Abstract

A multidecadal geopotential height pattern in the upper troposphere of the extratropical Northern Hemisphere (NH) is identified in this study. This pattern is characterized by the nearly zonal symmetry of geopotential height and temperature between 35° and 65°N and the equivalent barotropic vertical structure with the largest amplitude in the upper troposphere. This pattern is named the Eurasian–Pacific multidecadal oscillation (EAPMO) to describe its multidecadal time scale and the largest amplitudes over Eurasia and the North Pacific. Although nearly extending over the entire extratropics, the EAPMO exhibits larger amplitudes over western Europe, East Asia, and the North Pacific with a zonal scale equivalent to zonal wavenumbers 4 and 5. The zonally asymmetric perturbation tends to amplify over the major mountain ranges in the region, suggesting a significant topographic influence. The EAPMO has fluctuated concurrently with the Atlantic multidecadal oscillation (AMO) at least since the beginning of the twentieth century. The numerical simulation results suggest that the EAPMO could be induced by the AMO-like sea surface temperature anomaly and strengthened regionally by topography, especially over the Asian highland region, although the amplitude was undersimulated.

This study found that the multidecadal variability of the upper-tropospheric geopotential height in the extratropical NH is much more complicated than in the tropics and the Southern Hemisphere (SH). It takes both first (warming trend) and second (multidecadal) EOFs to explain the multidecadal variability in the extratropical NH, while only the first EOF, which exhibited a warming trend, is sufficient for the tropics and SH.

1. Introduction

The existence of a zonally symmetrical pattern has been one of the important issues toward understanding atmospheric low-frequency variability since the late 1930s. The well-known examples include the index cycle (e.g., Rossby and Willett 1948; Wallace and Hsu 1985), the tropical zonally symmetrical pattern associated with the El Niño–Southern Oscillation (ENSO) (e.g., Angell and Korshover 1978; Pan and Oort 1983; Yulaeva and Wallace 1994; Hsu 1994; Lu et al. 2008), and the northern/southern annular mode or Arctic/Antarctic Oscillation (AO/AAO) (e.g., Thompson and Wallace 1998, 2000; Thompson et al. 2000).

The relationship between the zonally averaged geopotential height and the AO and ENSO can be seen clearly in Fig. 1, which demonstrates 5-month running-averaged geopotential height at 200 hPa (H200) in the Northern Hemisphere (NH). In the tropics (0°–30°N), positive and negative anomalies of H200 tended to coincide with El Niño and La Niña, respectively, with a correlation coefficient of 0.46. The 65°–90°N averaged H200 variation and the AO are also well correlated with a significant correlation of −0.74. In addition to these two prominent features, a unique multidecadal fluctuation emerges in the extratropics, characterized by mostly negative anomalies during the late 1960s to late 1980s and positive anomalies before and after. To the best of our knowledge, this phenomenon has not been well documented and explored.

Fig. 1.

Zonal-mean geopotential height at 200 hPa from the equator to 90°N and the AO and Niño-3.4 index. Normalized geopotential heights are shown in colors. The AO and Niño-3.4 index are plotted at the top and bottom edge, respectively. Five-month running means are performed to all plotted variables. The multidecadal fluctuation in the extratropics is illustrated by the three dashed boxes. The middle box represents the negative phase from January 1968 to February 1988.

Fig. 1.

Zonal-mean geopotential height at 200 hPa from the equator to 90°N and the AO and Niño-3.4 index. Normalized geopotential heights are shown in colors. The AO and Niño-3.4 index are plotted at the top and bottom edge, respectively. Five-month running means are performed to all plotted variables. The multidecadal fluctuation in the extratropics is illustrated by the three dashed boxes. The middle box represents the negative phase from January 1968 to February 1988.

To further demonstrate the uniqueness of this phenomenon, the time series is compared with well-known large-scale fluctuations such as the global-mean surface temperature and decadal variations like the Pacific decadal oscillation (PDO), the North Atlantic Oscillation (NAO)/AO, and the Atlantic multidecadal oscillation (AMO). The 9-yr running and zonal means of annual-mean H200 averaged over 35°–60°N is illustrated in Fig. 2, along with indices of the aforementioned patterns/oscillations. The H200 exhibits a decreasing trend before the late 1960s and an increasing trend after the late 1980s, reflecting the multidecadal fluctuation seen in Fig. 1. The temporal evolution of the global-mean surface temperature is characterized by the well-known stepwise increasing trend. The PDO, which is the most significant decadal sea surface temperature (SST) fluctuation over the extratropical North Pacific (Latif and Barnett 1994; Mantua et al. 1997), switched from positive phase to negative phase in the 1940s and again, but in opposite sign, around 1976 (Parker et al. 2007). The NAO and AO can be viewed as the measure of westerly strength, except that the former represents local variability over the North Atlantic (Jones et al. 1997), and the latter describes a hemispheric-wide pattern. Both the NAO and AO fluctuated in decadal to multidecadal time scales with an increase from the 1960s to 1990s and a decrease after that (Scaife et al. 2005; Thompson and Wallace 2000). It is clear that the climate indices described above did not fluctuate consistently with the extratropical H200. Nevertheless, the AMO that is defined as the average annual-mean SST anomalies over the North Atlantic (25°–60°N, 75°–5.5°W) is an exception. The AMO exhibits a similar temporal evolution with H200 in the extratropical NH. It suggests a possible connection between the two.

Fig. 2.

