Abstract

The relationship between interannual variations of boreal winter North Atlantic Oscillation (NAO) and northern tropical Atlantic (NTA) sea surface temperature (SST) experienced obvious interdecadal changes during 1870–2012. Similar interdecadal changes are observed in the amplitude of NTA SST anomalies. The mean NTA SST change may be a plausible reason for several changes in the NAO–NTA SST connection. Under a higher mean NTA SST, NTA SST anomalies induce larger wind anomalies over the North Atlantic that produce a tripole SST anomaly pattern and amplify NTA SST anomalies. Comparison of the evolution of anomalies between 1970–86 and 1996–2012 unravels changing roles of El Niño–Southern Oscillation (ENSO) and extratropical atmospheric disturbances in the formation of NTA SST anomalies. During 1970–86, ENSO events play a key role in initiating NTA SST anomalies in the preceding spring through atmospheric circulation changes. With the decay of ENSO, SST anomalies in the midlatitude North Atlantic weaken in the following summer, whereas NTA SST anomalies are maintained up to winter. This leads to a weak NAO–NTA SST connection in winter. During 1996–2012, the preceding spring atmospheric circulation disturbances over the midlatitude North Atlantic play a dominant role in the genesis of a North Atlantic horseshoe (NAH)-like SST anomaly pattern in the following summer and fall. This NAH-like SST anomaly pattern contributes to the development of the NAO in late fall and early winter. The atmospheric circulation anomaly, in turn, is conducive to the maintenance of NTA SST anomalies to winter via changing surface latent heat flux and shortwave radiation. This leads to a close NAO–NTA SST connection in winter.

1. Introduction

Tropical Atlantic sea surface temperature (SST) variability displays several prominent modes, one of which is the northern tropical Atlantic (NTA) pattern (e.g., Enfield et al. 1999; Huang et al. 2004; Huang and Shukla 2005). The NTA mode features SST anomalies extending westward from the coast of northern Africa. Previous studies have demonstrated that NTA SST anomalies exert a substantial influence on precipitation, surface air temperature, and atmospheric circulation in the surrounding and remote regions (e.g., Folland et al. 1986; Uvo et al. 1998; Saravanan and Chang 2000; Cassou and Terray 2001; Cassou et al. 2004, 2005; Kushnir et al. 2010; Wu and Kirtman 2011; Wu et al. 2011; Hatzaki and Wu 2015). For example, Uvo et al. (1998) found that precipitation anomalies over northeast Brazil in April and May have a significant positive relation with SST anomalies in the NTA region. Kushnir et al. (2010) reported that precipitation tends to be less (more) than normal over most of the contiguous United States and Mexico when SST anomalies are positive (negative) in the NTA region. Wu et al. (2011) showed that a tripole SST anomaly pattern in the North Atlantic Ocean, featuring same-sign SST anomalies in the tropical and midlatitude North Atlantic Ocean and opposite-sign SST anomalies in the western subtropical North Atlantic Ocean, in preceding spring influences the interannual variability of summer temperature in northeast China. The NTA SST anomalies may also affect the Atlantic Niño (e.g., Zhu et al. 2012) and the Pacific El Niño–Southern Oscillation (ENSO) (e.g., Ham et al. 2013). Hence, a better understanding of the NTA SST variability as well as its factors is of great importance for improving the climate prediction in both the surrounding and remote regions.

The North Atlantic Oscillation (NAO) is the leading mode of atmospheric circulation variability over the North Atlantic Ocean (e.g., Walker 1924; Walker and Bliss 1932; Hurrell 1995; Hurrell and van Loon 1997). The spatial pattern of the NAO is characterized by a seesaw of sea level pressure (SLP) anomalies between two centers of action: one near Iceland and the other over the subtropical North Atlantic Ocean. A number of studies have demonstrated that weather and climate over the Northern Hemisphere are strongly influenced by the NAO (e.g., van Loon and Rogers 1978; Wallace and Gutzler 1981; Barnston and Livezey 1987; Hurrell 1995; Hurrell and van Loon 1997; Visbeck et al. 1998; Chang et al. 2001; Wu et al. 2009).

Studies found that the NAO influences the SST variability in the NTA region (e.g., Wallace et al. 1990; Tourre et al. 1999; Tanimoto and Xie 1999; Marshall et al. 2001; Walter and Graf 2002; Czaja et al. 2002). For instance, Czaja et al. (2002) showed that almost all strong SST anomalies in the NTA region from 1950 to 2000 can be accounted for by preceding ENSO or NAO events. Using a coupled ocean–atmosphere general circulation model, Huang and Shukla (2005) indicated that the NTA pattern in boreal winter–spring is closely associated with surface heat flux changes resulted from extratropical atmospheric disturbances, such as the NAO.

Previous studies have reported that the boreal winter NAO intensity and its variability display interdecadal changes (e.g., Hurrell 1995; Walter and Graf 2002). For example, Hurrell (1995) reported that the NAO index displays a significant increasing trend from the 1970s to the mid-1990s. Walter and Graf (2002) found that the boreal winter NAO index shows an enhanced decadal variability during the first three decades of the twentieth century and during the recent decades since the early-1970s. In contrast, the NAO index is characterized by a weak decadal variability from the 1930s to the early-1960s. As the influence of the NAO on NTA SST may depend on its intensity, an interesting question is whether the relationship between the NAO and NTA SST has changed in the past.

Our analysis shows that the connection between boreal winter NAO and NTA SST variations on an interannual time scale has experienced several marked interdecadal changes since 1870. The present study investigates the possible reasons for these interdecadal changes. It should be noted that the present study focuses on the relationship on interannual variations between NAO and NTA SST. The rest of the paper is organized as follows. Section 2 describes the datasets and analysis methods. Section 3 presents observational evidences for interdecadal changes in the boreal winter NAO–NTA SST relationship. Section 4 compares NTA SST-related anomalies of SST and atmospheric circulation in boreal winter and explores plausible reasons for interdecadal changes in the NAO–NTA SST relationship. Section 5 compares the evolution of NTA SST-related SST, wind, and surface heat flux anomalies between two recent periods with strong and weak NAO–NTA SST relationship. Section 6 gives a summary and discussion.

2. Data and methodology

This study uses monthly mean SST from the Met Office Hadley Centre Sea Ice and Sea Surface Temperature (HadISST) dataset (Rayner et al. 2003; http://www.metoffice.gov.uk/hadobs/hadisst/). This SST dataset has a horizontal resolution of 1° × 1° and is available from 1870 to present. The present study uses monthly mean horizontal winds at 850 hPa from the twentieth-century reanalysis (Compo et al. 2011) V2 data provided by the National Oceanic and Atmospheric Administration/Office of Oceanic and Atmospheric Research/Earth System Research Laboratory Physical Sciences Division (NOAA/OAR/ESRL PSD), Boulder, Colorado, from their website (http://www.esrl.noaa.gov/psd). The 850-hPa winds are available on 2° × 2° grids for the period 1871–2012. In addition, this study employs monthly mean horizontal winds at 850- and 200-hPa surface latent heat flux and surface net shortwave radiation from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis (Kalnay et al. 1996; ftp://ftp.cdc.noaa.gov/Datasets/). The horizontal winds are available on 2.5° × 2.5° grids, and surface latent heat flux and shortwave radiation are on T62 Gaussian grids from 1948 to the present.

