The distinct impact of tropical Indian Ocean (IO) and western Pacific (WP) sea surface temperatures (SSTs) after the El Niño winter has been investigated in relation to the summer North Pacific high (NPH) and western North Pacific subtropical high (WNPSH). After the El Niño winter, warming of the IO leads to a summer eastern Pacific (EP) SST anomaly distinct from the cooling of WP; EP cooling occurs in the extreme IO warming case and EP warming in the WP cooling case. Both the warming of the IO and cooling of the WP are responsible for the development of the WNPSH, whereas the summer EP cooling induces an enhanced NPH, especially if it coexists with IO warming. The IO warming triggers an abrupt termination of the El Niño event by causing the easterly anomaly in the WP, which leads to the coexistence of IO warming and EP cooling during the boreal summer. The tropical EP cooling may change the North Pacific SST anomalies via the atmospheric bridge and consequently strengthen the extratropical NPH. The experimental results, which have been obtained from the use of atmospheric general circulation model, support the distinct roles of EP cooling on the NPH and of IO warming and WP cooling on the WNPSH. This finding suggests that the combined effect of IO warming and EP cooling generates a coupled pattern of NPH and WNPSH.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Sea surface temperature (SST) warming of the eastern Pacific (EP) during the El Niño winter causes warming of the Indian Ocean (IO) and cooling of the western Pacific (WP) through the atmospheric bridge process [Klein et al., 1999; Alexander et al., 2002; Moon et al., 2011]. Excitation of midlatitude circulation by these tropical heat sources has been investigated as the teleconnection patterns of Pacific–North American (PNA) [Wallace and Gutzler, 1981] and Pacific–Japan (PJ) [Nitta, 1987]. In particular, it has been reported that the IO warming and WP cooling after the El Niño winter, which are maintained until the ensuing summer, play a significant role in modulating summertime extratropical circulation [Wang et al., 2000; Yun et al., 2008; Xie et al., 2009; Chowdary et al., 2010]. For example, Chowdary et al.  have shown the importance of both IO warming and WP cooling on rainfall prediction over the western North Pacific (WNP) after the mature phase of El Niño by using 11 coupled model hindcast results.
 For more accurate summer predictions, much attention has been paid to the mechanism and relative role of IO warming and WP cooling in relation to summer WNP rainfall and circulation. With delayed impact of El Niño/Southern Oscillation (ENSO), positive thermodynamic feedback [Wang et al., 2000; 2003] and Kelvin wave-induced Ekman divergence [Xie et al., 2009; Chowdary et al., 2010] have been roughly suggested as key mechanisms in determining the change in the WNP subtropical high (WNPSH). According to the previous mechanism proposed by Wang et al. , local WP cooling suppresses WNP convection, which generates a low-level WNPSH anomaly to its west. The WNPSH-induced easterly anomalies enhance mean trade winds and cause a reduction in WP SST through enhanced evaporative cooling and turbulent mixing. The latter mechanism of Xie et al.  is caused by tropical IO warming. This phenomenon causes a baroclinic Kelvin wave into the EP and resultant northeasterly wind anomalies in the WP, which consequently induces the suppressed convection and WNPSH. On the other hand, Wu et al.  have demonstrated that the impacts of IO warming and WP cooling on the WNPSH have a different subseasonal contribution during the El Niño decaying summer. The effect of IO warming is greater in the late summer, while WP cooling exerts a significant impact in the early summer.
 Consequently, both IO warming and WP cooling influence the summertime WNPSH, and their impacts are mainly restricted to the WNPSH. However, the summer mean anticyclone circulation occupies a much larger domain over the North Pacific, which is generally dominated by the North Pacific high (NPH). The WNPSH indicates an anomalous anticyclone over the WNP region [broadly 10°N–30°N, 100°E–160°E], which is caused by the westward extension of the NPH [Lu, 2001]. During the boreal summer, NPH dynamics is determined by topography and monsoonal heating [Rodwell and Hoskins, 2001], local air–sea interaction [Seager et al., 2003], and near-surface thermal contrast between a hotter land and a cooler ocean [Miyasaka and Nakamura, 2005]. Thus, the change in NPH plays an important role in modulating the Northern Hemisphere planetary waves and summer monsoon systems.