Nine-year running means of climate indices [i.e., global-mean surface temperature, the Atlantic multidecadal oscillation (AMO), the Pacific decadal oscillation (PDO), the Arctic Oscillation (AO), and the North Atlantic Oscillation (NAO)] and zonal-mean ERA-40 H200 averaged over 35°–60°N (red line). Climate indices are denoted by curves in various colors labeled in the upper left corner. All indices are annual means normalized by their own mean and standard deviation.

Fig. 2.

Nine-year running means of climate indices [i.e., global-mean surface temperature, the Atlantic multidecadal oscillation (AMO), the Pacific decadal oscillation (PDO), the Arctic Oscillation (AO), and the North Atlantic Oscillation (NAO)] and zonal-mean ERA-40 H200 averaged over 35°–60°N (red line). Climate indices are denoted by curves in various colors labeled in the upper left corner. All indices are annual means normalized by their own mean and standard deviation.

The AMO, a pronounced multidecadal SST fluctuation, is characterized by basinwide anomalies in the North Atlantic (Folland et al. 1984; Kushnir 1994; Schlesinger and Ramankutty 1994; Kerr 2000). Because of its impact on global climate, the AMO has been a major research focus for many years. Studies have proposed that the AMO could modulate the global and Northern Hemisphere mean surface temperature (e.g., Zhang et al. 2007; DelSole et al. 2011; Wu et al. 2011; Zhou and Tung 2013). Other impacts include the meridional shift of the intertropical convergence zone (Ting et al. 2011); the tropical rainfall over South America and Africa (Zhang and Delworth 2006; Sutton and Hodson 2007); the genesis of the tropical cyclone in the North Atlantic (Goldenberg et al. 2001); the North American monsoon strength, onset, and retreat (Arias et al. 2012); the tropical and extratropical SST variability in the North Pacific (Dong et al. 2006; Zhang and Delworth 2007); the East Asian summer monsoon (Lu et al. 2006); and so on. However, the connection between the AMO and the multidecadal fluctuation of the extratropical H200 in the NH has not been previously presented.

The aim of this study is to explore the characteristics and probable origin of this multidecadal phenomenon through circulation diagnostics and numerical simulations. Section 2 describes data and analysis procedures. Diagnostics results based on observational data and numerical simulations are presented in sections 3 and 4, respectively. A summary and discussion are given in section 5.

2. Data, model, and analysis procedure

The monthly data used in this study include the 1) 40-yr European Centre for Medium-Range Weather Forecasts Re-Analysis (ERA-40), 2.5° by 2.5° from September 1957 to August 2002 (Uppala et al. 2005); 2) the Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST1.1), 1° by 1° from 1870 to present (Rayner et al. 2003); 3) the Climate Research Unit (CRU) temperature (TS2.0), 0.5° by 0.5° from 1901 to 2002 (Mitchell et al. 2004); 4) British Atmospheric Data Centre (BADC) Northern Hemisphere (north of 15°N) 500-hPa geopotential height fields, 5° latitude by 10° longitude from 1945 to 2005 (available from http://badc.nerc.ac.uk/view/badc.nerc.ac.uk__ATOM__dataent_HEIGHT); and 5) Twentieth Century Reanalysis version 2, 2.5° by 2.5° from 1871 to 2008 (Compo et al. 2011). The National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) Reanalysis 1, which is a longer dataset, was not used in this study because of the dubious surface condition over Mongolia and its vicinity before the 1970s (Inoue and Matsumoto 2004).

Several multidecadal simulations were carried out using the ECHAM5 (Roeckner et al. 2003, 2006) to explore the effect of the North Atlantic SST anomaly in inducing the multidecadal fluctuation reported in this study. The resolution chosen for this study is a triangular truncation at wavenumber 42 (T42), which is equivalent to 2.8° by 2.8° with 19 vertical sigma levels from the surface to 10 hPa.

To isolate the signals on decadal and longer time scales, annual means of all data were calculated and then low-pass filtered through 9-yr running averaging. Empirical orthogonal function (EOF) analysis was performed to extract principal signals from the original data. Composite analysis was applied to various variables to identify the spatial structure. Statistical significance tests were applied to composite results. The null hypothesis of no relationship was rejected when the computed t score was larger than the t value at 95% in a Student's t distribution. Only those results rejecting the null hypothesis are plotted in most of the following figures. In addition, we also applied the empirical mode decomposition (EMD) (Huang et al. 1998; Wu et al. 2007) to more precisely isolate the climate signals in this study.

3. Characteristics of the EAMPO

As seen in Fig. 2, H200 in the extratropical NH exhibits a unique multidecadal oscillation. To further investigate its structure and mechanism, the EOF analysis was applied to the 9-yr running means of H200 over the NH from 30° to 70°N, the tropics from 30°S to 30°N, and the Southern Hemisphere (SH) from 70° to 30°S. Note that the same patterns can also be identified when unsmoothed data are used. Figure 3 illustrates the first EOFs of these three regions. Interestingly, the first EOFs of the three regions could be almost perfectly combined to form a global pattern. Moreover, their time evolutions [principal components (PCs)] also exhibit a similar increasing trend (Fig. 3b). This combined pattern, which can be fully reproduced by applying the EOF analysis to global H200, is characterized by an overall increasing height in the tropics, the SH, and the middle latitudes over North America and the North Atlantic. This spatial distribution is accompanied by the first EOF of SST (from 70°S to 70°N), which exhibits a nearly global warming structure except for the cooling in the North Pacific (figure not shown), similar to the cool ocean–warm land pattern (Zhang et al. 1997). These results suggest the first EOFs of H200 in the three regions reflect the climate variability associated with a warming tendency in the second half of the twentieth century. An interesting contrast between the NH and the remaining regions is that, while the first EOFs in the tropics and the SH explain nearly 90% (89.4% and 88.7%, respectively) of variance, only 46.5% is explained in the NH. That means the variability of H200 beyond the interannual time scale in the tropics and the SH could be well represented by one single quasi-linear trend. In contrast, a quasi-linear trend can only explain less than 50% of the variance in the NH. This suggests that the interdecadal variability of H200 in the NH is much more complicated than that in the tropics and the SH.