The monthly NAO index used in this study is obtained from the web page of the Climate Data Guide of the University Corporation for Atmospheric Research (UCAR) (https://climatedataguide.ucar.edu/climate-data/hurrell-north-atlantic-oscillation-nao-index-station-based) from 1864 to 2013. This NAO index was developed by Hurrell (1995), which is defined as the difference of normalized SLP between Lisbon, Portugal, and Stykkisholmur/Reykjavik, Iceland (Hurrell 1995). The present analysis focuses on the interannual variability. Hence, monthly mean data and the NAO index are subjected to a 9-yr high-pass Lanczos filter (Duchon 1979).

3. Interdecadal changes in the NAO–NTA SST connection

Following previous studies (e.g., Huang et al. 2004; Huang and Shukla 2005), we perform a rotated empirical orthogonal function (REOF) analysis to obtain the dominant modes of boreal winter [December–February or D(0)JF(+1)] tropical Atlantic SST anomalies for the period 1950–2012. Similar results are obtained when the analysis period is extended back to 1870. Here, 1950 D(0)JF(+1) refers to the boreal winter of 1949/50. The area we chose for REOF analysis extends from 30°S–30°N to 100°W–20°E in the Atlantic Ocean. As in Huang and Shukla (2005), 10 leading unrotated EOF modes, which explain about 89% of the total variance, are first calculated from the interannual anomalies of D(0)JF(+1) tropical Atlantic SST. Then, a varimax rotation is applied to these 10 leading unrotated EOF modes to obtain 10 leading rotated modes and their corresponding principle component (PC) time series. The above REOF analysis has been applied to original, decadal, and interannual SST anomalies, respectively. The three leading modes display very similar patterns though the order of these modes is different. Here, we only discuss the modes of interannual anomalies.

Our REOF analysis obtained leading modes similar to Huang et al. (2004) and Huang and Shukla (2005). The first REOF mode explains 18.7% of the total variance. It features SST anomalies extending from the Angola coast to the central equatorial Atlantic Ocean (not shown), resembling the southern tropical Atlantic (STA) pattern (Huang et al. 2004). The second REOF mode explains 17.4% of the total variance. This mode is dominated by SST anomalies centered at southern subtropical Atlantic (not shown) and is referred to as the southern subtropical Atlantic (SSA) pattern (Huang et al. 2004). The third REOF mode explains 16.1% of the total variance. This mode represents SST anomalies in the northern tropical Atlantic Ocean with the largest loading near the African coast (Fig. 1a) and is referred to as the NTA pattern (Huang et al. 2004). Based on the spatial distribution of loading in Fig. 1a, a NTA SST index is defined as area-averaged interannual SST anomalies over the region of 5°–25°N and 15°–60°W. The correlation coefficient between this NTA SST index and the PC time series corresponding to the third REOF mode (Fig. 1b) is as high as 0.97. Note that the amplitude of the PC time series is relatively small during the 1970s and 1980s compared to the other periods.

The statistical relationship between interannual variations of winter NAO and NTA SST appears unsteady. The changing relationship is obvious in Fig. 2 that displays the sliding correlation between the D(0)JF(+1) NAO and NTA SST with a 17-yr window. Note that years in Fig. 2 are labeled based on the central year of the 17-yr window. The correlation is negative and statistically significant at the 95% confidence level during the 1880s, the late 1920s, the 1950s, and after the early-1990s, while the correlation is weak during the 1900s to the mid-1910s, the mid-1930s to mid-1940s, and the 1970s to mid-1980s. We have calculated the sliding correlation for windows of 15-, 19-, 21-, and 23-yr (figures not shown) to examine the sensitivity of the correlation to the length of window. Although the correlation value at a specific year varies depending upon the length of window, the periods of high and low correlation are consistent among different windows.

In the following, we compare differences in SST and wind anomalies between high and low correlation periods to understand the change in the NAO–NTA SST relationship. To avoid overlapping and at the same time to achieve large contrast of the correlation between the neighboring periods, we select the high and low correlation periods based on the sliding correlation with a 17-yr window. We obtain four high correlation periods (1878–94, 1918–34, 1950–66, and 1996–2012) and three low correlation periods (1901–17, 1934–50, and 1970–86). The correlation coefficients are −0.55 (1878–94), −0.58 (1918–34), −0.63 (1950–66), −0.67 (1996–2012), 0.14 (1901–17), −0.19 (1934–50), and 0.17 (1970–86). According to the Fisher’s r-to-z transformation, the correlation during 1970–86 is significantly different from that during 1950–66 and 1996–2012 at the 95% confidence level, and the correlation during 1901–17 is significantly different from that during 1878–94 and 1918–34 at the 95% confidence level. The difference of the correlation during 1934–50 from that during 1918–34 and 1950–66, however, is insignificant.

The winter NAO-related SST anomalies show notable difference between the high and low correlation periods. Figure 3 displays interannual SST anomalies in D(0)JF(+1) as obtained by regression on the normalized D(0)JF(+1) NAO index for different periods. During 1878–94, 1918–34, 1950–66, and 1996–2012, a tripole SST anomaly pattern is present in the North Atlantic Ocean, with significant negative anomalies in the tropics and high latitudes and significant positive anomalies in the midlatitudes (Figs. 3a,c,e,g). During 1901–17, the SST anomalies are weak in the North Atlantic Ocean (Fig. 3b). During 1934–50, significant SST anomalies in the subtropics and midlatitudes are confined to the western part (Fig. 3d). During 1970, negative SST anomalies are confined to a narrow region in the eastern subtropics and significant positive SST anomalies are seen in the midlatitudes of the North Atlantic (Fig. 3f). However, SST anomalies are very weak in the NTA region. The above results suggest a strong connection during 1878–94, 1918–34, 1950–66, and 1996–2012, but a weak connection of the winter NAO with the NTA SST anomalies during 1901–17, 1934–50, and 1970–86.

