1. Introduction
El Niño–Southern Oscillation (ENSO) is a natural interannual fluctuation of the tropical Pacific characterized by anomalies in sea surface temperature (SST) and sea surface air pressure. ENSO leads to severe climatic disasters worldwide, with considerable impacts on ecosystems, agriculture, and economies (e.g., Philander 1990; Neelin et al. 1998; McPhaden et al. 2006; Deser et al. 2010). The response of ENSO to surface warming induced by increased greenhouse gases has attracted a wide level of attention for decades because it implies changes in global extremes under global warming (Meehl et al. 2007; Yeh et al. 2009; Collins et al. 2010; Vecchi and Wittenberg 2010; Christensen et al. 2013; Cai et al. 2015a; Huang and Xie 2015). However, the projected changes in ENSO-related SST in current state-of-the-art models contain large intermodel uncertainties, under a delicate balance among the impacts of the changes in multiple mean-state variables with complex mechanisms (Timmermann et al. 1999; Van Oldenborgh et al. 2005; Collins et al. 2010; Vecchi and Wittenberg 2010; DiNezio et al. 2012; Stevenson 2012; Watanabe et al. 2012; Kim et al. 2014; Cai et al. 2015a; Ham and Kug 2016).
Recently, some robust changes in ENSO-driven rainfall variability (ENSO rainfall) over the tropical Pacific have been identified in the state-of-the-art models participating in phases 3 and 5 of CMIP (CMIP3 and CMIP5; Power et al. 2013; Cai et al. 2014; Chung and Power 2014; Chung et al. 2014; Bonfils et al. 2015; Cai et al. 2015b). Specifically, ENSO rainfall is projected to intensify over the central-eastern Pacific but weaken over the western Pacific under global warming (Power et al. 2013; Cai et al. 2014; Huang and Xie 2015). This robustness in ENSO rainfall changes has been found despite the presence of large model-to-model and scenario-to-scenario differences in the projected changes in ENSO-related SST variability (ENSO SST; Power et al. 2013). These robust changes in ENSO rainfall have considerable implications for changes beyond the Pacific because the convection and rainfall anomalies over the tropical Pacific influence remote climate systems (Lau and Nath 1996; Klein et al. 1999; Alexander et al. 2002). For example, a number of previous studies have suggested that the ENSO teleconnection over the North Pacific and North America will move eastward under global warming in response to the eastward shift in ENSO rainfall in the tropical Pacific (Kug et al. 2010; Zhou et al. 2014; Bonfils et al. 2015).
A number of mechanisms have been proposed to explain how changes in ENSO rainfall could be robust despite the considerable model uncertainty in changes in ENSO SST (Power et al. 2013; Cai et al. 2014; Chung et al. 2014; Huang and Xie 2015). For instance, the absolute increase in mean-state SST could enhance the surface moisture and rainfall response to ENSO SST (Power et al. 2013; Chung et al. 2014). On the other hand, the El Niño–like pattern of changes in mean-state SST, with more warming in the tropical central-eastern Pacific (Xie et al. 2010; Huang and Ying 2015; Ying et al. 2016), could shift the ENSO-induced rainfall pattern to the equatorial eastern Pacific (Power et al. 2013; Cai et al. 2014), owing to the nonlinear relationship between tropical SST and rainfall (Graham and Barnett 1987; Johnson and Xie 2010).
Huang and Xie (2015, hereafter HX15) proposed a framework based on moisture budget decomposition to disentangle the nonlinear process involved in the formation of robust changes in ENSO rainfall (summarized in Fig. 1). It encompasses multiple terms related to moisture and vertical circulation and can ultimately be attributed to the absolute increase in mean-state SST, the spatial pattern of mean-state SST warming, and the changes in ENSO SST. The pattern of changes in ENSO rainfall is dominated by the thermodynamic and dynamic components induced by the increases in mean-state moisture and the changes in ENSO-driven circulation variability, respectively. The increase in mean-state moisture with the thermodynamic component is induced by the absolute increase in mean-state SST. The changes in ENSO-driven circulation variability with the dynamic component are mainly induced by the changes in the amplitude of ENSO SST, the changes in the structure of ENSO SST, and the pattern of mean-state SST changes.
Diagram showing the decomposition of changes in ENSO-driven rainfall variability. Gray-shaded boxes represent unchanged historical terms. Blue-shaded boxes represent terms with steadily increasing ENSO rainfall. Yellow-shaded boxes represent terms with steadily decreasing ENSO rainfall. Yellow–blue-shaded boxes represent terms with decreasing ENSO rainfall over the western Pacific but enhancing rainfall over the central-eastern Pacific. Dashed-line boxes represent terms where ENSO rainfall changes shift steadily eastward. The terms marked in the boxes are explicitly defined in the main text.
