Paleoclimate observations and modeling studies suggest that extratropical climate change affects the tropical Pacific. A global coupled general circulation model is used to investigate the equatorial Pacific response to extratropical surface heat flux forcing that is downward (upward) poleward of 40°N (S). The equatorial response consists of two distinct stages: the zonal sea surface temperature (SST) gradient strengthens for the first two to three decades and then weakens afterward. In the first stage, fast surface air–sea coupling feedback mechanism communicates the extratropical warming (cooling) from the North (South) Pacific toward the equator. The second stage is characterized by a basinwide shoaling of the tropical Pacific thermocline as the subtropical cell (STC) advects cold water from the South Pacific along the thermocline. This preference of Southern Hemisphere anomalies is due to the meridional asymmetry in the mean circulation: the interior pathway for STC is open south but partially blocked north of the equator. Paleoclimate implications are discussed.
Tropical Pacific SST change, such as El Niño–Southern Oscillation (ENSO), exerts significant impacts on global climate (Alexander et al. 2002; Chiang 2009). Paleoclimate proxy data show close linkages between high latitudes and the tropics (Clement and Peterson 2008; Chiang and Friedman 2012). Instrumental observations and model simulations indeed show that thermal forcing in the high latitudes can trigger tropical climate changes through atmospheric bridges and oceanic tunnels (Liu and Alexander 2007). The intertropical convergence zone (ITCZ) shifts toward the warmer hemisphere in paleoclimate scenarios in response to high-latitude ice cover (Chiang and Bitz 2005), interhemispheric antisymmetric heating in high latitudes (Broccoli et al. 2006; Kang et al. 2008), and changes in the Atlantic thermohaline circulation (Zhang and Delworth 2005; Wu et al. 2008).
The zonal SST gradient in the equatorial Pacific is coupled with atmospheric deep convection and zonal overturning circulation (e.g., Sun and Liu 1996). A recent study by Chiang et al. (2008) suggested that the zonal SST gradient in the tropical Pacific can be enhanced by a northward interhemispheric thermal gradient (ITG) forcing based on a hybrid coupled model consisting of an atmospheric general circulation model (AGCM) and a 1.5-layer reduced gravity ocean model. A fast surface ocean–atmosphere interaction mechanism, called the wind–evaporation–SST (WES; Xie and Philander 1994) footprinting (Vimont et al. 2001, 2003; Alexander et al. 2010), was identified as important for communicating the extratropical forcing to the equatorial Pacific (Wu and Li 2007; Ma and Wu 2011; Li and Wu 2012).
The ventilation of cold water from the midlatitude winter mixed layer by the shallow overturning circulation maintains the thermocline in the equatorial Pacific (McCreary and Lu 1994). The present study explores this extratropical-to-tropical teleconnection via the ventilated thermocline (Gu and Philander 1997; Liu 1999), which could become important on slower decadal time scales in regulating the tropical SST changes. Specifically, we investigate the tropical Pacific responses to a northward ITG forcing using a global fully coupled atmosphere–ocean model. We show that the model response comprises two distinct phases due to the fast surface air–sea coupling and slow thermocline ventilation, respectively.