The winter NAO-related atmospheric circulation anomalies display notable differences between the high and low correlation periods. Figure 4 shows 850-hPa wind anomalies in D(0)JF(+1) obtained by regression upon the normalized D(0)JF(+1) NAO index for different periods. Pronounced anticyclonic wind anomalies are seen over the subtropical North Atlantic and cyclonic wind anomalies are observed over the high latitudes for all the periods. Correspondingly, significant westerly wind anomalies appear over the midlatitudes and significant easterly wind anomalies over the high latitudes and subtropics of the North Atlantic. However, cyclonic wind anomalies near Iceland appear weaker during 1901–17 (Fig. 4b), and easterly wind anomalies over the subtropics are weaker and shift northward during 1934–50 and 1970–86 (Figs. 4d,f) compared to those during 1878–94, 1918–34, 1950–66, and 1996–2012 (Figs. 4a,c,e,g). Furthermore, large easterly wind anomalies over the subtropical North Atlantic extend more westward during the low correlation periods compared to those during the high correlation periods.

4. Plausible reasons for interdecadal changes in the NAO–NTA SST relationship

a. NTA SST-related anomalies of SST and atmospheric circulation

To help understand why the connection between the wintertime NTA SST and NAO index has experienced interdecadal changes during the analysis period, we first compare NTA SST-related anomalies between high and low correlation periods. Figures 5 and 6 shows anomalies of SST and winds at 850 hPa, respectively, in D(0)JF(+1) obtained by regression with respect to the normalized D(0)JF(+1) NTA SST index for different periods.

Significant and positive SST anomalies are seen in the NTA region in all the periods. However, SST anomalies display notable differences in both the subtropics and midlatitudes of the North Atlantic Ocean. During 1878–94, 1918–34, 1950–66, and 1996–2012, significant negative and positive SST anomalies are observed in the subtropics and midlatitudes, respectively (Figs. 5a,c,e,g). This indicates that the NTA SST anomalies have a strong connection with the SST anomalies in the subtropics and midlatitudes of the North Atlantic Ocean during the high correlation periods. During 1901–17, positive SST anomalies in the midlatitudes are mostly insignificant and shifted much farther to the east (Fig. 5b). During 1970–86, SST anomalies in the subtropics are less significant than in the high correlation periods (Fig. 5f). During 1934–50, the SST anomalies in the subtropics and midlatitudes are similar to those in the high correlation periods, but appear less significant in the midlatitudes (Fig. 5d).

The D(0)JF(+1) NTA SST-related wind anomalies at 850 hPa over the North Atlantic Ocean display pronounced differences between high and low correlation periods. During the high correlation periods, significant cyclonic wind anomalies are seen over the midlatitudes and significant anticyclonic wind anomalies are observed over the high latitudes although anomalous westerlies over the high latitudes do not seem so significant (Figs. 6a,c,e,g). This anomalous circulation pattern is similar to that associated with the NAO. It indicates that NTA SST anomalies have a close connection with the NAO during the high correlation periods. During 1901–17, the wind anomalies feature an obvious east–west contrast over the midlatitudes, with an anomalous cyclonic circulation over the ocean and an anomalous anticyclonic circulation over the east coast (Fig. 6b). During 1934–50, although a south–north contrast of wind anomalies is seen over the North Atlantic, the wind anomalies appear weaker than those in the high correlation periods (Fig. 6d). During 1970–86, the wind anomalies are mostly insignificant and feature a wave pattern that appears to be a part of internal atmospheric variability and does not seem to be connected to the NTA SST anomalies (Fig. 6f). In addition, it should be mentioned that the westerly anomalies in the subtropics are weaker and of smaller spatial extent during the low correlation periods than those during the high correlation periods (Fig. 6). The wind anomaly pattern at 200 hPa is similar to that at 850 hPa over the mid- and high latitudes in both high and low correlation periods (figures not shown), indicative of a barotropic structure.

During the high correlation periods, the North Atlantic SST anomaly pattern bears a close resemblance to the tripole SST anomaly pattern identified by previous studies (e.g., Watanabe et al. 1999; Marshall et al. 2001; Czaja and Frankignoul 2002; Wang et al. 2004; Pan 2005). Previous studies have demonstrated that the variability of winter NAO is closely connected with a tripole SST anomaly pattern in the North Atlantic Ocean (e.g., Wallace et al. 1990; Rodwell et al. 1999; Marshall et al. 2001; Czaja and Frankignoul 1999, 2002; Saunders and Qian 2002; Cassou et al. 2004; Wu et al. 2009; Wu et al. 2011). For example, Czaja and Frankignoul (2002) reported that a North Atlantic horseshoe SST anomaly pattern, with opposite SST anomalies between southeast of Newfoundland and near Ireland and the Morocco coast, in boreal summer and fall exert a significant influence on the following early winter NAO variability. The NAO-related atmospheric circulation anomalies, in turn, induce a tripolelike SST pattern in winter in the North Atlantic by modulating surface heat flux and oceanic advection (e.g., Wallace et al. 1990; Cayan 1992; Marshall et al. 2001; Visbeck et al. 2003). Hence, strong NAO–NTA SST connection may imply a close relationship between the tripolelike SST anomaly pattern over the North Atlantic and the NAO-like atmospheric anomalies.

The correspondence of SST and the wind anomaly pattern over the North Atlantic indicates a coupled variation during the high correlation periods. On one hand, positive NTA SST anomalies may induce an anomalous cyclonic circulation over the subtropics via a Rossby wave–type response. On the other hand, the anomalous cyclonic circulation may enhance oceanic upwelling and increase cloudiness. This may contribute to the SST cooling in the subtropics. The westerly wind anomalies to the south side reduce climatological northeasterly trade winds, suppressing surface evaporation. This provides a positive feedback to the NTA SST change. The easterly wind anomalies to the north side reduce climatological westerly winds, suppressing surface evaporation and inducing warm advection by anomalous northward Ekman current. These may contribute to the SST increase in the midlatitude North Atlantic Ocean. Thus, the wind effects on the ocean favor the development of the tripole SST anomaly pattern in the North Atlantic Ocean. The tripolelike SST anomaly pattern, in turn, may contribute to NAO-like atmospheric anomalies through transient eddy feedback processes (e.g., Peng et al. 2003; Cassou et al. 2004; Pan 2005). During 1901–17, no clear easterly wind anomalies are seen over the midlatitudes (Fig. 6b). Correspondingly, no positive SST anomalies develop to the east of Canada (Fig. 5b). During 1970–86, wind anomalies over the most of the North Atlantic and negative SST anomalies in the subtropics are small and insignificant (Figs. 5f and 6f). During 1934–50, the relatively less significant SST anomalies in the midlatitudes (Fig. 5d) appear to be consistent with relatively weak wind anomalies there (Fig. 6d).

b. Possible reasons for interdecadal changes in the NAO–NTA SST relationship

What are the plausible reasons for the interdecadal changes in the NAO–NTA SST relationship? The above analysis suggests that the NAO–NTA SST relationship reflects a coupling between the atmosphere and ocean. The atmospheric influence on the SST change depends on the magnitude of wind anomalies. The atmospheric wind response, in turn, depends on the magnitude of SST anomalies as well as the mean SST state. Previous studies have suggested that SST anomalies in the tropical Atlantic Ocean can be generated by remote forcing (e.g., ENSO, NAO) or triggered by regional air–sea interaction (e.g., Huang et al. 2002; Wu and Liu 2002; Huang et al. 2004). The influences of ENSO and NAO may depend on their intensity and the change in the mean state. In the following, we analyze the change in the relation of NTA SST anomalies and NAO with ENSO, the change in the amplitude of ENSO, NAO and NTA SST anomalies, and the change in the mean SST to understand the reasons for the change in the NAO–NTA SST relationship. The Niño-3.4 SST index, which is defined as area-averaged SST anomaly over the region of 5°S–5°N, 170°–120°W, is used to represent ENSO.