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
Although the projected changes in ENSO rainfall show a robust eastward shift under different scenarios (Power et al. 2013), the eastward shift by the end of the twenty-first century is more pronounced under the +8.5 W m−2 for representative concentration pathway 8.5 (RCP8.5) than under RCP4.5. As we know, the increase in global temperature by the end of the twenty-first century under RCP8.5 simulations is larger than under the RCP4.5 scenario. However, it is unclear whether the scenarios and global temperature increase influence the degree of the eastward shift of ENSO rainfall and if so, how.
The present study uses the moisture budget decomposition developed in HX15 to investigate the evolution of ENSO rainfall changes from 2006 to 2100, based on the historical and RCP8.5 outputs of 32 CMIP5 models. The reporting of the study is divided into two parts. In this paper (Part I), the results from investigating the characteristics of ENSO rainfall changes in the multimodel ensemble (MME) mean of the 32 CMIP5 models are reported. Unless stated otherwise, all the results presented herein are based on the MME from the RCP8.5 simulations of the 32 CMIP5 models. In Huang (2016, manuscript submitted to J. Climate, hereafter Part II), the results from examining the intermodel uncertainty of ENSO rainfall changes in the RCP8.5 simulations of the 32 CMIP5 models are reported.
In this first part, it is shown that the eastward shift of ENSO rainfall is steadily developing along with increasing global-mean temperature and that the time-varying ENSO rainfall changes can be attributed to two major factors: the steadily increased mean-state moisture and the El Niño–like changes in mean-state SST. The models and methods used are described in section 2. The formation mechanism in the projection is discussed in section 3. Conclusions are summarized in section 4.
2. Models and methods
a. Models
This study uses the outputs of 32 models of CMIP5 in historical runs from 1971 to 2010 and RCP8.5 runs from 2006 to 2100 (Taylor et al. 2012). A list of the 32 models is provided in Table 1. All outputs of the models are interpolated into a 2.5° × 2.5° grid before analysis. The variables of SST, precipitation, air specific humidity at the surface, and vertical pressure velocity at 500 hPa are used. The MME is defined as the simple average of all models.
List of the 32 CMIP5 models used in the present study. See http://cmip-pcmdi.llnl.gov/cmip5/availability.html for details. (Additional acronym expansions are available online at http://www.ametsoc.org/PubsAcronymList.)
b. Definition of the climatology and ENSO-related variability
The 1971–2000 mean in historical runs defines the historical climatology. The period 2006–2100 in the RCP8.5 runs is divided into 30-yr segments starting from 2006 with a 5-yr leap (i.e., 2006–35, 2011–40, 2016–45, …, 2071–2100). In previous studies (e.g., Power et al. 2013; HX15), longer (e.g., 40 yr) segments have often been used, but here, 30-yr segments are selected in order to decrease the overlap between adjacent segments and to enhance the separation of the segments. The length of 30 yr is sufficient to study ENSO variability with a typical period of 2–8 yr. Consequently, a 30-yr period of 1971–2000 is also selected in the historical runs. The conclusions are not dependent on this selection. The mean in one segment defines the climatology in this segment, and the difference from the historical climatology represents the mean-state change in the corresponding period. For the historical runs and each segment in the RCP8.5 runs, the annual cycle based on the 30-yr mean is first removed to define the interannual anomaly, and then a 13-yr running mean, including the decadal-scale variations, is removed to focus on the interannual variability (Power et al. 2013; Huang and Xie 2015).
As in HX15, EOF and regression analyses are performed on the interannual anomalies of SST and other variables to define ENSO SST and ENSO rainfall, circulation, and moisture. For the segment in the historical runs and each segment in the RCP8.5 runs, the first EOF mode is calculated and scaled by the standard deviation of the corresponding principal components (PCs) to define the ENSO SST pattern in this segment. The scaling of the EOF mode makes the defined ENSO SST contain information on both amplitude and structural changes in ENSO variability. The first PCs in each segment are standardized and regressed onto the interannual anomalies of precipitation, surface specific humidity, and 500-hPa vertical pressure velocity. The regression patterns define the ENSO-driven variability of these variables. The changes in ENSO-related variability in each segment of the RCP8.5 runs are defined by the differences between the variability in this segment and in the historical runs.
