Editorial Type:
Article
Article Type:
Research Article

The Large-Scale Ocean Dynamical Effect on Uncertainty in the Tropical Pacific SST Warming Pattern in CMIP5 Models

Jun Ying Center for Monsoon System Research, and LASG, Institute of Atmospheric Physics, Chinese Academy of Sciences, and University of Chinese Academy of Sciences, Beijing, China

Search for other papers by Jun Ying in
Current site
Google Scholar
PubMed Close
and
Ping Huang Center for Monsoon System Research, and LASG, Institute of Atmospheric Physics, Chinese Academy of Sciences, and Joint Center for Global Change Studies, Beijing, China

Search for other papers by Ping Huang in
Current site
Google Scholar
PubMed Close
Print Publication:
15 Nov 2016
DOI:
https://doi.org/10.1175/JCLI-D-16-0318.1
Page(s):
8051–8065
Full access

Abstract

This study investigates how intermodel differences in large-scale ocean dynamics affect the tropical Pacific sea surface temperature (SST) warming (TPSW) pattern under global warming, as projected by 32 models from phase 5 of the Coupled Model Intercomparison Project (CMIP5). The largest cause of intermodel TPSW differences is related to the cloud–radiation feedback. After removing the effect of cloud–radiation feedback, the authors find that differences in ocean advection play the next largest role, explaining around 14% of the total intermodel variance in TPSW. Of particular importance are differences in climatological zonal overturning circulation among the models. With the robust enhancement of ocean stratification across models, models with relatively strong climatological upwelling tend to have relatively weak SST warming in the eastern Pacific. Meanwhile, the pronounced intermodel differences in ocean overturning changes contribute little to uncertainty in the TPSW pattern. The intermodel differences in climatological zonal overturning are found to be associated with the intermodel spread in climatological SST. In most CMIP5 models, there is a common cold tongue associated with an overly strong overturning in the climatology simulation, implying a La Niña–like bias in the TPSW pattern projected by the MME of the CMIP5 models. This provides further evidence for the projection that the TPSW pattern should be closer to an El Niño–like pattern than the MME projection.

Corresponding author address: Dr. Ping Huang, Center for Monsoon System Research, Institute of Atmospheric Physics, Chinese Academy of Sciences, Bei-Er-Tiao 6, Zhong-Guan-Cun, Beijing 100190, China. E-mail: huangping@mail.iap.ac.cn

Abstract

This study investigates how intermodel differences in large-scale ocean dynamics affect the tropical Pacific sea surface temperature (SST) warming (TPSW) pattern under global warming, as projected by 32 models from phase 5 of the Coupled Model Intercomparison Project (CMIP5). The largest cause of intermodel TPSW differences is related to the cloud–radiation feedback. After removing the effect of cloud–radiation feedback, the authors find that differences in ocean advection play the next largest role, explaining around 14% of the total intermodel variance in TPSW. Of particular importance are differences in climatological zonal overturning circulation among the models. With the robust enhancement of ocean stratification across models, models with relatively strong climatological upwelling tend to have relatively weak SST warming in the eastern Pacific. Meanwhile, the pronounced intermodel differences in ocean overturning changes contribute little to uncertainty in the TPSW pattern. The intermodel differences in climatological zonal overturning are found to be associated with the intermodel spread in climatological SST. In most CMIP5 models, there is a common cold tongue associated with an overly strong overturning in the climatology simulation, implying a La Niña–like bias in the TPSW pattern projected by the MME of the CMIP5 models. This provides further evidence for the projection that the TPSW pattern should be closer to an El Niño–like pattern than the MME projection.

Corresponding author address: Dr. Ping Huang, Center for Monsoon System Research, Institute of Atmospheric Physics, Chinese Academy of Sciences, Bei-Er-Tiao 6, Zhong-Guan-Cun, Beijing 100190, China. E-mail: huangping@mail.iap.ac.cn
Keywords: Sea surface temperature; Climate change; Numerical analysis/modeling; Climate models; Model evaluation/performance

1. Introduction

Obtaining accurate projections of the tropical Pacific sea surface temperature (SST) warming (TPSW) pattern under global warming is one of the most important issues in regional climate change research (Meehl and Washington 1996; Collins et al. 2010; Xie et al. 2010; Ying et al. 2016), but it is associated with large uncertainty (Ma and Yu 2014; Huang and Ying 2015; Xie et al. 2015; Zhou and Xie 2015). Models from phases 3 and 5 of the Coupled Model Intercomparison Project (CMIP3 and CMIP5, respectively) both show great discrepancies in simulating the TPSW pattern (DiNezio et al. 2009; Zhang and Li 2014; Huang and Ying 2015). Such intermodel spreads are a dominant source of uncertainty in the projections of regional climate change (Ma and Xie 2013; Long and Xie 2015) and can influence projections of precipitation and atmospheric circulation locally and globally (Ma et al. 2012; Huang et al. 2013; Ma and Xie 2013; Huang 2014; Chadwick 2016; Long et al. 2016).

Uncertainty in projecting the zonal SST warming pattern could arise from multifarious mechanisms that give rise to the TPSW pattern (Xie et al. 2010; Lu and Zhao 2012; Luo et al. 2015; Ying et al. 2016). Some mechanisms have been proposed to explain an El Niño–like warming pattern, such as the weakened Walker circulation (Held and Soden 2006; Vecchi and Soden 2007), the distribution of climatological evaporation cooling (Knutson and Manabe 1995; Xie et al. 2010), and the cloud–radiation feedback (Ramanathan and Collins 1991; Song and Zhang 2014), whereas the ocean dynamical thermostat has been deemed to favor a La Niña–like warming pattern (Clement et al. 1996; DiNezio et al. 2009; An and Im 2014). All of these mechanisms are potential sources of uncertainty in projecting the zonal pattern of TPSW.

Ying and Huang (2016) suggested that cloud–radiation feedback is the leading source of uncertainty in the TPSW pattern, which could be associated with the long-standing simulation bias in the parameterized cloud process and the distribution of climatological clouds in CMIP5 models. The intermodel difference in cloud–radiation feedback induces a large intermodel spread in TPSW in the central and western Pacific, and the common bias of cloud–radiation feedback in CMIP5 models could lead to a La Niña–like warming bias in the TPSW projected by the CMIP5 models. However, cloud–radiation feedback can only explain around one-quarter of the total intermodel variance in the TPSW pattern. The sources of the residual three-quarters of uncertainty are still unknown.

One of the possible sources of the residual uncertainty in the TPSW pattern could be the ocean dynamical effect because changes in ocean dynamical processes include complicated ocean–atmosphere interactions that are of great importance to the TPSW pattern formation (Ying and Huang 2016). On one hand, the ocean upper-level zonal circulation over the tropical Pacific is weakened in a warmer climate, which is projected by almost all models because of the Walker circulation slowdown (Vecchi and Soden 2007), which favors an El Niño–like warming pattern. On the other hand, the ocean vertical temperature gradient in the eastern Pacific could be enhanced under global warming, contributing to a La Niña–like warming pattern (Clement et al. 1996; Cane et al. 1997). Although both mechanisms occur in multimodel simulations, their magnitudes vary among models (DiNezio et al. 2009; Zheng et al. 2012). More importantly, the effects of these two mechanisms are opposite to each other in the eastern Pacific, where climatological upwelling prevails (Ying et al. 2016). The influence of the ocean dynamical effect on the intermodel uncertainty in the TPSW pattern is unclear.

A number of previous studies have investigated how ocean dynamics influences the biases and intermodel differences in climatological SST simulations. For example, the cold tongue bias in the eastern Pacific is thought to originate from biases in the ocean heat transport (Zheng et al. 2012), thermocline depth (Li and Xie 2012; Li et al. 2015), and Bjerknes feedback (Zheng et al. 2012; Li and Xie 2014). Recently, a study by Li et al. (2016) revealed that the intermodel differences in the cold tongue bias are significantly related to the intermodel uncertainty in the zonal gradient changes of SST in the tropical Pacific. As the biases in the climatological SST can influence the projection of the SST warming pattern (Huang and Ying 2015), the intermodel spread of the ocean dynamical effect and its impact on the cold tongue bias should be closely related to the uncertainty in the zonal TPSW pattern.

