Geophysical Research Letters

The role of ocean dynamics in tropical Pacific SST response to warm climate in a fully coupled GCM

Authors


Abstract

[1] A global coupled ocean-atmosphere GCM forced by a zonally uniform heat flux anomaly in the tropics is used to investigate the role of oceanic dynamics in regulating the tropical Pacific SST response to warm climate. Consistent with earlier CZ model simulations, the GCM simulation demonstrates a La-Niña like response in the tropical Pacific, with the equatorial upwelling playing a dominant role. While the mean upwelling tends to reduce the overall surface warming due to the strengthening of the equatorial thermocline, the anomalous upwelling associated with the strengthening of the easterlies leads to a weak warming in the east. This easterly wind anomaly can be partly attributed to the enhanced equator-subtropical SST gradient which forces a stronger trade wind. Application to global warming is discussed.

1. Introduction

[2] Tropical Pacific coupled ocean-atmosphere interaction exerts powerful control over the global climate (see review by Alexander et al. [2002]), thus the response of the tropical Pacific climate to elevation of concentration level of greenhouse gases has been receiving great attention over the past decade. Such a response not only provides a useful fingerprint of global warming detection, but also presents a great challenge to our understanding of both regional and global climate shift in response to global warming.

[3] One important aspect of the tropical Pacific response to global warming is the change of the zonal SST gradient along the equator. This has been heavily debated over recent two decades. Current climate models show either a La-Niña like(enhanced zonal SST gradient) or an El-Niño like (reduced zonal SST gradient) response in the tropical Pacific [Collins and CMIP Modeling Groups, 2005], while trends in some reconstructions of tropical Pacific SST tend to show a La-Niña like structure [Cane et al., 1997; Kaplan et al., 1998; Hansen et al., 2006]. The large inconsistency of tropical Pacific SST response to global warming in both observations and climate model simulations leads to a great uncertainty of future changes of the tropical Pacific ocean-atmosphere system.

[4] Several mechanisms have been proposed to explain the changes of the west-east SST contrast, but remain controversial. On one hand, the negative feedback associated with the latent heat cooling [Knutson and Manabe, 1995] or cloud radiation forcing [Meehl and Washington, 1996] over the warm pool has been suggested for generation of an El-Niño like SST response over the tropical Pacific. On the other hand, the equatorial upwelling has been proposed for generation of a La-Niña like tropical Pacific SST response [Clement et al., 1996; Seager and Murtugudde, 1997]. Liu [1998] suggested that the equilibrium response of the west-east SST contrast depends on the effective latitudinal heating difference, with warm anomaly in the subtropics affecting the eastern equatorial Pacific through subduction pathway.

[5] Given the large discrepancy of zonal SST trend in current climate model simulations and observations, it will be very important to evaluate these competitive mechanisms in determining the tropical SST response to warm climate. This study follows a school of early studies to examine the tropical Pacific SST response to a zonally uniform heating [Clement et al., 1996; Seager and Murtugudde, 1997; Liu, 1998], but in a fully coupled ocean-atmosphere general circulation model. This approach, although idealized, can allow us to assess not only the role of ocean dynamics, but also the ocean-atmosphere coupling in regulating the tropical SST response. The paper is organized as follows. Section 2 describes the model and the experiments. Section 3 presents the modeling results. The paper is concluded with a summary and some discussions.

2. Model and Experiment Description

[6] The model used in this study is the fully coupled, global model FOAM, version 1.5, which was developed at the University of Wisconsin [Jacob, 1997]. FOAM captures major features of the observed climatology, as in most state-of-the-art climate models, also produces reasonable climate variability, such as ENSO [Liu et al., 2000] and Pacific decadal climate variability [Wu et al., 2003].

[7] To elucidate the roles of ocean dynamics in regulating the tropical Pacific SST response to warm climate, we carry a set of coupled GCM experiments with an idealized heating imposed over the tropical oceans. In the model, a zonally uniform heating is added over the top of the model-generated heat flux within the entire tropical belt (15°S and 15°N), while the ocean-atmosphere remain coupled locally and elsewhere. The imposed heating peaks on the equator with a magnitude of 15 W/m2, and decreases to zero at 15°N/°S following the cosine distribution. This experiment is referred to as HEAT experiment, which is integrated for 100 years from an equilibrium state of a long control simulation. The contrast of the last 50-years average between the HEAT and the corresponding control experiments is taken as the response. In addition, a few sensitivity experiments embedded with “modeling surgery” strategies [Wu et al., 2003] will be conducted to isolate dynamic processes.

