series. Contour interval for the left panels is 0.1.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
b. Precipitation changes
In A1Bexp, the global annual mean precipitation increases by about 5 and 6 cm yr−1, or 5% and 6%, by the end of the twenty-first and twenty-second centuries, respectively, relative to the late twentieth century in our model simulations (figure not shown). By including the land-based ice loss, the global mean precipitation does not change much during the twenty-first century in all simulations relative to the A1Bexp, but it decreases in the twenty-second century in all simulations with a maximum annual reduction of 1 cm in the GrAnRf 3% simulation or 17% of the increase in A1Bexp (figure not shown). Therefore, less global warming due to land-based ice loss induces further weakening of the AMOC that could potentially affect the global hydrological cycle.
Regionally, the changes of precipitation in A1Bexp relative to the late twentieth century show an intensification of precipitation in the tropics and most parts of mid- to high-latitude land areas (Figs. 14b,d). In the subtropical regions, there is a reduction of precipitation. This precipitation reduction extends from the North Atlantic to the Mediterranean region. These general changes of the precipitation patterns in our simulation are consistent with those model simulations included in the IPCC Fourth Assessment Report and many previous studies (e.g., Meehl et al. 2007; Vellinga and Wood 2004; Zhang and Delworth 2005; Chiang and Bitz 2005). An exception is that, in North America, the multimodel ensemble indicates reduced precipitation in the entire southwest part of North America, but our model only shows a precipitation reduction in the west coast region.
Associated with the changes of the surface temperature, the regional precipitation pattern also changes noticeably in those three simulations, with a further slowdown of the AMOC compared to A1Bexp (Fig. 18). The general pattern of the precipitation change is characterized by a reduction of precipitation in the Northern Hemisphere and an increase in the Southern Hemisphere, which agrees with previous studies (e.g., Stouffer et al. 2006). There is also a southward movement of the ITCZ in both the Pacific and the Atlantic, especially by the late twenty-second century. These changes are due to the slower AMOC that induces less warming in the Northern Hemisphere in these simulations and more enhanced warming in the Southern Hemisphere. This leads to an intensified hydrological cycle in the Southern Hemisphere and a weakened hydrological cycle in the Northern Hemisphere.
As in Fig. 16, but for precipitation. Contour interval is 0.03 m yr−1.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
It is also notable that, in Fig. 18, the precipitation is reduced over the continental United States but increased in Australia (especially the western part). In these three simulations, this is associated with a further spindown of the AMOC relative to the A1Bexp in the next two centuries. This precipitation change pattern in the United States and Australia is associated with less warming over the equatorial tropics in the Pacific. As shown in Meehl and Hu (2006) and Meehl et al. (2010), precipitation in the United States and Australia, especially in the southwest United States and western Australia, is related to the IPO pattern in the Pacific. A negative IPO phase induces drought conditions in the southwest United States and wet conditions in western Australia. To demonstrate this more clearly, the first EOF PC time series of the SST anomalies in these three simulations is correlated with the precipitation anomalies (Fig. 19). These correlation patterns show a positive correlation of the first EOF with the precipitation in the United States and a negative correlation with the precipitation in Australia, agreeing well with previous studies (Meehl and Hu 2006; Meehl et al. 2010). This negative IPO-like SST response to a further slowdown of the AMOC seems to also induce a precipitation reduction in the Amazon region. Thus, our results suggest that the negative IPO-like SST changes in the Pacific in these three simulations could provide us with a key to predict future changes of precipitation over the United States and Australia if the AMOC weakens further because of land-based ice loss.
Correlation of the land precipitation anomalies relative to the A1Bexp with the first EOF of the Pacific SST anomalies relative to A1Bexp: (a) Gr 3% simulation, (b) GrAn 3% simulation, and (c) GrAnRf 3% simulation. Contour interval is 0.1.