Within the process by which the intermodel spread of cloud–radiation feedback influences the intermodel uncertainty in the TPSW pattern (Ying and Huang 2016), the ocean dynamical effect, driven by atmospheric sources, also contributes to the TPSW pattern. Models that have weaker negative cloud–radiation feedback over the central Pacific tend to induce a warm local SST deviation and a low-level convergence, producing a zonal warm (cold) deviation of oceanic advection in the western (eastern) Pacific. Because a part of the ocean dynamical effect is dependent on the intermodel spread of cloud–radiation feedback, the ocean dynamical effect associated with the intermodel spread of cloud–radiation feedback is removed first in the analysis below. Then, the intermodel uncertainty in the TPSW pattern directly originating from the intermodel spread of the large-scale ocean dynamical effect is explored based on the historical and +8.5 W m−2 representative concentration pathway (RCP8.5) runs in 32 CMIP5 models. We elaborate on the impact of ocean dynamics on the TPSW pattern formation based on the surface heat budget and the decomposition of ocean heat transport. Furthermore, the physical connections between the climatological SST and the SST warming pattern derived from the intermodel spreads of the ocean dynamical effect are considered.

2. Data and methods

a. CMIP5 data

Monthly mean outputs from 32 CMIP5 models are used in this study. Table 1 lists the model names and modeling centers [see http://www-pcmdi.llnl.gov/ for more details; Taylor et al. 2012]. We compute the long-term mean for the period 1981–2000 in the historical runs to represent the present-day climatology and that for 2081–2100 in the RCP8.5 runs to represent the future climatology. The variables used in this study include SST, latent heat flux, sensible heat flux, net surface longwave radiation, net surface shortwave radiation, ocean three-dimensional potential temperature, ocean zonal and meridional current, and vertical mass transport. The net longwave and shortwave radiation are defined as the difference between upward and downward longwave and shortwave radiation, respectively. Ocean vertical velocity is obtained from the ocean vertical mass transport. Ocean vertical mass transport is not archived in the GFDL-ESM2G, GISS-E2-H, MIROC-ESM, and MIROC-ESM-CHEM models and is not well described in the CSIRO Mk3.6.0, BNU-ESM, and MIROC5 models (Ying et al. 2016), and thus ocean vertical mass transport in these models is excluded when computing the vertical velocity. The sign of the fluxes is positive for ocean warming.

Table 1.

List of the 32 CMIP5 models used in this study. (Expansions of institutions and model names are available online at http://www.ametsoc.org/PubsAcronymList.)

Table 1.

b. Definition of the TPSW pattern

The change in each model under global warming is defined as the difference between the future and the current climatology, which is normalized by the respective SST change averaged between 60°S and 60°N to remove the influence of the global mean SST warming and highlight the spatial pattern of relative SST warming. Then, the regional mean of the normalized SST change over the tropical Pacific (10°S–10°N, 120°E–80°W) is further removed to define the original TPSW pattern.

In Ying and Huang (2016), it was argued that the cloud–radiation feedback is the leading source of the intermodel uncertainty in the TPSW pattern, which can explain part of the intermodel uncertainty in various surface energy budgets, not only limited to the shortwave radiation. Therefore, the contribution of cloud–radiation feedback to the TPSW pattern and the surface energy budgets are first removed. The calculation of the cloud–radiation feedback contribution is the same as in Ying and Huang (2016). First, the intermodel singular value decomposition (SVD) is performed on the TPSW pattern and the cloud–radiation feedback index defined in Ying and Huang (2016). The first SVD mode of the TPSW pattern is considered as the mode influenced by the cloud–radiation feedback and thus removed from the total intermodel variation of the TPSW pattern. Second, the surface energy budget terms are regressed on the principal component (PC) associated with the first SVD mode of the TPSW pattern, and the variances linearly correlated to the cloud–radiation feedback are removed (Ying and Huang 2016). For simplicity, the residual TPSW pattern is referred to as the TPSW pattern hereafter.

c. Decomposition of the ocean dynamical effect

For changes in climatology, the balance of the heat budget in the ocean mixed layer can be expressed as follows (Xie et al. 2010; Huang 2015):
e1
where denotes change in the future and , , , , and represent changes in latent heat flux, sensible heat flux, net longwave radiation, net shortwave radiation, and ocean heat transport effect, respectively. The term includes changes in the ocean three-dimensional advection and a residual term, which can be decomposed as
e2
where
e3
represent changes in the ocean zonal, meridional, and vertical heat transports integrated from to the surface, respectively, and represents the residual term, including changes in the heat transports due to subgrid-scale processes, such as vertical mixing and lateral entrainment (DiNezio et al. 2009). The term must also contain errors from the calculation of the other terms when heat budgets are calculated offline. For simplicity, the mixed layer depth is chosen as a constant of 30 m, as in Ying et al. (2016). The is seawater density; is specific heat at constant pressure; and , , and in Eq. (3) represent changes in the ocean zonal, meridional, and vertical temperature advection, respectively. The term calculated by the diagnostic relationship in Eq. (1) is similar to the sum of , , and (not shown), and thus the computed by Eq. (1) is used in the following discussion.
The changes in the ocean three-dimensional temperature advection contains both the effects of changes in ocean currents and changes in ocean temperature gradients, which can be further decomposed into two components (Ying et al. 2016):
e4
where , , and are the climatological ocean zonal, meridional, and vertical current, respectively, and is ocean temperature. The terms , , and represent the temperature advection changes due to changes in ocean currents; , , and represent the temperature advection changes due to changes in ocean temperature gradients.
We further linearize the nonlinear term of temperature advection in Eq. (4) as follows:
e5
where the overbar represents the multimodel ensemble mean (MME) result and the prime denotes the deviation in individual models from the MME. In the decomposition, and represent the effects of intermodel spreads in and , respectively, and is treated as a residual term.

Apart from the large-scale ocean heat transport, the ocean dynamical processes also include several subgrid-scale processes, such as vertical mixing and lateral entrainment, which turn out to not be negligible (DiNezio et al. 2009; Ying et al. 2016). However, it is impossible to analyze these processes with the CMIP5 dataset because the variables corresponding to the processes are not archived.

3. Results

a. Relationship between the intermodel uncertainties in the ocean dynamical effect and the TPSW pattern

Figure 1a shows the intermodel standard deviation of TPSW after the effect of uncertainty in the cloud–radiation feedback is removed. The residual uncertainty in the TPSW pattern is still large compared with the MME TPSW pattern. The residual uncertainty of the TPSW pattern is mainly located in the eastern Pacific, where the strongest MME SST warming occurs (Fig. 1a, contours). The largest intermodel standard deviation exceeds 0.15°C per 1°C global mean surface warming. The MME result of the residual , which is the part of linearly independent of the first SVD mode of the original TPSW pattern (Ying and Huang 2016) (hereafter for simplicity), acts to suppress the MME SST warming in the eastern Pacific and enhance warming in the western Pacific. The term has the largest intermodel uncertainty located in the eastern Pacific, the same as the TPSW pattern (Fig. 1b). A correlation analysis, shown in Fig. 1c, indicates that the intermodel spread of TPSW is significantly (Student’s t test, 99% confidence level) correlated to that of over the eastern Pacific (2.5°S–2.5°N, 150°–90°W; green box in Figs. 1a,b).

Fig. 1.
Fig. 1.

(a) The MME SST warming pattern (contours) and intermodel standard deviations (shading) of SST change in the 32 models. (b) The MME (contour) and intermodel standard deviations (shading) of ocean heat transport . In (a),(b), the impact of cloud–radiation feedback has been removed. (c) Linear relationship between SST changes and in the equatorial eastern Pacific [2.5°S–2.5°N, 150°–90°W, denoted by the green box in (a),(b)].

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

An intermodel empirical orthogonal function (EOF) analysis is performed on the multimodel . The first intermodel EOF mode (EOF1) explains 25.8% of the total variance of and exhibits a pronounced pattern of negative values in the eastern Pacific cold tongue region and weakly positive values in the western Pacific (Fig. 2a). Figure 2b shows the regression pattern of the TPSW patterns on the normalized PC associated with the EOF1 of (PC1). The regression pattern of TPSW shows a west–east dipole pattern. In models where there is a relatively large cooling effect from , the projected SST warming in the eastern Pacific tends to be relatively weak. Although the EOF1 pattern of with a zonal dipole pattern is roughly similar to the regression pattern of onto the intermodel spread of the cloud–radiation feedback index in Fig. 3c of Ying and Huang (2016), these two patterns are linearly independent because the two patterns are in different positions. This result implies the role of the first mode of the residual is independent of the role of cloud–radiation feedback in Ying and Huang (2016).

Fig. 2.
Fig. 2.