3. Model Result

3.1. Tropical Pacific Response in HEAT Experiment

[8] Figure 1a displays the annual mean SST response to the imposed tropical heating. Within the tropical Pacific (15°N–15°S), the heating induces a basin-wide warming with an average magnitude of about 1°C. One distinctive feature of the tropical Pacific warming is an increase of the equatorial zonal SST gradient, or a La-Niña like response. Associated with the enhanced west-east SST contrast, easterly anomalies, albeit weak, appear in the eastern equatorial Pacific. A stronger warming over the warm pool region is also evident by local positive precipitation anomalies, corresponding to an enhanced deep convection that favors a stronger walker circulation (not shown).

Figure 1.

Annual mean response of the tropical Pacific ocean-atmosphere to a zonally uniform heating. (a) SST (contours) and surface wind (vectors, m/s), (b) net surface heat flux (including the imposed heating), (c) latent and sensible heat flux, and (d) short wave radiation flux. Unit for heat flux is W/m2.

[9] The enhanced zonal SST gradient can not be explained by surface heat flux which favors stronger warming in the east. The net surface heat flux in the eastern equatorial Pacific (16 W/m2) is two times stronger than that in the west (Figure 1b). The imposed heating in the western Pacific is largely offset by larger latent and sensible heat loss (Figure 1c), which reaches to 10 W/m2 but is almost negligible in the east. The total cloud cover increases by about 10% in the western equatorial Pacific (not shown), leading to 10 W/m2 reduction of short wave radiation flux (Figure 1d). In contrast, both cloud cover and short wave radiation remain nearly unchanged in the east. In a word, the larger latent and sensible heat loss [Knutson and Manabe, 1995] and the SST-cloud albedo feedback [Ramanathan and Collins, 1991] constrain the western Pacific warming, tending to result in an El Niño-like SST response.

[10] The west-east response contrast can also be seen in the subsurface. Figure 2a shows the vertical section of equatorial ocean temperature and current anomalies. The warming can penetrate down to a depth of about 100 m in the western Pacific, but is largely trapped within the upper 50 m in the far eastern equatorial Pacific. A substantial cooling occurs under the surface warming layer, leading to an enhanced vertical temperature gradient and equatorial thermocline uplift. The subsurface cooling in the eastern equatorial Pacific appears to be associated with the enhanced equatorial upwelling due to the intensified easterly.

Figure 2.

(a) Longitude-depth section of the equatorial temperature and current anomalies averaged within (5°S, 5°N) with the vertical velocity (m/s) amplified by 1.5 × 105. Heavy lines denote the 20°C isotherm for the control simulation (dashed) and the HEAT experiment. (b) Upper 40 m heat budget terms for western (WP) and eastern (EP) equatorial Pacific of the same volume, the unit is 108 × °Cm3/s.

[11] To further demonstrate the processes involved in the tropical Pacific SST response, a heat budget analysis is carried for the upper western and eastern equatorial Pacific respectively (Figure 2b). It can be seen that the vertical advection plays a dominant role to balance the surface heat flux. While the enhanced vertical temperature gradient provides a damping effect through the mean upwelling (−WT′/∂z < 0) on both western and eastern equatorial Pacific SST, the intensified upwelling substantially reduces the warming in the east (−W′∂T/∂z < 0), leading to the La Niña-like SST response.

[12] In summary, a zonal uniform heating intensifies the zonal SST gradient in the equatorial Pacific. This appears to be primarily associated with intensification of equatorial upwelling in the east. In the following, this mechanism will be further assessed in sensitivity and spin-up experiments.

3.2. Sensitivity Experiment

[13] To further assess the role of equatorial upwelling in regulating equatorial SST response, a sensitivity experiment is carried, which is similar to HEAT experiment but with the equatorial wind stress (15°S–15°N) fixed to the model climatology. This experiment is named as FW-HEAT (Fixed Wind Stress + Heating). In this case, the equatorial dynamic coupling such as upwelling-Bjerkness feedback is eliminated. To prevent potential climatological drift, we performed a parallel experiment in which the equatorial wind stress is constrained but without heating added. Both experiments are integrated for 100 years. The differences between the two experiments in the last 50 years are taken as the response.

[14] In FW-HEAT, the average tropical Pacific SST (15°S–15°N) increases by 1.3°C in contrast to 1°C in the HEAT experiment, but the zonal SST gradient does not show either an El-Niño or a La-Niña like structure. Compared with the warming in HEAT, SST in FW-HEAT increases by 50% in the east while by only 10% in the west. This clearly indicates the important role of upwelling (−W′∂T/∂z < 0) in regulating the eastern equatorial Pacific SST and zonal SST gradient. However, given the latent heat loss and SST-cloud albedo negative feedbacks in the west and absence of upwelling-Bjerkness feedback in the east, a stronger warming in the east (El Niño-like SST warming) should be expected in FW-HEAT. Here, a nearly zonally-uniform warming appear in the equatorial Pacific. This is due to the joint effects of mean upwelling and intensified vertical temperature gradient (−W∂T′/∂z < 0), which tends to reduce the eastern equatorial surface warming as discussed above.