2. Model and experiment
We use the Fast Ocean–Atmosphere Model (FOAM) version 1.5, a fully coupled global model. The atmosphere model is a parallel version of the National Center for Atmospheric Research (NCAR) Community Climate Model version 2 (CCM2), with the atmospheric physics replaced with those of CCM3. The ocean model follows the Geophysical Fluid Dynamics Laboratory (GFDL) Modular Ocean Model (MOM). The FOAM used mixed resolution here, an atmospheric resolution of R15 with 18 vertical levels and an oceanic resolution of 1.4° latitude × 2.8° longitude with 32 vertical levels. To accommodate the different resolutions of the ocean and atmosphere models, the coupler uses an “overlap grid” obtained by laying the atmosphere grid on top of the ocean grid. The fluxes of heat, etc., are calculated by the coupler on each overlap grid cell using the corresponding atmosphere and ocean temperatures. The fluxes are then accumulated onto the appropriate grid for use by the ocean or atmosphere. In addition, the FOAM also has a thermodynamic sea ice component. Without flux correction, the fully coupled control simulation has been integrated for over 2000 years, without apparent climate shift. FOAM captures major features of the observed climatology (Jacob 1997) and variability modes, such as ENSO (Liu et al. 2000) and Pacific decadal climate variability (Wu et al. 2003). As in observation, in particular, ENSO in the FOAM model has periods of 2–7 years and phase locks in boreal winter [November–January (NDJ)], but its amplitude is weaker about 40% in FOAM than in observations because of the diffusive thermocline (Liu et al. 2000), cold tongue extending too far west (Wu et al. 2007a), and the weak zonal SST gradient (Table 1) (Schopf and Burgman 2006). Because of its low resolution, the FOAM model does not well capture western boundary currents, but it captures the major characteristics of large-scale oceanic circulation and the subtropical–tropical cells (STCs), similar to the data-derived characterizations of McPhaden and Zhang (2002, 2004), including the equatorial asymmetry in STCs (Lu and McCreary 1995). In addition, FOAM also displays a double ITCZ in the tropical Pacific, a far west extension of the cold tongue, and a weak zonal SST gradient, biases common to many coupled models (Lin 2007; de Szoeke and Xie 2008).
Following previous studies (Broccoli et al. 2006; Kang et al. 2008; Chiang et al. 2008), we prescribe an idealized ITG forcing with uniform net heat flux anomalies of +30 W m−2 and −30 W m−2 imposed at the ocean surface north of 40°N and south of 40°S, respectively. To reduce noise, ensemble experiments with five members are performed from different initial conditions of the control run. In addition to this fully coupled (FC) experiment, we perform a “partial blocking” (PB) experiment to assess the role of subduction from the southern Pacific. In the PB experiment, we prescribe the same ITG forcing but place a sponge wall (by restoring the ocean temperature and salinity toward the climatology with the annual cycle from the FOAM control run) from 50-m depth to the bottom in the 40°–30°S latitude band to block the southern Pacific ventilation but allow surface air–sea coupling feedback (Wu et al. 2003; Ma and Wu 2011). Thus, this PB scheme in FOAM1.5 almost does not change the oceanic surface circulation by comparing surface currents averaged over upper 50 m and vertical profiles of zonal and meridional currents averaged over 140°E–110°W (not shown). The model takes about 40–50 years to reach a new equilibrium under the ITG forcing (Figs. 1c,d). Here, the differences between the FC (PB) experiment and the corresponding control run are defined as the anomalous response, respectively.
a. SST response
To monitor the equatorial Pacific response, we define a zonal SST gradient index as the difference between SST averaged in a western (5°S–5°N, 130°E–180°) and an eastern (5°S–5°N, 150°–90°W) box. To focus on the long time scale response, all time series are smoothed by 9-yr running mean to suppress the interannual variability.
With surface warming in the northern and cooling in the southern extratropics, the equatorial zonal SST gradient responses in all five member experiments first strengthen and then weaken despite considerable differences among runs (Fig. 1c). For the five-member ensemble mean, the equatorial Pacific response features two distinct stages (Fig. 1; Table 1). For the first two to three decades, SST in the west and east equatorial Pacific warms and cools by about 0.06° and −0.12°C, respectively (Figs. 1a,b), resulting in a 0.18°C increase in equatorial Pacific zonal SST gradient (Fig. 1c). The initial intensification of SST gradient in response to the northward ITG forcing is consistent with Chiang et al. (2008) in their hybrid coupled model simulation. It should be noted that the response of equatorial zonal SST gradient is weaker than that in Chiang et al. (2008), possibly because of the coarse resolution of our model. After the first three decades, however, western equatorial Pacific SST switches from the initial warming to strong cooling with a magnitude of −0.50°C, while eastern equatorial Pacific SST cooling amplifies to −0.35°C. The zonal SST gradient varies from +0.18°C in the first stage to −0.15°C in the second, in response to ITG forcing (Fig. 1c). This transition in equatorial SST response is accompanied by a basinwide shoaling of the equatorial Pacific thermocline (Fig. 1d), suggesting that different mechanisms operate at the two stages.