(a) The EOF1 mode of changes in the ocean heat transport . The explained variance is shown in the top-right corner. (b) The regression pattern of SST changes onto the PC1 of (shading) and the percentages of explained local variances (contours; unit: %). The total percentage of the variance of SST changes is shown in the top-right corner. Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

The PC1 of explains 18.6% of the intermodel variance of the residual TPSW with the effect of cloud–radiation feedback removed, which is around 14% (18.6% of 76%) of the total intermodel variance of the original TPSW pattern. The local variance explained by PC1 is up to 20% in most areas of the eastern Pacific (Fig. 2b), with the maximum exceeding 40% in the equatorial region, where the most pronounced intermodel variability of occurs (Fig. 2a). The high explained variance indicates that is an important source of uncertainty in the eastern Pacific SST warming, though smaller than the variance explained by the cloud–radiation feedback, which is around 24% of the total variance of the original TPSW pattern (Ying and Huang 2016). To simplify the presentation, the results in the following figures follow the pair of patterns in Fig. 2.

b. Mechanism of impact of the ocean dynamical effect

The changes in the surface heat budget associated with the EOF1 of are analyzed to investigate the process by which influences the uncertainty in the TPSW pattern. Figures 3a,b show the regression patterns of and on the PC1 of , respectively ( and are omitted owing to relatively small values). Interestingly, both the and regressions are positive in the east. This indicates that models with less SST warming in the east (associated with ocean dynamical effects) actually tend to receive larger surface heat fluxes from evaporation and shortwave radiation, compared to models with enhanced SST warming. This result indicates that and do not contribute to the relatively weak SST warming in the eastern Pacific. The positive in the eastern Pacific is a response to the small SST warming due to the evaporation–latent heat–SST negative feedback, and the positive is also a response to the SST due to the cloud–shortwave–SST negative feedback (Song and Zhang 2014; Ying et al. 2016). Therefore, the ocean dynamical process associated with should be the direct process influencing the uncertainty in the eastern Pacific SST warming pattern.

Fig. 3.
Fig. 3.

Regression patterns of changes in surface energy budgets in the ocean mixed layer onto the PC1 of : (a) latent heat flux, (b) shortwave radiation, (c) and the residual term of the heat transport equation. Stippling indicates that regressions are significant at the 95% confidence level, based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

To understand the ocean dynamical effect associated with the zonal dipole pattern of SST warming, changes in ocean temperature to a depth of 200 m at the equator (mean of 2.5°S–2.5°N) are regressed onto the PC1 of (Fig. 4). The regression pattern of changes in ocean temperature also shows a dipole distribution near the surface similar to the regression pattern of SST warming (Fig. 2b), though the significant region in the east is smaller than that in the SST warming, which could be due to differences between model-produced SST and ocean potential temperature. The significance test on the regression pattern of ocean temperature (stippling in Fig. 4) shows that a more pronounced signal related to the ocean dynamical effect is located in the subsurface ocean and that the TPSW pattern is just a superficial reflection of ocean dynamical processes.

Fig. 4.
Fig. 4.

The regression pattern of changes in equatorial (mean of 2.5°S–2.5°N) ocean temperature onto the PC1 of . Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

The changes in the terms associated with ocean temperature advection in Eq. (3) are regressed onto the PC1 of , and the results at the equator are shown in Fig. 5. The meridional component of changes in ocean temperature advection is omitted owing to relatively small values (not shown). The most pronounced anomalies are located in the subsurface ocean, where the climatological equatorial undercurrent is situated (Figs. 5a,b, vectors). The regression patterns of and largely oppose each other, indicating that models with stronger -related cooling in the eastern Pacific tend to have enhanced subsurface warming due to zonal advection (Fig. 5a) and the subsurface cooling due to vertical advection (Fig. 5b) in the equatorial undercurrent region. The significant region of the regression pattern of is confined to the subsurface, whereas the significant region of the regression pattern of outcrops on the surface of the eastern Pacific. The sum of the regression patterns of and shows a negative sign along the core of the climatological equatorial undercurrent and in the eastern Pacific where climatological upwelling prevails, the same as for (Fig. 5c), illustrating the dominant role of in . In the western Pacific, neither the surface (Fig. 2b) nor the subsurface (Fig. 4) intermodel temperature differences seem to be related to ocean dynamical differences. Rather, they could be related to the ocean residual term (Fig. 3c) including the subgrid-scale processes, the calculation errors of the heat budget, and so on.

Fig. 5.
Fig. 5.

Regression patterns of changes in equatorial (a) zonal and (b) vertical temperature advection onto the PC1 of and (c) the sum of (a) and (b). Vectors in (a) and (b) are the climatological ocean zonal overturning circulation in the MME. The vertical velocity is multiplied by 105 for display purposes, and vectors less than 0.1 are omitted. Stippling indicates correlations are significant at the 90% confidence level based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

The changes in zonal and vertical temperature advection are divided into the contribution of changes in ocean current and changes in ocean temperature gradient, based on the decomposition in Eq. (4) (Fig. 6). The is mainly contributed by the changes in ocean current (Fig. 6a), whereas is caused by the changes in both ocean current and ocean temperature gradient (Figs. 6c,d). The regression pattern of , which outcrops in the eastern Pacific (Fig. 5b), is caused by (Fig. 6d). On one hand, the significant regions in the regression patterns of and in the equatorial undercurrent region largely offset each other (Figs. 6a,c). As a result, the sum of and is very small and not significant (Fig. 6e), indicating that the temperature advection changes due to changes in ocean equatorial overturning do not contribute to the intermodel uncertainty in the ocean dynamical effect. On the other hand, compared with , contributes the major part of the effect of changes in ocean temperature (Figs. 6b,d,f). This result indicates that the intermodel differences in the temperature advection changes due to the changes in ocean temperature gradient are the dominant factor contributing to the intermodel differences in the ocean dynamical effect and to the uncertainty in SST warming in the eastern Pacific.

Fig. 6.
Fig. 6.

Regression patterns of the terms in Eq. (4) onto the PC1 of . Changes in zonal temperature advection induced by (a) changes in zonal current and (b) changes in zonal temperature gradient . Changes in vertical temperature advection induced by (c) changes in vertical current and (d) changes in vertical temperature gradient . (e) The sum of (a) and (c). (f) The sum of (b) and (d). Stippling indicates regressions are significant at the 90% confidence level based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

The intermodel differences in , , , and can be generated from the intermodel differences in changes in ocean current, changes in ocean temperature, climatological ocean current, and climatological ocean temperature. Therefore, the four terms are further decomposed following Eq. (5) (Fig. 7). The temperature gradients calculated for Fig. 7 are shown in Fig. 8. The residual terms of the decomposition of , , and are negligible (Figs. 7c,f,i), whereas the residual of is relatively large (Fig. 7l), probably owing to strong large-scale nonlinear ocean processes in the eastern Pacific subsurface.

Fig. 7.
Fig. 7.

Regression patterns of the linearized terms of changes in equatorial ocean temperature advection onto the PC1 of : (a) , (b) , (c) , (d) , (e) , (f) , (g) , (h) , (i) , (j) , (k) , and (l) . Contours in (a),(g) are the MME climatological ocean temperature, contours in (j) are the MME ocean temperature changes, and contours in (k) are the regression pattern of changes in ocean temperature. Vectors in (a),(g) are the regression patterns of changes in ocean zonal overturning circulation, vectors in (j) are the regression pattern of climatological ocean zonal overturning circulation, and vectors in (k) are the climatological ocean zonal overturning circulation. Vertical velocity is multiplied by 105 for display purposes. Stippling indicates that regressions are significant at the 90% confidence level based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

Fig. 8.
Fig. 8.

The gradient of ocean temperature used in calculating Fig. 7. The MME of (a) , (c) , (e) , and (g) and the regression patterns of (b) , (d) , (f) , and (h) onto PC1 of . Stippling indicates that regressions are significant at the 90% confidence level based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

Figures 7a–c,g–i reveal that the intermodel differences in changes in ocean current (Figs. 7a,g) are the main source of the intermodel differences both in and . The two terms and also approximately cancel each other out, as with the two terms and in Fig. 6e. This result is consistent with the picture of intermodel differences in ocean current changes and the MME climatological ocean temperature in Figs. 7a,g. The intermodel differences in ocean current changes (vectors in Figs. 7a,g) are positioned almost along the isotherms of the MME climatological ocean temperature. In this situation, the intermodel differences in changes in ocean current cannot induce significant intermodel differences in changes in ocean temperature advection. This characteristic of changes in ocean current could be attributed to the fact that changes in ocean current are mainly associated with changes in the Walker circulation. The coupled changes in zonal ocean and atmospheric circulation imply that the directions of changes in zonal ocean current are almost the same as those for climatological ocean current, which are positioned along the isotherms of climatological ocean temperature. The intermodel differences in climatological ocean temperature are relatively small (Figs. 7b,h). Therefore, the changes in ocean current mainly exhibit pronounced intermodel differences in magnitude but not in direction. Consequently, the intermodel differences in changes in ocean current, which basically follow the MME climatological ocean isotherms, do not induce pronounced intermodel differences in changes in ocean temperature advection and thus contribute little to the uncertainty in the TPSW pattern.