c. Moisture budget decomposition
d. Decomposition of changes in ENSO circulation and ENSO SST
The changes in ENSO circulation Δω′, as one of the most important components in ΔP′, are decomposed into several parts in HX15. First, the Δω′ is divided into amplitude and structural changes: 
The structural changes in ENSO circulation Δωs′ can be attributed to the structural changes in ENSO SST ΔTs′ and the pattern of mean-state SST changes (Cai et al. 2014; Huang and Xie 2015). To distinguish the roles of ΔTs′ and the pattern of mean-state SST changes, HX15 used a statistical method to estimate the contribution of ΔTs′ in the MME on Δωs′. This method is briefly introduced here, but further details can be found in HX15. The method is based on the linear relationship between Δωs′ and ΔTs′ and the intermodel diversity of ΔTs′ and Δωs′ in multiple models. Single-field principal component analysis (SFPCA) is first performed on the intermodel diversity of ΔTs′ and Δωs′ in the 32 models (Bretherton et al. 1992; Huang and Xie 2015; Huang and Ying 2015). The SFPCA method involves first performing principal component analysis (PCA) on the ΔTs′ in multiple models and then obtaining a group of orthogonal modes of ΔTs′ and corresponding PCs. Next, the Δωs′ in multiple models are regressed onto these PCs one by one. Because of the connection of the PCs, each regression pattern of Δωs′ can be understood as the Δωs′ driven by the respective ΔTs′ mode. The MME ΔTs′ can then be represented by a set of expansion coefficients with the orthogonal modes of ΔTs′ by projecting the MME ΔTs′ on each orthogonal mode of ΔTs′. Finally, the Δωs′ driven by the MME ΔTs′ can be reconstructed by these expansion coefficients and the regression patterns of Δωs′. In the reconstruction, the first eight modes are truncated in the present study—the same as that in HX15. The estimated Δωs′ driven by the MME ΔTs′ is written as Δωs1′, whereas the residual of Δωs1′ is written as Δωs2′: 
3. Results
Figures 2a,b show the historical ENSO SST and ENSO rainfall based on the data of 1971–2000 in the MME from the CMIP5 models. The result based on the 30-yr period is consistent with that based on the 40-yr period (1961–2000) in HX15, indicating that the results are insensitive to the period length. The historical ENSO SST and ENSO rainfall are apparently located farther west than in the observation, as revealed in previous studies (Guilyardi et al. 2009; Collins et al. 2010; Vecchi and Wittenberg 2010; Christensen et al. 2013; Capotondi et al. 2015). This system bias possibly influences the projection of ENSO rainfall changes (Power et al. 2013), but it does not influence the discussion on the pattern formation of ENSO rainfall changes (Power et al. 2013; HX15). The implications of the model biases are discussed in the final section.
(a) The simulated ENSO-related SST anomalies derived from the EOF analysis and (b) ENSO-driven rainfall anomalies calculated as the regression pattern in historical runs for the period 1971–2000. Also shown are the changes relative to those in 1971–2000 for (c),(e) ENSO SST and (d),(f) ENSO rainfall simulated in the RCP8.5 runs over (c),(d) 2011–40 and (e),(f) 2071–2100.
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
Figures 2c–f show the MME changes in ENSO SST and ENSO rainfall in the early and late periods of the twenty-first century. The ENSO SST changes in the two periods are completely different—a weak enhancement in 2011–40 but a decrease in 2071–2100. The time-varying ENSO SST changes are consistent with the results in Kim et al. (2014) based on the RCP8.5 simulations of some high-performance models. Kim et al. (2014) attributed the time-varying changes in ENSO SST to a varying zonal different rate of surface warming over the Indo-Pacific basin. The ENSO rainfall changes in the two periods also show some discrepancies (Figs. 2d,f); there are apparent negative changes occurring over the western Pacific in 2071–2100 (Fig. 2f), and the positive changes shift farther east in 2071–2100 compared to 2011–40. The farther eastward shift of ENSO rainfall in 2071–2100 compared to 2011–40 implies that the ENSO rainfall changes will move farther east in a more warming climate, which is consistent with the farther eastward shift in RCP8.5 simulations compared to RCP4.5 simulations (Power et al. 2013). Comparing the changes in ENSO SST and rainfall in the two periods, it is found that the enhancement and eastward shift of ENSO rainfall changes are basically independent of the variations of the changes in ENSO SST.