Citation: Journal of Climate 26, 10; 10.1175/JCLI-D-12-00102.1
5. Conclusions
Here we use the NCAR CCSM3 to simulate the potential impact of land-based ice loss on the AMOC and features of surface climate for the twenty-first and twenty-second centuries under idealized land-based ice melting scenarios. Our results suggest that the AMOC is especially sensitive to land-based ice meltwater discharge directly into the North Atlantic basin, for example, from the Greenland Ice Sheet. The discharge of the West Antarctic Ice Sheet based on our idealized scenario does not much affect the AMOC, but our results also suggest that a potentially large discharge from the West Antarctic Ice Sheet could strengthen the AMOC. In addition, a low rate of discharge of land-based ice into the North Atlantic, such as melting of the Greenland Ice Sheet and the ice caps and mountain glaciers up to an equivalent 3 mm yr−1 global mean sea level rise (0.03 Sv), would not significantly affect the AMOC in comparison to the simulation, which does not include this land-based ice melt flux into the North Atlantic. This agrees with a few previous studies (e.g., Hu et al. 2009, 2011), but a moderate increase of the land-based ice discharge could slow the AMOC further, such as in the Gr 3%, GrAn 3%, and GrAnRf 3% simulations. A freshwater budget analysis shows a net freshwater deficit in the upper 1000-m ocean layer over the subpolar North Atlantic in A1Bexp, suggesting that this freshwater deficit works against the AMOC weakening in the twenty-first and twenty-second centuries. This result indicates that the slowing of the AMOC in A1Bexp is primarily caused by greenhouse gas–induced warming, which reduces the surface water density and strengthens the upper-ocean stratification, thus weakening the deep convection in the subpolar North Atlantic. To make the AMOC weaken further in the sensitivity simulations relative to A1Bexp, the meltwater discharge into the subpolar North Atlantic from the land-based ice has to be large enough to generate a net freshwater gain in the upper subpolar North Atlantic Ocean, such as in the GrAnRf 3% simulation.
Overall, even as the AMOC weakens further in the sensitivity simulations relative to A1Bexp, the global mean surface temperature is not cooler in comparison with that in the late twentieth century. But this does lessen global warming slightly, especially in the late twenty-first century and the twenty-second century. Regionally, this reduced warming in the northern high latitudes due to a further slowdown of the AMOC, caused by the land-ice discharge into the ocean, could reach a few degrees. The weaker AMOC, to some degree, also contributes to comparatively less warming in Europe since a weaker AMOC transports less heat meridionally from the tropics into the subpolar North Atlantic region. This lessened warming could reach more than 1°C in most parts of Europe.
Corresponding to these temperature changes (less warming) in our sensitivity simulations, precipitation shows a general pattern of reduction in many regions of the Northern Hemisphere and a pattern of increase in many regions of the Southern Hemisphere. This is especially evident in the twenty-second century in the three simulations with even more weakening of the AMOC. Our results also suggest that, associated with a further weakening of the AMOC, the Pacific SST shows a negative IPO-like pattern, leading to reduced precipitation in the United States and increased precipitation in Australia.
Acknowledgments
Portions of this study were supported by the Office of Science (BER), U.S. Department of Energy, Cooperative Agreement DE-FC02-97ER62402, and the National Science Foundation. The National Center for Atmospheric Research is sponsored by the National Science Foundation. Weiqing Han was supported by NSF OCE-0452917, NASA Ocean Vector Winds Program 1283568, and NASA Ocean Surface Topography Science Team NNX08AR62G.
REFERENCES
Arblaster, J. M., G. A. Meehl, and A. Moore, 2002: Interdecadal modulation of Australian rainfall. Climate Dyn., 18, 519–531.
Bindoff, N. L., and Coauthors, 2007: Observations: Oceanic climate change and sea level. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 385–432.
Broccoli, A. J., K. A. Dahl, and R. J. Stouffer, 2006: Response of the ITCZ to Northern Hemisphere cooling. Geophys. Res. Lett., 33, L01702, doi:10.1029/2005GL024546.