For the major contributing term , the decomposition (Figs. 7j–l) shows that the intermodel differences in the climatological ocean vertical current and changes in the vertical gradient of ocean temperature both contribute to the intermodel differences in . Under global warming, the vertical gradient of ocean temperature will be enhanced as a result of increasing CO2 inducing an increase in longwave radiation, which is largely absorbed near the ocean surface (Pierce et al. 2006; DiNezio et al. 2009; Capotondi et al. 2012; Long et al. 2014; Huang et al. 2015). Against such changes in ocean stratification, the isotherms of changes in ocean temperature are almost parallel to the ocean surface, which is robust among the models and occurs in the MME changes in ocean temperature (contours in Figs. 7j and 8g). Meanwhile, the intermodel differences in climatological overturning (vectors in Fig. 7j) prominently cross the isotherms of the MME changes in the subsurface, inducing a large contribution to the intermodel differences in changes in ocean temperature advection (shaded in Fig. 7j). Therefore, models that have a relatively strong dynamical cooling of in the eastern Pacific (Fig. 2a) tend to be associated with a negative deviation of changes in ocean temperature advection, induced by a relatively strong climatological overturning (vectors in Fig. 7j). The change associated with suggests that models that have a relatively strong climatological overturning tend to be associated with a relatively strong ocean dynamical thermostat, which will act to suppress projected warming in the eastern Pacific (Clement et al. 1996; Cane et al. 1997). The largest values are located at depths of around 50–100 m, where the largest vertical temperature gradient changes and vertical ocean current occurs. Additionally, the intermodel differences in climatological horizontal currents contribute little to the intermodel differences in changes in ocean temperature advection under the approximately flat isotherms of the MME changes in ocean temperature (Fig. 7d).

The intermodel differences in vertical temperature gradient changes also contribute much to the intermodel differences in . With the transportation of the MME upwelling of the climatological vertical current, shows a vertical dipole pattern in the eastern Pacific and a tripole pattern in the central Pacific. The pattern of associated with in Fig. 2a is induced by a relatively strong (weak) enhancement of ocean stratification in the subsurface eastern (central) Pacific, which can be observed in Figs. 4 and 8h. This result indicates that there is an interaction between the ocean temperature changes (Figs. 4 and 8h) and the changes in ocean temperature advection transported by the climatological current (Fig. 7k). We can also understand this process insofar as the MME climatological current transports the changes in ocean temperature originating from (Fig. 7j) to the near surface of the eastern Pacific (Fig. 7k). Under the redistribution of heat by the climatological current, the total changes in ocean temperature advection near the surface of the eastern Pacific are comparable to the changes in the subsurface of the central-eastern Pacific (Figs. 6d,f).

There is also a complex nonlinear interaction between the intermodel differences in climatological current transport and the changes in ocean temperature (Fig. 7l). However, this effect is mainly located in the subsurface ocean and does not contribute much to the pattern of surface warming.

c. Relationship between the intermodel spread of climatological SST and the uncertainty in the TPSW pattern

Energy budget analysis has mainly attributed the intermodel differences in to the intermodel differences in climatological ocean upwelling. The concept of observational constraint and the statistical relationship between the intermodel differences in climatological SST simulation and future SST changes motivate us to further investigate the ocean dynamical process connecting the climatological SST and the SST change (Whetton et al. 2007; Bracegirdle and Stephenson 2012, 2013; Huang and Ying 2015; Li et al. 2016).

Figure 9 shows the zonal and vertical components of climatological ocean current among the models regressed onto the PC1 of , which is equivalent to the vectors in Figs. 7j and 10b. The regressed patterns clearly reflect that a relatively large cooling of in the eastern Pacific (Fig. 2a) is related to a relatively strong climatological overturning. The regression patterns of climatological SST and ocean temperature among the models are shown in Fig. 10. It can be easily understood that models with stronger upwelling of cool water in the eastern Pacific will tend to have cooler SST (Fig. 10a) and therefore a more pronounced cold tongue bias. Meanwhile, the stronger climatological zonal overturning circulation is also accompanied by a steeper ocean thermocline in the eastern Pacific (red curve in Fig. 10b). The attribution of the intermodel spread in the climatological cold tongue SST to the climatological zonal overturning circulation and the thermocline is consistent with the result in Li and Xie (2012).

Fig. 9.
Fig. 9.

Regression patterns of equatorial climatological (a) zonal and (b) vertical currents onto the PC1 of . Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

Fig. 10.
Fig. 10.

Regression patterns of (a) climatological SST and (b) equatorial climatological ocean temperature onto the PC1 of . The red and black curves in (b) denote the composited 20°C isotherm in the models. The red curve represents the models in which the PC1 of is more than 1, and the black curve those in which the PC1 of is less than −1. Vectors in (b) are the regression pattern of climatological ocean zonal overturning circulation, in which the vertical velocity is multiplied by 105 for display purposes and vectors less than 0.02 are omitted. Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.

Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-16-0318.1

These analyses reveal the role of climatological zonal overturning circulation in connecting the climatological SST and SST change, demonstrating the significant relationship between the intermodel spread of simulated climatological SST and that of projected SST warming pattern in previous statistical studies (Huang and Ying 2015; Li et al. 2016). A relatively strong climatological zonal overturning circulation in a model not only leads to a relatively strong cold tongue but also contributes to a weak SST warming in the eastern Pacific. Because most models suffer from excessive cold tongue biases and overly strong zonal overturning circulations in the equatorial Pacific (Li and Xie 2012; Zheng et al. 2012; Li and Xie 2014; Li et al. 2015), the projected TPSW pattern in these models likely includes a common La Niña–like bias, as in Fig. 2b. The systematic excessive cold tongue biases and the underestimated cloud–radiation feedback revealed in Ying and Huang (2016) both lead to a La Niña–like bias in the MME projection. As a result, we suggest that the pattern of SST change in the tropical Pacific should likely be closer to an El Niño–like pattern.

4. Conclusions and discussion

The large-scale ocean dynamical effect as a direct source of uncertainty in the tropical Pacific SST warming pattern projected by 32 CMIP5 models is revealed in the present study. The results show that the first intermodel EOF mode of changes in the large-scale ocean heat transport explains 18.6% of the residual variance of the TPSW pattern when the contribution of cloud–radiation feedback is removed. The explained variance by the large-scale ocean heat transport—around 14% of the total variance of the original TPSW pattern before removal of the contribution of cloud–radiation feedback—indicates that the large-scale ocean dynamical effect is another important source of uncertainty in the TPSW pattern, whereas the cloud–radiation feedback as the leading source of uncertainty explains around 24% of the total variance (Ying and Huang 2016).

The mechanism by which the ocean dynamics influences the uncertainty in the TPSW pattern is investigated by analyzing the surface energy budget and the decomposed ocean temperature advections. The surface latent heat changes and shortwave radiation changes play a suppressive role in the uncertainty in the TPSW pattern induced by the ocean dynamical effect. The former is a result of the negative evaporation–latent heat–SST feedback, while the latter is a result of the negative cloud–shortwave–SST feedback.

The influence of the ocean dynamical effect on the uncertainty in the TPSW pattern is mainly due to the intermodel differences in the simulated climatological zonal overturning. Under global warming, ocean surface warming is larger than warming in the subsurface, which induces a robust change in the vertical gradient in ocean temperature. In a model with a relatively strong climatological overturning, the combination of the strong mean state overturning and the enhanced vertical stratification leads to a relatively strong ocean dynamical thermostat effect, which suppresses the eastern Pacific warming (Clement et al. 1996; Cane et al. 1997). This negative effect peaks in the subsurface ocean where the vertical overturning and the enhancement of ocean stratification are largest. Additionally, the MME oceanic overturning current can transport the negative changes in the ocean dynamical effect from the subsurface ocean into the near-surface layer in the eastern Pacific. As a result, one model with a relatively strong climatological zonal overturning tends to induce a relatively weak SST increase in the eastern Pacific.

Another pronounced intermodel difference in the ocean dynamical process is the different changes in the zonal overturning circulation, which is coupled with the changes in atmospheric circulation. Under global warming, the zonal overturning circulation will likely be weakened and is associated with a weakened Walker circulation (Held and Soden 2006; Vecchi and Soden 2007). The wind-driven change in the zonal overturning circulation acts to change the strength of the overturning, but not its direction. Therefore, the intermodel differences in changes in zonal overturning circulation are positioned approximately along the isotherms of the MME climatological ocean temperature and thus do not induce pronounced intermodel differences in changes in ocean heat transport. As a result, the changes in overturning circulation do not contribute much to the uncertainty in the TPSW pattern.