Would the influence of tropical Pacific SST anomalies play a role in the interdecadal changes in the NAO–NTA SST connection? Previous studies have demonstrated that SST anomalies in tropical central-eastern Pacific induce SST anomalies in the NTA region one to two seasons later via an atmospheric teleconnection pattern from the equatorial Pacific across North America to the North Atlantic or via an anomalous Walker circulation (e.g., Hastenrath 1984; Curtis and Hastenrath 1995; Nobre and Shukla 1996; Enfield and Mayer 1997; Klein et al. 1999; Saravanan and Chang 2000; Alexander et al. 2002; Czaja et al. 2002; Huang et al. 2004; Wu and Zhang 2010). Positive (negative) SST anomalies tend to be observed in the tropical North Atlantic when El Niño (La Niña) events occur in the tropical Pacific. The correlation between DJF Niño-3.4 SST and DJF NTA SST or NAO displays interdecadal changes during the analysis period (Fig. 7a). These changes, however, differ largely from those seen in the NAO–NTA SST correlation. The standard deviation of the DJF Niño-3.4 SST index is higher after the late 1960s and before 1900 than during the period in between (Fig. 7b). As such, the change in the amplitude of ENSO is different from the change in the NAO–NTA SST correlation. The mean DJF Niño-3.4 SST does not show changes that match the NAO–NTA SST correlation changes (Fig. 7c). Thus, the interdecadal changes in the NAO–NTA SST correlation cannot be explained by the ENSO variability.

Would the change in the amplitude of NAO and/or NTA SST anomalies be a factor in the interdecadal changes in the connection between NAO and NTA SST variations? Indeed, the 17-yr sliding standard deviation of DJF NAO and NTA SST shows interdecadal changes opposite to the NAO–NTA SST correlation changes (Fig. 7b). The standard deviation of DJF NTA SST is relatively high during the 1890s, the 1920s–mid-1930s, around 1960, and after the early-1990s, but relatively low during the 1900s to the early-1910s, the 1940s, and the 1970s–80s. The difference of DJF NTA SST standard deviation between 1878–94 (1918–34) and 1901–17 is significant at the 90% (95%) confidence level, and the difference between 1996–2012 and 1970–86 is significant at the 95% confidence level according to the F test. Standard deviation of DJF NAO during 1996–2012 (1878–94) is significantly larger than that during 1970–86 (1901–17) at the 95% confidence level. The correlation coefficient between the standard deviation of DJF NTA SST and NAO with the DJF NAO–NTA SST correlation is −0.80 and −0.48, respectively, for the whole analysis period. The former is significant at the 95% confidence level assuming a degree of freedom of 5 for a 17-yr sliding window. The latter, however, is insignificant with the same degree of freedom. The lower correlation for the NAO standard deviation compared to the NTA SST standard deviation appears to be related to the shift of low standard deviation period during the 1940s–50s (Fig. 7b). Thus, the change in the amplitude of the NTA SST anomalies and, to a lesser extent, NAO anomalies may be a dominant factor in the interdecadal changes in the NAO–NTA SST relationship.

The atmospheric response to SST anomalies may depend not only on the amplitude of the SST anomalies, but also on the mean SST. The mean NTA SST displays noticeable differences between some of the high and low correlation periods. The mean NTA SST during the 1900s–10s and during the late 1970s–80s is lower than the neighboring periods (Fig. 7c). The mean NAO tends to show long-term change opposite to that of mean NTA SST. Note that the mean NTA SST change is small during the 1930s–50s. As pointed earlier, the difference of NAO–NTA SST correlation during 1934–50 from that during 1918–34 and 1950–66 is insignificant.

To further demonstrate the changes in the mean SST, we show in Fig. 8 mean SST difference between high and low correlation periods. The linear trends of SST have been removed before constructing the difference maps since our focus is on interdecadal change, not on long-term trend. The mean SST during 1996–2012 and 1950–66 is significantly higher than that during 1970–86 by about 0.3°–0.5° and 0.2°–0.4°C, respectively, in the tropical North Atlantic Ocean (Figs. 8c,d). The mean SST during 1878–94 is higher than that during 1901–17 by about 0.2°–0.4°C in eastern tropical North Atlantic Ocean though the difference is not significant (Fig. 8a). The difference of mean SST between 1918–34 and 1901–17 is relatively small in the tropical North Atlantic Ocean (Fig. 8b). Thus, the mean SST in the NTA region is higher during some (though not all) of the high NAO–NTA SST correlation periods compared to the low correlation periods. Note that the mean SST differences may be subject to uncertainty because of spare observations in early years.

The above analysis shows that there are differences in the mean SST and the amplitude of SST anomalies in the NTA region between some of the high and low correlation periods. The difference in the amplitude of the NTA SST anomalies may be connected to the difference in the NTA mean SST. In addition, the difference in the amplitude of the NTA SST anomalies may also be connected to the difference in the amplitude of the NAO anomalies. Given initial NTA SST anomalies, the atmospheric wind response over the North Atlantic would be larger under a higher mean NTA SST due to the nonlinearity of dependence of atmospheric heating on mean temperature. The stronger wind anomalies, in turn, may act on the ocean, generating SST anomalies in the subtropics and midlatitudes. This may lead to the formation of a tripole SST anomaly pattern in the North Atlantic Ocean. At the same time, the NTA SST anomalies may be amplified through a positive wind–evaporation feedback. The atmosphere–ocean coupling would favor the development and maintenance of the NAO. Thus, the mean NTA SST change may be a reason for the change in the amplitude of the NTA SST anomalies and the NAO–NTA SST relationship.

5. Contrast of evolution of anomalies between 1970–86 and 1996–2012

The previous section focuses on DJF SST and wind anomalies. We have examined evolution of SST and wind anomalies during different high and low correlation periods. In the following, we present a comparison of the SST and wind anomalies between 1970–86 and 1996–2012 to gain a better understanding of the origin of NTA SST anomalies. These two periods show different roles of ENSO and regional atmospheric disturbances in the initiation of NTA SST anomalies. Figure 9 shows the evolution of monthly mean NTA and Niño-3.4 SST anomalies obtained by regression on the D(0)JF(+1) NTA SST index for 1970–86 and 1996–2012, respectively.