 Our primary goal is to investigate the distinct role of IO warming and WP cooling after the El Niño winter against the summer WNPSH and NPH by using a modern global observation dataset (1979–2010). Previous studies [Ding et al., 2010; Li et al., 2010] have alluded to the interdecadal change in the relationship between tropical SSTs and the WNPSH and NPH after the late 1970s. For example, Li et al.  reported a strengthened relationship after the late 1970s between the preceding winter ENSO and the East Asian summer monsoon, which is closely associated with the change in the relationship between the WNPSH and tropical SSTs. The analysis after 1979 helps to exclude the effect of interdecadal change on this relationship after the late 1970s. In addition, because EP warming does not persist into the ensuing summer, the impact of the EP SST change on the extratropical WNPSH and NPH is not yet firmly addressed. The role of EP SST change during the boreal summer will be discussed in relation to the IO–WP SST and North Pacific anticyclone circulation. For this, the responses of an atmospheric general circulation model (AGCM) to the specified IO, WP, and EP SST forcing are examined in the extratropical North Pacific. This paper is organized as follows. Section 2 describes the data, methods, numerical model, and experimental design. The effects of tropical IO warming, WP cooling, and EP SST change on the summer extratropical circulation are suggested in Section 3. Section 4 examines individual and combined impacts of IO, WP, and EP SSTs by using the AGCM numerical experiment. In the last section, we provide a discussion and conclusion of the main results.
2 Data and Model Experiment
 To analyze the characteristics of tropical SST and extratropical circulation, the seasonal mean fields obtained from National Centers for Environmental Prediction/Department of Energy (NCEP/DOE) reanalysis for the period 1979–2010 [Kanamitsu et al., 2002] were used. The SST data were obtained from the British Atmospheric Data Centre (BADC) Hadley Centre Sea Ice and Sea Surface Temperature data set (HadISST) [Rayner et al., 2003] for the period 1979–2010. For the detailed structure in the statistical significance, the significances are shown in the 90%, 95%, and 99% confidence levels of a two-tailed Student's t-distribution. To investigate the individual impact of tropical SSTs on summertime extratropical circulation, the EP, IO, and WP SSTs were defined by the anomaly area averaged over the EP region [Niño3.4 region; 5°S–5°N, 170°W–120°W], IO region [5°S–5°N, 50°E–100°E], and the WP region [5°S–5°N, 120°E–160°E], respectively. Since most tropical SSTs tend to peak in the boreal winter, the results are mainly displayed in the December–January–February (DJF) seasonal mean SST anomalies.
 As shown in Figure 1, the IO SST anomaly is well correlated with the EP SST anomaly (r ~ 0.70) during boreal winter. A close linkage is also shown in the WP–EP SST relationship (r ~ −0.66). However, the IO SST is not significantly related to the WP SST (r ~ −0.18), implying that extreme ENSO events can induce a different response in IO and WP SST regions. On the other hand, tropical winter SSTs show distinct features in long-term changes, including no trend in EP SST, an increasing trend in IO SST, and a decadal change in WP SST after the mid-1990s. A remarkable long-term change in IO and WP SSTs has been reported in numerous previous studies [e.g., Wang and Mehta, 2008]. Note that most warm IO SST (i.e., W–IO) years belong to the period after the mid-1990s, and all cold WP SST (i.e., C–WP) years belong to that before the mid-1990s. An investigation of the distinct impact of IO and WP SSTs on the summertime subtropical high may contribute to an understanding of the long-term change (i.e., an increasing trend or decadal change) in the Northern Pacific climate.
 To diagnose the impact of extreme IO warming and WP cooling after the El Niño winter, the W–IO years and C–WP years are categorized as follows: extreme W–IO years are defined as years in which the DJF IO SST anomaly is greater than 1 standard deviation, whereas extreme C–WP years are defined as those in which the DJF WP SST anomaly is less than −1 standard deviation. If both indices of IO and WP SST exceed this criterion, the year is classified into a case of the greater SST anomaly among W–IO and C–WP cases. Thus, five W–IO years are selected as 1988, 1998, 2003, 2007, and 2010, and five C–WP years are 1980, 1983, 1987, 1992, and 1993. The chosen extreme cases occur during the El Niño winter and are displayed in Figure 1.