The present study concludes that the main mechanism generating intermodel TPSW differences associated with the large-scale ocean dynamical effect relates to intermodel differences in the climatological zonal overturning circulation. Moreover, the intermodel differences in climatological overturning circulation are related to the intermodel differences of climatological SST. Models that have a relatively strong climatological overturning tend to be associated with a relatively strong cold tongue. This mechanism demonstrates the statistical connection between the intermodel differences in climatological SST simulation and the uncertainty in the TPSW pattern projection revealed in some previous studies (Huang and Ying 2015). As most models suffer from an excessive cold tongue bias of climatological SST, as well as overly strong zonal overturning circulation in the equatorial Pacific, it is reasonable to suppose that a La Niña–like warming bias exists in the projections of the TPSW pattern in most CMIP5 models and in the MME (Huang and Ying 2015), based on the concept of observational constraint. This result increases our confidence that the tropical Pacific SST changes under global warming should be closer to an El Niño–like pattern than the MME projection in CMIP5 models.

In the present study, only the first intermodel mode of the ocean dynamical effect is discussed, and the explained intermodel variance for the TPSW pattern is limited. Even if we combine the effect of the current EOF1 of large-scale ocean dynamics with that of cloud–radiation feedback as a leading source of the uncertainty in the TPSW pattern (Ying and Huang 2016), the total explained intermodel variance of the TPSW pattern is less than 40%, indicating that there are many other mechanisms impacting upon the intermodel spread of the TPSW pattern—for example, the convective cloud–SST negative feedback in the warm pool region (Ramanathan and Collins 1991), the stratus cloud–SST positive feedback in the cold tongue region (Meehl and Washington 1996), and the vertical resolution of the ocean model (Stockdale et al. 1998). In addition, some marine biochemical processes, such as the activity of phytoplankton in the eastern Pacific, can also impact the uncertainty in the TPSW pattern (Murtugudde et al. 2002). These processes may explain the regional spread of the TPSW pattern (Huang and Ying 2015) and are worthy of further attention in the future.

Acknowledgments

The work was supported by the National Basic Research Program of China (2014CB953904 and 2012CB955604), the National Natural Science Foundation of China (Grants 41575088 and 41461164005), and the Youth Innovation Promotion Association of CAS. The World Climate Research Programme’s Working Group on Coupled Modelling, which is responsible for CMIP5, and the climate modeling groups (listed in Table 1) are acknowledged for producing and making available their model output.

REFERENCES

  • An, S.-I., and S.-H. Im, 2014: Blunt ocean dynamical thermostat in response of tropical eastern Pacific SST to global warming. Theor. Appl. Climatol., 118, 173–183, doi:10.1007/s00704-013-1048-0.

    • Search Google Scholar
    • Export Citation

  • Bracegirdle, T. J., and D. B. Stephenson, 2012: Higher precision estimates of regional polar warming by ensemble regression of climate model projections. Climate Dyn., 39, 2805–2821, doi:10.1007/s00382-012-1330-3.

    • Search Google Scholar
    • Export Citation

  • Bracegirdle, T. J., and D. B. Stephenson, 2013: On the robustness of emergent constraints used in multimodel climate change projections of arctic warming. J. Climate, 26, 669–678, doi:10.1175/JCLI-D-12-00537.1.

    • Search Google Scholar
    • Export Citation

  • Cane, M. A., and Coauthors, 1997: Twentieth-century sea surface temperature trends. Science, 275, 957–960, doi:10.1126/science.275.5302.957.

    • Search Google Scholar
    • Export Citation

  • Capotondi, A., M. A. Alexander, N. A. Bond, E. N. Curchitser, and J. D. Scott, 2012: Enhanced upper ocean stratification with climate change in the CMIP3 models. J. Geophys. Res., 117, C04031, doi:10.1029/2011JC007409.

    • Search Google Scholar
    • Export Citation

  • Chadwick, R., 2016: Which aspects of CO2 forcing and SST warming cause most uncertainty in projections of tropical rainfall change over land and ocean? J. Climate, 29, 2493–2509, doi:10.1175/JCLI-D-15-0777.1.

    • Search Google Scholar
    • Export Citation

  • Clement, A. C., R. Seager, M. A. Cane, and S. E. Zebiak, 1996: An ocean dynamical thermostat. J. Climate, 9, 2190–2196, doi:10.1175/1520-0442(1996)009<2190:AODT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Collins, M., and Coauthors, 2010: The impact of global warming on the tropical Pacific Ocean and El Niño. Nat. Geosci., 3, 391–397, doi:10.1038/ngeo868.

    • Search Google Scholar
    • Export Citation

  • DiNezio, P. N., A. C. Clement, G. A. Vecchi, B. J. Soden, B. P. Kirtman, and S.-K. Lee, 2009: Climate response of the equatorial Pacific to global warming. J. Climate, 22, 4873–4892, doi:10.1175/2009JCLI2982.1.

    • Search Google Scholar
    • Export Citation

  • Held, I. M., and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. J. Climate, 19, 5686–5699, doi:10.1175/JCLI3990.1.

    • Search Google Scholar
    • Export Citation

  • Huang, P., 2014: Regional response of annual-mean tropical rainfall to global warming. Atmos. Sci. Lett., 15, 103–109, doi:10.1002/asl2.475.

    • Search Google Scholar
    • Export Citation

  • Huang, P., 2015: Seasonal changes in tropical SST and the surface energy budget under global warming projected by CMIP5 models. J. Climate, 28, 6503–6515, doi:10.1175/JCLI-D-15-0055.1.

    • Search Google Scholar
    • Export Citation

  • Huang, P., and J. Ying, 2015: A multimodel ensemble pattern regression method to correct the tropical Pacific SST change patterns under global warming. J. Climate, 28, 4706–4723, doi:10.1175/JCLI-D-14-00833.1.

    • Search Google Scholar
    • Export Citation

  • Huang, P., S.-P. Xie, K. Hu, G. Huang, and R. Huang, 2013: Patterns of the seasonal response of tropical rainfall to global warming. Nat. Geosci., 6, 357–361, doi:10.1038/ngeo1792.

    • Search Google Scholar
    • Export Citation

  • Huang, P., I.-I. Lin, C. Chou, and R. Huang, 2015: Change in ocean subsurface environment to suppress tropical cyclone intensification under global warming. Nat. Commun., 6, 7188, doi:10.1038/ncomms8188.

    • Search Google Scholar
    • Export Citation

  • Knutson, T. R., and S. Manabe, 1995: Time-mean response over the tropical Pacific to increased CO2 in a coupled ocean–atmosphere model. J. Climate, 8, 2181–2199, doi:10.1175/1520-0442(1995)008<2181:TMROTT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Li, G., and S.-P. Xie, 2012: Origins of tropical-wide SST biases in CMIP multi-model ensembles. Geophys. Res. Lett., 39, L22703, doi:10.1029/2012GL053777.

    • Search Google Scholar
    • Export Citation

  • Li, G., and S.-P. Xie, 2014: Tropical biases in CMIP5 multimodel ensemble: The excessive equatorial Pacific cold tongue and double ITCZ problems. J. Climate, 27, 1765–1780, doi:10.1175/JCLI-D-13-00337.1.

    • Search Google Scholar
    • Export Citation

  • Li, G., Y. Du, H. Xu, and B. Ren, 2015: An intermodel approach to identify the source of excessive equatorial Pacific cold tongue in CMIP5 models and uncertainty in observational datasets. J. Climate, 28, 7630–7640, doi:10.1175/JCLI-D-15-0168.1.

    • Search Google Scholar
    • Export Citation

  • Li, G., S.-P. Xie, Y. Du, and Y. Luo, 2016: Effects of excessive equatorial cold tongue bias on the projections of tropical Pacific climate change. Part I: The warming pattern in CMIP5 multi-model ensemble. Climate Dyn., doi:10.1080/13697137.2016.1194388, in press.

    • Search Google Scholar
    • Export Citation

  • Long, S.-M., and S.-P. Xie, 2015: Intermodel variations in projected precipitation change over the North Atlantic: Sea surface temperature effect. Geophys. Res. Lett., 42, 4158–4165, doi:10.1002/2015GL063852.

    • Search Google Scholar
    • Export Citation

  • Long, S.-M., S.-P. Xie, X.-T. Zheng, and Q. Liu, 2014: Fast and slow responses to global warming: Sea surface temperature and precipitation patterns. J. Climate, 27, 285–299, doi:10.1175/JCLI-D-13-00297.1.

    • Search Google Scholar
    • Export Citation

  • Long, S.-M., S.-P. Xie, and W. Liu, 2016: Uncertainty in tropical rainfall projections: Atmospheric circulation effect and the ocean coupling. J. Climate, 29, 2671–2687, doi:10.1175/JCLI-D-15-0601.1.

    • Search Google Scholar
    • Export Citation

  • Lu, J., and B. Zhao, 2012: The role of oceanic feedback in the climate response to doubling CO2. J. Climate, 25, 7544–7563, doi:10.1175/JCLI-D-11-00712.1.

