1. Introduction
Variations in the tropics can affect the variability of the extratropical atmospheric circulation in many ways. The most important source of tropical influence is the El Niño–Southern Oscillation (ENSO) phenomenon. In the Pacific region, ENSO is known to strongly influence the Pacific–North American pattern [PNA; see the review by Trenberth et al. (1998)]; in particular, the Aleutian low is deepened during warm ENSO events. The tropics can also influence other circulation patterns in the extratropics, such as the southern annular mode (SAM; L’Heureux and Thompson 2006; Ding et al. 2012, 2014a,b) and the East Asian monsoon (Wang et al. 2008; Sun et al. 2010; Gollan et al. 2012). These teleconnections are sometimes seasonally varying (L’Heureux and Thompson 2006), can exhibit nonstationary behavior on decadal time scales (e.g., Greatbatch et al. 2004; Wang et al. 2008; Ding et al. 2012, 2014b,c), and sometimes involve the stratosphere in their dynamics (e.g., Ineson and Scaife 2009).
Over the Euro-Atlantic region, the dominant mode of low-frequency variability in the atmospheric circulation is the North Atlantic Oscillation [NAO; Walker and Bliss 1932; see Greatbatch (2000) and Hurrell et al. (2003) for reviews]. It has been noted that on intraseasonal time scales, the NAO interacts with the tropics via the Madden–Julian oscillation (MJO; Cassou 2008; Lin et al. 2009; Ding et al. 2010). On the other hand, Greatbatch et al. (2003, 2012a) have found that on interannual time scales the tropical influence on the NAO is, in general, rather weak and other factors may be more important such as variability internal to the extratropical atmosphere, including both the troposphere and the stratosphere, whereas on decadal time scales the tropics seem to play a more important role (Hoerling et al. 2001; Greatbatch et al. 2012a,b).
However, previous studies, using model experiments with relaxation toward reanalysis data in different regions of the atmosphere, have also shown that in certain boreal winters such as 2005/06 (Jung et al. 2010a) and 1962/63 (Greatbatch et al. 2014), the tropics can be a significant driver of the extratropical circulation in the Northern Hemisphere (NH). These authors argue that both NH winters 1962/63 and 2005/06 were likely to have been influenced by the easterly phase of the quasi-biennial oscillation (QBO). The QBO describes vertically alternating zonal wind bands in the stratosphere along the equator, propagating downward with a time scale of slightly more than two years. The QBO has a predictability of several months and its impact on the extratropical stratosphere and on the troposphere in the Euro-Atlantic region has been noted by many authors (e.g., Holton and Tan 1980; Baldwin et al. 2001; Boer and Hamilton 2008; Marshall and Scaife 2009). When the zonal winds associated with the QBO are easterly in the lower stratosphere, as during the winters 1962/63 and 2005/06, the QBO favors negative NAO regimes and cold European winters. However, this factor alone is not enough to explain the particularly extreme winter of 1962/63 (Folland et al. 2012; Greatbatch et al. 2014). Greatbatch et al. (2014) note the occurrence of anomalously strong easterly upper-tropospheric zonal-mean zonal winds along the equator during winter 1962/63 and speculate that this could, in addition to the QBO, be responsible for the strong tropical impact as seen in their relaxation experiments. We shall investigate this possibility in a more quantitative way in this study.
Following Nigam (1990), tropospheric zonal-mean zonal wind [U] (where U is the zonal wind and the square brackets denote the zonal average) will first be examined in detail in the present paper by using a rotated empirical orthogonal function analysis applied to [U] at all latitudes. We find a mode of [U] variability centered over the equator in the upper troposphere, linearly independent of ENSO and the QBO, but related to the MJO, consistent with previous studies (Slingo et al. 1996; Hoskins et al. 1999; Kraucunas and Hartmann 2005). The corresponding time series is found to be highly correlated with [U] at 150 hPa averaged between 5°N and 5°S (hereafter denoted [U150]E, where the E subscript indicates “equatorial”). Then, regression analysis using [U150]E reveals significant influence on middle tropospheric circulation anomalies on monthly to interannual time scales in the extratropical NH in boreal winter. The results are confirmed using a model experiment in which the tropical atmosphere, and therefore also [U150]E, is relaxed toward the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40; Uppala et al. 2005) data, as in Greatbatch et al. (2012a) and Greatbatch et al. (2014). A parallel comparing analysis of the extratropical influence of the QBO, and of the linearly combined influences of the QBO and [U150]E is also carried out.
The structure of the present paper is as follows. In section 2 we describe the datasets and the model experiment we use. We show the results from the RPC analysis in section 3, together with some tests on their robustness and an assessment regarding the forcing of [U150]E. In section 4, we present a regression analysis, using [U150]E, on 500-hPa geopotential height to study the impact of this mode on the extratropical circulation and discuss possible mechanisms for this influence by investigating changes in the subtropical waveguides and Rossby wave sources associated with [U150]E. An application of the results to the winter of 1962/63 is then presented in section 5. We finish with a summary and discussion in section 6.
2. Data and model
To ensure the robustness of our results, three different reanalysis datasets are used here:
the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis starts in January 1948 and is continuously updated (Kalnay et al. 1996),
the ERA-40 reanalysis is a complete dataset that extends from September 1957 to August 2002 (Uppala et al. 2005), and
the ECMWF Interim Re-Analysis (ERA-Interim) starts with the beginning of the satellite era in 1979 and is also continuously updated (Dee et al. 2011).
Results from model simulations carried out with the ECMWF model are also discussed (see Jung et al. 2010b). These simulations use a relaxation technique, in which the dynamical variables (U, V, T, and ps)1 are relaxed toward the ERA-40 reanalysis within the tropical atmosphere at all vertical levels (see Greatbatch et al. 2012a). In this way, the zonal winds in the upper tropical troposphere, but also all other variability in the tropical atmosphere including the QBO, are effectively prescribed by the relaxation, and the extratropical response in the model can be compared with results obtained from the ERA-40 reanalysis. The relaxation in our experiment is confined laterally from 20°S to 20°N, including a smooth transition zone between relaxed and unrelaxed parts of the model domain (see Greatbatch et al. 2012a). Climatological SST and sea ice, computed for the period 1979–2002, are prescribed at the lower boundary and radiative forcing does not change interannually. This is the experiment CLIM-TROPICS in Greatbatch et al. (2012a) and we continue to use this label here. The simulations cover each boreal winter from 1960/61 to 2001/02 and have all been initialized on or around 1 November and then run until the end of February. Twelve realizations are integrated for each winter, each using different initial conditions from around November 1st. Having 12 realizations of the experiment for each winter, it is possible to extract the forced signal from the noise by averaging over the whole ensemble. Model anomalies refer to deviations from the model climatology to reflect anomalous circulation states rather than the difference between the model climate and the reanalysis climate.
