Geophysical Research Letters

Increase of South Pacific eastern subtropical mode water under global warming



[1] The response of South Pacific Eastern Subtropical Mode Water (SPESTMW) to global warming is investigated by comparing solutions from a set of Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) coupled models between a present-day climate and a future, warmer climate. Under the warmer climate scenario, the SPESTMW extends southwestward and is significantly increased in volume. This is because all the local surface forcing mechanisms (i.e., wind stress, heat and freshwater fluxes) in the eastern subtropical South Pacific tends to de-stratify the upper ocean and thus deepen the mixed layer. Further, a suite of process-oriented experiments with an ocean general circulation model suggest that it is the intensified southeast trade winds under the warmer climate that promotes more heat flux from the ocean into the atmosphere that then results in a deepening of the mixed layer in the eastern subtropics of the South Pacific.

1. Introduction

[2] Recent analyses of global warming projections simulated with climate models suggest that there are significant atmospheric and oceanic changes in the Pacific in response to increased greenhouse gases. The atmospheric changes include a weakening of the Walker circulation in the tropics [e.g., Vecchi and Soden, 2007] and a poleward expansion of the Hadley circulation in the extra-tropics [e.g., Lu et al., 2007]. The oceanic changes include a surface-warming pattern with a minimum in the eastern subtropics of the South Pacific [e.g., Meehl et al., 2007], an enhanced equatorial response [e.g., Liu et al., 2005], a banded warming structure in the subtropics of the North Pacific [e.g., Xie et al., 2010], and a maximum warming in the vicinity of the Kuroshio Extension [e.g., Overland and Wang, 2007].

[3] In addition to atmospheric and sea surface temperature (SST) patterns, robust changes of the ocean circulation in the Pacific have also been identified under global warming. In responding to the weakening of the Walker circulation, the equatorial ocean features a reduction of the mean depth of the thermocline, a minimum warming in the thermocline of the western Pacific, and a weakening of the system of surface currents [Vecchi and Soden, 2007; Luo et al., 2009a]. Consistent with the expansion of the Hadley circulation, the boundary between the North Pacific subtropical and subpolar gyres shifts northward. Another robust feature in the North Pacific Ocean is a significant reduction in the volume of the Mode Waters [Luo et al., 2009b]. This change is due to a more stratified upper ocean and thus a shoaling of the winter mixed layer depth (MLD) resulting mainly from a reduction of the ocean-to-atmosphere heat loss over the subtropical North Pacific. The reduced Mode Water leads to a weakening of the North Pacific subtropical countercurrent, leaving behind a northeast slanted, banded structure in SST warming (S.-P. Xie, personal communication, 2010).

[4] For this study we turn our attention to the South Pacific and investigate how eastern subtropical Mode Water responds to global warming. The South Pacific Eastern Subtropical Mode Water (SPESTMW) is a water mass that persists all year long and is geographically distinct from the western subtropical Mode Water found north of New Zealand [Wong and Johnson, 2003]. SPESTMW is characterized by low potential vorticity (PV), a temperature of 13–26°C, salinity greater than 34 psu, and a density range of 24.5–25.8 kg m−3 [Sato and Suga, 2009]. After formation, the SPESTMW is advected by the eastern limb of the South Pacific subtropical gyre, moving it northwestward towards the equator and eventually joining the South Equatorial Current. As the climate warms, in sharp contrast to the change of North Pacific Mode Waters, we will show that the SPESTMW tends to extend southwestward and is significantly increased in volume, resulting mainly from an intensification of the southeast trade winds.

2. Data and Model

[5] We examine the response of SPESTMW to global warming by comparing solutions from a set of Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) coupled models between a present-day climate and a future, warmer climate. Simulations from the IPCC 20C3M and the SRESA1B experiments are selected to represent the present-day climate and the future, warmer climate, respectively. The 20C3M experiments simulate the 20th century climate where atmospheric CO2 concentrations and other input data are based upon historical records; the SRESA1B scenarios represent a climate change in which atmospheric CO2 concentrations double their present-day level in the year 2100 and are then held fixed and maintained to years 2200 or 2300. For this study, the 1951–2000 means are used for 20C3M and the 2151–2200 means are used for SRESA1B. Ensemble means from eleven climate models (CGCM3.1, CNRM−CM3, CSIRO−MK3.0, FGOALS−g1.0, GFDL−CM2.0, GFDL−CM2.1, HADCM3, IPSL−CM4, MIROC3.2, MRI−CGCM2.3.2, and PCM1) provided by the Program for Climate Model Diagnosis and Intercomparison (PCMDI) IPCC data center are examined in this study, chosen because they have sufficient output available for the SRESA1B scenarios. One member run is analyzed for each of the models.