[15] To further test how the wind stress anomaly is set up in the initial stage, a set of spin-up experiments is carried to track the development of both SST and wind stress. The experiments consist of 30-members, with each member starting from a different state of a long control experiment and integrated for two years with the heating added over the tropics.

[16] The ensemble mean responses of SST and surface wind for the first year are demonstrated in Figure 3b. It can be seen that eastern equatorial Pacific SST grows slower than that in the west, with easterly wind anomalies prevailing in the east. At the end of the first year, SST reaches a maximum of 0.8°C in the west, but only about 0.3°C in the east. It is noticed that the easterly wind anomalies appear to initiate on April when the zonal SST gradient has not shown any significant change. The easterly wind anomalies may be attributed to the equatorial heating that enhances the equator-subtropical SST gradient to force a stronger trade wind (Figure 1a). The easterly wind anomaly due to the enhanced meridional SST gradient can be also seen in the FW-HEAT experiment (Figure 3a), in which the zonal SST gradient doesn't show a significant increase. The enhanced easterly winds intensify equatorial upwelling in the east to reduce the surface warming, which can further sustain the easterly wind anomalies. It is also noted that the warming in the east is somewhat intensified in summer. This appears to be associated with model cloud-radiation bias in summer climatology. This warming indeed offsets the west-east gradient. In reality, the response of the west-east SST contrast can be even more significant than the model simulated.

Figure 3.

(a) Annual mean response of SST (contours) and surface wind (vectors) in the FW-HEAT. (b) SST (contours) and surface wind (vectors, m/s) averaged over (5°S, 5°N) for the first year in the 30-ensemble experiment.

[17] The idealized study here may shed light on the tropical Pacific SST response to global warming. To further demonstrate the role of ocean dynamics in regulating the tropical Pacific SST response to global warming, we carried two additional experiments. In one experiment, the concentration of CO2 is doubled (referred to as 2CO2 experiment), and the other one is the same as 2CO2, but with the tropical wind stress constrained to the model control climatology (referred to as FW-2CO2 experiment). FOAM simulates a La-Niña like tropical Pacific SST response to a doubling of CO2 concentration, similar to that forced by a uniform tropical heating (Figure 4a versus Figure 1a). As the tropical wind stress is constrained, the warming over the tropical Pacific becomes nearly uniform along the equator (Figure 4b), with some modest enhancement (reduction) in the east (west), although the change of SST due to the upwelling effects is less significant than that with a forcing of uniform heating.

Figure 4.

Annual mean response of SST (contours) in global warming experiment, (a) 2CO2 and (b) FW-2CO2.

4. Summary and Discussion

[18] A global coupled ocean-atmosphere model forced by a zonally uniform heat flux anomaly in the tropics is used to investigate the role of oceanic dynamics in regulating the tropical Pacific SST response to warm climate. The model demonstrates a La-Niña like response in the tropical Pacific, with the equatorial upwelling playing a dominant role. While the mean upwelling tends to reduce the overall surface warming due to the strengthening of the equatorial thermocline, the anomalous upwelling associated with the strengthening of the easterlies leads to a weak warming in the east, resulting in a La-Niña like pattern. This anomalous upwelling can be partly attributed to the enhanced meridional SST gradient that accelerates the trade winds.

[19] The study here suggests a dominant role of equatorial upwelling in regulating the tropical Pacific west-east SST gradient response to warm climate, with the horizontal advection playing a minor role. This appears to exclude the oceanic subduction mechanism proposed by Liu and Huang [1997], but this may be due to a diffusive thermocline in our coupled model. However, the enhanced equator-subtropical SST gradient can increase the tropical zonal SST gradient by forcing a stronger trade wind to increase upwelling in the eastern equatorial Pacific. This meridional mechanism is different from the zonal coupled mechanism proposed by Clement et al. [1996]. The enhanced equatorial warming also holds in global warming [Liu et al., 2005].

[20] Although the tropical Pacific SST pattern in response to global warming in our coupled model simulation appears to be a La-Niña like structure, and consistent with some observations, it still remains unclear if the results can be applied to other models, given the large inconsistency of tropical Pacific SST trends among different coupled climate models and observations. But at least, our modeling results here suggest we need to improve realization of equatorial dynamics in climate models for a better understanding of the tropical oceanic response in the warming climate.

Acknowledgments

[21] This work is supported by the “Zhufeng Project” of Ocean University of China with funds provided by the Chinese Ministry of Education and Chinese National Science Foundation Outstanding Youth grant 40788002. Discussions with Zhengyu Liu and Shang-Ping Xie are helpful. Comments from two anonymous reviewers improved the paper substantially. The experiments were carried on the Computer Cluster of Ocean Data and Assimilation Center at Ocean University of China.

Ancillary