 For the purpose of demonstrating the relative role of different SST forcing, an AGCM experiment was performed using the ECHAM4.6 model from the Max Planck Institute for Meteorology in Hamburg, Germany, which was built on the weather forecast model of the ECMWF. ECHAM4.6 is a global spectral model that includes triangular truncation at wave number 42 and a 19-level hybrid sigma-pressure coordinate system. A detailed description of this model is listed in Roeckner et al., . After 30-year integration, the results from the last 25 years of the simulation were used. The control experiment (CTRL) is simply run in the climatological SST, 1979 to 2010. In each experiment (e.g., EXP_IO and EXP_WP), the EP, IO, and WP SST forcings were specified by the composite SST anomaly to the climatological SST, in the domain of EP, WP, and IO along the equatorial band [15°S–15°N]. In detail, the SST forcing in EXP_IO and EXP_IOEP was obtained from the composite anomaly for W–IO years, while that in EXP_WP and EXP_WPEP was given from the anomaly for C–WP years. The temporal evolution of prescribed SST forcing in each experiment is presented in Figure 6.
3 Effect of Tropical SSTs on Extratropical Circulation
 To investigate the distinct peculiarity of IO warming and WP cooling after the El Niño winter, we have shown the composite SST, precipitation, and low-level geopotential height anomaly fields for the chosen W–IO and C–WP years, respectively (Figures 2 and 3). The tropical SSTs show different seasonal evolutions from winter to summer between the W–IO and C–WP years. During the winter season, the SST composite anomaly for W–IO years presents evident IO and EP warming, while that for C–WP years displays significant WP cooling and EP warming (Figures 2a and 2b). The following springtime SST anomalies exhibit the persistent IO warming and WP cooling for W–IO and C–WP years, respectively (Figures 2c and 2d). Despite the analogous wintertime ENSO signal, a significant difference between W–IO and C–WP years appears in the EP SST anomaly such that decaying EP warming for W–IO years and sustained EP warming for C–WP years. The distinct EP SST anomaly between IO warming and WP cooling years is manifested stronger in the summer season (Figures 2e and 2f). For the C–WP years, the remarkable EP warming signal persists until the following summer season. This persistent EP warming is compared with significant EP cooling for the W–IO years. In relation to the abrupt decay of EP warming for the W–IO years, previous studies [e.g., Kug and Kang, 2006; Yoo et al., 2010] have revealed that the IO warming-induced easterly anomaly in the WP triggers the upwelling of oceanic Kelvin wave and consequently induces a rapid termination of the El Niño event. On the other hand, for the W–IO (C–WP) years, the IO warming (WP cooling) persists until the following summer. On the basis of the longer period of 1950–2007, Ding and Li  have also reported the significant ENSO-related IO warming and WP cooling. These results correspond strongly to the evolutionary change in IO, WP, and EP SSTs displayed in Figure 2. Therefore, for the W–IO years, the IO–Pacific dipole SST pattern of IO–WP warming and EP cooling prevails, while the Pacific dipole pattern of WP cooling and EP warming appears in the C–WP years.
 The abrupt phase change of ENSO from El Niño winter into La Niña summer is presented in Figures 2a and 2e. During the boreal summer, persistent IO warming is responsible for the increased and decreased rainfall amounts over the IO and WP regions, respectively (Figure 3a). The zonal dipole precipitation pattern is strongly connected with an ascent over the IO accompanied by a descent over the WP through changes in the Walker circulation [Watanabe and Jin, 2002; Yun et al., 2008]. However, the WP cooling induces a meridional dipole pattern of precipitation (Figure 3b). The remarkable differences in rainfall over the North Pacific are mainly shown in the WNP [10°N–20°N, 120°E–160°E], WP [5°S–5°N, 140°E–180°], and North eastern Pacific [NEP; 2°N–12°N, 160°W–120°W] regions (denoted by red boxes in Figures 3a and 3b). These results imply the relative importance of the reduced WNP rainfall for C–WP years and the weakened WP and NEP rainfall for W–IO years. In relation to the different rainfall structure caused by the IO warming and WP cooling, the anomalous rainfall may act as important sources in modulating the change in extratropical circulation.
 In the Gill-type Rossby wave response to the suppressed convection in the WNP, a significant WNPSH is evident in the northeast of the decreased rainfall (Figure 3c). The IO impact on the WNPSH is explained by the IO-induced Kelvin wave and the resultant northeasterly wind anomalies in the WP [Xie et al., 2009]. Along the southern edge of the WNPSH, a significant easterly anomaly, which contributes to the rapid phase transition of El Niño into La Niña, is manifested in the equatorial WP.