3. Analysis of the zonal-mean zonal wind
a. Results of the rotated empirical orthogonal functions analysis
Nigam (1990, hereafter N90) has analyzed tropospheric [U] at all latitudes ϕ during a 9-yr period (1980–88) using a rotated empirical orthogonal function (REOF) analysis [as described, e.g., by von Storch and Zwiers (2001)]. N90 used the varimax rotation method, which makes the structure of the resulting patterns more regionally confined and at the same time more physically meaningful. Doing a classic empirical orthogonal function (EOF) analysis on [U] reveals unrealistic and patchy patterns with signatures in both hemispheres, without the simple physical interpretation we are able to give the REOFs below. N90 found four dominant modes of variability, but also made the remark that the data record was too short to make any definite conclusions from findings regarding the third and fourth modes. The first two modes in N90 represent fluctuations of the midlatitude jets in the northern and southern extratropics respectively, while he associated the third and fourth modes with variations of deep convection in the tropics. We update his REOF analysis with nearly the same methods, but using the longer reanalysis datasets for the December, January, and February monthly-mean anomalies of [U] on standard pressure levels from 1000 to 100 hPa (see section 2). The [U] anomalies have been weighted before the REOF analysis, according to area 
Our Fig. 1 shows the first four modes (REOF1–REOF4 hereafter) of our analysis for NCEP–NCAR December–February (DJF) monthly-mean data from December 1949 to February 2013 (climatological monthly means removed). Also plotted in light contours is the climatological zonal-mean zonal wind, mainly showing the midlatitude westerly jets. Not visible because of the large contour interval for the climatology are the negative [U] values corresponding to the easterly trade winds in the lower troposphere on both sides of the equator. At the equator, the climatological mean upper-tropospheric [U] is near zero.
(left) The first four modes from an REOF analysis using DJF monthly-mean [U] anomalies from NCEP–NCAR. Red and blue contours show the loading patterns corresponding to plus one standard deviation of the RPC time series in descending order of explained variance in the original data. The contour interval is 0.5 m s−1 and red (blue) contours refer to positive (negative) values, while the zero contour is omitted. Additionally, the climatological [U] for DJF is shown in gray contours with a contour interval of 10 m s−1, again without zero contour. The explained variance is given in percent in the header of each panel in (left), and the explained variance of the mode within the tropics (25°N/S) is given in brackets. (right) The RPC time series in units of standard deviation and, instead of the monthly means used for the analysis, the DJF means are plotted; years are labeled according to the January dates. Also given in the header of each panel are the highest correlations between the monthly-mean time series and the climate indices listed in Table 1.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
(left) The first four modes from an REOF analysis using DJF monthly-mean [U] anomalies from NCEP–NCAR. Red and blue contours show the loading patterns corresponding to plus one standard deviation of the RPC time series in descending order of explained variance in the original data. The contour interval is 0.5 m s−1 and red (blue) contours refer to positive (negative) values, while the zero contour is omitted. Additionally, the climatological [U] for DJF is shown in gray contours with a contour interval of 10 m s−1, again without zero contour. The explained variance is given in percent in the header of each panel in (left), and the explained variance of the mode within the tropics (25°N/S) is given in brackets. (right) The RPC time series in units of standard deviation and, instead of the monthly means used for the analysis, the DJF means are plotted; years are labeled according to the January dates. Also given in the header of each panel are the highest correlations between the monthly-mean time series and the climate indices listed in Table 1.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
(left) The first four modes from an REOF analysis using DJF monthly-mean [U] anomalies from NCEP–NCAR. Red and blue contours show the loading patterns corresponding to plus one standard deviation of the RPC time series in descending order of explained variance in the original data. The contour interval is 0.5 m s−1 and red (blue) contours refer to positive (negative) values, while the zero contour is omitted. Additionally, the climatological [U] for DJF is shown in gray contours with a contour interval of 10 m s−1, again without zero contour. The explained variance is given in percent in the header of each panel in (left), and the explained variance of the mode within the tropics (25°N/S) is given in brackets. (right) The RPC time series in units of standard deviation and, instead of the monthly means used for the analysis, the DJF means are plotted; years are labeled according to the January dates. Also given in the header of each panel are the highest correlations between the monthly-mean time series and the climate indices listed in Table 1.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
Table 1 shows the correlations between the monthly-mean principal component time series (Fig. 1, right) and some prominent large-scale circulation indices. As in N90, the two leading modes represent meridional fluctuations of the extratropical jets with a quasi-barotropic structure. The first mode is related to the northern annular mode (NAM) and explains about 17% of the global variance in [U]; its RPC time series is strongly positively correlated with the NAM index defined by Thompson and Wallace (2000) (see Table 1). Similarly for the Southern Hemisphere extratropics, the second REOF mode represents 14% of the global variance in [U] and is strongly negatively correlated with the SAM index (Thompson and Wallace 2000).
Correlations r between the monthly DJF time series of the REOF analysis and selected climate indices: Niño-3.4 was computed using the recipe described at http://www.cgd.ucar.edu/cas/catalog/climind/TNI_N34/index.html and NOAA ERSST V3b data. The QBO indices are zonal-mean zonal wind anomalies averaged between 10°S and 10°N at 30 hPa (QBO30) and at 50 hPa (QBO50), respectively, both normalized by their standard deviation. The NAM index has been calculated using geopotential height at 850 hPa following Thompson and Wallace (2000) (SAM index analog for the Southern Hemisphere). All data are from NCEP–NCAR reanalysis for winters 1949–2013. All indices have been detrended in advance. The highest correlations for each mode are highlighted in boldface, which are all significantly different from zero at the 95% level.
Modes 3 and 4 have a much more distinct structure compared to those presented in N90. While N90’s modes 3 and 4 have signatures in both the tropics and in the extratropics, we find two exclusively tropical modes. It has to be noted that our modes 3 and 4 represent nearly the same amount of variance in [U], indicating that they are of roughly equal importance. However, as we show below, both modes have a physically distinct interpretation and hence, we believe, are not artifacts of the REOF analysis.
The third mode (REOF3) has a quasi-monopole pattern centered in the upper troposphere (between 150 and 200 hPa) over the equator and explains about 8% of the variance of [U] globally. Within the tropics between 25°N and 25°S it explains about 25% of the variance of [U] (shown in brackets in the title of Fig. 1e) and is the leading mode there (discussed below). Its structure is baroclinic with no amplitude at the surface and its time series (RPC3) is not correlated with any of the NAM, SAM, Niño-3.4, or QBO indices (see Table 1). This mode describes changes between easterlies and westerlies in a region where the climatology of [U] in DJF is near zero. As already stated in the introduction, the time series RPC3 is highly correlated with [U150]E, to be discussed in more detail later.