The formation mechanism of ENSO rainfall changes and the relationship with ENSO SST changes are investigated by decomposing ENSO rainfall changes into several components, as in Eqs. (1)–(3). The climatologies and ENSO-driven anomalies in surface moisture and 500-hPa vertical pressure velocity are shown in Fig. 3 along with their changes in Fig. 4 [note that negative ω means upward flow associated with positive rainfall, and thus the same color (blue) is used to present negative ω and positive rainfall]. The pattern of q′ (Fig. 3c) is similar to the pattern of ENSO SST (Fig. 2a), indicating the dominant role of SST variability in the variability of surface specific humidity. The pattern of ω′ (Fig. 3d) is similar to the pattern of ENSO rainfall (Fig. 2b), indicating the consistent variability of vertical motion and convective precipitation. The root-mean-square is calculated for each variable in Figs. 3 and 4 to compare the magnitudes of these variables in Eq. (1). The changes in moisture variability Δq′ are small and have no apparent difference in the two periods. The small Δq′ demonstrates that the simplification from Eq. (1) to Eq. (2) is reasonable not only for a specific period in HX15 but also for different warming stages. The changes in mean-state moisture Δq and circulation Δω are noticeably enhanced in a warmer climate but with a similar spatial structure in the two periods, respectively (Figs. 4a,b,e,f). The Δq is positive everywhere, whereas the Δω exhibits a pattern of being negative over the equatorial central-eastern Pacific and positive over other regions.
(a),(b) The climatologies and (c),(d) ENSO-driven anomalies in (a),(c) surface specific humidity and (b),(d) 500-hPa vertical pressure velocity in historical runs (1971–2000). The regional root-mean-square is shown in the upper-right corner of each panel to represent the magnitude.
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
Changes in (a),(b),(e),(f) climatology and (c),(d),(g),(h) ENSO-driven anomalies of (left) surface specific humidity and (right) 500-hPa vertical pressure velocity simulated in RCP8.5 runs for 2011–40 in (a)–(d) and 2071–2100 in (e)–(h).
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
Figure 5 shows the thermodynamic and dynamic components of ENSO rainfall changes. The sum of the two components accurately describes the pattern of ENSO rainfall changes in the two periods (Figs. 5e,f), implying the simplified moisture budget decomposition is a good approximation for the pattern of change in ENSO rainfall variability in both periods. The thermodynamic components 
(a),(b) The thermodynamic and (c),(d) dynamic components and (e),(f) their sum of ENSO rainfall changes for 2011–40 in (a),(c),(e) and for 2071–2100 in (b),(d),(f). Contours in (e),(f) are the changes in ENSO rainfall (contour interval is 0.05 mm day–1, and negative contours are dashed).
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
Relationship between the mean-state moisture changes and SST changes, with a correlation coefficient close to 1.
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
To investigate the formation of Δω′, the total Δω′ is decomposed into the amplitude changes and structural changes: 
Relationship between the amplitude change percentages in ENSO circulation and ENSO SST in four periods. The filled circles indicate the MME. The correlation coefficient is shown in the upper-right corner of each panel.
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
In both 2011–14 and 2071–2100 periods, the MME ΔTs′ shows a robust pattern with increased variability over the central Pacific and decreased variability over the western Pacific, implying more frequent central Pacific El Niño in a warming climate (Yeh et al. 2009). The structural changes in ENSO circulation induced by the MME ΔTs′, written as Δωs1′, are estimated by the aforementioned method and shown in Figs. 8c,g. The estimated patterns of Δωs1′ are closely related to the ΔTs′, with enhanced ENSO circulation located over enhanced ENSO SST. The Δωs1′ in 2071–2100 is much stronger than that in 2011–40 because of stronger ΔTs′ during 2071–2100 in the western Pacific with frequent convection, in which the convection anomaly is sensitive to the SST anomaly.
The (a),(e) structural changes in ENSO SST (ΔTs′), (b),(f) total structural changes in ENSO circulation (Δωs′), (c),(g) estimated structural changes in ENSO circulation driven by the structural changes in ENSO SST (Δωs1′), and (d),(h) residual of Δωs1′ in Δωs′ (Δωs2′) for 2011–40 in (a)–(d) and for 2071–2100 in (e)–(h). Contours in (d),(h) are the relative changes in mean-state SST (contour interval is 0.05°C, and negative contours are dashed).
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
The residual of Δωs1′ subtracted from the total ENSO circulation change Δωs′, 
The decomposition of Δωs′ in Fig. 8 suggests that both the structural changes in ENSO SST ΔTs′ and the relative changes in mean-state SST have a stronger effect on ENSO circulation changes in a more warming climate, but only the relative changes in mean-state SST can move the ENSO circulation farther east. To illustrate the zonal shift of Δωs′ more clearly, the equatorial (2.5°S–2.5°N mean) ΔT′, ΔTa′, ΔTs′, Δωs′, Δωs1′, and Δωs2′ are shown in Fig. 9. The nonunidirectional changes in the amplitude of ENSO SST are clear in Fig. 9b, contributing a large portion of the nonunidirectional changes in total ENSO SST (Fig. 9a; Kim et al. 2014). The structural changes in ENSO SST ΔTs′ gradually increase with the global-mean temperature increase, and a pronounced pattern with increased variability over the central Pacific and decreased variability over the western and eastern Pacific occurs from the period 2021–50 (Fig. 9c), inducing stronger Δωs1′ (Fig. 9e).
