Abstract

The role of the wind–evaporation–sea surface temperature (WES) feedback in the propagation of the high-latitude cooling signal to the tropical oceans using the NCAR atmospheric Community Climate Model (CCM3) coupled thermodynamically to a slab-ocean model (SOM) is studied. Abruptly imposed additional Northern Hemispheric sea ice cover equivalent to the Last Glacial Maximum (LGM; 18 kyr BP) in the model causes a Northern Hemisphere–wide cooling, as well as the generation and amplification of an anomalous cross-equatorial meridional SST dipole associated with a southward migration of the intertropical convergence zone (ITCZ) stabilizing within a period of 5 yr. In experiments where the WES feedback is switched off explicitly by modifying the sensible and latent heat flux bulk aerodynamic formulations over the oceans in CCM3, imposed Northern Hemispheric sea ice also results in widespread northern cooling at the same rate as the unmodified run, suggesting that the WES feedback is not essential in the propagation of the high-latitude cooling signal to the deep tropics. However, the WES-off experiment generates a weaker cross-equatorial SST dipole with a modest southward movement of the ITCZ, suggesting that the WES feedback is responsible for amplifying SST and atmospheric anomalies in the deep tropics during their transition to the new equilibrium state. The propagation of high-latitude cooling to the deep tropics is proposed to be caused by the decrease of near-surface specific humidity in the northern tropics.

1. Introduction

Recently, interest has grown over the possibility of higher-latitude climate change forcings impacting the tropics. Studies have particularly focused on the impacts of high-latitude cooling on the tropics, such as that associated with the Last Glacial Maximum (LGM) and Heinrich events, when the Atlantic meridional overturning circulation (AMOC) is believed to be weaker than the current state (e.g., Manabe and Broccoli 1985; Broccoli 2000; Chiang et al. 2003; Chiang and Bitz 2005; Cheng et al. 2007; Dahl et al. 2005; Dong and Sutton 2002; Vellinga and Wood 2002; Zhang and Delworth 2005).

While a weaker AMOC implies a weaker poleward ocean heat transport, the atmosphere, which is a more potent transporter of heat, has been proposed to be the stronger teleconnection medium for the global impacts of AMOC slowdown (Dong and Sutton 2002; Seager et al. 2002). Further, sea ice formation in the North Atlantic initiated by a weaker AMOC and amplified by ice-associated feedbacks also results in cooling over the North Atlantic (e.g., Maykut and Untersteiner 1971; Li et al. 2005). Changes in sea ice have then been shown in GCMs (Li et al. 2005) to influence dominantly the climate of entire high- and midlatitude regions via atmospheric heat transports, consistent with the paleoclimate records of Greenland, which in particular help to explain the warming episodes known as the Dansgaard–Oeschger events (Dansgaard et al. 1993) during the last glacial period (50–10 kyr BP).

Paleoclimate observational studies correlating tropical proxy records (e.g., Peterson et al. 2000; Kennett and Ingram 1995) with Greenland ice-core records (Stuiver and Grootes 2000) indicate that during cold events in the northern high latitudes, the ITCZ moved southward and exhibited more symmetry about the equator (Lynch-Stieglitz 2004). Fully coupled GCM studies that estimate the climate response to AMOC slowdown simulated by large freshwater input into the subpolar North Atlantic (Broccoli 2000; Vellinga and Wood 2002; Dahl et al. 2005; Stouffer et al. 2006; Timmermann et al. 2005; Cheng et al. 2007; Zhang and Delworth 2005) generate such conditions. The southern extratropics also exhibit a tangible warming in response to AMOC shutdown in most studies (e.g., Vellinga and Wood 2002; Dahl et al. 2005; Zhang and Delworth 2005).

Chiang et al. (2003) and Chiang and Bitz (2005) find that just the thermodynamic coupled interaction between the atmosphere and the upper ocean can act as a mechanism for the communication between the high latitudes and the tropics within a short period of 5 yr. Northward atmospheric heat transport by the altered Hadley circulation transporting heat from the southern tropics to the Northern Hemisphere and transient and stationary eddies over the midlatitudes, caused by the increased temperature gradient between the poles and the tropics, have also been suggested (Cheng et al. 2007; Broccoli et al. 2006). A role for humidity (Chiang and Bitz 2005) and changes in cloud cover (Kang et al. 2008) have also been proposed in the propagation of high-latitude signals to the tropics.

Chiang and Bitz (2005) further conjecture that once the high-latitude cooling signal reaches the northern tropics, further southward propagation over global oceans is carried out by the wind–evaporation–sea surface temperature (WES) feedback. The WES feedback is also proposed to be responsible for the southward propagation of the ITCZ, and the development of a cross-equatorial SST gradient (CESG) with cooling on the northern side and warming on the southern side in the deep tropics.