Both NTA SST and Niño-3.4 SST anomalies show different evolution between 1970–86 and 1996–2012. During 1970–86, positive NTA SST anomalies develop in previous winter, weaken in summer, reintensify in autumn, and decay in the following spring–summer (Fig. 9a). During 1996–2012, positive NTA SST anomalies develop in previous spring, intensify in autumn, and decay in the following spring–summer (Fig. 9a). Thus, the NTA SST anomalies develop earlier during 1970–86 than during 1996–2012. During 1970–86, the Niño-3.4 SST anomalies are significantly positive before May in the previous year [May(0)] and become weak afterward (Fig. 9b). Together with Fig. 9a, it suggests that tropical Pacific SST anomalies during the previous winter may contribute to the positive NTA SST anomalies in the previous spring during 1970–86 via atmospheric teleconnection (e.g., Klein et al. 1999). During 1996–2012, the Niño-3.4 SST anomalies are positive but insignificant in previous winter. This indicates that the influence of tropical Pacific SST anomalies on the genesis of NTA SST anomalies is weak in the latter period. The positive Niño-3.4 SST anomalies increase from the previous fall to winter and decrease after April in the following year (Fig. 9b). It implies that SST anomalies in the tropical Pacific may have a supplementary contribution to the NTA SST anomalies in the ENSO decaying year.

To understand further the origins of the SST anomalies in the NTA region, we contrast evolutions of SST and wind anomalies between 1970–86 and 1996–2012. Figures 10 and 11 show evolutions of SST and 850-hPa wind anomalies obtained by regression on the D(0)JF(+1) NTA SST index for 1970–86 and 1996–2012, respectively. The above figures are displayed as consecutive 3-month means with an interval of 2 months. To help understand the evolutions of SST anomalies in the NTA region, we also display in Fig. 12 the evolution of area-mean surface latent heat flux and shortwave radiation anomalies averaged over the NTA region for the two periods, respectively. Surface sensible heat flux anomalies are much smaller than latent heat flux anomalies and longwave radiation anomalies are smaller (with opposite sign) than shortwave radiation anomalies, and thus they are not shown. The convention for surface heat fluxes is that positive fluxes act to warm the ocean surface. Note that the 850-hPa winds and surface heat fluxes in Figs. 1012 are from the NCEP–NCAR reanalysis.

Remarkable differences are seen in the evolution of SST and wind anomalies between the two periods. During 1970–86, significant and positive SST anomalies are observed in the tropical central-eastern Pacific in D(−1)JF(0) and (February–April) FMA(0) (Figs. 10a,b). These positive SST anomalies weaken in (April–June) AMJ(0) and almost disappear in (June–August) JJA(0) (Figs. 10c,d). The evolution of positive SST anomalies in the tropical central-eastern Pacific features a decaying El Niño event (Figs. 10a–d). The North Atlantic SST anomalies are weak in D(−1)JF(0) (Fig. 10a). A wave train–like teleconnection pattern, resembling the Pacific–North Atlantic (PNA) or the tropical–Northern Hemisphere (TNH) teleconnection pattern (Wallace and Gutzler 1981; Barnston and Livezey 1987; Lee et al. 2008), is observed from the tropical Pacific across North America to the North Atlantic and an anomalous cyclonic circulation is seen over subtropical North Atlantic in D(−1)JF(0) and FMA(0) (Figs. 10a,b). The southwesterly wind anomalies in the south side of the anomalous cyclonic circulation weaken the climatological northeasterly trade winds and reduce upward surface latent heat flux (Fig. 12a). This contributes to the development of positive SST anomalies in the tropical North Atlantic in FMA(0) (Fig. 10b). The anomalous cyclonic circulation over the subtropics may increase cloudiness and reduce downward shortwave radiation and induce anomalous upwelling. These processes contribute to negative SST anomalies in the subtropical North Atlantic Ocean. Warm SST anomalies in the western part of the midlatitude North Atlantic Ocean may be attributed in part to reduced upward surface latent heat flux as easterly wind anomalies weaken the surface wind speed. Associated with the decay of El Niño events (Figs. 10d,e), wind anomalies over the North Atlantic Ocean weaken, and so are surface heat flux anomalies, resulting in the loss of the tripole SST anomaly pattern in summer (Figs. 10d,e). SST anomalies in the NTA region are maintained through (October–December) OND(0) and D(0)JF(+1) though weaken slightly after summer likely due to negative but small latent heat flux and shortwave radiation anomalies (Fig. 12a) in relation to weak wind anomalies (Figs. 10d,e).

During 1996–2012, SST anomalies in the equatorial central-eastern Pacific are weak in D(−1)JF(0) (not shown) and FMA(0) (Fig. 11a). Correspondingly, the NTA SST anomalies are small during the preceding spring (Fig. 11b). An anomalous cyclonic circulation, which is likely due to internal atmospheric variability, is seen over the midlatitude North Atlantic in FMA(0) (Fig. 11a). This anomalous cyclonic circulation shifts eastward from AMJ(0) to (August–October) ASO(0) (Figs. 11b–d). The anomalous southwesterly winds on the southeastern side of the anomalous cyclonic circulation weaken climatological northeasterly winds, reducing upward surface latent heat flux (Fig. 12b). The anomalous easterly winds over the midlatitudes weaken climatological westerly winds, reducing upward surface latent heat flux as well as inducing northward Ekman transport (Marshall et al. 2001; Visbeck et al. 2003). Anomalous cyclonic circulation may induce upwelling and increase cloudiness, contributing to the SST cooling in the subtropics (Figs. 11c–e). These processes lead to the development of a North Atlantic horseshoe (NAH)-like SST anomaly pattern (Czaja and Frankignoul 1999, 2002; Hu and Huang 2006a,b), with colder SST in the midlatitudes and warmer SST in the eastern subtropics, along the east coast, and the high latitudes (Figs. 11b–d).

Previous studies (Czaja and Frankignoul 1999, 2002; Hu and Huang 2006a,b) have demonstrated that the NAH SST anomaly pattern in preceding summer and fall can be an effective precursor for following early winter NAO. Using an atmospheric general circulation model (AGCM), Cassou et al. (2004) indicated the influence of preceding summer NAH SST anomaly pattern on early winter NAO variability through changes in the eddy–mean flow interaction from October onward. Hence, the NAH SST anomaly pattern identified in AMJ(0), JJA(0), and ASO(0) (Figs. 11b–d) during 1996–2012 may contribute to the formation of NAO-like atmospheric circulation anomalies in late fall and early winter (Figs. 11e,f). The NAO in early winter further contributes to the southward development and westward extension of subtropical warm SST anomalies (Figs. 11e–g) via reducing upward surface latent heat flux and increasing downward shortwave radiation (Fig. 12b). The NTA SST anomalies start to decay after FMA(+1) (Figs. 11g,h), which is attributed to the reversal of both latent heat flux and shortwave radiation anomalies (Fig. 12b).

