A number of studies have noticed that inverse relationship between the Indian summer monsoon (ISM) and El Niño–Southern Oscillation (ENSO) has weakened rapidly since the late 1970s (e.g. Webster and Palmer, 1997; Kumar et al., 1999; Chang et al., 2001). The explanations to such a phenomenon can be classified into three categories. Webster and Palmer (1997) attributed it to the chaotic nature of the ISM. Kumar et al. (1999) suggested a southeastward shift in the Walker circulation anomalies associated with ENSO events may lead to a reduced subsidence over the Indian region, thus favouring normal monsoon conditions. Besides that, they also suggested increased surface temperatures over Eurasia in winter and spring may favour the enhanced land-ocean thermal gradient conducive to a strong monsoon. Chang et al. (2001) believed the weakened ISM–ENSO relationship is most likely due to the strengthening and pole-ward shift of the jet stream over the North Atlantic, which led to the recent development of a significant correlation between winter time western European surface air temperatures and the ensuing ISM rainfall. This western European winter signal extended eastward over most of northern Eurasia, enhanced the meridional thermal contrast and disrupted the influence of ENSO on the ISM. This point has been further verified by other studies (e.g. Yang et al., 2004; Goswami et al., 2006).
As the other principal component of the Asian monsoon system, the East Asian summer monsoon (EASM) has many distinct features due to unique tectonic forcing: huge thermal contrasts between the world's largest continent, Eurasia, and the largest ocean basin, the Pacific, and is strongly influenced by the world's highest land feature, the Tibetan Plateau (Ding, 1992; Lau and Yang, 1997; Wang et al., 2008a; Wu et al., 2009). In contrast to the breakdown of the ISM–ENSO relationship, the EASM exhibits an enhanced relationship with ENSO (Wang et al., 2009; Li et al., 2010; Ding et al., 2010). Wang et al. (2008b) attributed such an interdecadal change in EASM–ENSO relationship to increased magnitude and periodicity of ENSO and the strengthened local monsoon–ocean interactions in low latitudes. Since the EASM has complex space and time structures that encompass tropics, subtropics, and mid-latitudes, its variability is influenced not only by tropical systems (Wu et al., 2006; Li et al., 2010), but also by systems from mid-high latitudes (i.e. North Atlantic Oscillation or NAO) (e.g. Ding and Wang, 2007; Wu et al., 2009, 2010).
The NAO has long been recognized as a major circulation pattern over northern mid-high latitudes influencing weather and climate from eastern North America to Europe (Hurrell, 1995). Recently, NAO has also been found to be able to influence the interannual variations of the EASM through changing the low boundary forcing over North Atlantic and the associated subpolar tele-connections (Wu et al., 2009). Nevertheless, it is still not clear whether the NAO–EASM relationship has an inter-decadal change or not in the past decades. If so, how does it change and to what extent does it contribute to the strengthened relationship between the EASM and ENSO? In this paper, we attempt to answer these questions. The main datasets used in this study include: (1) European Centre for Medium-Range Weather Forecasts 40-year reanalysis datasets (ERA-40; Uppala et al., 2005) and ERA-interim reanalysis datasets; (2) The Niño 3 sea surface temperature index calculated from the improved Extended Reconstructed sea surface temperature Version 2 (ERSST V2; Smith and Reynolds, 2004); (3) The EASM index (EASMI) is defined by the U850 averaged in (22.5°N–32.5°N, 110°E–140°E) minus U850 in (5°N–15°N, 90°E–130°E), where U850 denotes the zonal wind at 850 hPa (Wang et al., 2008a); (4) The NAO index used in this study (hereafter NAOI) is defined as the difference in the normalized monthly sea level pressure zonal-averaged over the North Atlantic sector from 80°W to 30°E between 35°N and 65°N (Li and Wang, 2003).
Sliding correlations on a 21-year moving window between the EASMI and preceding winter (December–Feburary, DJF) Niño 3 SST index have become significant since the 1970s, exceeding the 95% confidence level (Figure 1). It indicates an obvious strengthening relationship between the EASM and ENSO. It is interesting to notice that the EASM also exhibits an enhanced relationship with prior spring (March–May, MAM) NAO, their correlation coefficients having been beyond the 95% confidence level since 1985. If NAO signals were removed from the Niño 3 SST anomaly (SSTA), the ENSO–EASM relationship would break down with their correlation coefficients becoming insignificant (Figure 1). Such a breakdown implies that spring NAO might impact on the linkage between the EASM and ENSO.