Cai, M., 2006: Dynamical greenhouse-plus feedback and polar warming amplification. Part I: A dry radiative-transportive climate model. Climate Dyn., 26, 661–675, doi:10.1007/s00382-005-0104-6.
Cazenave, A., 2006: How fast are the ice sheets melting? Science, 314, 1250–1252.
Cazenave, A., K. Dominh, S. Guinehut, E. Berthier, W. Llovel, G. Ramillien, M. Ablain, and G. Larnicol, 2009: Sea level budget over 2003–2008: A reevaluation from GRACE space gravimetry, satellite altimetry and Argo. Global Planet. Change, 65, 83–88, doi:10.1016/j.gloplacha.2008.10.004.
Chiang, J. C. H., and C. M. Bitz, 2005: Influence of high latitude ice cover on the marine intertropical convergence zone. Climate Dyn., 25, 477–496, doi:10.1007/s00382-005-0040-5.
Church, J. A., J. M. Gregory, P. Huybrechts, M. Kuhn, K. Lambeck, M. T. Nhuan, D. Qin, and P. L. Woodworth, 2001: Changes in sea level. Climate Change 2001: The Scientific Basis, J. T. Houghton, Eds., Cambridge University Press, 639–693.
Church, J. A., and Coauthors, 2011: Revisiting the Earth’s sea-level and energy budgets from 1961 to 2008. Geophys. Res. Lett., 38, L18601, doi:10.1029/2011GL048794.
Collins, W. D., and Coauthors, 2006: The Community Climate System Model version 3 (CCSM3). J. Climate, 19, 2122–2143.
Crowley, T. J., 1992: North Atlantic Deep Water cools the Southern Hemisphere. Paleoceanography, 7, 489–497.
Dahl, K., 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.
Dansgaard, W., and Coauthors, 1993: Evidence for general instability for past climate from a 250-kyr ice-core record. Nature, 364, 218–220.
De Boer, A. M., and D. Nof, 2004a: The exhaust valve of the North Atlantic. J. Climate, 17, 417–422.
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.
Delworth, T., S. Manabe, and R. J. Stouffer, 1993: Interdecadal variations of the thermohaline circulation in a coupled ocean–atmosphere model. J. Climate, 6, 1993–2011.
Dickson, R. R., and J. Brown, 1994: The production of North Atlantic Deep Water: Sources, rates, and pathways. J. Geophys. Res., 99, 12 319–12 341.
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.
Dixon, K., T. Delworth, M. Spelman, and R. Stouffer, 1999: The influence of transient surface fluxes on North Atlantic overturning in a coupled GCM climate change experiment. Geophys. Res. Lett., 26, 2749–2752.
Fichefet, T., and Coauthors, 2003: Implications of changes in freshwater flux from the Greenland Ice Sheet for the climate of the twenty-first century. Geophys. Res. Lett., 30, 1911, doi:10.1029/2003GL017826.
Ganachaud, A., and C. Wunsch, 2000: Improved estimates of global ocean circulation, heat transport and mixing from hydrographic data. Nature, 408, 453–457.
Gerdes, R., W. Hurlin, and S. M. Griffies, 2006: Sensitivity of a global ocean model to increased run-off from Greenland. Ocean Modell., 12, 416–435.
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, doi:10.1029/2005GL023209.
Hegerl, G. C., and Coauthors, 2007: Understanding and attributing climate change. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 747–846.
Heinrich, H., 1988: Origin and consequences of cyclic ice rafting in the northeast Atlantic Ocean during the past 130,000 years. Quat. Res., 29, 143–152.
Hemming, S. R., 2004: Heinrich events: Massive late Pleistocene detritus layers of the North Atlantic and their global climate imprint. Rev. Geophys., 42, RG1005, doi:10.1029/2003RG000128.