[6] In addition, we setup a suite of numerical experiments with the Hybrid Coordinate Ocean Model (HYCOM) to investigate the relative importance of wind stress and buoyancy forcing for changing the SPESTMW under global warming. Here we configure HYCOM for the Pacific Ocean from 124°E to 72°W and 55°S to 60°N, with a constant horizontal resolution of 0.5° × 0.5° and 20 layers in the vertical. Based on the eleven IPCC model solutions from the 20C3M and SRESA1B experiments, we create the monthly climatology of temperature and salinity to use as the HYCOM's initial and open boundary conditions, and of atmospheric fields (wind stress, air temperature, longwave and shortwave radiation, specific humidity, and precipitation) to use as the HYCOM's surface forcing. Note that in HYCOM evaporation as well as latent and sensible heat fluxes are calculated using bulk formula in which wind speed as an input atmospheric variable is derived from wind stress. Five experiments are performed and their atmospheric forcing fields are listed in Table 1. In addition, experiments CRTL and WIND are initialized from the January 20C3M climatology, and their temperature and salinity near the southern boundary at 55°S are relaxed to the monthly 20C3M climatology; while experiments FULL, FLUX, and SPED are initialized from the January SRESA1B climatology in order to reduce the model's spin-up time, and their temperature and salinity near the southern boundary are relaxed to the monthly SRESA1B climatology. All the experiments are integrated for 25 years, and the winter MLD presented in Section 4 is an average from July to September of the last 5 model years.

Table 1. HYCOM Experiments
NameAtmospheric ForcingMotivation
Wind StressWind SpeedOther Fields
CTRL20C3M20C3M20C3MControl run
FULLSRESA1BSRESA1BSRESA1BResponse to global warming
WINDSRESA1B20C3M20C3MRole of wind stress
FLUX20C3MSRESA1BSRESA1BRole of buoyancy flux
SPED20C3M20C3MSRESA1BRole of wind speed


[7] A key process for formation of Mode Waters is subduction through which mixed layer waters enter the permanent thermocline. The subduction of Mode Waters can be accomplished by either Ekman pumping or lateral induction, with the latter found to play a more important role for producing SPESTMW [e.g., Qu et al., 2008]. Since lateral induction, defined as horizontal advection across a sloping mixed layer base, becomes large where strong convection produces considerable mixed layer deepening due to strong wintertime cooling [Toyoda et al., 2004], we first examine the winter MLD. Here the MLD is defined as the depth at which the water density is 0.1 kg m−3 denser than the sea surface. For the South Pacific, the MLD reaches its seasonal maximum in austral winter (Figure 1a), and the winter MLD is generally shallow along the northern rim of the subtropical gyre but gets increasingly deeper toward high latitudes. In the eastern subtropics, there is a local maximum (∼100 m) centered at (105°W, 25°S), corresponding to the formation region of SPESTMW [Wong and Johnson, 2003]. These features of MLD in the model compare reasonably well with the observation [e.g., Qu et al., 2008].

Figure 1.

(a) Mean depth (color in m) and density (contour interval (CI) = 0.5 Kg m−3) of winter mixed layer in 20C3M. (b) Mean differences of depth (color in m) and density (CI = 0.1 Kg m−3) of winter mixed layer between the two scenarios (SRESA1B minus 20C3M). (c) Number of models simulating an increase of winter mixed layer depth in SRESA1B.

[8] As the climate warms (Figure 1b), the MLD becomes shallower in most parts of the South Pacific, resulting from a more stratified upper layer due to ocean warming that is greater near the surface and decreases with depth [Luo et al., 2009a]. However, the local mixed layer in the eastern subtropics is seen to deepen, with the maximum MLD area (the SPESTMW formation area) extending southwestward, and this change is robust among the eleven models (Figure 1c). In addition, it can also be seen that this deepening is not uniform spatially, with its maximum ∼30 m near (112°W, 28°S). Such a change in the mixed layer is associated with larger MLD horizontal gradients which in turn will result in stronger lateral induction and subduction, suggesting an increase of SPESTMW in the warmer climate.