Evolution of (a) ΔT′, (b) ΔTa′, (c) ΔTs′, (d) Δωs′, (e) Δωs1′ and (f) Δωs2′ at the equator (2.5°S–2.5°N mean). The red-dashed arrow in (d),(f) illustrates the eastward shift of the negative–positive demarcation. Contours in (f) are the relative changes in mean-state SST (contour interval is 0.05°C, and negative contours are dashed).
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
Although HX15 and Figs. 8d,h show that the Δωs2′ is closely correlated with the relative changes in mean-state SST, the evolution of equatorial relative changes in mean-state SST apparently differs from that of the Δωs2′. With the global-mean temperature increase, the gradients of mean-state SST changes gradually increase, but the demarcation between positive and negative relative changes in mean-state SST does not show an apparent zonal shift (contours in Fig. 9f). On the other hand, the strength of Δωs2′ gradually increases from 2006 to 2041–70 but remains unchanged after then, whereas the positive–negative demarcation of Δωs2′ shifts steadily eastward from 150°E to 180° at the end of the twenty-first century. The positive–negative demarcation of Δωs2′ at the end of the twenty-first century in the present analysis, based on the MME from the RCP8.5 runs, is farther east than that in HX15, based on the MME from RCP4.5 runs, which is around 165°E.
The evolutions of Δωs2′ and the relative changes in mean-state SST exhibit a characteristic of the impacts of relative changes in mean-state SST on Δωs2′: when the relative SST increase over the equatorial central-eastern Pacific reaches a threshold, it only moves the position of the Δωs2′ response to SST anomalies; it does not linearly modify the strength of Δωs2′. This result could be attributable to the nonlinear relationship between tropical convection and/or precipitation and SST (Graham and Barnett 1987; Waliser et al. 1993; Johnson and Xie 2010): 1) in the tropics, convection is insensitive to SST anomalies when the mean-state SST is less than a lower threshold, which is dependent on the tropical mean surface temperature—around 26°C in the present-day climate; 2) convection becomes sensitive to SST anomalies and increases when the mean-state SST is larger than the lower threshold and less than a higher threshold—around 29.5°C in the present-day climate; and 3) finally, the sensitivity of convection to SST anomalies will decrease again when the mean-state SST is larger than the higher threshold (Graham and Barnett 1987; Waliser et al. 1993; Johnson and Xie 2010). This nonlinear response of convection to SST anomalies indicates that the high-sensitivity regions of convection to SST anomalies should be located over the regions with mean-state SST larger than the lower threshold and smaller than the higher threshold, which is the reason why the largest SST anomalies are located over the eastern Pacific (Fig. 2a) but the largest rainfall responses are over the western-central Pacific (Fig. 2b). Under global warming, the changes in mean-state SST with an El Niño–like pattern will promote the SST to be larger than the lower threshold and increase the sensitivity of convection to SST anomalies in the central-eastern Pacific, whereas the relative changes in mean-state SST will promote the SST in the western Pacific to be larger than the higher threshold and decrease the convection sensitivity (Ham and Kug 2012; Cai et al. 2014; Ham and Kug 2015). In other words, the El Niño–like changes in mean-state SST will shift the regions with high convection sensitivity eastward and, consequently, so too the response pattern of rainfall to ENSO SST. With the increase in global-mean temperature, the steadily enhanced relative changes in mean-state SST will gradually shift the regions with high convection sensitivity eastward to low mean-state SST regions. This mechanism can explain the effect of relative changes in mean-state SST in the eastward shift of the ENSO circulation changes.
Figure 10 shows all components of ENSO rainfall at the equator. The sum of thermodynamic and dynamic components (Fig. 10a) accurately describes the evolution of ENSO rainfall changes, exhibiting two pronounced characteristics: the increase in magnitude and the eastward shift along with the global-mean temperature increase. The time-varying increase in ENSO rainfall changes is mainly contributed by the thermodynamic component, whereas the eastward shift of ENSO rainfall changes is mainly contributed by the dynamic component (Fig. 10c).
Evolution of the components of ENSO rainfall changes at the equator (2.5°S–2.5°N mean). The red-dashed arrow in (a),(c),(e),(f) illustrates the eastward shift of the negative–positive demarcation. Contours in (a) are the changes in rainfall anomalies (contour interval is 0.1 mm day−1, and negative contours are dashed).