Hu, A., 2001: Changes in the Arctic and their impact on the oceanic meridional overturning circulation. Ph.D. dissertation, University of Miami, 171 pp.
Hu, A., and G. A. Meehl, 2005: Bering Strait throughflow and the thermohaline circulation. Geophys. Res. Lett., 32, L24610, doi:10.1029/2005GL024424.
Hu, A., G. A. Meehl, and W. Han, 2004a: Detecting thermohaline circulation changes from ocean properties in a coupled model. Geophys. Res. Lett., 31, L13204, doi:10.1029/2004GL020218.
Hu, A., G. A. Meehl, W. M. Washington, and A. Dai, 2004b: Response of the Atlantic thermohaline circulation to increased atmospheric CO2 in a coupled model. J. Climate, 17, 4267–4279.
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, doi: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.
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 twenty-first century. Geophys. Res. Lett., 36, L10707, doi: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, doi: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, doi:10.1016/j.dsr2.2010.10.069.
Hu, A., G. A. Meehl, W. Han, A. Abe-Ouchi, C. Morrill, Y. Okazaki, and M. O. Chikamoto, 2012a: The Pacific–Atlantic seesaw and the Bering Strait. Geophys. Res. Lett., 39, L03702, doi:10.1029/2011GL050567.
Hu, A., and Coauthors, 2012b: 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, doi:10.1073/pnas.1116014109.
Hughes, T. M. C., and A. J. Weaver, 1994: Multiple equilibria of an asymmetric two-basin ocean model. J. Phys. Oceanogr., 24, 619–637.
Huybrechts, P., and Coauthors, 2002: The response of the Greenland Ice Sheet to climate changes in the twenty-first century by interactive coupling of an AOGCM with a thermomechnical ice-sheet model. Ann. Glaciol., 35, 409–415.
Jacob, T., J. Wahr, W. Tad Pfeffer, and S. Swenson, 2012: Recent contributions of glaciers and ice caps to sea level rise. Nature, 482, 514–518, doi:10.1038/nature10847.
Jungclaus, J. H., and Coauthors, 2006: Will Greenland melting halt the thermohaline circulation? Geophys. Res. Lett., 33, L17708, doi:10.1029/2006GL026815.
Kamenkovich, I. V., A. Sokolov, and P. H. Stone, 2003: Feedbacks affecting the response of the thermohaline circulation to increasing CO2. A study with a model of intermediate complexity. Climate Dyn., 21, 119–130.
Kaser, G., J. G. Cogley, M. B. Dyurgerov, M. F. Meier, and A. Ohmura, 2006: Mass balance of glaciers and ice caps: Consensus estimates for 1961–2004. Geophys. Res. Lett., 33, L19501, doi:10.1029/2006GL027511.
Keigwin, L. D., and G. A. Jones, 1994: Western North Atlantic evidence for millennial-scale changes in ocean circulation and climate. J. Geophys. Res., 99, 12 397–12 410.
Large, W. G., and G. Danabasoglu, 2006: Attribution and impacts of upper-ocean biases in CCSM3. J. Climate, 19, 2325–2346.
Lemke, P., and Coauthors, 2007: Observations: Changes in snow, ice and frozen ground. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 337–383.
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.
Levitus, S., J. I. Antonov, T. P. Boyer, R. A. Locarnini, H. E. Garcia, and A. V. Mishonov, 2009: Global ocean heat content 1955–2008 in light of recently revealed instrumental problems. Geophys. Res. Lett., 36, L07608, doi:10.1029/2008GL037155.
Lu, J., and M. Cai, 2010: Quantifying contributions to polar warming amplification in an idealized coupled circulation model. Climate Dyn., 34, 669–687, doi:10.1007/s00382-009-0673-x.
Manabe, S., and R. J. Stouffer, 1988: Two stable equilibria of a coupled ocean–atmosphere model. J. Climate, 1, 841–866.