[9] To demonstrate the change of the SPESTMW under global warming, the distributions of winter PV are shown in Figure 2 along both isopycnal surfaces and sections chosen as guided by the distinct cores of minimum PV in 20C3M and SRESA1B. It can clearly be seen that the region with low-PV water in the eastern subtropics is expanded both horizontally and vertically under the future warming scenario. Since the low-PV water in the region corresponds to the SPESTMW, this indicates that the warmer climate induces a significant increase in volume of the SPESTMW. (If the SPESTMW is represented by PV < 2.5 × 10−10 m−1 s−1 and density classes of 24.5 < σt < 25.8 Kg m−3 in 20C3M and of 24.2 < σt < 25.4 Kg m−3 in SRESA1B, we find that its volume within the region between 80°−150°W and 15°−35°S is increased by 1.5 × 1014 m3 or 24% from 6.3 × 1014 m3 in 20C3M to 7.8 × 1014 m3 in SRESA1B). However, the SPESTMW appears to be lighter in the warming climate, with a core density range of 24.7−25.1 Kg m−3 in SRESA1B as opposed to a density range 25.0−25.4 Kg m−3 in 20C3M. In the next section, we will look into why the response of the SPESTMW to global warming is different than that in the North Pacific Mode Waters, which are significantly weakened in the warmer climate [Luo et al., 2009b].

Figure 2.

Mean winter potential vorticity (PV; color in 10−11 m−1 s−1) in (left) 20C3M and (right) SRESA1B. (top) PV along isopycnal surfaces; (middle and bottom) PV along sections and density superimposed (CI = 0.2 Kg m−3).

4. Forcing Mechanisms

[10] As mentioned previously a spatially non-uniform deepening of the mixed layer is associated with a strengthening of lateral induction. To illustrate the change of lateral induction in the South Pacific eastern subtropics under global warming, we estimate the lateral induction by equation imagembequation imagehm where equation imagemb is the horizontal velocity at the base of the mixed layer and hm is the depth of the mixed layer. For the central area between 85°−125°W and the outcropping lines of the core isopycnals (25.0−25.4 Kg m−3 in 20C3M and 24.7−25.1 Kg m−3 in SRESA1B, respectively), the lateral induction transport is found to increase by 0.78 Sv, or 40%, from 1.94 Sv in 20C3M to 2.72 Sv in SRESA1B. This significantly strengthened lateral induction corresponds to the significantly increased SPESTMW described above in Section 3, demonstrating that the increased MLD horizontal gradients, a result of spatially non-uniform deepening of the mixed layer, plays a major role for increasing SPESTMW under global warming.

[11] Since the mixed layer change is typically related to the surface forcing change, the deepening of the mixed layer in the eastern subtropical South Pacific is most likely due to changes of local surface buoyancy fluxes (i.e., heat and freshwater fluxes) and wind stress under global warming. We will show below that both of these surface fluxes appear to change significantly in the warmer climate and that these changes indeed tend to deepen the mixed layer, resulting in more low-PV waters formed in the eastern subtropics of the South Pacific.

[12] For the changes in wind stress (Figure 3a), the southeast trade winds are intensified across the basin with the region in the eastern subtropics most significantly affected, leading directly to local mixed layer deepening due to stronger wind stirring. In addition, the minimum surface warming in the South Pacific eastern subtropics is linked to the intensified trade winds, suggestive of a wind-evaporation-SST (WES) feedback between them [Xie et al., 2010; Timmermann et al., 2004].

Figure 3.

Mean winter differences of (a) wind stress (Pa) and Ekman pumping (color in 10−8 m s−1) and (b) ocean-to-atmosphere heat flux (color in W m−2) and freshwater flux (CI = 5 × 10−6 Kg m−2 s−1) between the two scenarios (SRESA1B minus 20C3M).

[13] For the changes in heat flux (Figure 3b), there is a net heat loss from the ocean into the atmosphere over most regions of the subtropics, with a maximum in the eastern subtropical region nearing the tropics. In addition, a localized maximum of heat loss (∼15 W m−2) is also found in the area around (105°W, 30°S), the center of the mixed layer deepening under global warming. This change of heat flux produces a less stratified upper ocean and a resultant deepening of the MLD in the eastern subtropics. Among the four physical components included in the heat flux (i.e., solar radiation, longwave radiation, turbulent fluxes of sensible and latent heat), a further analysis indicates that the contribution from the latent heat flux is predominant (not shown). This change in the latent heat flux can be caused by the intensified southeast trade winds under the future climate scenario, which promotes more heat loss from the ocean into the atmosphere through the WES feedback mechanism. In terms of the changes in the freshwater flux (Figure 3b), the warmer climate drives an increase of evaporation minus precipitation over the entire subtropical South Pacific with values exceeding 5 × 10−6 Kg m−2 s−1 coinciding with the high heat loss region over the eastern subtropics. This change of freshwater flux is also favorable for decreasing the upper ocean stratification and deepening the mixed layer. Therefore, an analysis of the IPCC model outputs suggests that it is the local changes in wind stress and buoyancy fluxes in the eastern subtropics of the South Pacific that leads to the mixed layer deepening and thus increased SPESTMW under global warming.