Citation: Journal of Climate 29, 16; 10.1175/JCLI-D-16-0058.1
The evolutions of the parts of the dynamic component (Figs. 10c–f) are dominated by the respective ENSO circulation changes (Fig. 9) because of the stronger spatial gradients of ENSO circulation changes than those of q. The amplitude changes in the dynamic component 
4. Summary
The present study investigates the evolution of changes in ENSO-driven tropical Pacific rainfall variability based on the multimodel ensemble mean of 32 CMIP5 models under the RCP8.5 scenario. The MME changes in ENSO-driven rainfall variability generally exhibit a robust pattern, with increased variability over the central-eastern Pacific and decreased variability over the western Pacific in all warming stages, as revealed in previous studies (Power et al. 2013; HX15). A characteristic of the temporal evolution of ENSO rainfall changes is reported in this study, insofar as the pattern of ENSO rainfall changes is steadily enhanced and shifts eastward along with the increase in global-mean surface temperature. The demarcation point between positive and negative ENSO rainfall changes on the equator moves from around 150°E in the early twenty-first century to around 180° at the end of the twenty-first century, based on the RCP8.5 scenario.
The mechanism of the time-varying ENSO rainfall changes is studied based on a moisture budget decomposition developed in HX15. The results illustrate that all simplifications and decompositions in HX15 developed for the ENSO rainfall changes at the end of the twenty-first century are practicable for the ENSO rainfall changes in different warming stages (Fig. 1). The ENSO rainfall changes in all periods can be accurately described by the sum of thermodynamic and dynamic components. The thermodynamic component 
In the MME from 32 CMIP5 model RCP8.5 simulations, the thermodynamic component 
The present projection of ENSO rainfall changes is based on the MME result from 32 CMIP5 model RCP8.5 simulations. The projected characteristics of the evolution of ENSO rainfall changes are quite robust among the models, but some apparent intermodel differences still exist, which are clear to see in Figs. 6 and 7. The projected ENSO rainfall changes will also possibly differ from the results based on the simulations under other scenarios (Power et al. 2013). However, the simplification and decomposition, developed in HX15 and practicable in all warming periods, are reasonable for all models. This method provides a useful tool to further investigate the source of intermodel differences in ENSO rainfall changes, which will be reported on in a future study.
The mechanism of ENSO rainfall changes shows that the present projection of ENSO rainfall changes is dependent on the pattern of historical ENSO circulation and ENSO circulation changes simulated in the models. It has been widely demonstrated that there are apparent biases in the simulation of historical ENSO circulation associated with ENSO SST in current models (Guilyardi et al. 2009; Collins et al. 2010; Vecchi and Wittenberg 2010; Christensen et al. 2013; Capotondi et al. 2015). The historical ENSO circulation and ENSO SST could extend excessively westward relative to observations (Fig. 2). Power et al. (2013) has illustrated that the ENSO rainfall changes should shift farther east than the projections in current CMIP5 models when the bias of historical ENSO SST is corrected. Additionally, some previous studies suggest that the relative changes in mean-state SST should be closer to an El Niño–like pattern than the MME projection from CMIP5 models because of the common bias of historical mean-state SST simulations (Huang and Ying 2015; Ying and Huang 2016). This result implies that the ENSO rainfall changes could shift farther east than the current MME projection because of the role of mean-state SST changes in the ENSO circulation changes. Removing the biases of historical ENSO SST and historical mean-state SST could both lead to a farther eastward shift of ENSO rainfall changes than the MME projection from CMIP5 models. The implications of this should be verified using upgraded models in future.
The temporal evolution of the ENSO-driven tropical Pacific rainfall changes has some important implications for the changes in the teleconnection patterns driven by ENSO, such as the patterns over the North Pacific and North America (Kug et al. 2010; Zhou et al. 2014; Bonfils et al. 2015). As the variation of ENSO rainfall changes, the ENSO-driven changes in these teleconnection patterns are likely to gradually enhance and shift steadily eastward along with the increasing global-mean temperature.
Acknowledgments
The work was supported by the National Natural Science Foundation of China (Grants 41575088 and 41461164005), the National Basic Research Program of China (2014CB953903 and 2012CB955604), and the Youth Innovation Promotion Association of the Chinese Academy of Sciences. 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. Thanks are extended to Mr. Jun Ying for preparing part of the CMIP5 data. I also thank four anonymous reviewers for their constructive suggestions.
REFERENCES
Alexander, M. A., I. Blade, M. Newman, J. R. Lanzante, N.-C. Lau, and J. D. Scott, 2002: The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans. J. Climate, 15, 2205–2231, doi:10.1175/1520-0442(2002)015<2205:TABTIO>2.0.CO;2.