Manabe, S., R. J. Stouffer, M. J. Spelman, and K. Bryan, 1991: Transient responses of a coupled ocean–atmosphere model to gradual changes of atmospheric CO2. Part I: Annual mean response. J. Climate, 4, 785–818.
Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78, 1069–1079.
Meehl, G. A., and A. Hu, 2006: Megadroughts in the Indian monsoon region and southwest North America and a mechanism for associated multidecadal Pacific sea surface temperature anomalies. J. Climate, 19, 1605–1623.
Meehl, G. A., and Coauthors, 2007: Global climate projections. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 747–845.
Meehl, G. A., A. Hu, and C. Tebaldi, 2010: Decadal prediction in the Pacific region. J. Climate, 23, 2959–2973.
Meier, M., and Coauthors, 2007: Glaciers dominate eustatic sea-level rise in the twenty-first century. Science, 317, 1064–1067.
Mikolajewicz, U., and R. Voss, 2000: The role of the individual air–sea flux components in CO2-induced changes of the ocean’s circulation and climate. Climate Dyn., 16, 627–642.
Peltier, W. R., 2009: Closure of the budget of global sea level rise over the GRACE era: The importance and magnitudes of the required corrections for global glacial isostatic adjustment. Quat. Sci. Rev., 28, 1658–1674, doi:10.1016/j.quascirev.2009.04.004.
Peltier, W. R., and A. M. Tushingham, 1989: Global sea level rise and the greenhouse effect: Might they be connected? Science, 244, 806–810.
Power, S., T. Casey, C. Folland, A. Colman, and V. Mehta, 1999: Inter-decadal modulation of the impact of ENSO on Australia. Climate Dyn., 15, 319–324.
Pritchard, H. D., and D. G. Vaughan, 2007: Widespread acceleration of tidewater glaciers on the Antarctic Peninsula. J. Geophys. Res., 112, F03S29, doi:10.1029/2006JF000597.
Pritchard, H. D., R. H. Arthern, D. G. Vaughan, and L. A. Edwards, 2009: Extensive dynamic thinning on the margins of the Greenland and Antarctic ice sheets. Nature, 461, 971–975, doi:10.1038/nature08471.
Rahmstorf, S., 1996: On the freshwater forcing and transport of the Atlantic thermohaline circulation. Climate Dyn., 12, 799–811.
Ridley, J. K., P. Huyberechts, J. M. Gregory, and J. A. Lowe, 2005: Elimination of the Greenland Ice Sheet in a high CO2 climate. J. Climate, 18, 3409–3427.
Rignot, E., and P. Kanagaratnam, 2006: Changes in the velocity structure of the Greenland Ice Sheet. Science, 311, 986–990.
Rignot, E., I. Velicogna, M. R. van den Broeke, A. Monaghan, and J. Lenaerts, 2011: Acceleration of the contribution of the Greenland and Antarctic ice sheets to sea level rise. Geophys. Res. Lett., 38, L05503, doi:10.1029/2011GL046583.
Saenko, O. A., A. J. Weaver, and A. Schmittner, 2003: Atlantic deep circulation controlled by freshening in the Southern Ocean. Geophys. Res. Lett., 30, 1754, doi:10.1029/2003GL017681.
Sakai, K., and W. R. Peltier, 1997: Dansgaard–Oeschger Oscillation in a coupled atmosphere–ocean climate model. J. Climate, 10, 949–970.
Schmittner, A., and T. F. Stocker, 1999: The stability of the thermohaline circulation in global warming experiments. J. Climate, 12, 1117–1133.
Schmittner, A., M. Latif, and B. Schneider, 2005: Model projections of the North Atlantic thermohaline circulation for the twenty-first century assessed by observations. Geophys. Res. Lett., 32, L23710, doi:10.1029/2005GL024368.
Shaffer, G., and J. Bendtsen, 1994: Role of the Bering Strait in controlling North Atlantic ocean circulation and climate. Nature, 367, 354–357.