The WES feedback as proposed by Xie and Philander (1994) works as follows in the deep tropics: an initial southward CESG would result in the development of a northward pressure gradient due to hydrostatic adjustment (Lindzen and Nigam 1987), which generates southward wind anomalies. The induced wind anomalies would, however, turn westward (eastward) on the northern (southern) side of the equator due to the Coriolis force, exhibiting a C-shaped profile, adding to the background northeasterly (southeasterly) winds and increasing (decreasing) the wind speed locally. The increased (decreased) winds on the northern (southern) side increase (decrease) the latent heat flux release from the oceans to the atmosphere, cooling (warming) the SSTs and further amplifying the initial CESG, forming a positive feedback. The ITCZ follows the warming SSTs on the southern side, migrating southward. A new steady state is reached by the balancing effects of the convective cloud–SST negative feedback with cooler SSTs on the northern side, warmer SSTs on the southern side, and a more symmetric ITCZ (Chiang and Bitz 2005).

A manifestation of the WES feedback in the trade wind region, when cooling occurs in the high and midlatitudes, works as follows. When the midlatitude cold SST “front” reaches the trade wind region, it induces a northward pressure gradient that generates easterlies in the trade wind region south of the front, which add to the background easterly winds, increasing the latent heat release from the oceans and cooling the mixed layer immediately south of the front, hence propagating the front southward into the deep tropical northern oceans (Chiang and Bitz 2005).

In this study, it is our intent to clearly identify the role of the WES feedback in the propagation of cold anomalies from the higher latitudes that are associated with abrupt climate change to the tropics. For this purpose, we modify an AGCM coupled to a slab-ocean model (SOM) such that the model WES feedback is turned off when it is integrated. Comparative studies between imposed sea ice experiments with the WES-off model and the unmodified model isolate the role of the WES feedback. Imposed LGM land ice or sea ice anomalies produce similar responses in the tropics, suggesting a climate response independent of the longitudinal placement of the forcing and similar mechanisms in action under the two forcing scenarios (Chiang et al. 2003). In this study, we thus only focus on the impacts of imposed sea ice, expecting the effects of imposed land ice to be similar. We find a strong role for the WES feedback in generating and maintaining the anomalous cross-equatorial dipole when the high latitudes are cold. However, a propagation of cold anomalies to the tropics is observed even in the absence of the WES feedback.

We proceed as follows. Section 2, which describes the model setup and simulations, is followed by an analysis of the results of the experiments in section 3. We conclude by summarizing our findings and discussing the implications of our study in section 4.

2. Model setup

We use the National Center for Atmospheric Research (NCAR) Community Climate Model (CCM3) coupled to an SOM at the standard spectral resolution of T42 (corresponding to a spatial resolution of about 2.8° × 2.8°), to isolate the role of the WES feedback in the communication of the high-latitude cooling signal to the tropics. CCM3 is a full-physics complex atmospheric GCM that simulates the current climate mean state and variability reasonably well (Hurrell et al. 1998; Saravanan 1998). It has been used in previous studies (Chiang et al. 2003; Chiang and Bitz 2005; Li et al. 2005) to understand the global impacts of changes in sea ice cover in higher latitudes.

The SOM model is a static ocean model that interacts with the CCM3 only thermodynamically and lacks ocean dynamics. The SOM exhibits a finite thermal heat capacity dependent on the prescribed mixed layer depth. Our prescribed SOM exhibits a spatially varying mixed layer depth that remains constant in time. Since our focus is on understanding a thermodynamic process in isolation, the AGCM coupled to an SOM appears to be the right tool, which allows us to study thermodynamic air–sea coupling in the absence of complex oceanic circulation. The lack of heat transport by ocean circulation is corrected by artificially adding heat fluxes to the mixed layer, namely Q fluxes, at model ocean grid points (Kiehl et al. 1996).

The basic SOM also includes a thermodynamic sea ice model. For our experiments, we have modified the sea ice model, such that sea ice becomes noninteractive with the atmosphere above and the surrounding ocean. The prescribed sea ice model permits us to prescribe steady cold high-latitude forcings, which allows us to focus completely on the response of the forcing. Sea ice is prescribed to be 2 m thick with 5 cm of snow cover across all sea ice–covered grid points. In addition, new sea ice is prevented from forming over open-ocean grid points by restricting the ocean mixed layer temperature from dropping below the freezing threshold of salty water (−1.8°C). This experimental setup is similar to that used by Chiang and Bitz (2005).