RPC4, on the other hand, is highly correlated with the Niño-3.4 index (r = 0.74) and REOF4 has a signature in [U] in the regions of the subtropical jets on both sides of the equator. Warm ENSO events (positive Niño-3.4 index) cause more upper-tropospheric divergence in the equatorial region and upper-tropospheric poleward flow, transporting angular momentum poleward in both hemispheres and therefore accelerating the subtropical jets (e.g., Held and Hou 1980; Oort and Yienger 1996; Seager et al. 2003). Furthermore, the structure of REOF4 is consistent with the response to a zonally symmetric deep equatorial heating in a model with a climatological zonal-mean basic flow (Hoskins et al. 1999).
Dommenget and Latif (2002) note that the results of EOF analyses, including rotated EOF analyses, have to be interpreted with caution as they need not be physically meaningful, because of the requirements of the methods, such as the orthogonality of EOF patterns (also the case for the varimax method used here). Since REOF1, REOF2, and REOF4 all represent known phenomena with a physical interpretation, we compare REOF3 to the simple index [U150]E in Fig. 2. In this case, the REOF analysis was applied to detrended monthly-mean data over the full year and gives a similar REOF3 pattern as in Fig. 1. It can also be deduced from Table 2 that the RPC3 time series based on the REOF analysis applied to the whole year is very similar in DJF to the RPC3 time series obtained from the REOF analysis applied to DJF monthly-mean data. Figure 2 shows both the time series of [U150]E, normalized by its standard deviation, and its regression onto the global [U] in comparison with the REOF3 pattern. For both indices in Fig. 2, detrended monthly-mean anomalies of [U] over the full year were used from NCEP–NCAR data between 1949 and 2012. The regression pattern associated with [U150]E has a very similar shape, with slightly enhanced amplitude compared to the REOF3 pattern. Furthermore, the correlation between the two time series is very high (r = 0.91) and also the characteristics of the autocorrelation functions (Fig. 2c) of [U150]E and RPC3 are very similar. Using the e-folding time scale of the autocorrelation function, we get a decorrelation time of about two months for both the RPC3 and the [U150]E indices, implying at least some potential for medium-range predictability. There is also a hint of an annual memory/periodicity in the autocorrelation function, but possible reasons for this are not discussed in the present paper. Given the similarity between [U150]E and RPC3 and the simple definition of [U150]E compared to that of RPC3, the former is used for the analysis in the rest of the paper.
(a) The REOF3 loading pattern as in Fig. 1, but computed using monthly-mean anomalies from the whole year (black) and the regression of [U] from NCEP–NCAR onto the monthly-mean [U150]E index (red). The contour interval is 0.5 m s−1. (b) The corresponding monthly time series of RPC3 ([U150]E) in black (red) normalized to one standard deviation. (c) The autocorrelation functions for the two time series. The e-folding time scales are reached when the autocorrelation functions cut the horizontal black line.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
(a) The REOF3 loading pattern as in Fig. 1, but computed using monthly-mean anomalies from the whole year (black) and the regression of [U] from NCEP–NCAR onto the monthly-mean [U150]E index (red). The contour interval is 0.5 m s−1. (b) The corresponding monthly time series of RPC3 ([U150]E) in black (red) normalized to one standard deviation. (c) The autocorrelation functions for the two time series. The e-folding time scales are reached when the autocorrelation functions cut the horizontal black line.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
(a) The REOF3 loading pattern as in Fig. 1, but computed using monthly-mean anomalies from the whole year (black) and the regression of [U] from NCEP–NCAR onto the monthly-mean [U150]E index (red). The contour interval is 0.5 m s−1. (b) The corresponding monthly time series of RPC3 ([U150]E) in black (red) normalized to one standard deviation. (c) The autocorrelation functions for the two time series. The e-folding time scales are reached when the autocorrelation functions cut the horizontal black line.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
REOF modes that are highly correlated to [U150]E, based on REOF analysis using the datasets ERA-40, ERA-Interim, and NCEP–NCAR with different time periods: monthly anomalies, seasonal anomalies, and an REOF analysis carried out on monthly anomalies over the whole year (i.e., full year). For each combination the rank of the mode is given (first column), as well as the explained variance (second column) and the correlation r between the corresponding RPC time series and [U150]E from ERA-40 (third column) during the overlapping period. Shown in brackets are the respective numbers for the same REOF analysis restricted to tropical latitudes ϕ ≤ 25°N/S). For each dataset, the correlation between [U150]E from ERA-40 and [U150]E from each dataset is also given, referring to the full year monthly time series during each overlapping period. REOF3 from Fig. 1 is highlighted in boldface.
Figure 3 shows a comparison between the [U150]E indices from NCEP–NCAR, ERA-40, and ERA-Interim data showing winter (DJF) seasonal mean anomalies for convenience. QBO indices (computed at 30 hPa) are also shown, since these will be referred to later in this study. As noted in Table 1, [U150]E is weakly anticorrelated with the Niño-3.4 index in the NCEP–NCAR reanalysis, which is also true for the ECMWF reanalyses. To remove any unwanted effect of ENSO in our following composite and regression analyses, we remove variations from the [U150]E index that are linearly dependent of ENSO, as measured by the Niño-3.4 index. All time series are normalized to have unit standard deviation σ—note that σ([U150]E) is about 1.7 m s−1 and σ(QBO30) is about 10 m s−1. The normalized time series with ENSO effects removed is hereafter denoted as 
Time series of the DJF mean 
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
Time series of the DJF mean
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
Time series of the DJF mean
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
We further analyze the robustness of our REOF analysis with respect to the different reanalysis datasets and to slightly different methods or time periods used for the analysis. Data from ERA-40, ERA-Interim, and NCEP–NCAR are tested, as well as using DJF mean anomalies or monthly-mean anomalies of [U] over the full year for the REOF analysis. Furthermore, we have performed the same set of REOF analyses, but restricted to tropical latitudes (ϕ ≤ 25°N/S). For each case, the correlation between the RPC time series and the reference index [U150]E from ERA-40 during the overlapping period is also computed. The results are noted in Table 2 for the REOF mode whose RPC time series has the highest correlation with the [U150]E index from ERA-40. These correlations are generally very high (note the variable overlap; i.e., 23 yr for ERA-Interim and 45 yr for NCEP–NCAR reanalysis). The results of Table 2 can be summarized by saying that the mode related to [U150]E always occurs as one of the second to fifth REOF modes and that the explained variance of this mode is always between 8% and 13%. When restricting the REOF analysis to the tropics, this mode occurs as the first or second mode and explains about 20%–25% of the variance there. Additionally the correlations between the reference index and [U150]E from ERA-Interim and NCEP–NCAR reanalysis are also given in Table 2.