 Of particular interest is the fact that the IO warming is related to the enhancement of a broadly expanded subtropical high over the North Pacific, which is often referred to as the NPH. Meanwhile, the combination of WP cooling and EP warming is associated only with strengthened WNPSH and is more closely related to the meridional wave train pattern over the WNP than the combination of IO warming and EP cooling (Figure 3d). The WNP reduced rainfall is responsible for the anomalous anticyclone circulation over the WNP and, in turn, the cyclonic anomaly to the north of the WNPSH, through the Rossby wave train emanating northeastward into the extratropical region. The prominent effect of WP cooling on the WNPSH could be explained by positive thermodynamic feedback suggested by Wang et al. [2000; 2003], in which the cooling suppresses convection and latent heating to further excite the Rossby waves from the anomalous anticyclone. This anticyclonic anomaly cools the ocean to its east because of enhanced evaporative cooling and turbulent mixing.
 Consequently, the summertime extratropical circulation for the W–IO years differs significantly from that for the C–WP years. To prove the different role of IO warming and WP cooling after the El Niño winter on summertime circulation, we apply a singular value decomposition (SVD) analysis to the DJF IO–WP SST anomaly over [20°S–20°N, 50°E–160°E] and the June–July–August (JJA) geopotential height anomaly at 850 hPa over [10°N–55°N, 100°E–100°W] (Figure 4). The SVD result extending into EP domain interrupts in part to show the impact of IO–WP SST on the extratropical circulation due to the large variability in EP. To exclude the EP variability in the SVD analysis, the SST domain is restricted to the IO–WP region (i.e., 50°E–160°E). The first two SVD modes explain approximately 81% of the total variance. The first SVD mode (SVD1) captures the IO–WP warming and coupled NPH–WNPSH pattern. On the other hand, the second SVD mode (SVD2) exhibits the WP cooling, moderate IO warming, and WNPSH. These results effectively represent the distinct relationship between IO–WP SST and extratropical circulation. The SVD patterns of the summertime low-level circulation show similarities to the composite anomaly patterns for W–IO and C–WP years (Figures 3c and 3d). The SVD analysis to the JJA IO–WP SST anomaly and JJA geopotential height anomaly at 850 hPa also suggests a significant simultaneous relationship between IO–WP warming and coupled NPH–WNPSH pattern and between WP cooling and WNPSH (figure not shown).
 The principal component (PC) time series of the SST patterns of SVD1 and SVD2 is strongly correlated to those of low-level circulation at the correlation coefficient equal to approximately 0.68 and 0.74, respectively. It should be noted that the associated time series shows notable decadal changes after the late 1990s. It is of particular interest that the SVD1 increases after the late 1990s, while the SVD2 decreases after the late 1990s. The decadal change after the late 1990s may be related to the recent SST change in the tropical Pacific region, such as the IO–WP SST warming and the shift of the Pacific warming center [e.g., Wang and Mehta, 2008; Yun et al., 2010; Xiang et al., 2012], which will be investigated in the future study.
 In comparison with WP cooling, IO warming is significantly linked to summertime NPH. Could the difference in the extratropical circulation be immediately induced by the IO warming? In the boreal summer, the main SST difference between W–IO years and C–WP years is the zonal deviation between WP and EP (i.e., WP warming and EP cooling) for W–IO years (Figure 2e). What role do these EP cooling and WP warming play in the relationship between tropical SSTs and extratropical circulation? In fact, the NPH is more closely correlated with the summer EP SST anomaly (r ~ −0.50) than with the summer WP SST anomaly (r ~ 0.35). The summer WP warming may be a response to the EP cooling via the atmospheric bridge. To investigate the role of summer EP cooling on this relationship, we have shown a scatter plot of DJF IO SST (or WP SST) and JJA EP SST in terms of the magnitude of NPH and WNPSH (Figure 5) in which the NPH and WNPSH are simply defined as the area-averaged sea level pressure anomalies over the region [30°N − 40°N, 170°W − 140°W] and [15°N − 25°N, 110°E − 150°E], respectively. Here, the strong (weak) highs indicate normalized anomaly of highs greater than 0.7 (less than −0.7). The threshold of 0.7 supplies an adequate sample size for the extreme NPH and WNPSH cases; that is, the extreme cases chosen for both strong and weak years are nearly half the total sample size.