The correlation between the EASMI and 500 hPa stream function during 1985–2008 (Figure 2) also displays a pronounced dipole pattern over North Atlantic, which is indicative of the strong relationship with NAO. Such a NAO pattern is not clear during 1958–1984. Another striking feature is that in Figure 2(b), an Atlantic-Eurasian wave-train controls the mid-high latitudes north of 40°N, extending from the North Atlantic to the Okhotsk Sea. Associated with this wave-train pattern, three positive geo-potential height anomaly centres are located at North Atlantic, the Ural Mountains and the Okhotsk Sea. Such a high-latitude tele-connection pattern is not clear during 1958–1984 (Figure 2(a)). The interdecadal changes in circulations over the northern Eurasia indicate that NAO has stronger and further downstream impacts. In addition, large areas of positive correlations prevail over the western Pacific and expand to the southern China, Indo-China Peninsular and the Bay of Bengal during 1985–2008 (Figure 2(b)), whereas during 1958–1984, the areas of positive correlations over the western Pacific are relatively smaller and weaker (Figure 2(a)). The expanding of the positive correlation areas over the western Pacific from 1958 to 2008 indicates that western Pacific subtropical high (WPSH) is strengthening and its linkage with NAO-like circulation anomaly is enhanced, too.
Wang et al. (2000, 2008a) attributed such a WPSH change to ‘prolonged’ impacts of ENSO or a ‘delayed’ response of the EASM to ENSO. They pointed out the critical role of the interaction between the off-equatorial moist atmospheric Rossby waves and the underlying SSTA in the western Pacific warm pool region. The moist Rossby wave–SST interaction can maintain both the Philippine Sea anticyclone (or WPSH) anomalies and the dipole-like SSTA in the western Pacific during the decaying El Niño. Thus, even if the SSTA disappears during the summer after the peak El Niño, the EASM remains to be affected by the WPSH anomalies significantly. In light of the strengthened linkage between the NAO-like circulation anomalies and the WPSH from 1958 to 2008 (Figure 2(a) and (b)), it is reasonable to speculate spring NAO might also affect the EASM via altering the WPSH.
Due to lack of persistence, the NAO can only prolong its influence through coupled mechanisms that involve low boundary forcing such as SST (Charney and Shukla, 1981; Shukla, 1998). Figure 2(c) and (d) show SST correlation patterns with the EASMI for 1958–1984 and 1985–2008, respectively. A pronounced feature during 1985–2008 is that the EASM is connected to a tripole SSTA pattern over North Atlantic. Previous studies proved that the tripole SSTA is coupled with atmospheric NAO-like anomalies (e.g. Wu et al., 2009). To quantitatively depict the tripole pattern during boreal summer, a simple tripole SST index (TSSTI) is defined as the difference between the sum of averaged SST in two positive correlation boxes and averaged SST in the negative correlation box (positive domain minus negative domain). The TSSTI displays a nearly consistent variability with the EASMI during 1985–2008 (not shown), their correlation coefficient reaching − 0.64 (beyond 95% confidence level). The tripole SSTA is not detectable during 1958–1984 (Figure 2(c)) and the correlation coefficient between the TSSTI and the EASMI is only − 0.2 for the 1958–1984 period.
The enhancement of such a tripole SSTA pattern over North Atlantic may result from the interdecadal change in spring NAO, which is supported by Figure 3. Figure 3 takes the summer TSSTI as a reference and computes the lead correlation with 500 hPa stream function for 1958–1984 and 1985–2008. In the earlier period (Figure 3(a) and (c)), NAO-like circulation anomalies are undetectable. In the latter period (Figure 3(b) and (d)), a notable NAO-like circulation anomaly, namely, a meridional dipole pattern, emerges over North Atlantic from − 2 month to − 1 month. This indicates that tripole SSTA in ensuing summer arises from spring NAO-like circulation anomalies.
How could the atmosphere respond to such a tripole SSTA forcing in summer in North Atlantic? Figure 4 presents 500 hPa zonal winds and stream function anomalies regressed to the summer TSSTI for 1958–1984 and 1985–2008. For the earlier period (Figure 4(a)), the mid-latitude jets are strong and shallow, whereas for the latter period (Figure 4(b)), the mid-latitude jets are weak and wide. This feature is particularly true over the central Eurasia. Associated with such an interdecadal change of jet streams, for the latter period, a pronounced Rossby wave train is excited over the northern Eurasia as responses to the tripole SSTA forcing in North Atlantic. The northern wave train propagates along the subpolar region with two anomalous high value centres over the Ural Mountain and the Okhotsk Sea. Its spatial pattern resembles an inverted East Atlantic Western Russia (EAWR) pattern (Barnston and Livezey, 1987). It raises (or weakens) pressure over these two regions which have important effects on the extra-tropical components of the EASM (e.g. Ding and Sikka, 2006; Wu et al., 2009).