The first two modes from an EOF analysis applied to monthly-mean data for DJF of U at 200 hPa (U200) between 10°N and 10°S (hereafter U200-EOF1 and U200-EOF2), using ERA-40 data, are shown in Fig. 4 together with the corresponding time series U200-PC1 and U200-PC2. Notably, the first mode represents the modulation of the Walker circulation due to ENSO variability, explaining about 30% of the total variance of U200 between 10°N and 10°S. The correlation between U200-PC1 and the Niño-3.4 index is −0.81 (see Table 3). On the other hand, U200-EOF2 represents variations of U200 (about 20% of the total variance) with the same sign over the whole domain and with strongest variations over the eastern tropical Pacific. The time series U200-PC2 is highly correlated (r = 0.92) with [U150]E, which in turn indicates that significant zonal variation is related to variability of the [U150]E index and in particular that the wind anomalies associated with [U150]E are largest over the eastern tropical Pacific. In section 4b we will further discuss the importance of wind anomalies in the eastern tropical Pacific for the extratropical influence of [U150]E.
The first two EOF modes using DJF monthly anomalies of U200 from ERA-40 at all longitudes between 10°N and 10°S. (a),(c) The first and second loading patterns with red (blue) contours indicating positive (negative) anomalies. The contour intervals (ci) are different for the two loading patterns and are indicated in the title of the panel, while the zero contour is always omitted; the explained variances in percent are also given in the titles. (b),(d) The corresponding time series in units of standard deviations. The correlations between U200-PC1 and Niño-3.4 and between U200-PC2 and [U150]E are −0.81 and 0.92, respectively (see Table 3).
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
The first two EOF modes using DJF monthly anomalies of U200 from ERA-40 at all longitudes between 10°N and 10°S. (a),(c) The first and second loading patterns with red (blue) contours indicating positive (negative) anomalies. The contour intervals (ci) are different for the two loading patterns and are indicated in the title of the panel, while the zero contour is always omitted; the explained variances in percent are also given in the titles. (b),(d) The corresponding time series in units of standard deviations. The correlations between U200-PC1 and Niño-3.4 and between U200-PC2 and [U150]E are −0.81 and 0.92, respectively (see Table 3).
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
The first two EOF modes using DJF monthly anomalies of U200 from ERA-40 at all longitudes between 10°N and 10°S. (a),(c) The first and second loading patterns with red (blue) contours indicating positive (negative) anomalies. The contour intervals (ci) are different for the two loading patterns and are indicated in the title of the panel, while the zero contour is always omitted; the explained variances in percent are also given in the titles. (b),(d) The corresponding time series in units of standard deviations. The correlations between U200-PC1 and Niño-3.4 and between U200-PC2 and [U150]E are −0.81 and 0.92, respectively (see Table 3).
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
Correlations r between the PC time series for an EOF analysis applied to ERA-40 U200 between 10°N and 10°S (shown in Fig. 4) and the Niño-3.4 and [U150]E indices. The highest correlation for each mode is highlighted in boldface, which is also significantly different from zero at the 95% level.
b. Mechanisms for upper-tropospheric anomalies of [U] at the equator
As shown by previous studies (e.g., Hoskins et al. 1999; Lee 1999; Kraucunas and Hartmann 2005), zonally asymmetric tropical diabatic heating can lead to westerly acceleration of [U] in the upper equatorial troposphere if the heating is located so that there is an associated eddy flux of westerly momentum toward the equator. The westerly acceleration is balanced by downward fluxes of momentum and by the zonal-mean meridional overturning circulation, both removing westerly momentum from the equatorial upper troposphere (Kraucunas and Hartmann 2005). A localized heating is provided, on intraseasonal time scales, by the heating centers of the MJO, which can both accelerate and decelerate upper equatorial [U], depending on the phase of the MJO, resulting from the interaction of the circulation generated by the heating anomaly and the climatological stationary waves (Hoskins et al. 1999).
A simple analysis is shown in Fig. 5, using the daily real-time multivariate MJO (RMM) indices by Wheeler and Hendon (2004) (indices are taken from http://cawcr.gov.au/staff/mwheeler/maproom/RMM/RMM1RMM2.74toRealtime.txt), where RMM1 and RMM2 are the time series of the first two multivariate EOFs (using normalized tropical OLR, zonal wind at 850 hPa, and U200) and 
Evaluation of monthly 
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
Evaluation of monthly
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
Evaluation of monthly
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
First in Fig. 5a, the activity of the MJO is evaluated against the phase of 
Furthermore, hemispherically asymmetric tropical heating leads to an easterly acceleration of [U] over the equator due to the advection of easterly angular momentum by the associated cross-equatorial flow (e.g., Lee 1999). This is evident during boreal summer, when the tropical heating zone is displaced into the Northern Hemisphere and climatological [U] is strongly easterly in the upper troposphere over the equator, whereas during boreal winter, when the heating zone is only slightly south of the equator, climatological [U] is near zero there. Figure 6 shows the regression map of monthly-mean OLR anomalies from ERA-40 for DJF onto the reference 
(left) Regressions onto the 
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
(left) Regressions onto the
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
(left) Regressions onto the
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
4. Circulation anomalies associated with [U150]E
a. Regression analysis for ERA-40 data and model output
As seen from its autocorrelation function, [U150]E may be predictable to some extent, or at least persists on longer time scales than synoptic variability in the extratropics. In this section we discuss a possible impact on the extratropical circulation with potential implications for medium-range to seasonal forecasting. Previous studies showed that the stratospheric equatorial zonal wind (i.e., the QBO) has good predictability because of its periodicity and also has an impact on the extratropics (e.g., Holton and Tan 1980; Marshall and Scaife 2009). Indeed, the QBO has been shown to be valuable for medium-range to seasonal predictions of the extratropical circulation (e.g., Boer and Hamilton 2008; Folland et al. 2012). Here we want to compare the influence of the QBO and variability associated with [U150]E by means of a regression analysis while noting that the QBO and [U150]E indices are uncorrelated with each other (see Table 1). Since [U150]E is weakly (not significantly) anticorrelated with the Niño-3.4 index and is therefore not completely independent, we use the index 
The linear regression of ERA-40 DJF monthly-mean anomalies of Z500 onto the reference 
(a) Regression of the DJF monthly-mean anomalies of Z500 onto the monthly 
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
(a) Regression of the DJF monthly-mean anomalies of Z500 onto the monthly
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