Fig. 3.

(a) First EOF of 9-yr running H200 during 1960–97 for the NH, the tropics, and the SH and (b) the corresponding principal component. The explained variance is 46.5% for the NH, 89.4% for the tropics, and 88.7% for the SH.

Fig. 3.

(a) First EOF of 9-yr running H200 during 1960–97 for the NH, the tropics, and the SH and (b) the corresponding principal component. The explained variance is 46.5% for the NH, 89.4% for the tropics, and 88.7% for the SH.

The second EOF for the NH shown in Fig. 4 reveals another important feature that contributed 29.8% of variance, compared to the second EOFs in the tropics and SH, which explain only 5.6% and 4.6% of variance, respectively. Evidently, it took two EOFs in the NH to explain the equivalent amount (about 76%) of variance as the first EOF did in the tropics and SH. This contrast between the NH and other regions demonstrates more complicated climate variability and processes in the NH. The EOF2 in the NH is characterized by a nearly circumpolar perturbation of the same sign between 30° and 60°N, except in eastern Canada and the North Atlantic north of 40°N where signals are weak. Embedded in this nearly zonally symmetrical pattern are the maxima over Europe, East Asia, and the North Pacific. The distance between maxima suggests the existence of an extratropical eddy (defined as the deviation from the zonal mean) with a spatial scale equivalent to zonal wavenumbers 4 and 5. This zonal scale was identified in a zonal wavenumber spectral analysis (not shown). The temporal evolution of EOF2 reveals a decreasing trend in the 1960s and an increasing trend after the late 1980s (Fig. 4b). This temporal variation is almost identical to the multidecadal fluctuation of the H200 in the extratropical NH shown in Fig. 2. The strong zonal symmetry of the EOF2 explains the zonal-mean feature seen in Fig. 1. This pattern is named the Eurasian–Pacific multidecadal oscillation (EAPMO) to characterize its multidecadal time scale and spatial distribution. The EAPMO appears to play an important role in shaping the multidecadal variability in the extratropical NH. The major characteristics of this pattern will be further explored below. The period when the positive (negative) PC2 amplitude was larger than 0.5 standard deviations is defined as the positive (negative) phase. As a result, there are two positive-phase periods from 1958 to 1966 and 1990 to 2001, and one negative-phase period from 1970 to 1985. Composites for each phase were calculated from the unsmoothed fields of the chosen years. To enhance the characteristics of this pattern, composite differences were further calculated by subtracting the negative phase from the positive phase. The resultant composite of H200 shown in Fig. 5a resembles the spatial pattern of the second EOF (Fig. 4a). The positive anomalies of H200 were located to the north of the jet stream core. The meridional gradient of the vertically integrated 600–250-hPa temperature shown in Fig. 5b reveals a weakened temperature gradient near the jet stream core and enhanced gradient in the north of the positive H200 anomaly. These features reveal the weakened and strengthened zonal wind speed in the core and poleward side of the jet stream during the positive phase, respectively. This indicates a weakened zonal wind shear near the jet stream core.

Fig. 4.

As in Fig. 3, but for the second EOF in the NH. Explained variance is 29.8%.

Fig. 4.

As in Fig. 3, but for the second EOF in the NH. Explained variance is 29.8%.

Fig. 5.

Composite differences between the positive and negative phases of the second PC of H200 in the NH: (a) H200 (m), (b) meridional 600–250-hPa temperature gradient (10−5 K km−1), and (c) land and sea surface temperature (K). The shading areas represent the regions exceeding the 95% confidence level. The dotted line segments in (a) and (b) indicates the location of climatological jet stream at 200 hPa.

Fig. 5.

Composite differences between the positive and negative phases of the second PC of H200 in the NH: (a) H200 (m), (b) meridional 600–250-hPa temperature gradient (10−5 K km−1), and (c) land and sea surface temperature (K). The shading areas represent the regions exceeding the 95% confidence level. The dotted line segments in (a) and (b) indicates the location of climatological jet stream at 200 hPa.

Surface temperature anomalies (Fig. 5c) exhibited widespread warming during the positive phase with larger amplitude in the extratropical NH, especially in those regions where positive H200 anomalies were found. In contrast, the signals in the SH are relatively weak. The warming surrounding the Mediterranean, East Asia, and western coast of North America was associated with the weakened jet stream. The warming in the Mediterranean and East Asia was similar to the warming pattern associated with the climate shift occurring in the late 1980s as documented by Lo and Hsu (2010). Another interesting feature in Fig. 5c is the basinwide warming in the North Atlantic, while the positive H200 anomaly was only observed in the southeastern corner of the North Atlantic (Fig. 5a). The basinwide warming in the North Atlantic reflected the basic feature of the AMO. The temporal evolution of PC2 (Fig. 4b) also exhibits behavior similar to the AMO. The existence of this warm feature suggests that the EAPMO might be excited by surface temperature anomalies in the North Atlantic.