With the fixed sea ice model we integrate a control run for 100 yr in CCM3-SOM, which is prescribed with current sea ice conditions. The control run is termed the CCM3-SOM run for further reference. In addition, we perform another 45-yr run with a modified CCM3 coupled to an SOM where the WES feedback is turned off. This run with the modified model is termed the WES-off-SOM run. The climate of the CCM3-SOM and WES-off-SOM simulations is discussed in more detail in Mahajan et al. (2009).

For the WES-off runs, modifications are made in the bulk aerodynamic formulations of surface turbulent fluxes over the oceans. The parameterizations of surface latent and sensible heat fluxes in CCM3-SOM can be simplified for our purposes as follows (Kiehl et al. 1998):

 
formula
 
formula

where Qlh and Qsh are, respectively, the latent and sensible heat fluxes defined to be positive for an upward transfer from the ocean to the atmosphere; u* is a product of the wind speed at the lowest atmospheric level and a neutral momentum exchange coefficient; Δq is the difference between the surface saturation specific humidity qs and the specific humidity q of the lowest atmospheric surface (qsq); Ce is a neutral tracer exchange coefficient; ΔT is the difference in the potential temperature of the ocean surface temperature Ts and the lowest atmospheric surface Ta (TsTa); Cd is a neutral heat exchange coefficient; ρ is the density of air at the lowest atmospheric surface; Cp is the specific heat capacity of moist air; and Lvap is the latent heat of the vaporization of water.

The winds are related to the surface latent and sensible heat fluxes through these parameterizations. To remove the influence of winds on the SSTs, we prescribe u* over global oceans in the above formulations to a control run monthly climatology at each ocean grid point. In such a setup, while anomalies in the SSTs would still be influencing the winds, the model wind anomalies would have no direct impact on the latent and sensible heat fluxes and hence the SSTs, thus breaking the WES feedback loop. The climate mean state of the modified run is ensured to be similar to that of the control run, at least to first order, as u* is prescribed as the climatological u* of the control run.

In addition to the control runs CCM3-SOM and WES-off-SOM, we also perform two sets of experiments with additional sea ice in the northern latitudes. In the first experiment, we prescribe a monthly climatology of the sea ice cover in CCM3-SOM over the northern high latitudes corresponding to the Last Glacial Maximum as interpolated from the February and August reconstructions of LGM sea ice cover derived from Climate: Long-range Investigation, Mapping, and Prediction (CLIMAP) data (Hostetler and Mix 1999), as seen in Fig. 1. It is noteworthy that most of the additional sea ice occurs over the North Atlantic Ocean, consistent with the local response of a weakening of the AMOC. The prescribed LGM sea ice run is referred to as the CCM3-SICE run. Following Chiang et al. (2003), other boundary conditions like orbital parameters, ice topography, and greenhouse gases are kept the same as the modern climate to study the impacts of additional sea ice in isolation. The second experiment features exactly the same setup as the first experiment, but with the modified WES-off model, and is termed the WES-off-SICE run. To study the equilibrium state of the sea-ice–forced run, we make one run of 45 yr for each of these experiments.

Fig. 1.

Prescribed sea ice cover climatology for sea ice experiments replicating the Northern Hemisphere LGM (18 kyr BP) sea ice cover derived from CLIMAP data for the month of January. Gray shading represents sea ice. Darker gray shading represents additional sea ice as compared to the modern-day sea ice cover.

Fig. 1.

Prescribed sea ice cover climatology for sea ice experiments replicating the Northern Hemisphere LGM (18 kyr BP) sea ice cover derived from CLIMAP data for the month of January. Gray shading represents sea ice. Darker gray shading represents additional sea ice as compared to the modern-day sea ice cover.

In addition, we make three more ensemble runs, each with a different set of atmospheric initial conditions, of both the experiments to study the transient responses of the models. All the runs were prescribed with Q fluxes in the SOM that force the model to follow present-day SST climatology. Most paleoclimate AMOC slowdown experiments use today’s climate as the mean state of their models for intercomparisons, even though the LGM climate might be cooler than the current climate (Stouffer et al. 2006). Since our intention is to isolate a particular thermodynamic mechanism, modest differences in the background mean state are likely to have little impact on our results. Hence, we also use the current climate as our mean state, similar to other AMOC slowdown experiments.

3. Mean state response

Both CCM3-SICE and WES-off-SICE, which are forced with northern high-latitude LGM-equivalent sea ice boundary conditions, reach a new equilibrium state within a period of 5 yr. Figures 2a and 2b show the changes in the annual mean state SSTs averaged over 40 yr of simulation for the two sea-ice-perturbed runs after reaching equilibrium, as compared to the mean state of their control runs (CCM3-SOM and WES-off-SOM). A cooling of comparable magnitude is observed over the Northern Hemisphere north of the deep tropics in both experiments, suggesting other mechanisms in the northern tropical oceans that work to cool the Northern Hemisphere even in the absence of the WES feedback.