(a) Regression of the DJF monthly-mean anomalies of Z500 onto the monthly
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
In the ERA-40 regression, associated with the westerly phase of 
To get more insight into the tropically forced extratropical circulation associated with 
We also note here that in the lower stratosphere (e.g., represented by 50-hPa geopotential height; not shown), there is the suggestion of a slight shift of the polar vortex in the NH toward the Eastern Hemisphere associated with the positive 
Now we turn to the extratropical influence of the QBO for comparison. The canonical influence of the QBO on the extratropical troposphere in boreal winter favors a positive (negative) NAO pattern during the westerly (easterly) phase of the QBO. The mechanism for this effect is thought to be as follows [e.g., as described by Holton and Tan (1980)]. During the westerly phase, the zero wind line of [U] in the stratosphere is shifted to the south, causing fewer stationary waves to be reflected toward the pole, leading to a cooler and stronger polar vortex in the stratosphere (Holton and Tan 1980; Baldwin et al. 2001). A stronger polar stratospheric vortex in turn favors a positive NAM/NAO regime in the troposphere (e.g., Baldwin and Dunkerton 2001). This is confirmed by our Fig. 7c, where the regression of ERA-40 Z500 onto the ERA-40 QBO30 index shows, corresponding to the westerly phase of QBO30, a pattern of deepened geopotential over the North Pole extending from Greenland to the Aleutians and enhanced geopotential over the North Atlantic where it projects onto the positive NAO pattern, even though the regression is not significantly different from zero at the 90% level. Reasons for the lack of significance could be that the downward control of the polar vortex only works efficiently in winters with stratospheric warming events. There is also large interannual variability in the winter tropospheric circulation and decadal variations in the link between QBO and the stratospheric polar vortex (see Lu et al. 2014), which can weaken the link between the QBO and the NAO. For the ensemble mean of the experiment CLIM-TROPICS (Fig. 7d), the regression has a similar pattern as for the reanalysis, but with a strongly reduced amplitude compared to the reanalysis case. Overall, the canonical extratropical influence of the QBO is confirmed by our linear regression analysis.
b. Mechanisms for the influence of [U150]E on the extratropical circulation
In this section we want to discuss possible mechanisms for the extratropical influence of [U150]E and, in particular, for the wave train identified over the North Atlantic. In a theoretical barotropic case, representative of the upper troposphere, the propagation of Rossby waves is controlled by so-called waveguides, which are determined by the background zonal flow (Hoskins and Ambrizzi 1993). Middle tropospheric heating anomalies (e.g., related to precipitation) can then affect the extratropical circulation by initiating Rossby wave trains in these waveguides (e.g., Simmons 1982; Hoskins and Ambrizzi 1993). Furthermore, quasi-stationary Rossby waves can be reflected by critical layers associated, for example, with zero-wind lines (Killworth and McIntyre 1985).
The 200-hPa Ks, giving the number of waves that would fit around each corresponding latitude band (see color bar); Ks serves as a refractive index for the propagation of Rossby waves [see Eq. (1)]. (a) Value of Ks computed using U200 averaged over months from DJF when 
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
The 200-hPa Ks, giving the number of waves that would fit around each corresponding latitude band (see color bar); Ks serves as a refractive index for the propagation of Rossby waves [see Eq. (1)]. (a) Value of Ks computed using U200 averaged over months from DJF when
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
The 200-hPa Ks, giving the number of waves that would fit around each corresponding latitude band (see color bar); Ks serves as a refractive index for the propagation of Rossby waves [see Eq. (1)]. (a) Value of Ks computed using U200 averaged over months from DJF when
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
The North Atlantic waveguide, originating in the eastern subtropical North Pacific, can be seen in both Figs. 8a and 8b and it is clear that it is on this waveguide that the North Atlantic wave train, noted when discussing Fig. 7, sits. There are also some differences in Ks between the westerly 
Honda et al. (2001) have investigated the connection between the Aleutian low and the Icelandic low by regressing late winter 250-hPa geopotential height onto their so-called seesaw index [see Figs. 6e,f in Honda et al. (2001)]. They find a pattern over the North Atlantic that is similar to the Z500 regression patterns associated with 
Rossby wave source at 200 hPa (RWS200). (a) DJF climatology from ERA-40 data; (b) the linear regression pattern of RWS200 from ERA-40 onto the 
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
Rossby wave source at 200 hPa (RWS200). (a) DJF climatology from ERA-40 data; (b) the linear regression pattern of RWS200 from ERA-40 onto the
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
Rossby wave source at 200 hPa (RWS200). (a) DJF climatology from ERA-40 data; (b) the linear regression pattern of RWS200 from ERA-40 onto the
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
For ERA-40 data, the regression pattern associated with 
The RWS anomalies presented in Fig. 9 are located precisely in the regions of the waveguides provided by the zonal-mean flow shown in Fig. 8. In particular, the North American RWS anomalies can initiate Rossby waves that propagate toward Europe via the North Atlantic waveguide and subsequently disturb the circulation there. This is true for both westerly and easterly phase of 
5. Application to the anomalous winter of 1962/63
The DJF mean values of both indices in the particular winter of 1962/63 were extremely negative: 
The Z500 DJF mean anomalies for winter 1962/63. (a) The ERA-40 anomaly with respect to the climatological winter mean from 1960/61 to 2001/02, and (b) the component R(x, y, 1962/63) and (c) the component Q(x, y, 1962/63) according to Eqs. (4) and (5), respectively. (d) The linear combination R(x, y, 1962/63) + Q(x, y, 1962/63). Contour intervals are 16 m for (a) and 8 m for (b)–(d), with red (blue) contours indicating positive (negative) anomalies and the zero lines being omitted.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
The Z500 DJF mean anomalies for winter 1962/63. (a) The ERA-40 anomaly with respect to the climatological winter mean from 1960/61 to 2001/02, and (b) the component R(x, y, 1962/63) and (c) the component Q(x, y, 1962/63) according to Eqs. (4) and (5), respectively. (d) The linear combination R(x, y, 1962/63) + Q(x, y, 1962/63). Contour intervals are 16 m for (a) and 8 m for (b)–(d), with red (blue) contours indicating positive (negative) anomalies and the zero lines being omitted.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
The Z500 DJF mean anomalies for winter 1962/63. (a) The ERA-40 anomaly with respect to the climatological winter mean from 1960/61 to 2001/02, and (b) the component R(x, y, 1962/63) and (c) the component Q(x, y, 1962/63) according to Eqs. (4) and (5), respectively. (d) The linear combination R(x, y, 1962/63) + Q(x, y, 1962/63). Contour intervals are 16 m for (a) and 8 m for (b)–(d), with red (blue) contours indicating positive (negative) anomalies and the zero lines being omitted.