To explore the vertical structure of the EAPMO, the zonal cross sections of geopotential height and temperature averaged over 40°–60°N are presented in Fig. 6. The EAPMO shows an equivalent barotropic structure in geopotential height with the largest amplitude in the upper troposphere. Maxima were identified over western Europe, East Asia, the North Pacific, and western North America. Interestingly, these anomalies extended all the way down to the surface over East Asia and North America where mountain ranges exist. This regional contrast suggests that mountains might play a role in enhancing the eddy amplitude. The vertical profile of temperature anomalies in Fig. 6b shows a close association with the geopotential heights. Another interesting feature is the two pronounced warming regions in the vertical profile over the mountainous regions: one in the middle upper troposphere and another near the surface. The latter may be attributed to mountain forcing.

Fig. 6.

Zonal cross sections of (a) geopotential height (m) and (b) temperature differences (K) between the positive and negative phases of the second PC of H200 in the NH, averaged over 40°–60°N. Only the signals exceeding the 95% confidence level are plotted. The topography profile is plotted at the bottom of each panel.

Fig. 6.

Zonal cross sections of (a) geopotential height (m) and (b) temperature differences (K) between the positive and negative phases of the second PC of H200 in the NH, averaged over 40°–60°N. Only the signals exceeding the 95% confidence level are plotted. The topography profile is plotted at the bottom of each panel.

Another question is why eddies embedded in the EAPMO exhibited maximum amplitudes on the poleward side of the jet stream. To answer this, the climatological stationary Rossby wavenumber was calculated. The stationary Rossby wavenumber can written as

 
formula

in which Ks is valid if both the zonal wind and meridional gradient of absolute vorticity are positive. As shown in Fig. 7a, the stationary wavenumber distribution seemed to be separated by the jet stream into two groups: regions of lower (<5) and higher (≥5) wavenumbers on the poleward and equatorward sides of the jet stream, respectively. It follows that stationary waves of lower wavenumber tended to exist at the poleward side of the jet stream, while the shorter waves were likely to occur on the southern flank of the jet stream. As mentioned in previous sections, the EAPMO was characterized by wavenumbers 4 and 5 and therefore tended to appear on the poleward side of the jet stream.

Fig. 7.

(a) Climatological Rossby stationary wavenumber (shaded color) and H200 differences between the positive and negative phases of the second PC of H200 (red contours: the three contours for the H200 are 15, 24, and 33 m). (b) The 200–300-hPa averaged wave activity flux (vector) and 200-hPa geopotential height eddy (shaded color). (c) Zonal cross section of wave activity flux (vector) and geopotential height eddy averaged over 40°–60°N (shaded color). The eddy is defined as the deviation from zonal mean.

Fig. 7.

(a) Climatological Rossby stationary wavenumber (shaded color) and H200 differences between the positive and negative phases of the second PC of H200 (red contours: the three contours for the H200 are 15, 24, and 33 m). (b) The 200–300-hPa averaged wave activity flux (vector) and 200-hPa geopotential height eddy (shaded color). (c) Zonal cross section of wave activity flux (vector) and geopotential height eddy averaged over 40°–60°N (shaded color). The eddy is defined as the deviation from zonal mean.

Wave activity flux (WAF) ( Takaya and Nakamura 2001) averaged over 300–200 hPa, shown in Fig. 7b, revealed a tendency of wave energy propagation from the extratropical North Atlantic southeastward to North Africa and also eastward across the extratropical Eurasian continent to East Asia along the path where eddies were observed. The zonal vertical cross section along 40°–60°N shown in Fig. 7c reveal this tendency in more details. The eastward propagation of wave activity, although not consistently strong across the whole Eurasian continent, was observed mainly in the upper troposphere. Strong upward WAF was also observed over the Asian highland regions where the eddy exhibited the largest amplitude and westward tilting with increasing height. Another region of upward WAF occurred over the North Pacific where the warm ocean surface is observed in Fig. 5c. It is interesting to note that the geopotential height anomalies in these two upward WAF regions tilted westward with increasing height, a characteristic consistent with the upward energy dispersion of a Rossby wave. This result suggests that regional topography and SST anomaly might act to enhance the wave activity and eddy amplitude associated with the EAPMO.

The above analysis was based on the annual means. The seasonal dependence of the EAPMO was also diagnosed following the same analysis procedures. While the same pattern can be obtained in all four seasons, the EAPMO was weaker in summer and fluctuated in different tempos from the other three seasons and annual mean. The time series of H200 averaged between 35° and 60°N, presented in Fig. 8, shows that the annual mean and all four seasons, except the summer, fluctuated concurrently through the whole period. The EAPMO in the summer stayed in the negative phase longer than the other seasons until the mid-1990s when it sharply switched to the positive phase and increased dramatically. Because the EAPMO existed in all seasons, the following discussion will continue focusing on annual means.

Fig. 8.

Time series of 9-yr low-pass filtered annual and seasonal means of the zonal-mean H200 anomaly averaged over 35°–60°N.

Fig. 8.

Time series of 9-yr low-pass filtered annual and seasonal means of the zonal-mean H200 anomaly averaged over 35°–60°N.