Fig. 2.

Difference (CCM3-SICE − CCM3-SOM and WES-off-SICE − WES-off-SOM) in the equilibrium state of the LGM sea-ice–forced simulation CCM3-SICE and the control run CCM3-SOM and between the WES-off-SICE and WES-off-SOM annual mean of (a),(b) SST, (c),(d) convective precipitation, (e),(f) surface wind speed and wind vectors, and (g),(h) latent heat flux. The hatched regions represent differences that are statistically significant at the 95% confidence level based on a two-tailed t test for the SST and the latent heat flux. Only wind vector differences that are statistically significant at the 95% confidence level based on a two-tailed t test are shown in (e) and (f).

Fig. 2.

Difference (CCM3-SICE − CCM3-SOM and WES-off-SICE − WES-off-SOM) in the equilibrium state of the LGM sea-ice–forced simulation CCM3-SICE and the control run CCM3-SOM and between the WES-off-SICE and WES-off-SOM annual mean of (a),(b) SST, (c),(d) convective precipitation, (e),(f) surface wind speed and wind vectors, and (g),(h) latent heat flux. The hatched regions represent differences that are statistically significant at the 95% confidence level based on a two-tailed t test for the SST and the latent heat flux. Only wind vector differences that are statistically significant at the 95% confidence level based on a two-tailed t test are shown in (e) and (f).

Consistent with the results of Chiang and Bitz (2005), an anomalous cross-equatorial SST dipole across global oceans is observed in the CCM3-SICE experiment, with cooling (warming) on the northern (southern) side, particularly over the Pacific Ocean. The annual and seasonal averages of the dipolar response over the tropical Atlantic and Indian Oceans are not statistically significant in the experiment in the equilibrium state of CCM3-SICE. The SST dipole pattern seen over the central and eastern equatorial Pacific Ocean in the CCM3-SICE experiment is noticeably weaker in the WES-off-SICE response, suggesting a strong role for the WES feedback in the region in maintaining the equilibrium state. A strong cooling observed over the central equatorial Pacific in the CCM3-SICE experiment is also absent in the WES-off-SICE experiment, again indicative of a dominant role for the WES feedback. From an energy balance viewpoint, the equatorial dipole of SST in the presence of the WES feedback is replaced by a more uniform cooling of the equatorial region in its absence.

The equilibrium states of the two experiments also exhibit an anomalous southward shift of tropical precipitation over global oceans as compared to their respective control runs, with the CCM3-SICE anomalies being much stronger than the WES-off-SICE anomalies. The strongest response in the shift of tropical convective precipitation in the CCM3-SICE experiment is seen in the central Pacific basin, where the strongest cross-equatorial SST response is observed (Figs. 2c and 2d).

The southward shift of the ITCZ is also reflected in the trade wind response of the CCM3-SICE experiment. In the equilibrium state, anomalous northeasterlies (northwesterlies) are observed over the northern (southern) equatorial ocean basins, as seen in Fig. 2e. This C-shaped profile of equatorial trade wind anomalies, as discussed in the introduction, is suggestive of the participation of the WES feedback in maintaining the equilibrium state of the southward-shifted ITCZ in the CCM3-SICE experiment, as proposed by Chiang and Bitz (2005). A cross-equatorial dipole pattern is also observed in the latent heat flux anomalies in the CCM3-SICE experiment over the ocean basins (Fig. 2g), with increased (decreased) cooling of the ocean mixed layer occurring over the north (south) equatorial oceans, with the strongest response seen over the central Pacific Ocean. The collocated dipole patterns of SST, winds, and latent heat flux suggest the presence of the positive WES feedback.

Weaker responses from the surface winds and latent heat fluxes are seen in the WES-off-SICE experiment (Figs. 2f and 2h) over global equatorial oceans, particularly over the central and eastern Pacific Ocean. By design, the latent heat flux response in the WES-off-SICE experiment is mainly caused by changes in the near-surface humidity. The appearance of a dipole, although weaker as compared to that in the CCM3-SICE experiment, suggests a role for humidity in the thermodynamic feedback in CCM3-SOM. Mahajan et al. (2010) suggest that near-surface humidity can sustain a meridional mode of variability in the deep tropical Atlantic, in the absence of the WES feedback. Nonetheless, the weak cross-equatorial dipole of SST, winds, and the latent heat fluxes in the absence of the WES feedback in the WES-off-SICE run conclusively indicates that the WES feedback is essential for generating a cross-equatorial dipole response in CCM3-SOM.