    • Search Google Scholar
    • Export Citation

  • Luo, Y., J. Lu, F. Liu, and W. Liu, 2015: Understanding the El Niño-like oceanic response in the tropical Pacific to global warming. Climate Dyn., 45, 1945–1964, doi:10.1007/s00382-014-2448-2.

    • Search Google Scholar
    • Export Citation

  • Ma, J., and S.-P. Xie, 2013: Regional patterns of sea surface temperature change: A source of uncertainty in future projections of precipitation and atmospheric circulation. J. Climate, 26, 2482–2501, doi:10.1175/JCLI-D-12-00283.1.

    • Search Google Scholar
    • Export Citation

  • Ma, J., and J.-Y. Yu, 2014: Linking centennial surface warming patterns in the equatorial Pacific to the relative strengths of the Walker and Hadley circulations. J. Atmos. Sci., 71, 3454–3464, doi:10.1175/JAS-D-14-0028.1.

    • Search Google Scholar
    • Export Citation

  • Ma, J., S.-P. Xie, and Y. Kosaka, 2012: Mechanisms for tropical tropospheric circulation change in response to global warming. J. Climate, 25, 2979–2994, doi:10.1175/JCLI-D-11-00048.1.

    • Search Google Scholar
    • Export Citation

  • Meehl, G. A., and W. M. Washington, 1996: El Niño-like climate change in a model with increased atmospheric CO2 concentrations. Nature, 382, 56–60, doi:10.1038/382056a0.

    • Search Google Scholar
    • Export Citation

  • Murtugudde, R., J. Beauchamp, C. R. McClain, M. Lewis, and A. J. Busalacchi, 2002: Effects of penetrative radiation on the upper tropical ocean circulation. J. Climate, 15, 470–486, doi:10.1175/1520-0442(2002)015<0470:EOPROT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Pierce, D. W., T. P. Barnett, K. M. AchutaRao, P. J. Gleckler, J. M. Gregory, and W. M. Washington, 2006: Anthropogenic warming of the oceans: Observations and model results. J. Climate, 19, 1873–1900, doi:10.1175/JCLI3723.1.

    • Search Google Scholar
    • Export Citation

  • Ramanathan, V., and W. Collins, 1991: Thermodynamic regulation of ocean warming by cirrus clouds deduced from observations of the 1987 El Niño. Nature, 351, 27–32, doi:10.1038/351027a0.

    • Search Google Scholar
    • Export Citation

  • Song, X., and G. J. Zhang, 2014: Role of climate feedback in El Niño–like SST response to global warming. J. Climate, 27, 7301–7318, doi:10.1175/JCLI-D-14-00072.1.

    • Search Google Scholar
    • Export Citation

  • Stockdale, T. N., A. J. Busalacchi, D. E. Harrison, and R. Seager, 1998: Ocean modeling for ENSO. J. Geophys. Res., 103, 14 325–14 355, doi:10.1029/97JC02440.

    • Search Google Scholar
    • Export Citation

  • Taylor, K. E., R. J. Stouffer, and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485–498, doi:10.1175/BAMS-D-11-00094.1.

    • Search Google Scholar
    • Export Citation

  • Vecchi, G. A., and B. J. Soden, 2007: Global warming and the weakening of the tropical circulation. J. Climate, 20, 4316–4340, doi:10.1175/JCLI4258.1.

    • Search Google Scholar
    • Export Citation

  • Whetton, P., I. Macadam, J. Bathols, and J. O’Grady, 2007: Assessment of the use of current climate patterns to evaluate regional enhanced greenhouse response patterns of climate models. Geophys. Res. Lett., 34, L14701, doi:10.1029/2007GL030025.

    • Search Google Scholar
    • Export Citation

  • Xie, S.-P., C. Deser, G. A. Vecchi, J. Ma, H. Teng, and A. T. Wittenberg, 2010: Global warming pattern formation: Sea surface temperature and rainfall. J. Climate, 23, 966–986, doi:10.1175/2009JCLI3329.1.

    • Search Google Scholar
    • Export Citation

  • Xie, S.-P., and Coauthors, 2015: Towards predictive understanding of regional climate change. Nat. Climate Change, 5, 921–930, doi:10.1038/nclimate2689.

    • Search Google Scholar
    • Export Citation

  • Ying, J., and P. Huang, 2016: Cloud–radiation feedback as a leading source of uncertainty in the tropical Pacific SST warming pattern in CMIP5 models. J. Climate, 29, 3867–3881, doi:10.1175/JCLI-D-15-0796.1.

    • Search Google Scholar
    • Export Citation

  • Ying, J., P. Huang, and R. Huang, 2016: Evaluating the formation mechanisms of the equatorial Pacific SST warming pattern in CMIP5 models. Adv. Atmos. Sci., 33, 433–441, doi:10.1007/s00376-015-5184-6.

    • Search Google Scholar
    • Export Citation

  • Zhang, L., and T. Li, 2014: A simple analytical model for understanding the formation of sea surface temperature patterns under global warming. J. Climate, 27, 8413–8421, doi:10.1175/JCLI-D-14-00346.1.

    • Search Google Scholar
    • Export Citation

  • Zheng, Y., J.-L. Lin, and T. Shinoda, 2012: The equatorial Pacific cold tongue simulated by IPCC AR4 coupled GCMs: Upper ocean heat budget and feedback analysis. J. Geophys. Res., 117, C05024, doi:10.1029/2011JC007746.

    • Search Google Scholar
    • Export Citation

  • Zhou, Z.-Q., and S.-P. Xie, 2015: Effects of climatological model biases on the projection of tropical climate change. J. Climate, 28, 9909–9917, doi:10.1175/JCLI-D-15-0243.1.

    • Search Google Scholar
    • Export Citation
Save
Cover Journal of Climate
Journal of Climate

  • An, S.-I., and S.-H. Im, 2014: Blunt ocean dynamical thermostat in response of tropical eastern Pacific SST to global warming. Theor. Appl. Climatol., 118, 173–183, doi:10.1007/s00704-013-1048-0.

    • Search Google Scholar
    • Export Citation

  • Bracegirdle, T. J., and D. B. Stephenson, 2012: Higher precision estimates of regional polar warming by ensemble regression of climate model projections. Climate Dyn., 39, 2805–2821, doi:10.1007/s00382-012-1330-3.

    • Search Google Scholar
    • Export Citation

  • Bracegirdle, T. J., and D. B. Stephenson, 2013: On the robustness of emergent constraints used in multimodel climate change projections of arctic warming. J. Climate, 26, 669–678, doi:10.1175/JCLI-D-12-00537.1.

    • Search Google Scholar
    • Export Citation

  • Cane, M. A., and Coauthors, 1997: Twentieth-century sea surface temperature trends. Science, 275, 957–960, doi:10.1126/science.275.5302.957.

    • Search Google Scholar
    • Export Citation

  • Capotondi, A., M. A. Alexander, N. A. Bond, E. N. Curchitser, and J. D. Scott, 2012: Enhanced upper ocean stratification with climate change in the CMIP3 models. J. Geophys. Res., 117, C04031, doi:10.1029/2011JC007409.

    • Search Google Scholar
    • Export Citation

  • Chadwick, R., 2016: Which aspects of CO2 forcing and SST warming cause most uncertainty in projections of tropical rainfall change over land and ocean? J. Climate, 29, 2493–2509, doi:10.1175/JCLI-D-15-0777.1.

    • Search Google Scholar
    • Export Citation

  • Clement, A. C., R. Seager, M. A. Cane, and S. E. Zebiak, 1996: An ocean dynamical thermostat. J. Climate, 9, 2190–2196, doi:10.1175/1520-0442(1996)009<2190:AODT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Collins, M., and Coauthors, 2010: The impact of global warming on the tropical Pacific Ocean and El Niño. Nat. Geosci., 3, 391–397, doi:10.1038/ngeo868.

    • Search Google Scholar
    • Export Citation

  • DiNezio, P. N., A. C. Clement, G. A. Vecchi, B. J. Soden, B. P. Kirtman, and S.-K. Lee, 2009: Climate response of the equatorial Pacific to global warming. J. Climate, 22, 4873–4892, doi:10.1175/2009JCLI2982.1.

    • Search Google Scholar
    • Export Citation

  • Held, I. M., and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. J. Climate, 19, 5686–5699, doi:10.1175/JCLI3990.1.

    • Search Google Scholar
    • Export Citation

  • Huang, P., 2014: Regional response of annual-mean tropical rainfall to global warming. Atmos. Sci. Lett., 15, 103–109, doi:10.1002/asl2.475.

    • Search Google Scholar
    • Export Citation

  • Huang, P., 2015: Seasonal changes in tropical SST and the surface energy budget under global warming projected by CMIP5 models. J. Climate, 28, 6503–6515, doi:10.1175/JCLI-D-15-0055.1.