The above results suggest that, during 1996–2012, an internal atmospheric disturbance over the midlatitude North Atlantic in the preceding spring results in the formation of a NAH-like SST anomaly pattern in the following summer through changes in surface latent heat flux and oceanic processes. The NAH-like SST anomaly pattern induces a NAO-like atmospheric circulation anomaly in late fall and early winter over the North Atlantic (Czaja and Frankignoul 2002). The NAO-like atmospheric circulation anomaly, in turn, contributes partly to the maintenance of the NTA SST anomalies in the following winter. The above processes can explain the strong connection of the NAO with NTA SST in the simultaneous winter during 1996–2012. In contrast, during 1970–86, the formation of NTA SST anomalies in spring is attributed mainly to the preceding winter ENSO events. Extratropical atmospheric circulation anomalies appear to contribute little to the development of NTA SST anomalies from spring to the following winter. Hence, the winter NTA SST variability has a weak connection with the NAO variability during 1970–86.

The evolution of SST and wind anomalies during other periods has been compared to that during 1970–86 and 1996–2012. The development of NTA SST anomalies during 1950–66 displays features similar to that during 1996–2012 with weak ENSO effect. One difference from 1996–2012 is that the development of NTA SST anomalies is delayed by 1–2 months. During 1878–94 and 1918–34, the development of positive NTA SST anomalies follows El Niño events that start in summer and decay in the succeeding spring. Thus, ENSO may play a role in the initiation and maintenance of NTA SST anomalies during these two periods. In addition, significant cyclonic wind anomalies are observed over the midlatitude North Atlantic in the preceding spring during 1878–94. This indicates that extratropical atmospheric disturbance may also be important in the development of NTA SST anomalies during this period. Compared to 1996–2012, the development of NTA SST anomalies is delayed to September during 1878–94 and November during 1918–34. During 1901–17, significant positive Niño-3.4 SST anomalies develop in summer and decay in the succeeding spring. Thus, ENSO may contribute to the wind anomalies over the North Atlantic. However, the wind anomaly pattern in winter differs from the NAO, as pointed out in section 4. During 1934–50, positive Niño-3.4 SST anomalies develop in summer and decay in the succeeding spring, and positive NTA SST anomalies and the anomalous cyclonic circulation over the subtropical North Atlantic start in November. From the temporal relation of Niño-3.4 SST and NTA SST anomalies, the ENSO effect in the formation of NTA SST anomalies is more common in the periods before than after 1950 (Fig. 7a).

6. Summary and discussion

The present study has identified obvious interdecadal changes in the relationship between interannual variations of boreal winter NAO and NTA SST since 1870. It is found that the NAO has a strong connection with the NTA SST anomalies during the 1880s, the late 1920s, the 1950s, and after the early-1990s. In contrast, the connection between the NAO and NTA SST anomalies is weak during the 1900s to mid-1910s, the mid-1930s to mid-1940s, and the 1970s to mid-1980s.

The interdecadal change in the NAO–NTA SST connection appears not related to the change in the ENSO variability. The amplitude of NTA SST anomalies shows interdecadal changes opposite to the NAO–NTA SST correlation changes during the analysis period. The amplitude of NAO also shows interdecadal changes opposite to the NAO–NTA SST correlation changes except for the period during the 1940s–50s. The mean SST in the NTA region is higher during the 1880s than during the 1900s–10s, during the 1950s–60s and the mid1990s–2000s than during the 1970s–80s. Thus, the mean NTA SST change appears to be a reason for several interdecadal changes in the NAO–NTA SST relationship. Under a warmer background, the NTA SST anomalies, after they are initiated, may induce larger wind response over the North Atlantic. The impacts of the wind changes on the ocean lead to the formation of a tripole SST anomaly pattern in the North Atlantic Ocean and amplify the NTA SST anomalies through a positive wind–evaporation feedback. The tripolelike SST anomaly pattern over the North Atlantic, in turn, contributes to NAO-like atmospheric anomalies, as demonstrated by previous studies (e.g., Peng et al. 2003; Cassou et al. 2004; Pan 2005). This may explain why a strong NAO–NTA SST connection is more likely to occur during periods with higher mean SST than with lower mean SST in the NTA region.

The initiation of the NTA SST anomalies is related to different factors between 1970–86 and 1996–2012. During 1970–86, ENSO events in the preceding winter may lead to the formation of NTA SST anomalies in the following spring via atmospheric circulation changes over the North Atlantic. Following the decay of ENSO after the spring, wind anomalies over the midlatitude North Atlantic weaken and the associated SST anomalies decrease. The tropical Atlantic SST anomalies are maintained until winter though there is a weakening in the magnitude. As such, the connection of NTA SST and NAO is weak during this period. In contrast, during 1996–2012, a NAH-like SST anomaly pattern forms in summer, which is attributed to an atmospheric disturbance over the midlatitude North Atlantic in the preceding spring. This NAH-like SST anomaly pattern may force a NAO-like atmospheric circulation anomaly pattern in the late fall and early winter over the North Atlantic. Atmospheric circulation anomalies further contribute to the maintenance of the NTA SST anomalies in the following winter. Hence, boreal winter NAO has a strong connection with the NAT SST during the latter period.

Previous studies have demonstrated that the SST variability in the NTA region is influenced by ENSO and extratropical atmospheric disturbance (e.g., NAO) (e.g., Wu and Liu 2002; Huang et al. 2002; Huang et al. 2004; Huang and Shukla 2005). However, discrepancies exist in previous studies regarding the relative roles of the extratropical atmospheric disturbance and ENSO in the NTA SST variability. For example, Wu and Liu (2002) argued that the NAO is not a necessary precondition for the genesis of tropical Atlantic variability. Huang and Shukla (2005) indicated that the NTA SST anomalies in boreal winter–spring are usually associated with surface heat flux changes that result from extratropical atmospheric disturbances, such as the NAO. Huang et al. (2002) showed that SST anomalies in the northern tropical Atlantic Ocean are significantly influenced by ENSO. Huang et al. (2004) reported that ENSO is the leading external source in interannual variability of tropical Atlantic SST. The present study suggests that the relative roles of ENSO and extratropical atmospheric disturbances over the midlatitude North Atlantic in the formation of the NTA SST anomalies may vary with time period. The ENSO in the preceding winter plays an important role in initiating the NTA SST anomalies in the following spring during 1970–86, whereas an extratropical atmospheric disturbance in the preceding spring is dominant in generating the NTA SST anomalies during 1996–2012. The finding of the present study provides a clue for understanding discrepancies in previous studies regarding the roles of extratropical atmospheric disturbance and ENSO in the formation of the NTA SST anomalies.