To examine the signal of the EAPMO more carefully and also to explore whether the pattern existed before the 1950s, EMD was applied to various datasets such as BADC 500-hPa height (H500) and the Twentieth Century Reanalysis in addition to the ERA-40. EMD was also applied to the North Atlantic SST anomaly (SSTA) averaged over 25°–60°N, 75°W–10°E where the observed SSTA exhibited largest amplitude (Fig. 5c). This SSTA index is referred to as the extratropical AMO hereafter. The intrinsic mode functions (IMF) representing multidecadal fluctuations in the aforementioned datasets are shown in Fig. 9. These IMFs explain 31.1%, 42.1%, 20.7%, and 43.8% of total variance in ERA-40, BADC, the Twentieth Century Reanalysis, and extratropical AMO, respectively. The multidecadal swing between the 1950s and early 2000s that was seen in ERA-40 H200 was also evident in BADC H500 and the Twentieth Century Reanalysis, although discrepancies existed. The periodicity was longer in BADC H500 and the Twentieth Century Reanalysis than the one in ERA-40. While the BADC H500 index dropped concurrently with the ERA-40 in the 1960s, it increased and changed from negative to positive signs in the late 1980s and early 1990s, 5–6 years later than the ERA-40 but more consistent with the Twentieth Century Reanalysis. Different methodologies and data sources in producing these datasets might cause the discrepancies. Nevertheless, the overall consistency between analyses confirms the existence of the multidecadal nature of the EAMPO. The oscillation revealed in the Twentieth Century Reanalysis can be traced back to the early twentieth century with a clear synchronization with the extratropical AMO. The similar temporal behavior between the multidecadal variability in the North Atlantic SST and the Twentieth Century Reanalysis H200 indicates that the AMO-like SSTA might play an important role in inducing the EAPMO, at least since the early twentieth century. Many studies suggested the existence of the AMO over the last several centuries (Delworth and Mann 2000; Knight et al. 2005) and even throughout the Holocene (Wei and Lohmann 2012). Whether the EAPMO has existed in the earth climate system for a period as long as the AMO is an intriguing question. Although such a question may not be answered by data analysis, idealized numerical simulation may provide hints.

Fig. 9.

Intrinsic mode function derived from zonal means of the 200-hPa geopotential height of ERA-40 and Twentieth Century Reanalysis and the BADC 500-hPa geopotential height averaged over 35°–60°N and SST averaged over the North Atlantic (25°–60°N, 75°W–10°E). The intrinsic mode function is the third for ERA-40 H200 but is the fourth for others. All indices are normalized by their own mean and standard deviation. Explained variance is marked in the upper part of figure.

Fig. 9.

Intrinsic mode function derived from zonal means of the 200-hPa geopotential height of ERA-40 and Twentieth Century Reanalysis and the BADC 500-hPa geopotential height averaged over 35°–60°N and SST averaged over the North Atlantic (25°–60°N, 75°W–10°E). The intrinsic mode function is the third for ERA-40 H200 but is the fourth for others. All indices are normalized by their own mean and standard deviation. Explained variance is marked in the upper part of figure.

4. Numerical experiments

The association of the EAPMO with the SSTA in the North Atlantic suggests that the EAPMO may be forced by the AMO-like SSTA. To test this hypothesis, numerical experiments were carried out to force the models by adding an idealized AMO-like SSTA to the climatological monthly SST. The SSTA was obtained by regressing 9-yr running annual-mean SST onto the AMO index. Following Sutton and Hodson (2007), a cosine-squared smoothing with weight zero outside the region 5°S–65°N, 82.5°W–5°E was applied to the prescribed SSTA for isolating the North Atlantic part of the pattern and avoiding discontinuities in the SST gradient. Because the northern boundary of the imposed SSTA only extends to 65°N, the effect from sea ice was ignored. The prescribed AMO-like SSTA fluctuated with a period of 70 yr. Two sets of experiments forced by 2 × and 1 × observed SSTA, named double and realistic SSTA experiments, were carried out. While both double and realistic SSTA experiments produced similar results, those of double SSTA experiments are shown here since amplitudes are larger. This practice of enhanced forcing has been widely applied in previous studies (e.g., Sutton and Hodson 2007). Although the SSTA is exaggerated, such a simulation can still provide important physical insights. Simulation results with different SSTA strength will be further discussed in the conclusions. Four simulations starting from different initial conditions were performed. The length of each simulation was 107 years, equivalent to 1.5 prescribed SSTA cycles. The ensemble means of four members between year 38 and 107 were calculated and used for the following diagnostics and comparison with the observation.

Simulated 9-yr low-pass filtered and interannual H200 anomaly averaged over three different latitudinal bands are shown in Fig. 10. Interannual variability is significant in all three latitudinal bands, although no interannual variation in the forcing was prescribed. It is clearly seen in Fig. 10 that interannual noises were embedded in a multidecadal fluctuation in the NH. The ratios between multidecadal and total variance are 0.39, 0.31, and 0.14 at 35°–60°N, 35°S–35°N, and 35°–60°S, respectively. The 9-yr low-pass filtered H200 anomalies in the NH tend to be positive (negative) during the warm (cool) SST phase. In contrast, this relationship between H200 and the imposed SST anomalies is weak in the tropics and not evident in the SH. This result indicates that the AMO-like SST forcing induces multidecadal fluctuations in the NH upper troposphere, although its variability appears to be weaker than the inherent interannual variability in the model. A statistical measure appears to be necessary to extract the multidecadal variability. Same diagnostics applied to the observation data were applied to the 9-yr running means of the simulated anomalies.

Fig. 10.

Time series of simulated 9-yr low-pass filtered (solid and thick lines) and interannual (dashed and thin lines) H200 anomalies averaged from (a) 35° to 60°N, (b) 35°S to 35°N, and (c) 60° to 35°S. The prescribed SST fluctuation in the North Atlantic in the double SSTA experiment is also shown in color shading. Left and right ordinates are the scales for the H200 and SST, respectively.

Fig. 10.