Another notable circulation anomaly associated with the tripole SSTA is the anomalous WPSH over northwestern Pacific (Figure 4(b)). It implies that the tripole SSTA in North Atlantic might also contribute to the WPSH anomalies. Then, what kind of physical mechanism might be responsible for bridging the tripole SSTA and the WPSH anomalies? Kucharski i. (2009) found SSTAs in the tropical Atlantic can influence African and Indian monsoon rainfall and a Gill-Matsuno-type quadrupole response (Matsuno 1966; Gill 1980) is proposed to explain the tele-connection between the tropical Atlantic and the Indian basin. Whether this physical process also works for bridging the tripole SSTA and the WPSH largely depends on whether the tropical component of the tripole SSTA can extend its influence further to East Asia. To answer this question, we perform an idealized numerical experiment with a simplified general circulation model (SGCM) (Hoskins and Simmons, 1975; Lin and Derome, 1996). In the control run, the SGCM is forced by climatological SSTs. In the sensitivity experiments, we superimposed a heating (cooling) source in tropical Atlantic (its position centred in the bottom box in Figure 4(d)) on the climatological values to mimic a warm (cold) SSTA in tropical Atlantic during a high (low) TSSTI summer; each source has an elliptical squared cosine distribution in latitude and longitude with a vertically integrated heating rate of 1 K/day. To make the numerical results more robust, the control run was integrated for 12 years and the last 10 years' integration was used to derive a reference state. The two sensitivity tests were integrated for 10 years each. The 10-year integrations were used to construct a 10-member ensemble (arithmetic) mean to reduce the uncertainties arising from differing initial conditions.
Figure 5 presents the SGCM responses to the above tropical Atlantic forcing. The results of the numerical experiments are basically similar with those obtained by Kucharski et al. (2009). The atmosphere responds a Gill-Matsuno-type pattern. A warming (cooling) over the tropical Atlantic during a high (low) TSSTI summer induces a surface pressure (SP) gradient extending from the northeast to the southwest over Africa (Figure 5(a)). The vertical structure of the response is baroclinic with a positive (negative) SP and low-level stream function anomaly (Figure 5(a) and (b)) and a negative (positive) high-level stream function anomaly (Figure 5(c)) over western Pacific. This usually favours a strong (weak) WPSH anomaly (e.g. He et al., 2011). Some studies have already pointed out that a strong (weak) WPSH anomaly is usually associated with El Niño (La Niña) events (Wang et al., 2000). Therefore, the tropical component associated with the tripole SSTA tends to strengthen the linkage between the WPSH and ENSO. On the other hand, the anomalously high (low) pressure over the Okhotsk Sea, together with a strengthened (or weakened) and westward extended (or eastward withdrawn) WPSH, would favour an enhanced (or weakened) subtropical Meiyu-Baiu-Changma front, the primary rain-bearing system of the EASM, thus, leading to a strong (weak) EASM (Wang et al., 2008a).
3. Summary and Discussion
Most previous studies focus on the weakened ISM–ENSO relationship, while fewer studies investigate the strengthened EASM–ENSO relationship which is also of great importance. This work investigates contribution of prior spring NAO to the enhanced linkage between the EASM and ENSO from observational and numerical perspectives. The relevant physical processes can be summarized as following: The anomalous spring NAO induces a tripole SSTA pattern in North Atlantic which persists into ensuring summer and excites downstream development of a distinct Rossby wave train prevailing over the northern and a simple Gill-Matsuno-type quadrupole response over western Pacific. The former modulates the blocking highs over the Ural Mountain and the Okhotsk Sea, while the latter enhances the linkage between the WPSH and ENSO. The co-effects of the two tele-connection patterns help to strengthen (or weaken) the subtropical Meiyu-Baiu-Changma front, the primary rain-bearing system of the EASM. As such, spring NAO is tied to the strengthened connection between ENSO and the EASM.
This study focuses on the potential roles of the interdecadal change in spring NAO in strengthening the EASM-ENSO relationship. Some previous studies found that ENSO, itself, may exert remote influences on the NAO variability (e.g. Lin et al., 2002, 2005), while others believed that occurrences of particular modes of ENSO with those of the NAO seem to occur by chance (e.g. Hurrel, 1996; Rogers, 2000; Zanchettin et al., 2008). Therefore, whether the interdecadal change of the NAO arises from ENSO is still an open question.
Some recent papers have reported the influence of the tropical Atlantic SSTs (e.g. Atlantic Niño) on the Indian summer monsoon (Kucharski et al., 2007, 2008, 2009; Losada et al., 2010). It should be pointed out that many factors may influence the tropical Atlantic SST variability besides the NAO (Latif and Grötzner, 2000; Huang et al., 2004; Chang et al., 2006; Hu and Huang, 2007). The contribution from spring NAO provides another physical background to understand the strengthened EASM–ENSO relationship and to predict the long-term variations of the EASM.
Zhiwei Wu is supported by the National Basic Research Program ‘973’ of China (Grant No. 2010CB950401), the Sustainable Agriculture Environment Systems (SAGES) research initiative of Agriculture and Afri-Food Canada through the Natural Sciences and Engineering Research Council of Canada (NSERC) Fellowship Program and the Special Research Program for Public Welfare (Meteorology) of China under Grant No. GYHY200906016.