Smith, L. T., E. P. Chassignet, and R. Bleck, 2000: The impact of lateral boundary conditions and horizontal resolution on North Atlantic water mass transformations and pathways in an isopycnic coordinate ocean model. J. Phys. Oceanogr., 30, 137–159.
Stammer, D., 2008: Response of the global ocean to Greenland and Antarctic ice melting. J. Geophys. Res., 113, C06022, doi:10.1029/2006JC004079.
Stocker, T. F., and W. S. Broecker, 1994: Observation and modeling of North Atlantic deep water formation and its variability: Introduction. J. Geophys. Res., 99, 12 317.
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.
Stouffer, R. J., D. Seidov, and B. J. Haupt, 2007: Climate response to external sources of freshwater: North Atlantic versus the Southern Ocean. J. Climate, 20, 436–448.
Sutton, R. T., B. Dong, and J. M. Gregory, 2007: Land/sea warming ratio in response to climate change: IPCC AR4 model results and comparison with observations. Geophys. Res. Lett., 34, L02701, doi:10.1029/2006GL028164.
Swingedouw, D., P. Braconnot, and O. Marti, 2006: Sensitivity of the Atlantic meridional overturning circulation to the melting from northern glaciers in climate change experiments. Geophys. Res. Lett., 33, L07711, doi:10.1029/2006GL025765.
Swingedouw, D., P. Braconnot, P. Delecluse, E. Guilyardi, and O. Marti, 2007: Quantifying the AMOC feedbacks during a 2×CO2 stabilization experiment with land-ice melting. Climate Dyn., 29, 521–534, doi:10.1007/s00382-007-0250-0.
Thorpe, R. B., J. M. Gregory, T. C. Johns, R. A. Wood, and J. F. B. Mitchell, 2001: Mechanisms determining the Atlantic thermohaline circulation response to greenhouse gas forcing in a non-flux-adjusted coupled climate model. J. Climate, 14, 3102–3116.
Timmermann, A., M. Latif, R. Voss, and A. Groetzner, 1998: Northern Hemispheric interdecadal variability: A coupled air–sea mode. J. Climate, 11, 1906–1931.
Timmermann, A., S.-I. An, U. Krebs, and H. Goosse, 2005a: ENSO suppression due to weakening of the North Atlantic thermohaline circulation. J. Climate, 18, 3122–3139.
Timmermann, A., U. Krebs, F. Justino, H. Goosse, and T. Ivanochko, 2005b: Mechanisms for millennial-scale global synchronization during the last glacial period. Paleoceanography, 20, PA4008, doi:10.1029/2004PA001090.
van den Broeke, M., and Coauthors, 2009: Partitioning recent Greenland mass loss. Science, 326, 984–986, doi:10.1126/science.1178176.
Velicogna, I., 2009: Increasing rates of ice mass loss from the Greenland and Antarctic ice sheets revealed by GRACE. Geophys. Res. Lett., 36, L19503, doi:10.1029/2009GL040222.
Vellinga, M., and R. A. Wood, 2002: Global climatic impacts of a collapse of the Atlantic thermohaline circulation. Climatic Change, 54, 251–267.
Vizcaino, M., and Coauthors, 2008: Long-term ice sheet–climate interactions under anthropogenic greenhouse forcing simulated with a complex Earth System Model. Climate Dyn., 31, 665–690, doi:10.1007/s00382-008-0369-7.
Woodgate, R. A., and K. Aagaard, 2005: Revising the Bering Strait freshwater flux into the Arctic Ocean. Geophys. Res. Lett., 32, L02602, doi:10.1029/2004GL021747.
Yin, J., J. T. Overpeck, S. M. Griffies, A. Hu, J. L. Russell, and R. J. Stouffer, 2011: Different magnitudes of projected subsurface ocean warming around Greenland and Antarctica. Nat. Geosci., 4, 524–528, doi:10.1038/ngeo1189.