    • Search Google Scholar
    • Export Citation

  • Huang, P., and J. Ying, 2015: A multimodel ensemble pattern regression method to correct the tropical Pacific SST change patterns under global warming. J. Climate, 28, 4706–4723, doi:10.1175/JCLI-D-14-00833.1.

    • Search Google Scholar
    • Export Citation

  • Huang, P., S.-P. Xie, K. Hu, G. Huang, and R. Huang, 2013: Patterns of the seasonal response of tropical rainfall to global warming. Nat. Geosci., 6, 357–361, doi:10.1038/ngeo1792.

    • Search Google Scholar
    • Export Citation

  • Huang, P., I.-I. Lin, C. Chou, and R. Huang, 2015: Change in ocean subsurface environment to suppress tropical cyclone intensification under global warming. Nat. Commun., 6, 7188, doi:10.1038/ncomms8188.

    • Search Google Scholar
    • Export Citation

  • Knutson, T. R., and S. Manabe, 1995: Time-mean response over the tropical Pacific to increased CO2 in a coupled ocean–atmosphere model. J. Climate, 8, 2181–2199, doi:10.1175/1520-0442(1995)008<2181:TMROTT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Li, G., and S.-P. Xie, 2012: Origins of tropical-wide SST biases in CMIP multi-model ensembles. Geophys. Res. Lett., 39, L22703, doi:10.1029/2012GL053777.

    • Search Google Scholar
    • Export Citation

  • Li, G., and S.-P. Xie, 2014: Tropical biases in CMIP5 multimodel ensemble: The excessive equatorial Pacific cold tongue and double ITCZ problems. J. Climate, 27, 1765–1780, doi:10.1175/JCLI-D-13-00337.1.

    • Search Google Scholar
    • Export Citation

  • Li, G., Y. Du, H. Xu, and B. Ren, 2015: An intermodel approach to identify the source of excessive equatorial Pacific cold tongue in CMIP5 models and uncertainty in observational datasets. J. Climate, 28, 7630–7640, doi:10.1175/JCLI-D-15-0168.1.

    • Search Google Scholar
    • Export Citation

  • Li, G., S.-P. Xie, Y. Du, and Y. Luo, 2016: Effects of excessive equatorial cold tongue bias on the projections of tropical Pacific climate change. Part I: The warming pattern in CMIP5 multi-model ensemble. Climate Dyn., doi:10.1080/13697137.2016.1194388, in press.

    • Search Google Scholar
    • Export Citation

  • Long, S.-M., and S.-P. Xie, 2015: Intermodel variations in projected precipitation change over the North Atlantic: Sea surface temperature effect. Geophys. Res. Lett., 42, 4158–4165, doi:10.1002/2015GL063852.

    • Search Google Scholar
    • Export Citation

  • Long, S.-M., S.-P. Xie, X.-T. Zheng, and Q. Liu, 2014: Fast and slow responses to global warming: Sea surface temperature and precipitation patterns. J. Climate, 27, 285–299, doi:10.1175/JCLI-D-13-00297.1.

    • Search Google Scholar
    • Export Citation

  • Long, S.-M., S.-P. Xie, and W. Liu, 2016: Uncertainty in tropical rainfall projections: Atmospheric circulation effect and the ocean coupling. J. Climate, 29, 2671–2687, doi:10.1175/JCLI-D-15-0601.1.

    • Search Google Scholar
    • Export Citation

  • Lu, J., and B. Zhao, 2012: The role of oceanic feedback in the climate response to doubling CO2. J. Climate, 25, 7544–7563, doi:10.1175/JCLI-D-11-00712.1.

    • Search Google Scholar
    • Export Citation

  • Luo, Y., J. Lu, F. Liu, and W. Liu, 2015: Understanding the El Niño-like oceanic response in the tropical Pacific to global warming. Climate Dyn., 45, 1945–1964, doi:10.1007/s00382-014-2448-2.

    • Search Google Scholar
    • Export Citation

  • Ma, J., and S.-P. Xie, 2013: Regional patterns of sea surface temperature change: A source of uncertainty in future projections of precipitation and atmospheric circulation. J. Climate, 26, 2482–2501, doi:10.1175/JCLI-D-12-00283.1.

    • Search Google Scholar
    • Export Citation

  • Ma, J., and J.-Y. Yu, 2014: Linking centennial surface warming patterns in the equatorial Pacific to the relative strengths of the Walker and Hadley circulations. J. Atmos. Sci., 71, 3454–3464, doi:10.1175/JAS-D-14-0028.1.

    • Search Google Scholar
    • Export Citation

  • Ma, J., S.-P. Xie, and Y. Kosaka, 2012: Mechanisms for tropical tropospheric circulation change in response to global warming. J. Climate, 25, 2979–2994, doi:10.1175/JCLI-D-11-00048.1.

    • Search Google Scholar
    • Export Citation

  • Meehl, G. A., and W. M. Washington, 1996: El Niño-like climate change in a model with increased atmospheric CO2 concentrations. Nature, 382, 56–60, doi:10.1038/382056a0.

    • Search Google Scholar
    • Export Citation

  • Murtugudde, R., J. Beauchamp, C. R. McClain, M. Lewis, and A. J. Busalacchi, 2002: Effects of penetrative radiation on the upper tropical ocean circulation. J. Climate, 15, 470–486, doi:10.1175/1520-0442(2002)015<0470:EOPROT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Pierce, D. W., T. P. Barnett, K. M. AchutaRao, P. J. Gleckler, J. M. Gregory, and W. M. Washington, 2006: Anthropogenic warming of the oceans: Observations and model results. J. Climate, 19, 1873–1900, doi:10.1175/JCLI3723.1.

    • Search Google Scholar
    • Export Citation

  • Ramanathan, V., and W. Collins, 1991: Thermodynamic regulation of ocean warming by cirrus clouds deduced from observations of the 1987 El Niño. Nature, 351, 27–32, doi:10.1038/351027a0.

    • Search Google Scholar
    • Export Citation

  • Song, X., and G. J. Zhang, 2014: Role of climate feedback in El Niño–like SST response to global warming. J. Climate, 27, 7301–7318, doi:10.1175/JCLI-D-14-00072.1.

    • Search Google Scholar
    • Export Citation

  • Stockdale, T. N., A. J. Busalacchi, D. E. Harrison, and R. Seager, 1998: Ocean modeling for ENSO. J. Geophys. Res., 103, 14 325–14 355, doi:10.1029/97JC02440.

    • Search Google Scholar
    • Export Citation

  • Taylor, K. E., R. J. Stouffer, and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485–498, doi:10.1175/BAMS-D-11-00094.1.

    • Search Google Scholar
    • Export Citation

  • Vecchi, G. A., and B. J. Soden, 2007: Global warming and the weakening of the tropical circulation. J. Climate, 20, 4316–4340, doi:10.1175/JCLI4258.1.

    • Search Google Scholar
    • Export Citation

  • Whetton, P., I. Macadam, J. Bathols, and J. O’Grady, 2007: Assessment of the use of current climate patterns to evaluate regional enhanced greenhouse response patterns of climate models. Geophys. Res. Lett., 34, L14701, doi:10.1029/2007GL030025.

    • Search Google Scholar
    • Export Citation

  • Xie, S.-P., C. Deser, G. A. Vecchi, J. Ma, H. Teng, and A. T. Wittenberg, 2010: Global warming pattern formation: Sea surface temperature and rainfall. J. Climate, 23, 966–986, doi:10.1175/2009JCLI3329.1.

    • Search Google Scholar
    • Export Citation

  • Xie, S.-P., and Coauthors, 2015: Towards predictive understanding of regional climate change. Nat. Climate Change, 5, 921–930, doi:10.1038/nclimate2689.

    • Search Google Scholar
    • Export Citation

  • Ying, J., and P. Huang, 2016: Cloud–radiation feedback as a leading source of uncertainty in the tropical Pacific SST warming pattern in CMIP5 models. J. Climate, 29, 3867–3881, doi:10.1175/JCLI-D-15-0796.1.

    • Search Google Scholar
    • Export Citation

  • Ying, J., P. Huang, and R. Huang, 2016: Evaluating the formation mechanisms of the equatorial Pacific SST warming pattern in CMIP5 models. Adv. Atmos. Sci., 33, 433–441, doi:10.1007/s00376-015-5184-6.

    • Search Google Scholar
    • Export Citation

  • Zhang, L., and T. Li, 2014: A simple analytical model for understanding the formation of sea surface temperature patterns under global warming. J. Climate, 27, 8413–8421, doi:10.1175/JCLI-D-14-00346.1.