4. Transient response

To isolate the role of the WES feedback as a meridional pathway participating in the transition of the model’s tropical climate from one equilibrium state to the other, we compare winter season anomalies, when the high-latitude forcing is strongest, for the first 4 yr of the CCM3-SICE and WES-off-SICE experiments. Our results overall show little sensitivity to different seasons in the northern tropics. Figure 3 shows winter season (averaged over December–February) SST and convective precipitation anomalies averaged over four ensemble member runs of CCM3-SICE and WES-off-SICE for the first 4 yr as the two systems reach equilibrium. The anomalies are computed as deviations from their control run equilibrium state (CCM3-SOM and WES-off-SOM control runs, respectively). Statistically significant SST anomalies reach the trade wind region before the winter of the second year and then the deep northern tropics by the third year in both runs, indicating that processes other than the WES feedback are active in the propagation of cold anomalies southward into the trade wind region. Chiang and Bitz (2005) suggest that humidity plays a key role in the equatorward propagation of cold anomalies. Transient eddy transport mechanisms (Cheng et al. 2007; Broccoli et al. 2006) and the role of clouds (Kang et al. 2008) have also been proposed.

Fig. 3.

Anomalies of (a),(b) SST and (c),(d) convective precipitation, averaged over four ensemble runs, for (left) CCM3-SICE and (right) WES-off-SICE with respect to their unperturbed control runs in the winter (December–February) for years 1–4 of the simulations. The hatched regions in (a) and (b) represent differences that are statistically significant at the 95% confidence level based on a two-tailed t test.

Fig. 3.

Anomalies of (a),(b) SST and (c),(d) convective precipitation, averaged over four ensemble runs, for (left) CCM3-SICE and (right) WES-off-SICE with respect to their unperturbed control runs in the winter (December–February) for years 1–4 of the simulations. The hatched regions in (a) and (b) represent differences that are statistically significant at the 95% confidence level based on a two-tailed t test.

In the CCM3-SICE run, however, there is a development of significant CESG, with warming in the southern deep tropics, in the second year that is amplified in the third and fourth years in the deep tropical Pacific. The strong anomalous dipolar SST pattern is absent in the WES-off-SICE run in all seasons, suggesting a strong role for the WES feedback in the deep tropical response, as proposed by Chiang and Bitz (2005). The development and amplification of a dipole pattern is also seen in the precipitation anomalies in the CCM3-SICE run with the ITCZ following warmer SSTs. While the cooling of the northern deep tropics pushes the ITCZ into the southern deep tropics, the precipitation anomalies in the WES-off-SICE run are weaker due to the lack of strong CESG caused by the WES feedback. It should be noted that the ITCZ serves as the axis of meridional asymmetry over the equatorial oceans, across which the WES feedback mechanizes. Also, the WES feedback is strongest when the ITCZ is closer to the equator. In the CCM3-SICE experiment, the location and the strength of the anomalous dipole about the ITCZ hence follows the climatological location of the ITCZ, moving to its southernmost position and exhibiting the strongest anomalies in the spring across the tropics, which are also significant over the Atlantic Ocean (not shown).

Figure 4 shows the latent heat flux and wind speed winter season anomalies for the two experiments. Only the wind vectors that are statistically significant at the 95% confidence level are shown. Latent heat flux anomalies appear in the trade wind region by the second year in both runs. Over the deep tropical oceans the latent heat flux anomalies amplify in the third and fourth years, forming a dipolelike pattern across the equator, and are stronger in the CCM3-SICE run in the tropical Pacific and stronger in the WES-off-SICE over the tropical Atlantic. The wind anomalies reflect the dipole in latent heat flux anomalies in the CCM3-SICE experiment with stronger winds in the northern deep tropics and weaker winds in the southern deep tropics, which is suggestive of the WES feedback. The weaker dipole pattern of the latent heat flux and wind anomalies in the WES-off-SICE in the tropical Pacific and Indian Oceans further suggest the role of the WES feedback in the deep tropics.

Fig. 4.

As in Fig. 3, but for (a),(b) latent heat flux and (c),(d) surface wind speed and wind vectors. Only wind vector differences that are statistically significant at the 95% confidence level based on a two-tailed t test are shown in (c) and (d).

Fig. 4.

As in Fig. 3, but for (a),(b) latent heat flux and (c),(d) surface wind speed and wind vectors. Only wind vector differences that are statistically significant at the 95% confidence level based on a two-tailed t test are shown in (c) and (d).