Citation: Journal of Climate 28, 1; 10.1175/JCLI-D-14-00185.1
The influence of the QBO in this winter (Fig. 10c) has a signature over the North Atlantic that is very similar to the negative NAO pattern and therefore contributes to a more meridional dipole there that gives some extra skill to the combination R + Q. According to the Holton and Tan (1980) mechanism, it is possible that the easterly phase of the QBO in 1962/63 favored the stratospheric warming that occurred in late January 1963, which in turn impacted the troposphere [see Greatbatch et al. (2014), their Figs. 7 and 11]. The combination R + Q (Fig. 10d) thus results in an anomaly pattern qualitatively very similar to the reanalysis anomaly over the whole NH extratropics. This suggests that the extreme negative value of 
6. Summary and discussion
Updating the rotated EOF analysis of Nigam (1990) applied to DJF monthly-mean zonal-mean zonal wind ([U]) anomalies at all latitudes, we find four dominant modes in the reanalysis datasets NCEP–NCAR, ERA-40, and ERA-Interim. Three of the four modes correspond to widely known circulation modes of the atmosphere: The first two modes are the NAM and SAM, explaining most of the variance of [U] in the extratropics of their respective hemispheres. Modes 3 and 4 are confined to the tropics and, while noting that both explain slightly more than 20% of the variance of tropical [U], both can be associated with physically distinct mechanisms. One of the two is closely related to ENSO and is therefore, as for the first two modes, not investigated in detail in the present paper. Indeed, it is the mode that describes the equatorial upper-tropospheric [U], which has been investigated in detail here by using the time series of monthly anomalies of [U] at 150 hPa averaged between 5°S and 5°N [U150]E, as this time series is highly correlated with the principal component time series of the REOF mode. The [U150]E mode was shown to be related to the MJO, since the propagating diabatic heating anomalies of the MJO favor upper-tropospheric westerly (easterly) anomalies, especially in late (early) MJO phases. Furthermore, the absence (presence) of MJO activity has been shown to favor easterly (westerly) [U150]E anomalies, consistent with previous studies (e.g., Hoskins et al. 1999; Kraucunas and Hartmann 2005).
Long-term trends and low-frequency variability of [U150]E are not discussed in the present paper, as all data have been detrended prior to the analysis and we focus on the seasonal to interannual time scales. Interestingly though, we note here a correlation, albeit weak, of r = 0.22 between [U150]E and the normalized time series of sunspot activity (as downloaded from http://www.esrl.noaa.gov/psd/gcos_wgsp/Timeseries/Data/sunspot.long.data, an estimate of variations in total solar incoming radiation, mainly representing the 11-yr solar cycle). This is consistent with the observational results of Haigh et al. (2005, see their Fig. 1c), who use a different measure for solar activity (i.e., the 10.7-cm flux) and perform a multiple linear regression on [U], but do not discuss the equatorial signal in their results. A possible explanation for the positive correlation is that during solar maxima, the mean meridional overturning circulation is weaker and therefore can remove less westerly momentum from the equatorial upper troposphere (Kraucunas and Hartmann 2005). Indeed, given that anomalies of [U150]E can impact the North Atlantic sector (see Figs. 7 and 10), this mechanism could be another pathway for the connection between solar activity and European winter weather that has been noted by previous studies (e.g., by Lockwood et al. 2010). These authors noted that periods of low solar activity go along with periods of more frequent Atlantic blockings and hence more severe winter weather in Europe, of which 1962/63 (at the time of a solar minimum) is a striking example. We also note here that the undetrended time series of [U150]E has a strong trend toward westerly values (between +0.27 and +0.46 m s−1 decade−1 in the different reanalyses). However, Allen and Sherwood (2008) find exactly the opposite trend in tropical upper-tropospheric [U] by using direct radiosonde measurements, which points to possible inconsistencies in the reanalyses regarding long-term trends in this region.
The second part of the paper focuses on the extratropical influence of [U150]E on monthly-to-seasonal time scales. We make use of an experiment with the ECMWF model, in which the tropical atmosphere has been relaxed toward ERA-40 data, which is why we focus on this reanalysis dataset in this part of the study. We use linear regression onto the time series 
While the positive QBO index is associated with a positive NAO pattern, consistent with many other studies, the westerly phase of 
Analysis of the zonal stationary wavenumber diagnostic (Ks; Hoskins and Ambrizzi 1993; Barnes and Hartmann 2011) reveals a stronger than normal waveguide especially over the North Pacific during the easterly phase of [U150]E with a slightly stronger link to the North Atlantic and also the hint of a link between the extratropical and tropical Atlantic waveguides. The former change suggests a role of the equatorial upper-tropospheric winds for explaining the “seesaw” between the Aleutian low and the Icelandic low as noted by Honda et al. (2001). Rossby wave source (RWS; Sardeshmukh and Hoskins 1988) anomalies associated with [U150]E were found to be strongest and most robust in the linking region of the North Pacific and the North Atlantic waveguide, and are probably associated with the Rossby wave train over the North Atlantic as found in the regression of Z500 against 
Over the western North Pacific, [U150]E has been found to be important for the winter circulation, and can therefore be useful for statistical seasonal prediction. It might be that the signal associated with [U150]E is relatively weak over the North Atlantic and Europe compared to other influences, as can be seen when comparing the winters of 1962/63 and 1983/84, both of which were associated with an extremely easterly phase of [U150]E and the easterly phase of the QBO. The differences in detail of the regression patterns shown in Figs. 7a,b also suggest that the impact of [U150]E is not strong over Europe, although in certain winters (e.g., 1962/63) the impact could be important. Indeed, the circulation anomalies of winter 1962/63 agree well with the combined influence from both [U150]E and the QBO over large parts of the NH, including the European sector, while those of 1983/84 (when 
Acknowledgments
We want to thank the ECMWF for the provision of the model data and especially Prof. Dr. Thomas Jung and Dr. Soumia Serrar who carried out the model runs reported here. We also thank Dr. Lisa Neef and Dr. Remi Thieblemont from GEOMAR for discussions that drew our attention to the tropospheric [U] along the equator and Prof. Adrian Matthews, who made very useful suggestions about the Ks diagnostic. Additionally, we thank three anonymous reviewers for helpful comments that led to a substantial improvement of this manuscript. GG has been funded by the DFG under ISOLAA, a project within the Priority Programme 1158, and RJG has been supported by GEOMAR, Kiel.