[14] To differentiate the wind stress versus the buoyancy flux forcing for determining their contributions to the mixed layer deepening, we have performed five process-oriented experiments with HYCOM and highlight the winter MLD patterns (Figure 4). In experiments CTRL and FULL (Figures 4a and 4b), HYCOM produces spatial patterns of MLD similar to the IPCC simulations in 20C3M and SRESA1B (Figures 1a and 1b), i.e. the local maximum MLD in the eastern subtropics expands southwestward and deepens under the warmer climate scenario. Note that the MLD from the IPCC output appears to be shallower compared with that from the HYCOM experiments. This is because the IPCC results are means from an eleven-model ensemble over a period of 50 years and are thus very much smoothed. By comparing the MLD patterns of experiments WIND (Figure 4c) and FLUX (Figure 4d) with experiment FULL (Figure 4b), it can be seen that the mixed layer deepening over the eastern subtropics results mainly from the change of buoyancy fluxes in the warmer climate. However, this deepening pattern in experiment FLUX (Figure 4d) does not appear in experiment SPED (Figure 4e), implying that the change in the wind speed is the major factor for deepening the mixed layer in the eastern subtropical South Pacific. Therefore, a comparison of the MLD spatial patterns from a suite of HYCOM experiments suggests that it is the intensified southeast trade winds, via promoting the generation of anomalous buoyancy fluxes from the ocean into the atmosphere, which results in a deepening of the MLD in the eastern subtropics of the South Pacific.

Figure 4.

(a-e) Winter mixed layer depth (color in m) from process-oriented experiments with HYCOM. Design of these experiments can be found in the text and Table 1.

5. Discussions

[15] In addition to lateral induction, the subduction of SPESTMW can also be accomplished through Ekman pumping [e.g., Huang and Qiu, 1998]. Comparing the SRESA1B with the 20C3M experiments, the wind changes result in an enhanced Ekman pumping (downward is negative) in the eastern subtropical South Pacific (Figure 3a), favoring the SPESTMW formation. It should be noted that, although subduction is important, it is not the only process responsible for Mode Water formation. Changes in other processes, e.g. those associated with mass sinks (dispycnal mixing), might also play a role in changing SPESTMW in a warmer climate.

[16] Because model results depend upon how well the models represent the surface forcing as well as the pycnocline structure, the formation region of the SPESTMW varies among the models. As the climate warms, however, a deepening of winter MLD in the eastern subtropics of the South Pacific (Figure 1c) and an increase of the volume of SPESTMW are seen in all of the eleven models. In addition, the changes of surface heat flux, freshwater flux and winds over the eastern subtropics under global warming are common to these models, strongly supporting our findings above that are based upon a model ensemble mean.

[17] The increased volume of the SPESTMW possibly contributes to a minimum warming in the thermocline of the western equatorial Pacific under global warming scenarios, which is one of the more robust features that appear in climate models. There are three good reasons for this possibility. First, the equatorial thermocline water originates from the subtropics in both hemispheres, but with a more significant contribution from the southern subtropics [e.g., Yeager and Large, 2004]. Second, as the climate warms the SST warming is much less in the southern subtropics than in the northern subtropics, with a minimum found in the eastern subtropics of the South Pacific [Xie et al., 2010]. Lastly, the North Pacific Mode Waters are significantly reduced in a warmer climate [Luo et al., 2009b] and the contribution from the northern subtropics to the minimum warming in the western equatorial thermocline is thus excluded.


[18] We acknowledge the international modeling groups for providing their data for analysis and the Program for Climate Model Diagnosis and Intercomparison and the IPCC Data Archive at Lawrence Livermore National Laboratory for collecting, archiving, and making the data available. The authors wish to thank anonymous reviewers for their numerous comments that helped to improve the original manuscript. This work is supported by NSF grant OCE-0752194, the NSFC 40830106, Ministry of Science and Technology of China (National Key Programs for Developing Basic Science 2007CB411803 and 2010CB428904), and the 111 Project of China (B07036).