Time series of simulated 9-yr low-pass filtered (solid and thick lines) and interannual (dashed and thin lines) H200 anomalies averaged from (a) 35° to 60°N, (b) 35°S to 35°N, and (c) 60° to 35°S. The prescribed SST fluctuation in the North Atlantic in the double SSTA experiment is also shown in color shading. Left and right ordinates are the scales for the H200 and SST, respectively.

The first EOFs of the simulated H200 and 2-m air temperature (T2m), explaining 73.1% and 78% of the interdecadal variance (Figs. 11a,b), respectively, exhibited close resemblance to the observed patterns (Figs. 5a,c), for example, the positive H200 anomaly in the 40°–60°N band over the Eurasian continent, the North Pacific, and the subtropical North Atlantic and also the positive T2m anomaly in the Eurasian continent. The positive T2m over the North Atlantic was mainly due to the prescribed SSTA. The corresponding time series of the first EOFs for both H200 and T2m (Fig. 11c) exhibited a multidecadal fluctuation synchronizing with the imposed SSTA. These results confirmed the proposed hypothesis; that is, the EAPMO can be driven by AMO-like SST fluctuations.

Fig. 11.

First EOF of 9-yr running means of simulated (a) H200 and (b) T2m that explains 73.1% and 78% of variance, respectively, in the double SSTA experiment. (c) The amplitude of the corresponding principal component for H200 (blue line, left ordinate) and T2m (dashed red line, right ordinate). The gray curve in (c) marks the prescribed SST fluctuation.

Fig. 11.

First EOF of 9-yr running means of simulated (a) H200 and (b) T2m that explains 73.1% and 78% of variance, respectively, in the double SSTA experiment. (c) The amplitude of the corresponding principal component for H200 (blue line, left ordinate) and T2m (dashed red line, right ordinate). The gray curve in (c) marks the prescribed SST fluctuation.

The corresponding vertical structures of the geopotential height and temperature averaged over 40°–60°N shown in Fig. 12 are the composite differences between the positive and negative SSTA phases, that is, prescribed SST anomaly greater than 0.5 and less than −0.5 in Fig. 10c, respectively. The good agreement with the observed perturbation is evident. The overall positive geopotential associated with warming in the troposphere and cooling in the lower stratosphere was reasonably simulated. Larger geopotential amplitudes were simulated over Europe, East Asia, and the North Pacific, as observed. The observed large temperature amplitude over the Asian mountains and the near-surface and tropospheric warming over the North Atlantic were reasonably simulated. The experiments also realistically simulated the wave energy propagation from the extratropical North Atlantic to the North Pacific (not shown here) as that shown in Figs. 7b and 7c.

Fig. 12.

Zonal cross section of simulated (a) geopotential height and (b) temperature averaged over 40°–60°N in the double SSTA experiment. The composite is defined as differences between the positive and negative phases of the prescribed SST anomaly.

Fig. 12.

Zonal cross section of simulated (a) geopotential height and (b) temperature averaged over 40°–60°N in the double SSTA experiment. The composite is defined as differences between the positive and negative phases of the prescribed SST anomaly.

The model also simulated some unrealistic features, for example, a lack of positive T2m in western North America and a negative T2m anomaly in Canada. Lack of T2m response over the oceanic regions besides the North Atlantic is expected because T2m is driven mostly by the local SST in the AGCM and no SSTA was prescribed except in the North Atlantic. Furthermore, while the simulated features are statistically significant, the amplitudes are only half of the observed. It is interesting to note that the model also undersimulated overall variability, that is, smaller variance than the observed. A better approach than simply examining the amplitude itself is to compare the results in terms of the relative amplitude of multidecadal fluctuation, for example, the multidecadal amplitude divided by total standard deviation, for both simulated and observed fields. Cross sections of relative amplitude for geopotential height and temperature are presented in Fig. 13. The relative amplitudes range from 0.7 to 0.9 in the major signal regions. That means the relative strength of model responses (in this case, signals) in the double SSTA experiments are about the same as in the real world. In contrast, the model responses in the realistic SSTA experiments are still weaker than the observed. The weaker amplitudes may be due to the deficiencies of the model and/or the experiment designs. Forcing other than SST (e.g., the declining sea ice) that may have a strong effect on the phenomenon was not considered in the experiments. Nevertheless, the results presented above support our hypotheses that the EAPMO could be forced by the AMO-like SSTA. This is likely why the pattern existed only in the northern extratropics. Much weaker multidecadal signals in the SH are likely due to the lack of similar forcing. The prescribed SSTA in the North Atlantic was not able to force the observed anomalies in North America. It is likely that the observed anomalies in North America were induced by other factors that were not considered in our simulations.

Fig. 13.

Zonal cross section of observed (a) geopotential height and (b) temperature averaged over 40°–60°N. Both variables were divided by the standard deviation of unsmoothed data. The composite is defined as the differences between the positive and negative phase of the second PC of H200 in the NH. (c),(d) As in (a),(b), but for the double SSTA experiment with topography; the positive (negative) phase is defined as the prescribed SST anomaly greater (less) than 0.5 (−0.5) in Fig. 10c.

Fig. 13.

Zonal cross section of observed (a) geopotential height and (b) temperature averaged over 40°–60°N. Both variables were divided by the standard deviation of unsmoothed data. The composite is defined as the differences between the positive and negative phase of the second PC of H200 in the NH. (c),(d) As in (a),(b), but for the double SSTA experiment with topography; the positive (negative) phase is defined as the prescribed SST anomaly greater (less) than 0.5 (−0.5) in Fig. 10c.