 In the scatter plot, the strong WNPSHs (denoted by green squares) are mostly shown in winter IO warming and WP cooling, regardless of the summer EP SST anomaly. On the contrary, the weak WNPSHs (indicated by yellow squares) tend to occur in winter IO cooling and WP warming. As noted in previous studies [e.g., Wu et al., 2010; Chowdary et al., 2010], the correlation coefficient between the summer WNPSH and winter IO SST (WP SST) is approximately 0.53 (−0.52), which reflects the close relationship between the winter IO warming (or WP cooling) and strong WNPSH. On the other hand, strong NPHs (shown by red triangles) roughly occur in summer EP cooling, and the weak NPHs (displayed by blue triangles) occur in EP warming. The winter IO and WP SST anomalies exhibit no linear relationship with the magnitude of summer NPH. As a result, the NPH is more closely associated with the summer EP SST anomaly (r ~ −0.50) than the winter IO SST (r ~ 0.25) or WP SST anomaly (r ~ −0.18) (Figure 5). The summer EP cooling warms the tropical WP and extratropical central North Pacific and cools the water along the west coast of North America through the atmospheric bridge [Alexander et al., 2002], which may be significantly linked to the enhanced NPH.
4 Numerical Experiment on Distinct Impact of Tropical SSTs
 In section 3, we suggest the effects of tropical IO, WP, and EP SSTs on extratropical circulation. Both IO SST and WP SST after the El Niño winter are related to WNPSH; IO warming-induced summer EP cooling forces NPH variability. In relation to a different seasonal evolution of SST, the winter IO warming results in EP cooling in the following summer, while the winter WP cooling accompanies moderate EP warming in the ensuing summer. Because of the time-lagged relationship between IO SST (or WP SST) and EP SST, it is difficult to separate the effects of these SSTs in the composite analysis. To examine the distinct role between IO, WP, and EP SST anomalies on extratropical circulation, we have performed numerical experiments with the observed SST forcing by using ECHAM4.6 AGCM. As shown in Figure 6, the SST forcing in each experiment is prescribed by the monthly anomaly of composite SST for the W–IO and C–WP years. As an example, EXP_IOEP has IO warming and EP cooling during the boreal summer, while EXP_WPEP exhibits the WP cooling and the EP warming.
 Figure 7 displays the differences in low-level JJA geopotential height anomalies between the CTRL and each experiment (i.e., EXP minus CTRL). As expected, both IO warming and WP cooling induce a strong WNPSH (Figures 7c and 7d). However, IO warming is not responsible for the enhanced NPH. The EP SST anomaly for W–IO years induces an enhanced summer NPH (Figure 7a). The strong NPH anomaly is compared with the cyclonic anomaly for C–WP years (Figure 7b). On the other hand, the experiment results indicate that both IO and EP impacts (Figure 7e) generate a more strengthened NPH anomaly with the WNPSH than that with EP impact only (Figure 7a). The coupled pattern of NPH and WNPSH in EXP_IOEP is similar to the observed pattern for W–IO years shown in Figure 3c. Although the NPH is more strongly related to the summer EP cooling, the combined effect (or interaction) between IO warming and EP cooling could play an important role in generating the coupled pattern of NPH and WNPSH. The combined effect of WP cooling and EP warming (Figure 7f) also forms a more strengthened WNPSH anomaly than that of WP cooling only (Figure 7d) or EP warming only (Figure 7b), indicating the importance of the interaction between WP and EP SSTs. Considering EP SST anomaly affects the NPH variability, the combination of WP cooling and EP warming may cause variability both in WNPSH and NPH. However, the result of numerical experiment in Figure 7f is not well compared with the observed low-level structure in Figure 3d, in particular for the NPH variability (at about 40°N, 180). In detail, the combination of WP cooling and EP warming in Figure 3d induces North Pacific cyclonic anomaly, while the simulated pattern in Figure 7f does not show the North Pacific cyclonic anomaly. The inconsistency between numerical experiment and observation results makes it difficult to draw a clear conclusion for the combined effect on the NPH variability, which should be investigated in future study.