REFERENCES
Allen, R. J., and S. C. Sherwood, 2008: Warming maximum in the tropical upper troposphere deduced from thermal winds. Nat. Geosci., 1, 399–403, doi:10.1038/ngeo208.
Baldwin, M. P., and T. J. Dunkerton, 2001: Stratospheric harbingers of anomalous weather regimes. Science, 294, 581–584, doi:10.1126/science.1063315.
Baldwin, M. P., and Coauthors, 2001: The quasi-biennial oscillation. Rev. Geophys., 39, 179–229, doi:10.1029/1999RG000073.
Barnes, E. A., and D. L. Hartmann, 2011: Rossby wave scales, propagation, and the variability of eddy-driven jets. J. Atmos. Sci., 68, 2893–2908, doi:10.1175/JAS-D-11-039.1.
Barnston, A., and R. Livezey, 1987: Classification, seasonality, and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev., 115, 1083–1126, doi:10.1175/1520-0493(1987)115<1083:CSAPOL>2.0.CO;2.
Boer, G. J., and K. Hamilton, 2008: QBO influence on extratropical predictive skill. Climate Dyn., 31, 987–1000, doi:10.1007/s00382-008-0379-5.
Cassou, C., 2008: Intraseasonal interaction between the Madden–Julian oscillation and the North Atlantic Oscillation. Nature, 455, 523–527, doi:10.1038/nature07286.
Dawson, A., A. J. Matthews, and D. P. Stevens, 2011: Rossby wave dynamics of the North Pacific extra-tropical response to El Niño: Importance of the basic state in coupled GCMs. Climate Dyn., 37, 391–405, doi:10.1007/s00382-010-0854-7.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, doi:10.1002/qj.828.
Ding, H., R. J. Greatbatch, and G. Gollan, 2014a: Tropical influence independent of ENSO on the austral summer southern annular mode. Geophys. Res. Lett., 41, 3643–3648, doi:10.1002/2014GL059987.
Ding, H., R. J. Greatbatch, and G. Gollan, 2014b: Tropical impact on the interannual variability and long-term trend of the southern annular mode during austral summer from 1960/1961 to 2001/2002. Climate Dyn., doi:10.1007/s00382-014-2299-x, in press.
Ding, H., R. J. Greatbatch, W. Park, M. Latif, V. A. Semenov, and X. Sun, 2014c: The variability of the East Asian summer monsoon and its relationship to ENSO in a partially coupled climate model. Climate Dyn., 42, 367–379, doi:10.1007/s00382-012-1642-3.
Ding, Q., E. J. Steig, D. S. Battisti, and J. M. Wallace, 2012: Influence of the tropics on the southern annular mode. J. Climate, 25, 6330–6348, doi:10.1175/JCLI-D-11-00523.1.
Ding, R., J. Li, and K.-H. Seo, 2010: Predictability of the Madden–Julian oscillation estimated using observational data. Mon. Wea. Rev., 138, 1004–1013, doi:10.1175/2009MWR3082.1.
Dommenget, D., and M. Latif, 2002: A cautionary note on the interpretation of EOFs. J. Climate, 15, 216–225, doi:10.1175/1520-0442(2002)015<0216:ACNOTI>2.0.CO;2.
Folland, C. K., A. A. Scaife, J. Lindesay, and D. B. Stephenson, 2012: How potentially predictable is northern European winter climate a season ahead? Int. J. Climatol., 32, 801–818, doi:10.1002/joc.2314.
Gollan, G., R. J. Greatbatch, and T. Jung, 2012: Tropical impact on the East Asian winter monsoon. Geophys. Res. Lett., 39, L17801, doi:10.1029/2012GL052978.
Greatbatch, R. J., 2000: The North Atlantic Oscillation. Stochastic Environ. Res. Risk Assess.,14, 213–242, doi:10.1007/s004770000047.
Greatbatch, R. J., H. Lin, J. Lu, K. A. Peterson, and J. Derome, 2003: Tropical/extratropical forcing of the AO/NAO: A corrigendum. Geophys. Res. Lett., 30, 1738, doi:10.1029/2003GL017406.
Greatbatch, R. J., J. Lu, and K. A. Peterson, 2004: Nonstationary impact of ENSO on Euro-Atlantic winter climate. Geophys. Res. Lett., 31, L02208, doi:10.1029/2003GL018542.
Greatbatch, R. J., G. Gollan, T. Jung, and T. Kunz, 2012a: Factors influencing Northern Hemisphere winter mean atmospheric circulation anomalies during the period 1960/61 to 2001/02. Quart. J. Roy. Meteor. Soc., 138, 1970–1982, doi:10.1002/qj.1947.
Greatbatch, R. J., G. Gollan, and T. Jung, 2012b: An analysis of trends in the boreal winter mean tropospheric circulation during the second half of the 20th century. Geophys. Res. Lett.,39, L13809, doi:10.1029/2012GL052243.
Greatbatch, R. J., G. Gollan, T. Jung, and T. Kunz, 2014: Tropical origin of the severe European winter of 1962/63. Quart. J. Roy. Meteor. Soc., doi:10.1002/qj.2346, in press.
Haigh, J. D., M. Blackburn, and R. Day, 2005: The response of tropospheric circulation to perturbations in lower-stratospheric temperature. J. Climate, 18, 3672–3685, doi:10.1175/JCLI3472.1.
Held, I. M., and A. Y. Hou, 1980: Nonlinear axially symmetric circulations in a nearly inviscid atmosphere. J. Atmos. Sci., 37, 515–533, doi:10.1175/1520-0469(1980)037<0515:NASCIA>2.0.CO;2.
Hoerling, M. P., J. W. Hurrell, and T. Xu, 2001: Tropical origins for recent North Atlantic climate change. Science, 292, 90–92, doi:10.1126/science.1058582.
Holton, J. R., and H.-C. Tan, 1980: The influence of the equatorial quasi-biennial oscillation on the global circulation at 50 mb. J. Atmos. Sci., 37, 2200–2208, doi:10.1175/1520-0469(1980)037<2200:TIOTEQ>2.0.CO;2.
Honda, M., H. Nakamura, J. Ukita, I. Kousaka, and K. Takeuchi, 2001: Interannual seesaw between the Aleutian and Icelandic lows. Part I: Seasonal dependence and life cycle. J. Climate, 14, 1029–1042, doi:10.1175/1520-0442(2001)014<1029:ISBTAA>2.0.CO;2.
Hoskins, B. J., and T. Ambrizzi, 1993: Rossby wave propagation on a realistic longitudinally varying flow. J. Atmos. Sci., 50, 1661–1671, doi:10.1175/1520-0469(1993)050<1661:RWPOAR>2.0.CO;2.