    • Search Google Scholar
    • Export Citation

  • Zheng, Y., J.-L. Lin, and T. Shinoda, 2012: The equatorial Pacific cold tongue simulated by IPCC AR4 coupled GCMs: Upper ocean heat budget and feedback analysis. J. Geophys. Res., 117, C05024, doi:10.1029/2011JC007746.

    • Search Google Scholar
    • Export Citation

  • Zhou, Z.-Q., and S.-P. Xie, 2015: Effects of climatological model biases on the projection of tropical climate change. J. Climate, 28, 9909–9917, doi:10.1175/JCLI-D-15-0243.1.

    • Search Google Scholar
    • Export Citation
  • <sc>Fig</sc>. 1.
    View in gallery
    Fig. 1.

    (a) The MME SST warming pattern (contours) and intermodel standard deviations (shading) of SST change in the 32 models. (b) The MME (contour) and intermodel standard deviations (shading) of ocean heat transport . In (a),(b), the impact of cloud–radiation feedback has been removed. (c) Linear relationship between SST changes and in the equatorial eastern Pacific [2.5°S–2.5°N, 150°–90°W, denoted by the green box in (a),(b)].

  • <sc>Fig</sc>. 2.
    View in gallery
    Fig. 2.

    (a) The EOF1 mode of changes in the ocean heat transport . The explained variance is shown in the top-right corner. (b) The regression pattern of SST changes onto the PC1 of (shading) and the percentages of explained local variances (contours; unit: %). The total percentage of the variance of SST changes is shown in the top-right corner. Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.

  • <sc>Fig</sc>. 3.
    View in gallery
    Fig. 3.

    Regression patterns of changes in surface energy budgets in the ocean mixed layer onto the PC1 of : (a) latent heat flux, (b) shortwave radiation, (c) and the residual term of the heat transport equation. Stippling indicates that regressions are significant at the 95% confidence level, based on the Student’s t test.

  • <sc>Fig</sc>. 4.
    View in gallery
    Fig. 4.

    The regression pattern of changes in equatorial (mean of 2.5°S–2.5°N) ocean temperature onto the PC1 of . Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.

  • <sc>Fig</sc>. 5.
    View in gallery
    Fig. 5.

    Regression patterns of changes in equatorial (a) zonal and (b) vertical temperature advection onto the PC1 of and (c) the sum of (a) and (b). Vectors in (a) and (b) are the climatological ocean zonal overturning circulation in the MME. The vertical velocity is multiplied by 105 for display purposes, and vectors less than 0.1 are omitted. Stippling indicates correlations are significant at the 90% confidence level based on the Student’s t test.

  • <sc>Fig</sc>. 6.
    View in gallery
    Fig. 6.

    Regression patterns of the terms in Eq. (4) onto the PC1 of . Changes in zonal temperature advection induced by (a) changes in zonal current and (b) changes in zonal temperature gradient . Changes in vertical temperature advection induced by (c) changes in vertical current and (d) changes in vertical temperature gradient . (e) The sum of (a) and (c). (f) The sum of (b) and (d). Stippling indicates regressions are significant at the 90% confidence level based on the Student’s t test.

  • <sc>Fig</sc>. 7.
    View in gallery
    Fig. 7.

    Regression patterns of the linearized terms of changes in equatorial ocean temperature advection onto the PC1 of : (a) , (b) , (c) , (d) , (e) , (f) , (g) , (h) , (i) , (j) , (k) , and (l) . Contours in (a),(g) are the MME climatological ocean temperature, contours in (j) are the MME ocean temperature changes, and contours in (k) are the regression pattern of changes in ocean temperature. Vectors in (a),(g) are the regression patterns of changes in ocean zonal overturning circulation, vectors in (j) are the regression pattern of climatological ocean zonal overturning circulation, and vectors in (k) are the climatological ocean zonal overturning circulation. Vertical velocity is multiplied by 105 for display purposes. Stippling indicates that regressions are significant at the 90% confidence level based on the Student’s t test.

  • <sc>Fig</sc>. 8.
    View in gallery
    Fig. 8.

    The gradient of ocean temperature used in calculating Fig. 7. The MME of (a) , (c) , (e) , and (g) and the regression patterns of (b) , (d) , (f) , and (h) onto PC1 of . Stippling indicates that regressions are significant at the 90% confidence level based on the Student’s t test.

  • <sc>Fig</sc>. 9.
    View in gallery
    Fig. 9.

    Regression patterns of equatorial climatological (a) zonal and (b) vertical currents onto the PC1 of . Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.

  • <sc>Fig</sc>. 10.
    View in gallery
    Fig. 10.

    Regression patterns of (a) climatological SST and (b) equatorial climatological ocean temperature onto the PC1 of . The red and black curves in (b) denote the composited 20°C isotherm in the models. The red curve represents the models in which the PC1 of is more than 1, and the black curve those in which the PC1 of is less than −1. Vectors in (b) are the regression pattern of climatological ocean zonal overturning circulation, in which the vertical velocity is multiplied by 105 for display purposes and vectors less than 0.02 are omitted. Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.

Fig. 1.

(a) The MME SST warming pattern (contours) and intermodel standard deviations (shading) of SST change in the 32 models. (b) The MME (contour) and intermodel standard deviations (shading) of ocean heat transport . In (a),(b), the impact of cloud–radiation feedback has been removed. (c) Linear relationship between SST changes and in the equatorial eastern Pacific [2.5°S–2.5°N, 150°–90°W, denoted by the green box in (a),(b)].
Fig. 2.

(a) The EOF1 mode of changes in the ocean heat transport . The explained variance is shown in the top-right corner. (b) The regression pattern of SST changes onto the PC1 of (shading) and the percentages of explained local variances (contours; unit: %). The total percentage of the variance of SST changes is shown in the top-right corner. Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.
Fig. 3.

Regression patterns of changes in surface energy budgets in the ocean mixed layer onto the PC1 of : (a) latent heat flux, (b) shortwave radiation, (c) and the residual term of the heat transport equation. Stippling indicates that regressions are significant at the 95% confidence level, based on the Student’s t test.
Fig. 4.

The regression pattern of changes in equatorial (mean of 2.5°S–2.5°N) ocean temperature onto the PC1 of . Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.
Fig. 5.

Regression patterns of changes in equatorial (a) zonal and (b) vertical temperature advection onto the PC1 of and (c) the sum of (a) and (b). Vectors in (a) and (b) are the climatological ocean zonal overturning circulation in the MME. The vertical velocity is multiplied by 105 for display purposes, and vectors less than 0.1 are omitted. Stippling indicates correlations are significant at the 90% confidence level based on the Student’s t test.
Fig. 6.

Regression patterns of the terms in Eq. (4) onto the PC1 of . Changes in zonal temperature advection induced by (a) changes in zonal current and (b) changes in zonal temperature gradient . Changes in vertical temperature advection induced by (c) changes in vertical current and (d) changes in vertical temperature gradient . (e) The sum of (a) and (c). (f) The sum of (b) and (d). Stippling indicates regressions are significant at the 90% confidence level based on the Student’s t test.
Fig. 7.

Regression patterns of the linearized terms of changes in equatorial ocean temperature advection onto the PC1 of : (a) , (b) , (c) , (d) , (e) , (f) , (g) , (h) , (i) , (j) , (k) , and (l) . Contours in (a),(g) are the MME climatological ocean temperature, contours in (j) are the MME ocean temperature changes, and contours in (k) are the regression pattern of changes in ocean temperature. Vectors in (a),(g) are the regression patterns of changes in ocean zonal overturning circulation, vectors in (j) are the regression pattern of climatological ocean zonal overturning circulation, and vectors in (k) are the climatological ocean zonal overturning circulation. Vertical velocity is multiplied by 105 for display purposes. Stippling indicates that regressions are significant at the 90% confidence level based on the Student’s t test.
Fig. 8.

The gradient of ocean temperature used in calculating Fig. 7. The MME of (a) , (c) , (e) , and (g) and the regression patterns of (b) , (d) , (f) , and (h) onto PC1 of . Stippling indicates that regressions are significant at the 90% confidence level based on the Student’s t test.
Fig. 9.

Regression patterns of equatorial climatological (a) zonal and (b) vertical currents onto the PC1 of . Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.
Fig. 10.

Regression patterns of (a) climatological SST and (b) equatorial climatological ocean temperature onto the PC1 of . The red and black curves in (b) denote the composited 20°C isotherm in the models. The red curve represents the models in which the PC1 of is more than 1, and the black curve those in which the PC1 of is less than −1. Vectors in (b) are the regression pattern of climatological ocean zonal overturning circulation, in which the vertical velocity is multiplied by 105 for display purposes and vectors less than 0.02 are omitted. Stippling indicates that regressions are significant at the 95% confidence level based on the Student’s t test.
All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 225 81 9
PDF Downloads 162 63 9