 Similar to the differences in low-level circulation, the EP impacts in EXP_EPio and EXP_EPwp runs affect mainly the tropical Pacific (i.e., WP and NEP) rainfall anomaly which may be closely related to the change in NPH (Figures 8a and 8b). Impacts of IO SST and WP SST lead to enhanced (reduced) rainfall over the IO (WNP) region (Figures 8c and 8d). Eventually, the experiment results of both IO and EP impacts (Figure 8e) reveal a weakened rainfall anomaly in the WNP, WP, and NEP regions, while that in both WP and EP impacts shows reduced WNP rainfall and enhanced WP and NEP rainfall anomalies (Figure 8f). The difference in precipitation strongly resembles the observed rainfall pattern in the composite analysis for the W–IO and C–WP years (Figures 3a and 3b). This resemblance suggests that both IO warming and EP cooling are important for the coupled structure of NPH–WNPSH, while both WP cooling and EP warming are significantly connected with the WNPSH. The strong connection between extratropical pressure variability and tropical SST corresponds well to the tropical–extratropical relationship suggested in previous studies [e.g., Yu et al., 2010; Yu and Kim, 2011].
5 Discussion and Conclusion
 We investigate the distinct impact of tropical SSTs, including IO, WP, and EP SST, on the summer NPH and WNPSH. Unlike the close relationship of IO–WP SST with EP SST, the IO SST anomaly is not significantly correlated with the WP SST anomaly. The IO warming and WP cooling after the El Niño winter introduce different summer SST structures including IO–WP SST warming and EP cooling for W–IO years and WP cooling and EP warming for C–WP years. Consequently, the IO warming and WP cooling are responsible for the WNPSH variability, whereas the summer EP cooling forces the NPH variability. The SVD analysis supports a notable relationship between IO warming and the coupled pattern of NPH–WNPSH, which is strongly associated with the decadal change after the late 1990s. This change may be attributed to the recent IO–WP SST warming and the shift of Pacific warming center [Wang and Mehta, 2008; Xiang et al., 2012]. It should be noted that the decadal change in IO–WP SST differs from that in EP SST after the late 1990s. The change in IO–WP SST occurs in the global warming mode, while that in EP SST is exceedingly modulated by interdecadal Pacific oscillation (IPO) [Dai, 2012]. The detailed dynamics should be investigated in a future study.
 The IO warming causes the summer EP SST cooling by inducing an easterly anomaly in the WP [Kug and Kang, 2006], which leads to the coexistence of IO warming and EP cooling. Tropical EP cooling may change the North Pacific SST anomalies through the atmospheric bridge and, consequently, strengthen the extratropical NPH. The experimental results verify the distinct roles of EP cooling on the NPH and of IO warming and WP cooling on the WNPSH. The coupled pattern of NPH and WNPSH is significantly related to the combined effect of IO warming and EP cooling. In relation to the combined effect of WP cooling and EP warming, the strengthened WNPSH variability is shown. Considering the impact of EP SST anomaly on the NPH variability, the combination of WP cooling and EP warming may cause variability both in WNPSH and NPH. However, there is an inconsistency between numerical experiment and observation results for the NPH variability. This discrepancy might be a result of the lack of midlatitude SST variability in model experiments, as AGCM is forced by tropical SST variability. The combined effect of WP cooling and EP warming on the NPH variability should be investigated in the further study, using more detailed and careful numerical experiments. The WP SST anomaly, which shows a remarkable decadal change in background mean state, may contribute to some extent to the relationship between tropical SSTs and extratropical circulation [Moon et al., 2012]. However, the question of the relative role of WP SST anomaly on the extratropical circulation remains unsolved. In addition, response of midlatitude atmosphere to cool WP SST may be different from that to warm WP SST. The issue on asymmetry of WP SST forcing and effect of decadal change in background mean state will be explored in the near future.
 Despite the impact of tropical SSTs on the extratropics, strengthening of the extratropical NPH can induce changes in tropical SSTs [e.g., Yu et al., 2010]. However, it is difficult to confirm the detailed dynamical process of the tropical–extratropical relationship on the coupling process between EP SST and extratropical circulation over the North Pacific, which will be investigated in a future study, through coupled model experiment. This result provides new insights into the impact of tropical SSTs on extratropical summer climate. Therefore, this finding may help us to more effectively predict summer climates.
 This work was supported by GRL grant of the National Research Foundation (NRF) funded by the Korean Government (MEST 2011–0021927).