Hoskins, B. J., R. Neale, M. Rodwell, and G.-Y. Yang, 1999: Aspects of the large-scale tropical atmospheric circulation. Tellus, 51B, 33–44, doi:10.1034/j.1600-0889.1999.00004.x.
Hurrell, J. W., Y. Kushnir, G. Ottersen, and M. Visbeck, 2003: An overview of the North Atlantic oscillation. The North Atlantic Oscillation: Climatic Significance and Environmental Impact, Geophys. Monogr., Vol. 134, Amer. Geophys. Union, 1–36.
Ineson, S., and A. A. Scaife, 2009: The role of the stratosphere in the European climate response to El Niño. Nat. Geosci., 2, 32–36, doi:10.1038/ngeo381.
Jung, T., T. N. Palmer, M. J. Rodwell, and S. Serrar, 2010a: Understanding the anomalously cold European winter of 2005/06 using relaxation experiments. Mon. Wea. Rev., 138, 3157–3174, doi:10.1175/2010MWR3258.1.
Jung, T., and Coauthors, 2010b: The ECMWF model climate: Recent progress through improved physical parametrizations. Quart. J. Roy. Meteor. Soc., 136, 1145–1160, doi:10.1002/qj.634.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
Killworth, P. D., and M. E. McIntyre, 1985: Do Rossby-wave critical layers absorb, reflect, or over-reflect? J. Fluid Mech., 161, 449–492, doi:10.1017/S0022112085003019.
Kraucunas, I., and D. L. Hartmann, 2005: Equatorial superrotation and the factors controlling the zonal-mean zonal winds in the tropical upper troposphere. J. Atmos. Sci., 62, 371–389, doi:10.1175/JAS-3365.1.
Lee, S., 1999: Why are the climatological zonal winds easterly in the equatorial upper troposphere? J. Atmos. Sci., 56, 1353–1363, doi:10.1175/1520-0469(1999)056<1353:WATCZW>2.0.CO;2.
L’Heureux, M. L., and D. W. J. Thompson, 2006: Observed relationships between the El Niño–Southern Oscillation and the extratropical zonal-mean circulation. J. Climate, 19, 276–287, doi:10.1175/JCLI3617.1.
Lin, H., G. Brunet, and J. Derome, 2009: An observed connection between the North Atlantic Oscillation and the Madden–Julian oscillation. J. Climate, 22, 364–380, doi:10.1175/2008JCLI2515.1.
Lockwood, M., R. G. Harrison, T. Woollings, and S. K. Solanki, 2010: Are cold winters in Europe associated with low solar activity? Environ. Res. Lett.,5, 024001, doi:10.1088/1748-9326/5/2/024001.
Lu, H., T. J. Bracegirdle, T. Phillips, A. Bushell, and L. Gray, 2014: Mechanisms for the Holton–Tan relationship and its decadal variation. J. Geophys. Res. Atmos., 119, 2811–2830, doi:10.1002/2013JD021352.
Marshall, A. G., and A. A. Scaife, 2009: Impact of the QBO on surface winter climate. J. Geophys. Res.,114, D18110, doi:10.1029/2009JD011737.
Moore, G. W. K., and I. A. Renfrew, 2012: Cold European winters: Interplay between the NAO and the East Atlantic mode. Atmos. Sci. Lett., 13, 1–8, doi:10.1002/asl.356.
Nigam, S., 1990: On the structure of variability of the observed tropospheric and stratospheric zonal-mean zonal wind. J. Atmos. Sci., 47, 1799–1813, doi:10.1175/1520-0469(1990)047<1799:OTSOVO>2.0.CO;2.
Oort, A. H., and J. J. Yienger, 1996: Observed interannual variability in the Hadley circulation and its connection to ENSO. J. Climate, 9, 2751–2767, doi:10.1175/1520-0442(1996)009<2751:OIVITH>2.0.CO;2.
Sardeshmukh, P. D., and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45, 1228–1251, doi:10.1175/1520-0469(1988)045<1228:TGOGRF>2.0.CO;2.
Seager, R., N. Harnik, Y. Kushnir, W. Robinson, and J. Miller, 2003: Mechanisms of hemispherically symmetric climate variability. J. Climate, 16, 2960–2978, doi:10.1175/1520-0442(2003)016<2960:MOHSCV>2.0.CO;2.
Simmons, A. J., 1982: The forcing of stationary wave motion by tropical diabatic heating. Quart. J. Roy. Meteor. Soc., 108, 503–534, doi:10.1002/qj.49710845703.
Slingo, J. M., and Coauthors, 1996: Intraseasonal oscillations in 15 atmospheric general circulation models: Results from an AMIP diagnostic subproject. Climate Dyn., 12, 325–357, doi:10.1007/BF00231106.
Sun, X., R. J. Greatbatch, W. Park, and M. Latif, 2010: Two major modes of variability of the East Asian summer monsoon. Quart. J. Roy. Meteor. Soc., 136, 829–841, doi:10.1002/qj.635.
Thompson, D. W. J., and J. M. Wallace, 2000: Annular modes in the extratropical circulation. Part I: Month-to-month variability. J. Climate, 13, 1000–1016, doi:10.1175/1520-0442(2000)013<1000:AMITEC>2.0.CO;2.
Trenberth, K. E., G. W. Branstator, D. Karoly, A. Kumar, N.-C. Lau, and C. Ropelewski, 1998: Progress during TOGA in understanding and modeling global teleconnections associated with tropical sea surface temperatures. J. Geophys. Res.,103, 14 291–14 324, doi:10.1029/97JC01444.
Uppala, S. M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131, 2961–3012, doi:10.1256/qj.04.176.
von Storch, H., and F. W. Zwiers, 2001: Statistical Analysis in Climate Research. Cambridge University Press, 484 pp.
Walker, G., and E. Bliss, 1932: World weather V. Mem. Roy. Meteor. Soc., 4 (36), 53–84.
Wang, L., W. Chen, and R. Huang, 2008: Interdecadal modulation of PDO on the impact of ENSO on the East Asian winter monsoon. Geophys. Res. Lett.,35, L20702, doi:10.1029/2008GL035287.
Wheeler, M. C., and H. H. Hendon, 2004: An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132, 1917–1932, doi:10.1175/1520-0493(2004)132<1917:AARMMI>2.0.CO;2.
Zonal velocity is U, meridional velocity is V, air temperature is T, and surface pressure is ps.
To compute the significance levels Monte Carlo simulations are done. For this, 10 000 random time series X of the same length as 