As the NAO–NTA SST connection changes, the circulation pattern over downstream Eurasia may differ among strong and weak NAO–NTA SST connection periods. This may affect the NTA SST-related climate anomalies over Eurasia. Our preliminary analysis shows notable differences in the NTA SST-related precipitation and surface air temperature anomalies over many regions of Eurasia and North America between 1970–86 and 1996–2012. Further studies are needed to understand these differences and the roles of atmospheric circulation changes in connecting the NTA SST anomalies to precipitation and surface air temperature anomalies in various regions, which may help to improve the seasonal climate prediction.

Acknowledgments

The comments of three anonymous reviewers have led to a significant improvement of this paper. This study is supported by the National Key Basic Research Program of China Grant 2014CB953902, the National Natural Science Foundation of China Grants 41275081 and 41230527, and the Hong Kong Research Grants Council Grant CUHK403612.

REFERENCES

REFERENCES
Alexander
,
M. A.
,
I.
Bladé
,
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:.
Barnston
,
A. G.
, and
R. E.
Livezey
,
1987
:
Classification, seasonality, and persistence of low-frequency atmospheric circulation patterns
.
Mon. Wea. Rev.
,
115
,
1083
1126
, doi:.
Cassou
,
C.
, and
L.
Terray
,
2001
:
Dual influence of Atlantic and Pacific SST anomalies on the North Atlantic/Europe winter climate
.
Geophys. Res. Lett.
,
28
,
3195
3198
, doi:.
Cassou
,
C.
,
C.
Deser
,
L.
Terray
,
J. W.
Hurrell
, and
M.
Drévillon
,
2004
:
Summer sea surface temperature conditions in the North Atlantic and their impact upon the atmospheric circulation in early winter
.
J. Climate
,
17
,
3349
3363
, doi:.
Cassou
,
C.
,
L.
Terray
, and
A. S.
Phillips
,
2005
:
Tropical Atlantic influence on European heat waves
.
J. Climate
,
18
,
2805
2811
, doi:.
Cayan
,
D. R.
,
1992
:
Latent and sensible heat flux anomalies over the northern oceans: Driving the sea surface temperature
.
J. Phys. Oceanogr.
,
22
,
859
881
, doi:.
Chang
,
C.-P.
,
P.
Harr
, and
J.
Ju
,
2001
:
Possible roles of Atlantic circulations on the weakening Indian monsoon rainfall–ENSO relationship
.
J. Climate
,
14
,
2376
2380
, doi:.
Compo
,
G. P.
, and Coauthors
,
2011
:
The Twentieth Century Reanalysis Project
.
Quart. J. Roy. Meteor. Soc.
,
137
,
1
28
, doi:.
Curtis
,
S.
, and
S.
Hastenrath
,
1995
:
Forcing of anomalous sea surface temperature evolution in the tropical Atlantic during Pacific warm events
.
J. Geophys. Res.
,
100
,
15 835
15 847
, doi:.
Czaja
,
A.
, and
C.
Frankignoul
,
1999
:
Influence of the North Atlantic SST on the atmospheric circulation
.
Geophys. Res. Lett.
,
26
,
2969
2972
, doi:.
Czaja
,
A.
, and
C.
Frankignoul
,
2002
:
Observed impact of Atlantic SST anomalies on the North Atlantic Oscillation
.
J. Climate
,
15
,
606
623
, doi:.
Czaja
,
A.
,
P.
Van der Vaart
, and
J.
Marshall
,
2002
:
A diagnostic study of the role of remote forcing in tropical Atlantic variability
.
J. Climate
,
15
,
3280
3290
, doi:.
Duchon
,
C.
,
1979
:
Lanczos filtering in one and two dimensions
.
J. Appl. Meteor.
,
18
,
1016
1022
, doi:.
Enfield
,
D. B.
, and
D. A.
Mayer
,
1997
:
Tropical Atlantic sea surface temperature variability and its relation to El Niño–Southern Oscillation
.
J. Geophys. Res.
,
102
,
929
945
, doi:.
Enfield
,
D. B.
,
A. M.
Mestas-Nuñez
,
D. A.
Mayer
, and
L.
Cid-Serrano
,
1999
:
How ubiquitous is the dipole relationship in tropical Atlantic sea surface temperatures?
J. Geophys. Res.
,
104
,
7841
7848
, doi:.
Folland
,
C.
,
T.
Palmer
, and
D.
Parker
,
1986
:
Sahel rainfall and worldwide sea temperatures, 1901–85
.
Nature
,
320
,
602
607
, doi:.
Ham
,
Y.-G.
,
J.-S.
Kug
,
J.-Y.
Park
, and
F.-F.
Jin
,
2013
:
Sea surface temperature in the north tropical Atlantic as a trigger for El Niño/Southern Oscillation events
.
Nat. Geosci.
,
6
,
112
116
, doi:.
Hastenrath
,
S.
,
1984
:
Interannual variability and annual cycle: Mechanisms of circulation and climate in the tropical Atlantic sector
.
Mon. Wea. Rev.
,
112
,
1097
1107
, doi:.
Hatzaki
,
M.
, and
R.
Wu
,
2015
:
The south-eastern Europe winter precipitation variability in relation to the North Atlantic SST
.
Atmos. Res.
, 152, 61–68, doi:.
Hu
,
Z.-Z.
, and
B.
Huang
,
2006a
:
Air–sea coupling in the North Atlantic during summer
.
Climate Dyn.
,
26
,
441
457
, doi:.
Hu
,
Z.-Z.
, and
B.
Huang
,
2006b
:
On the significance of the relationship between the North Atlantic Oscillation in early winter and Atlantic sea surface temperature anomalies
.
J. Geophys. Res.
,
111
,
D12103
, doi:.
Huang
,
B.
, and
J.
Shukla
,
2005
:
Ocean–atmosphere interactions in the tropical and subtropical Atlantic Ocean
.
J. Climate
,
18
,
1652
1672
, doi:.
Huang
,
B.
,
P. S.
Schopf
, and
Z.
Pan
,
2002
:
The ENSO effect on the tropical Atlantic variability: A regionally coupled model study
.
Geophys. Res. Lett.
,
29
,
2039
, doi:.
Huang
,
B.
,
P. S.
Schopf
, and
J.
Shukla
,
2004
:
Intrinsic ocean–atmosphere variability of the tropical Atlantic Ocean
.
J. Climate
,
17
,
2058
2077
, doi:.
Hurrell
,
J. W.
,
1995
:
Decadal trends in the North Atlantic oscillation
.
Science
,
269
,
676
679
, doi:.
Hurrell
,
J. W.
, and
H.
van Loon
,
1997
: Decadal variations in climate associated with the North Atlantic Oscillation. Climatic Change at High Elevation Sites, H. F. Diaz, M. Beniston, and R. Bradley, Eds., Springer, 69–94.