Bonfils, C. J. W., B. D. Santer, T. J. Phillips, K. Marvel, L. R. Leung, C. Doutriaux, and A. Capotondi, 2015: Relative contributions of mean-state shifts and ENSO-driven variability to precipitation changes in a warming climate. J. Climate, 28, 9997–10 013, doi:10.1175/JCLI-D-15-0341.1.
Bretherton, C. S., C. Smith, and J. M. Wallace, 1992: An intercomparison of methods for finding coupled patterns in climate data. J. Climate, 5, 541–560, doi:10.1175/1520-0442(1992)005<0541:AIOMFF>2.0.CO;2.
Cai, W., and Coauthors, 2014: Increasing frequency of extreme El Niño events due to greenhouse warming. Nat. Climate Change, 4, 111–116, doi:10.1038/nclimate2100.
Cai, W., and Coauthors, 2015a: ENSO and greenhouse warming. Nat. Climate Change, 5, 849–859, doi:10.1038/nclimate2743.
Cai, W., and Coauthors, 2015b: Increased frequency of extreme La Niña events under greenhouse warming. Nat. Climate Change, 5, 132–137, doi:10.1038/nclimate2492.
Capotondi, A., Y.-G. Ham, A. Wittenberg, and J.-S. Kug, 2015: Climate model biases and El Niño Southern Oscillation (ENSO) simulation. U.S. CLIVAR Variations, Vol. 13, 21–25. [Available online at http://www.gfdl.noaa.gov/~atw/yr/2015/capotondi_etal_variations2015.pdf.]
Chou, C., J. D. Neelin, C. Chen, and J. Tu, 2009: Evaluating the “rich-get-richer” mechanism in tropical precipitation change under global warming. J. Climate, 22, 1982–2005, doi:10.1175/2008JCLI2471.1.
Christensen, J. H., and Coauthors, 2013: Climate phenomena and their relevance for future regional climate change. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 1217–1308.
Chung, C. T. Y., and S. B. Power, 2014: Precipitation response to La Niña and global warming in the Indo-Pacific. Climate Dyn., 43, 3293–3307, doi:10.1007/s00382-014-2105-9.
Chung, C. T. Y., S. B. Power, J. M. Arblaster, H. A. Rashid, and G. L. Roff, 2014: Nonlinear precipitation response to El Niño and global warming in the Indo-Pacific. Climate Dyn., 42, 1837–1856, doi:10.1007/s00382-013-1892-8.
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.
Deser, C., M. A. Alexander, S.-P. Xie, and A. S. Phillips, 2010: Sea surface temperature variability: Patterns and mechanisms. Annu. Rev. Mar. Sci., 2, 115–143, doi:10.1146/annurev-marine-120408-151453.
DiNezio, P. N., B. P. Kirtman, A. C. Clement, S.-K. Lee, G. A. Vecchi, and A. Wittenberg, 2012: Mean climate controls on the simulated response of ENSO to increasing greenhouse gases. J. Climate, 25, 7399–7420, doi:10.1175/JCLI-D-11-00494.1.
Graham, N., and T. P. Barnett, 1987: Observations of sea surface temperature and convection over tropical oceans. Science, 238, 657–659, doi:10.1126/science.238.4827.657.
Guilyardi, E., and Coauthors, 2009: Understanding El Niño in ocean–atmosphere general circulation models: Progress and challenges. Bull. Amer. Meteor. Soc., 90, 325–340, doi:10.1175/2008BAMS2387.1.
Ham, Y.-G., and J.-S. Kug, 2012: How well do current climate models simulate two types of El Niño? Climate Dyn., 39, 383–398, doi:10.1007/s00382-011-1157-3.
Ham, Y.-G., and J.-S. Kug, 2015: Improvement of ENSO simulation based on intermodel diversity. J. Climate, 28, 998–1015, doi:10.1175/JCLI-D-14-00376.1.
Ham, Y.-G., and J.-S. Kug, 2016: ENSO amplitude changes due to greenhouse warming in CMIP5: Role of mean tropical precipitation in the 20th century. Geophys. Res. Lett., 43, 422–430, doi:10.1002/2015GL066864.
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.
Huang, P., 2014: Regional response of annual-mean tropical rainfall to global warming. Atmos. Sci. Lett., 15, 103–109, doi:10.1002/asl2.475.
Huang, P., and S.-P. Xie, 2015: Mechanisms of change in ENSO-induced tropical Pacific rainfall variability in a warming climate. Nat. Geosci., 8, 922–926, doi:10.1038/ngeo2571.
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.
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.