Propagation to the deep tropics

In the CCM3-SICE experiment, while the northern tropical Atlantic latent heat flux anomalies are associated with increased wind speeds, as predicted by the WES feedback proposed by Chiang and Bitz (2005), the latent heat flux anomalies over the northern tropical Pacific are not clearly associated with increased winds, further suggesting that the WES feedback is not essential for the propagation of SST anomalies from the high latitudes to the deep tropics. The development of latent heat flux anomalies in the northern tropical oceans in the WES-off-SICE run in the second year suggests a role for surface humidity, as winds are not allowed to influence the surface heat fluxes in the experiment.

Figure 5 shows surface specific humidity and longwave heat flux winter anomalies for the two experiments. A near-surface drying spreads into the northern tropics by the second year. In the deep tropics, an amplification of the cross-equatorial dipole in the anomalies is observed in the CCM3-SICE experiment with increased surface specific humidity in the southern deep tropics. The drying of the near-surface humidity is associated with an increase in the latent heat flux from the surface to the atmosphere. The greenhouse effect of water vapor implies that a drying of the atmosphere would also result in an increase in the longwave radiation flux from the surface to the atmosphere, causing further surface cooling in both experiments. An increase in longwave radiation flux from the surface to the atmosphere could also result from a decrease in cloud cover. Figures 6a and 6b show the total cloud cover fraction anomalies for the two experiments. A comparison with longwave radiation anomalies reveals that a part of the longwave radiation anomalies could be explained by the change in cloud cover. Also, Kang et al. (2008) suggest that cloud parameterizations in climate models play an important role in the propagation of high-latitude signals to the tropics. Further, decreases in cloud cover would also increase the net surface shortwave radiative heat flux, causing surface warming that would counteract the evaporative and radiative cooling.

Fig. 5.

As in Fig. 3, but for (a),(b) near-surface humidity and (c),(d) net surface longwave radiation.

Fig. 5.

As in Fig. 3, but for (a),(b) near-surface humidity and (c),(d) net surface longwave radiation.

Fig. 6.

As in Fig. 3, but for (a),(b) total cloud fraction and (c),(d) vertical velocity in pressure coordinates, with positive values defined downward.

Fig. 6.

As in Fig. 3, but for (a),(b) total cloud fraction and (c),(d) vertical velocity in pressure coordinates, with positive values defined downward.

However, a decrease in the near-surface humidity, and a corresponding increase in the latent heat flux and longwave radiation heat flux from the surface to the atmosphere in the CCM3-SICE and WES-off-SICE experiments, suggest a strong role for humidity, as suggested by Chiang and Bitz (2005). We conjecture that the following mechanism also participates in the propagation of high-latitude anomalies to the tropics even in the absence of the WES feedback. Climatological trade winds transport dry cold air from the midlatitude regions into the northern tropical regions. In addition to anomalous evaporative cooling, the mixing of dry air with the tropical air also causes surface cooling by anomalously increasing the loss of surface longwave radiation due to a weakened greenhouse effect caused by decreased atmospheric water vapor. The mean trade winds transport the dry air farther south, propagating cool surface anomalies into the deep tropics. Near-surface humidity anomalies in the northern tropics could also be caused by the subsidence of dry air masses. Figures 6c and 6d, which show the vertical velocities in pressure coordinates for the two experiments, reveal that anomalous subsidence mostly occurs in the deep tropical regions associated with the shift of the ITCZ and the anomalous near-surface humidity in the northern tropics shows little dependence on subsidence.

5. Summary and discussion

We investigate the role of the WES feedback in the propagation of cold high-latitude anomalies to the tropics and in the southward migration of the ITCZ using CCM3 coupled to a slab-ocean model by imposing anomalous sea ice. Rapid cooling of the tropics despite the absence of the WES feedback suggests the presence of other large-scale meridional heat transport mechanisms that allow for the communication of high-latitude signals to the tropics. We conjecture that the greenhouse effect of humidity, along with its influence on the latent heat flux, plays an important role in the southward propagation of SSTs, with the southward moisture transport being carried out by mean climatological winds. Further studies are required to distinctly identify and isolate these possible other mechanisms, like increased subsidence, the development of low-level clouds, and meridional eddy transports. However, strong SST dipolar anomalies are observed in the deep tropics in response to high-latitude sea ice anomalies, which disappear in the absence of the WES feedback, suggesting a dominant role of the WES feedback in the deep tropics, as proposed by Chiang and Bitz (2005).