To further investigate the cause for the amplification of the EAPMO over the Asian highlands, which were evident in both observation and simulation, an additional series of experiments with no topography were performed to explore the topographic effect. These additional experiments were designed in the same way as the previous ones except that the topographic height was set to zero over the Eurasian and North American continents. Through a careful comparison between the original and the no topography experiments, we were able to demonstrate the topographic effect on the EAPMO. As shown in Fig. 14, distinct regional maxima in H200 and T2m are less evident. For example, both H200 and T2m anomalies over Eurasia in the no topography experiments exhibited a more uniform spatial distribution. The interesting feature is the maximum T2m in East Asia occurring in northeast Asia instead of occurring in southeast China as seen in the observation and simulations with topography. It is well known that the existence of the high-rising topography in East Asia plays an important role in steering the cold air southward and results in cold surges in southern East Asia (Hsu and Wallace 1985; Hsu 1987). The difference between the original and no topography simulations is consistent with the findings of previous studies. A smoother zonal distribution and much weaker amplitude in the zonal cross section of both geopotential height and temperature were also evident over the region where the Asian highland should be (Fig. 15). The results obtained from the no topography experiments provide the evidence that the EAPMO induced by the AMO-like SSTA is strengthened regionally by topography.

Fig. 14.

As in Figs. 11a,b, but for no topography experiments. Explained variances are 73.9% for H200 and 77.6% for T2m.

Fig. 14.

As in Figs. 11a,b, but for no topography experiments. Explained variances are 73.9% for H200 and 77.6% for T2m.

Fig. 15.

As in Fig. 12, but for no topography experiments.

Fig. 15.

As in Fig. 12, but for no topography experiments.

5. Summary and discussion

This study identified a unique and new multidecadal pattern that is characterized by a nearly zonally symmetrical geopotential height anomaly in the upper troposphere of the extratropical NH between 40° and 60°N. Its southern edge coincides with the climatological location of the jet stream core. The pattern reveals an equivalent barotropic vertical structure with the largest amplitude in the upper troposphere. This pattern is named the Eurasian–Pacific multidecadal oscillation (EAPMO) to depict its multidecadal time scale and the largest amplitudes over Eurasia and the North Pacific. Although nearly extending over the whole northern extratropics, the EAPMO exhibits larger amplitudes over western Europe, East Asia, and the North Pacific with a zonal scale equivalent to zonal wavenumbers 4 and 5. The eddy tends to become amplified over major mountain ranges, suggesting the significant influence of topography. During the positive (negative) phase of the EAPMO, a weakened (strengthened) zonal wind occurs near the jet stream core in the NH, and the surface temperature anomaly is positive (negative) in the North Atlantic, western Europe, East Asia, North Pacific, and western North America. Although the pattern was identified based on the reanalysis data of recent decades (e.g., ERA-40 and BADC H500), similar signals were also identified in the Twentieth Century Reanalysis and century-long sea surface temperature data (e.g., HadISST), suggesting that the EAMPO has very likely been present for at least a century.

Evidences from observational data indicate the close temporal association of the EAPMO with the AMO-like SSTA in the twentieth century, suggesting that AMO-like SSTA might be the major driver of the EAPMO. The connection between them could be explained as follows. For the sake of simplicity, the following discussion is based mainly on the positive AMO phase. Reversed conditions occur during the negative AMO phase. The positive AMO phase, characterized by basinwide warming over the North Atlantic with the largest amplitude in midlatitudes, induces overall warming and a positive height anomaly in the troposphere of the extratropical NH and also Rossby wavelike activity propagating from the North Atlantic to the North Pacific. The induced wavelike perturbation exhibits a zonal scale equivalent to wavenumbers 4 and 5 embedding in a nearly zonally symmetrical perturbation. The existence of the wavenumber 4 and 5 eddies in the observed latitudinal band is consistent with the stationary Rossby wavenumber calculation based on the long-term climatological flow at 200 hPa. This may explain why the induced perturbations downstream of the North Atlantic exhibited such a zonal scale. The eddies were enhanced over the Asian highland region and the extratropical North Pacific. The observed upward wave activity flux over the Asian highland region suggests that topographically induced Rossby wavelike disturbances may act to enhance the local perturbation and farther downstream propagation. The warmer SST in the extratropical North Pacific may also have a similar effect on the eddy.

To examine the above interpretation, numerical experiments were performed to explore the atmospheric responses to the AMO-like SSTA. Major characteristics of the EAMPO were reproduced well, although with amplitudes smaller than the observed, by adding an idealized AMO-like SSTA to the climatological monthly SST. The result confirmed our hypothesis that the EAMPO, which is a unique multidecadal signal found only in the extratropical NH, was likely forced by the AMO-like SSTA. The contrast between the original forcing experiment and no topography experiment also supports the hypothesis that the topographic forcing acted to enhance the EAMPO in the vicinity of mountain ranges, especially the Asian highland.

The finding reported here can also explain other large-scale phenomena. For example, the late 1980s climate shift in the wintertime surface temperature reported in previous literature may be attributed to the AMO changing from cold to warm phases. Lo and Hsu (2010) found that the late 1980s climate shift occurred synchronously in a widespread area in the extratropical NH (e.g., northern East Asia, the Kuroshio Extension, Europe, and the southeastern United States), whereas cooling occurred in eastern Canada. These warming and cooling features can also be found in the simulated T2m anomaly shown in Fig. 11b. This result suggests that the AMO warming phase occurring in the late 1980s and early 1990s may be responsible, at least partially, for the synchronized climate shift and the observed NH warming that started in the late 1980s.