Johnson, N. C., and S.-P. Xie, 2010: Changes in the sea surface temperature threshold for tropical convection. Nat. Geosci., 3, 842–845, doi:10.1038/ngeo1008.
Kim, S. T., W. Cai, F.-F. Jin, A. Santoso, L. Wu, E. Guilyardi, and S.-I. An, 2014: Response of El Niño sea surface temperature variability to greenhouse warming. Nat. Climate Change, 4, 786–790, doi:10.1038/nclimate2326.
Klein, S. A., B. J. Soden, and N. C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, 917–932, doi:10.1175/1520-0442(1999)012<0917:RSSTVD>2.0.CO;2.
Kug, J.-S., S.-I. An, Y.-G. Ham, and I.-S. Kang, 2010: Changes in El Niño and La Niña teleconnections over North Pacific–America in the global warming simulations. Theor. Appl. Climatol., 100, 275–282, doi:10.1007/s00704-009-0183-0.
Lau, N. C., and M. J. Nath, 1996: The role of the “atmospheric bridge” in linking tropical Pacific ENSO events to extratropical SST anomalies. J. Climate, 9, 2036–2057, doi:10.1175/1520-0442(1996)009<2036:TROTBI>2.0.CO;2.
McPhaden, M. J., S. E. Zebiak, and M. H. Glantz, 2006: ENSO as an integrating concept in Earth science. Science, 314, 1740–1745, doi:10.1126/science.1132588.
Meehl, G. A., and Coauthors, 2007: Global climate projections. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 747–845.
Neelin, J., D. Battisti, A. Hirst, F. Jin, Y. Wakata, T. Yamagata, and S. Zebiak, 1998: ENSO theory. J. Geophys. Res., 103, 14 261–14 290, doi:10.1029/97JC03424.
Philander, S. G., 1990: El Niño, La Niña, and the Southern Oscillation. Academic Press, 293 pp.
Power, S., F. Delage, C. Chung, G. Kociuba, and K. Keay, 2013: Robust twenty-first-century projections of El Niño and related precipitation variability. Nature, 502, 541–545, doi:10.1038/nature12580.
Seager, R., N. Naik, and G. A. Vecchi, 2010: Thermodynamic and dynamic mechanisms for large-scale changes in the hydrological cycle in response to global warming. J. Climate, 23, 4651–4668, doi:10.1175/2010JCLI3655.1.
Seager, R., N. Naik, and L. Vogel, 2012: Does global warming cause intensified interannual hydroclimate variability? J. Climate, 25, 3355–3372, doi:10.1175/JCLI-D-11-00363.1.
Stevenson, S. L., 2012: Significant changes to ENSO strength and impacts in the twenty-first century: Results from CMIP5. Geophys. Res. Lett., 39, L17703, doi:10.1029/2012GL052759.
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.
Timmermann, A., J. Oberhuber, A. Bacher, M. Esch, M. Latif, and E. Roeckner, 1999: Increased El Niño frequency in a climate model forced by future greenhouse warming. Nature, 398, 694–697, doi:10.1038/19505.
Van Oldenborgh, G., S. Philip, and M. Collins, 2005: El Niño in a changing climate: A multi-model study. Ocean Sci., 1, 81–95, doi:10.5194/os-1-81-2005.
Vecchi, G. A., and A. T. Wittenberg, 2010: El Niño and our future climate: Where do we stand? Wiley Interdiscip. Rev.: Climate Change, 1, 260–270, doi:10.1002/wcc.33.
Waliser, D. E., N. E. Graham, and C. Gautier, 1993: Comparison of the highly reflective cloud and outgoing longwave radiation datasets for use in estimating tropical deep convection. J. Climate, 6, 331–353, doi:10.1175/1520-0442(1993)006<0331:COTHRC>2.0.CO;2.
Watanabe, M., J.-S. Kug, F.-F. Jin, M. Collins, M. Ohba, and A. T. Wittenberg, 2012: Uncertainty in the ENSO amplitude change from the past to the future. Geophys. Res. Lett., 39, L20703, doi:10.1029/2012GL053305.
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.
Yeh, S., J. Kug, B. Dewitte, M. Kwon, B. Kirtman, and F. Jin, 2009: El Niño in a changing climate. Nature, 461, 511–514, doi:10.1038/nature08316.
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.
Ying, J., P. Huang, and R.-H. 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.
Zhou, Z.-Q., S.-P. Xie, X.-T. Zheng, Q. Liu, and H. Wang, 2014: Global warming–induced changes in El Niño teleconnections over the North Pacific and North America. J. Climate, 27, 9050–9064, doi:10.1175/JCLI-D-14-00254.1.