The initial adjustment is largely accomplished by surface ocean–atmosphere coupling, (Chiang and Bitz 2005) (Fig. 2a, left). The downward (upward) heat flux anomalies forces oceanic surface warming (cooling) in the northern (southern) high-latitude Pacific, which feeds back to the atmosphere above through surface turbulent heat flux (both latent and sensible) and warms (cools) the atmosphere. This warming (cooling) in the northern (southern) high-latitude Pacific extends into the midlatitude Pacific through heat transport and mixing, and further spreads to the regions of northeasterly (southeasterly) trades (Chiang and Bitz 2005). In the subtropical North Pacific, the SST warming induces anomalous southwesterlies and weakens the trade winds on the south flank, reducing oceanic turbulent heat loss and causing warm SST anomalies to propagate southwestward to the equatorial west Pacific (Wu and Li 2007; Li and Wu 2012). Likewise in the southern subtropical Pacific, the cooling induces anomalous southeasterlies and intensifies the trades, causing cold SST anomalies to propagate equatorward to the equatorial east Pacific (Ma and Wu 2011). Near the equator, the warming in the north and the cooling in the south set up a meridional pressure gradient, intensifying the southerly cross-equatorial wind anomalies and causing anomalous rainfall concentrated north of the model’s ITCZ and a northward migration of the ITCZ (Chiang and Bitz 2005; Broccoli et al. 2006; Kang et al. 2008). This surface air–sea coupling feedback mechanism has been used to explain the tropical Pacific response to extratropical perturbations in the North Pacific (Vimont et al. 2001, 2003; Wu et al. 2007b; Wu and Li 2007; Alexander et al. 2010; Li and Wu 2012) and South Pacific (Ma and Wu 2011).
Anomalous wind curls associated with the air–sea coupling feedback excite upwelling (downwelling) around 15°N (15°S) (Fig. 2b, left). In addition, the reduced (enhanced) trade winds weaken (intensify) the subtropical–tropical cell in the subtropical North (South) Pacific (Fig. 3b). The antisymmetric change in the STC reinforces the thermocline cooling (warming) in the north (south). The antisymmetric change in the STC, however, provides little net effect on the equatorial thermocline (Fig. 1d). The upwelling in the subtropical North Pacific is stronger than the downwelling one in the subtropical South Pacific. The former appears responsible for the cooling anomalies in the thermocline of the equatorial west Pacific (Fig. 2c, left) (Liu 1999; Yang et al. 2005).
After the first several decades, the tropical Pacific Ocean experiences a cooling from the surface to 800 m (Figs. 2c,d). The equilibrium minus first-stage temperature difference suggests that the surface cooling propagates into the equatorial region from the South Pacific via the thermocline ventilation (Fig. 2d, right). To confirm the thermocline ventilation, we calculate meridional temperature advection (Figs. 4a,b). The thermocline cooling in the second stage is mainly caused by mean subduction () from the southern Pacific, while the southern STC () enhances the subsurface cooling in the equatorial to north subtropical Pacific regions (10°N). The meridional temperature advection is consistent with Figs. 2d and 3c.