Cooling of the higher latitudes and the formation of sea ice has been associated with a weakening of the AMOC, one of the proposed impacts of global warming induced by greenhouse gases (Gregory et al. 2005; Schmittner et al. 2005). Our study is aimed at understanding the physical mechanisms that cause a change in the tropical climate, a better knowledge of which would help improve climate models, with the eventual goal of improving predictability. Finally, a couple of other caveats need to be mentioned. Studies show that in addition to the tropical Pacific, the tropical Atlantic climate can also potentially modulate the North Atlantic climate (Okumura et al. 2001; Rajagopalan et al. 1998). Hence, the response of the tropical oceans to high-latitude forcings could potentially feed back upon the forcing itself, a phenomenon only partly constrained in our experiments by making the sea ice noninteractive. In addition, as discussed by Chiang et al. (2003), the SOM does not exhibit mechanisms like the equatorial El Niño or Atlantic Niño phenomena, which themselves respond dynamically as well as thermodynamically to high-latitude forcings. As mentioned above, our experiments exclude these mechanisms by design, and focus on purely thermodynamical interactions. Simulations of an AGCM coupled to dynamical ocean models would be needed to completely study the role of the WES feedback in the response of tropical oceans to sea ice anomalies.

Acknowledgments

This work was supported by research grants from NOAA’s CLIVAR Program (Project NA050AR4311136) and NSF’s Climate Dynamics Program (ATM-0337846). We are grateful for the improvements suggested by anonymous reviewers.