Kalnay
,
E.
, and Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project
.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
, doi:.
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:.
Kushnir
,
Y.
,
R.
Seager
,
M.
Ting
,
N.
Naik
, and
J.
Nakamura
,
2010
:
Mechanisms of tropical Atlantic SST influence on North American precipitation variability
.
J. Climate
,
23
,
5610
5628
, doi:.
Lee
,
S.-K.
,
D.-B.
Enfield
, and
C.
Wang
,
2008
:
Why do some El Niños have no impact on tropical North Atlantic SST?
Geophys. Res. Lett.
,
35
,
L16705
, doi:.
Marshall
,
J.
, and Coauthors
,
2001
:
North Atlantic climate variability: Phenomena, impacts and mechanisms
.
Int. J. Climatol.
,
21
,
1863
1898
, doi:.
Nobre
,
P.
, and
J.
Shukla
,
1996
:
Variations of sea surface temperature, wind stress, and rainfall over the tropical Atlantic and South America
.
J. Climate
,
9
,
2464
2479
, doi:.
Pan
,
L.-L.
,
2005
:
Observed positive feedback between the NAO and the North Atlantic SSTA tripole
.
Geophys. Res. Lett.
,
32
,
L06707
, doi:.
Peng
,
S.
,
W. A.
Robinson
, and
S.
Li
,
2003
:
Mechanisms for the NAO responses to the North Atlantic SST tripole
.
J. Climate
,
16
,
1987
2004
, doi:.
Rayner
,
N.
,
D.
Parker
,
E.
Horton
,
C.
Folland
,
L.
Alexander
,
D.
Rowell
,
E.
Kent
, and
A.
Kaplan
,
2003
:
Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century
.
J. Geophys. Res.
,
108
,
4407
, doi:.
Rodwell
,
M.
,
D.
Rowell
, and
C.
Folland
,
1999
:
Oceanic forcing of the wintertime North Atlantic Oscillation and European climate
.
Nature
,
398
,
320
323
, doi:.
Saravanan
,
R.
, and
P.
Chang
,
2000
:
Interaction between tropical Atlantic variability and El Niño–Southern Oscillation
.
J. Climate
,
13
,
2177
2194
, doi:.
Saunders
,
M. A.
, and
B.
Qian
,
2002
: Seasonal predictability of the winter NAO from North Atlantic sea surface temperatures. Geophys. Res. Lett.,29, 2049, doi:.
Tanimoto
,
Y.
, and
S.-P.
Xie
,
1999
:
Ocean-atmosphere variability over the Pan-Atlantic basin
.
J. Meteor. Soc. Japan
,
77
,
31
46
.
Tourre
,
Y. M.
,
B.
Rajagopalan
, and
Y.
Kushnir
,
1999
:
Dominant patterns of climate variability in the Atlantic Ocean during the last 136 years
.
J. Climate
,
12
,
2285
2299
, doi:.
Uvo
,
C. B.
,
C. A.
Repelli
,
S. E.
Zebiak
, and
Y.
Kushnir
,
1998
:
The relationships between tropical Pacific and Atlantic SST and northeast Brazil monthly precipitation
.
J. Climate
,
11
,
551
562
, doi:.
van Loon
,
H.
, and
J. C.
Rogers
,
1978
:
The seesaw in winter temperatures between Greenland and northern Europe. Part I: General description
.
Mon. Wea. Rev.
,
106
,
296
310
, doi:.
Visbeck
,
M.
,
H.
Cullen
,
G.
Krahmann
, and
N.
Naik
,
1998
:
An ocean model’s response to North Atlantic Oscillation–like wind forcing
.
Geophys. Res. Lett.
,
25
,
4521
4524
, doi:.
Visbeck
,
M.
,
E. P.
Chassignet
,
R. G.
Curry
,
T. L.
Delworth
,
R. R.
Dickson
, and
G.
Krahmann
,
2003
: The ocean’s response to North Atlantic Oscillation variability. The North Atlantic Oscillation: Climatic Significance and Environmental Impact, Geophys. Monogr., Vol. 134, Amer. Geophys. Union, 113–145.
Walker
,
G. T.
,
1924
:
Correlation in seasonal variations of weather, IX. A further study of world weather
. Mem. Ind. Meteor. Dept.,
24
,
275
332
.
Walker
,
G. T.
, and
E.
Bliss
,
1932
:
World weather V
.
Mem. Roy. Meteor. Soc.
,
4
,
53
84
.
Wallace
,
J. M.
, and
D. S.
Gutzler
,
1981
:
Teleconnections in the geopotential height field during the Northern Hemisphere winter
.
Mon. Wea. Rev.
,
109
,
784
812
, doi:.
Wallace
,
J. M.
,
C.
Smith
, and
Q.
Jiang
,
1990
:
Spatial patterns of atmosphere-ocean interaction in the northern winter
.
J. Climate
,
3
,
990
998
, doi:.
Walter
,
K.
, and
H.-F.
Graf
,
2002
: On the changing nature of the regional connection between the North Atlantic Oscillation and sea surface temperature. J. Geophys. Res.,107, 4338, doi:.
Wang
,
W.
,
B. T.
Anderson
,
R. K.
Kaufmann
, and
R. B.
Myneni
,
2004
:
The relation between the North Atlantic Oscillation and SSTs in the North Atlantic basin
.
J. Climate
,
17
,
4752
4759
, doi:.
Watanabe
,
M.
,
M.
Kimoto
,
T.
Nitta
, and
M.
Kachi
,
1999
:
A comparison of decadal climate oscillations in the North Atlantic detected in observations and a coupled GCM
.
J. Climate
,
12
,
2920
2940
, doi:.
Wu
,
L.
, and
Z.
Liu
,
2002
:
Is tropical Atlantic variability driven by the North Atlantic Oscillation?
Geophys. Res. Lett.
,
29
,
1653
, doi:.
Wu
,
R.
, and
L.
Zhang
,
2010
:
Biennial relationship of rainfall variability between Central America and equatorial South America
.
Geophys. Res. Lett.
,
37
,
L08701
, doi:.
Wu
,
R.
, and
B.-P.
Kirtman
,
2011
:
Caribbean Sea rainfall variability during the rainy season and relationship to the equatorial Pacific and tropical Atlantic SST
.
Climate Dyn.
,
37
,
1533
1550
, doi:.
Wu
,
R.
,
S.
Yang
,
S.
Liu
,
L.
Sun
,
Y.
Lian
, and
Z.
Gao
,
2011
:
Northeast China summer temperature and North Atlantic SST
.
J. Geophys. Res.
,
116
,
D16116
, doi:.
Wu
,
Z.
,
B.
Wang
,
J.
Li
, and
F. F.
Jin
,
2009
:
An empirical seasonal prediction model of the East Asian summer monsoon using ENSO and NAO
.
J. Geophys. Res.
,
114
,
D18120
, doi:.
Zhu
,
J.
,
B.
Huang
, and
Z.
Wu
,
2012
:
The role of ocean dynamics in the interaction between the Atlantic meridional and equatorial modes
.
J. Climate
,
25
,
3583
3598
, doi:.