The tropical Pacific climate is asymmetric about the equator, and the northward displaced ITCZ limits the interior pathway from the northern midlatitude to the equator but allows such an interior pathway to operate efficiently in the Southern Hemisphere. Thus, the surface cooling in the midlatitude South Pacific can reach the equator via the mean STC while the surface warming in the North Pacific cannot. Compared to the initial stage, the northern STC changes little while the southern STC strengthens significantly in the equilibrium (Fig. 3c). This asymmetry in STC intensity change also helps cool the equatorial thermocline. In the equilibrium with a shoaling thermocline (Fig. 1d), the surface cooling is stronger in the west than in the east along the equator, accompanied by easterly wind anomalies in the west that enhance upwelling cooling (Figs. 2a and 3). This suggests that the change in zonal SST gradient involves interaction with zonal wind on the equator, which is confirmed by heat budget analyses (Fig. 4c). In the west equatorial Pacific, horizontal advection by the mean flow, anomalous meridional advection, and vertical entrainment contribute to the SST cooling, while other heat budget terms damp it. In the east equatorial Pacific, meridional advection and vertical advection by the mean flow contribute SST cooling while other heat budget terms damp the SST cooling. The dominant terms for SST cooling are meridional advection by the mean flow and vertical entrainment in the west equatorial Pacific while meridional advection by the mean flow in the east equatorial Pacific.
By blocking subsurface ocean pathways, the PB experiment tests the mechanism that the southern STC advects cold water from the extratropical South Pacific, cooling the tropical Pacific. The equilibrium response in PB resembles the FC response at the initial stage (Fig. 5). The cooling of the equatorial Pacific is much weaker than the equilibrium response in FC, and the surface cooling is stronger in the eastern than western basin. The failure for the second stage to develop in PB supports our hypothesis that the advection by the southern STC holds the key to the second-stage response in FC characterized by a basinwide shoaling of the tropical thermocline.
4. Summary and discussion
We have investigated the tropical Pacific response to a northward thermal gradient forcing that is itself limited to 40°N/S using a global coupled ocean–atmosphere model. The results show that the zonal SST gradient in the equatorial Pacific goes through two distinct stages: it intensifies initially and then relaxes after two to three decades. The second stage has not been discussed in the literature and is due to subsurface oceanic advection.
In the initial stage, the equatorial Pacific SST response is due to the surface air–sea coupling feedback mechanism, which propagates the warming (cooling) SST in the subtropical North (South) Pacific to the equator. The second stage takes place on the equator in two to three decades after the extratropical forcing is imposed. Cold temperature anomalies are subducted into the equatorial thermocline from the South Pacific (Fig. 2d, right), resulting in a basinwide shoaling of the thermocline in the tropical Pacific (Fig. 1d). The preference for a subsurface ocean pathway from the Southern Hemisphere has to do with the latitudinal asymmetry of Pacific climate: the southern STC connects to the equator through the interior ocean south of the equator while the interior pathway is largely blocked in the North Pacific by the barrier of low potential vorticity under the ITCZ (Lu and McCreary 1995). At the surface on the equator, the SST cooling is stronger in the western basin, accompanied by easterly wind anomalies.
Tropical variability goes through three stages, called the enhanced, weakened, and re-enhanced stages in our forcing experiments (Fig. 6). For the first two to three decades, tropical variability strengthens immediately after the surface thermal forcing is imposed, suggesting the important role of the fast surface air–sea coupling. Competition between surface air–sea coupling and subduction mechanisms from the southern Pacific seems to weaken the tropical variability. After 50 years, the tropical variability reintensifies, suggesting that role of extratropical subduction is more important than surface air–sea coupling at the equilibrium stage. The time delay of tropical response to the southern Pacific ventilation is consistent with Sun et al. (2004) and Sun and Zhang (2006).
Our study identifies the advection by the southern STC as important for tropical Pacific response to extratropical surface heat flux forcing, in addition to the surface air–sea coupling feedback mechanism identified in previous studies. Our PB experiment supports this southern STC mechanism. Our results indicate that under interhemispherically antisymmetric forcing, the equatorial Pacific preferentially responds to the Southern Hemispheric forcing, a result that should be tested in other coupled GCMs. If confirmed, this result forms an important element in the study of tropical–extratropical interaction and can help interpret sparse paleoclimate observations.
This work is supported by Natural Science Foundation of China grants (40906003, 40830106 and 41176006), the National Key Basic Research Project (2012CB955600), and the Changjiang Scholar and Qianren Projects. All experiments were run in the Ocean University of China Supercomputing Center.