REFERENCES

REFERENCES
Broccoli
,
A. J.
,
2000
:
Tropical cooling at the Last Glacial Maximum: An atmosphere–mixed layer ocean model simulation.
J. Climate
,
13
,
951
976
.
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
.
Cheng
,
W.
,
C. M.
Bitz
, and
J. C. H.
Chiang
,
2007
:
Adjustment of the global climate to an abrupt slowdown of the Atlantic meridional overturning circulation.
Ocean Circulation: Mechanisms and Impacts, Geophys. Monogr., Vol. 173, Amer. Geophys. Union, 295–314
.
Chiang
,
J. C. H.
, and
C.
Bitz
,
2005
:
Influence of high latitude ice cover on the marine intertropical convergence zone.
Climate Dyn.
,
25
,
477
496
.
Chiang
,
J. C. H.
,
M.
Biasutti
, and
D. S.
Battisti
,
2003
:
Sensitivity of the Atlantic intertropical convergence zone to Last Glacial Maximum boundary conditions.
Paleoceanography
,
18
,
1094
.
doi:10.1029/2003PA000916
.
Dahl
,
K. A.
,
A. J.
Broccoli
, and
R. J.
Stouffer
,
2005
:
Assessing the role of North Atlantic freshwater forcing in millennial scale climate variability: A tropical Atlantic perspective.
Climate Dyn.
,
24
,
325
346
.
Dansgaard
,
W.
, and
Coauthors
,
1993
:
Evidence for general instability of past climate from a 250-kyr ice-core record.
Nature
,
364
,
218
220
.
Dong
,
B.-W.
, and
R. T.
Sutton
,
2002
:
Adjustment of the coupled ocean–atmosphere system to a sudden change in the thermohaline circulation.
Geophys. Res. Lett.
,
29
,
1728
.
doi:10.1029/2002GL015229
.
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
.
Hostetler
,
S. W.
, and
A. C.
Mix
,
1999
:
Reassessment of ice-age cooling of the tropical ocean and atmosphere.
Nature
,
399
,
673
676
.
Hurrell
,
J. W.
,
J. J.
Hack
,
B. A.
Boville
,
D. L.
Williamson
, and
J. T.
Kiehl
,
1998
:
The dynamical simulation of the NCAR Community Climate Model version 3 (CCM3).
J. Climate
,
11
,
1207
1236
.
Kang
,
S. M.
,
I. M.
Held
,
D. M. W.
Frierson
, and
M.
Zhao
,
2008
:
The response of the ITCZ to extratropical thermal forcing: Idealized slab-ocean experiments with a GCM.
J. Climate
,
21
,
3521
3532
.
Kennett
,
J. P.
, and
B. L.
Ingram
,
1995
:
A 20,000-year record of ocean circulation and climate change from the Santa Barbara basin.
Nature
,
377
,
510
514
.
Kiehl
,
J. T.
,
J. J.
Hack
,
G. B.
Bonan
,
B. A.
Boville
,
B. P.
Briegleb
,
D. L.
Williamson
, and
P. J.
Rasch
,
1996
:
Description of the NCAR Community Climate Model.
NCAR Tech. Note TN-420+SR, 152 pp
.
Kiehl
,
J. T.
,
J. J.
Hack
,
G. B.
Bonan
,
B. A.
Boville
,
D. L.
Williamson
, and
P. J.
Rasch
,
1998
:
The National Center for Atmospheric Research Community Climate Model: CCM3.
J. Climate
,
11
,
1131
1149
.
Li
,
C.
,
D. S.
Battisti
,
D. P.
Schrag
, and
E.
Tziperman
,
2005
:
Abrupt climate shifts in Greenland due to displacements of the sea ice edge.
Geophys. Res. Lett.
,
32
,
L19702
.
doi:10.1029/2005GL023492
.
Lindzen
,
R. S.
, and
S.
Nigam
,
1987
:
On the role of sea surface temperature gradients in forcing low-level winds and convergence in the tropics.
J. Atmos. Sci.
,
44
,
2418
2436
.
Lynch-Stieglitz
,
J.
,
2004
:
Ocean science: Hemispheric asynchrony of abrupt climate change.
Science
,
304
,
1919
1920
.
Mahajan
,
S.
,
R.
Saravanan
, and
P.
Chang
,
2009
:
The role of the wind–evaporation–sea surface temperature (WES) feedback in air–sea coupled tropical variability.
Atmos. Res.
,
94
,
19
36
.
Mahajan
,
S.
,
R.
Saravanan
, and
P.
Chang
,
2010
:
Free and forced variability of the tropical atlantic ocean: Role of the wind–evaporation–sea surface temperature (WES) feedback.
J. Climate
,
23
,
5958
5977
.
Manabe
,
S.
, and
A. J.
Broccoli
,
1985
:
The influence of continental ice sheets on the climate of an ice age.
J. Geophys. Res.
,
90
, (
D1
).
2167
2190
.
Maykut
,
G. A.
, and
N.
Untersteiner
,
1971
:
Some results from a time-dependent thermodynamic model of sea ice.
J. Geophys. Res.
,
76
,
1550
1575
.
Okumura
,
Y.
,
S.-P.
Xie
,
A.
Numaguti
, and
Y.
Tanimoto
,
2001
:
Tropical Atlantic air–sea interaction and its influence on the NAO.
Geophys. Res. Lett.
,
28
,
1507
1510
.
Peterson
,
L. C.
,
G. H.
Haug
,
K. A.
Hughen
, and
U.
Rohl
,
2000
:
Rapid changes in the hydrologic cycle of the tropical Atlantic during the last glacial.
Science
,
290
,
1947
1951
.
Rajagopalan
,
B.
,
Y.
Kushnir
, and
Y. M.
Tourre
,
1998
:
Observed decadal midlatitude and tropical Atlantic climate variability.
Geophys. Res. Lett.
,
25
,
3967
3970
.
Saravanan
,
R.
,
1998
:
Atmospheric low-frequency variability and its relationship to midlatitude SST variability: Studies using the NCAR Climate System Model.
J. Climate
,
11
,
1386
1404
.
Schmittner
,
A.
,
M.
Latif
, and
B.
Schneider
,
2005
:
Model projections of the North Atlantic thermohaline circulation for the 21st century assessed by observations.
Geophys. Res. Lett.
,
32
,
L23710
.
doi:10.1029/2005GL024368
.
Seager
,
R.
,
D.
Battisti
,
M.
Gordon
,
N.
Naik
,
A.
Clement
, and
M.
Cane
,
2002
:
Is the gulf stream responsible for Europe’s mild winters?
Quart. J. Roy. Meteor. Soc.
,
128
,
2563
2586
.
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
.
Stuiver
,
M.
, and
P. M.
Grootes
,
2000
:
GISP2 oxygen isotope ratios.
Quat. Res.
,
53
,
277
284
.
Timmermann
,
A.
,
S. I.
An
,
U.
Krebs
, and
H.
Goosse
,
2005
:
ENSO suppression due to weakening of the North Atlantic thermohaline circulation.
J. Climate
,
18
,
3122
3139
.
Vellinga
,
M.
, and
R. A.
Wood
,
2002
:
Global climatic impacts of a collapse of the Atlantic thermohaline circulation.
Climatic Change
,
54
,
251
267
.
Xie
,
S.-P.
, and
S.
Philander
,
1994
:
A coupled ocean-atmosphere model of relevance to the ITCZ in the eastern Pacific.
Tellus
,
46A
,
340
350
.
Zhang
,
R.
, and
T. L.
Delworth
,
2005
:
Simulated tropical response to a substantial weakening of the Atlantic thermohaline circulation.
J. Climate
,
18
,
1853
1860
.

Footnotes

* Current affiliation: Oak Ridge National Laboratory, Oak Ridge, Tennessee

Corresponding author address: Salil Mahajan, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN 37830-6301. Email: mahajans@ornl.gov