This study found that multidecadal variability in the extratropical NH is much more complicated than in the tropics and the SH. It takes both first (warming trend) and second (multidecadal) EOFs to explain the multidecadal variability in the extratropical NH, while only the first EOF, which exhibited a warming trend, is sufficient for the tropics and the SH. This complexity can also be seen in Fig. 16, which presents a time series of area-mean surface temperatures in the extratropical NH and SH, the tropics, and the whole globe. The surface temperature in the SH is characterized by a stronger centennial warming trend and a weaker multidecadal fluctuation, whereas the surface temperature is characterized by a stronger multidecadal fluctuation and weaker centennial warming trend in the NH. This contrast was particularly evident between the 1970s and 1990s. The cooling trend embedded in the multidecadal fluctuation in the Northern Hemisphere was partially offset by the continuing warming trend in the SH. As a result, the global-mean temperature fluctuated with averaged characteristics of the NH and SH and showed greatest resemblance to the tropical surface temperature, which has a much weaker multidecadal signal.

Fig. 16.

Time series of area-mean surface temperature in the extratropical NH (20°–90°N), SH (90°–20°S), the tropics (20°S–20°N), and the whole globe.

Fig. 16.

Time series of area-mean surface temperature in the extratropical NH (20°–90°N), SH (90°–20°S), the tropics (20°S–20°N), and the whole globe.

It has been recently proposed that the multidecadal characteristic of the global-mean temperature may be attributed to the AMO (e.g., Zhang et al. 2007; DelSole et al. 2011; Wu et al. 2011; Zhou and Tung 2013). This study further suggests that the multidecadal characteristic in the global-mean surface temperature is mainly a NH contribution, as shown in Fig. 16. The observational and simulated evidences revealed in this study confirmed the finding of previous studies (e.g., Zhang et al. 2007) that the AMO can have a strong effect on the land surface temperature of the extratropical NH. The combination of the AMO itself and the AMO-induced fluctuation might contribute, at least partially, to the distinct multidecadal variation in surface temperature in the NH, a phenomenon that is much less notable in the tropics and the Southern Hemisphere.

One unrealistic part of our simulations is the weak model response to the prescribed SSTA. This common deficiency in AGCMs may be due to limitations in the current AGCMs or Atmospheric Model Intercomparison Project (AMIP)-type simulations, especially when applied in a multidecadal variability simulation that is usually much weaker than interannual variability. One possibility is the model top effect. As demonstrated in a recent study by Omrani et al. (2013), a high-top ECAHM5 produces much stronger signals in response to the SSTA in the North Atlantic. The ECHAM5 used in this study is the original version with the top at 10 hPa. Whether the high-top ECHAM5 will enhance the model response in our simulations will be explored in a future study. Other factors, such as spatial resolution and forcing other than SST (e.g., declining sea ice), that may have a strong effect on the phenomenon will also need to be explored.

Acknowledgments

The authors are grateful to the valuable comments of anonymous reviewers. This study was supported by the National Science Council, Taiwan, under Grant NSC-100-2119-M-001-029-MY5.

REFERENCES

REFERENCES
Angell
,
J. K.
, and
J.
Korshover
,
1978
:
Estimate of global temperature variations in the 100–30-mbar layer between 1958 and 1977
.
Mon. Wea. Rev.
,
106
,
1422
1432
.
Arias
,
P.
,
R.
Fu
, and
K.
Mo
,
2012
:
Decadal variation of rainfall seasonality in the North American monsoon region and its potential causes
.
J. Climate
,
25
,
4258
4274
.
Compo
,
G. P.
, and
Coauthors
,
2011
:
The Twentieth Century Reanalysis Project
.
Quart. J. Roy. Meteor. Soc.
,
137
,
1
28
,
doi:10.1002/qj.776
.
DelSole
,
T.
,
M. K.
Tippett
, and
J.
Shukla
,
2011
:
A significant component of unforced multidecadal variability in the recent acceleration of global warming
.
J. Climate
,
24
,
909
926
.
Delworth
,
T. L.
, and
M. E.
Mann
,
2000
:
Observed and simulated multidecadal variability in the Northern Hemisphere
.
Climate Dyn.
,
16
,
661
676
.
Dong
,
B.
,
R. T.
Sutton
, and
A. A.
Scaife
,
2006
:
Multidecadal modulation of El Niño–Southern Oscillation (ENSO) variance by Atlantic Ocean sea surface temperatures
.
Geophys. Res. Lett.
,
33
, L08705,
doi:10.1029/2006GL025766
.
Folland
,
C. K.
,
D. E.
Parker
, and
F. E.
Kates
,
1984
:
Worldwide marine temperature fluctuations 1856-1981
.
Nature
,
310
,
670
673
.
Goldenberg
,
S. B.
,
C. W.
Landsea
,
A. M.
Mestas-Nuñez
, and
W. M.
Gray
,
2001
:
The recent increase in Atlantic hurricane activity: Causes and implications
.
Science
,
293
,
474
479
.
Hsu
,
H.-H.
,
1987
:
Propagation of low-level circulation patterns in the vicinity of mountain ranges
.
Mon. Wea. Rev.
,
115
,
1864
1892
.
Hsu
,
H.-H.
,
1994
:
Relationship between tropical heating and global circulation: Interannual variability
.
J. Geophys. Res.
,
99
(
D5
),
10 473
10 489
.
Hsu
,
H.-H.
, and
J. M.
Wallace
,
1985
:
Vertical structure of wintertime teleconnection patterns in Northern Hemisphere
.
J. Atmos. Sci.
,
42
,
1693
1710
.
Huang
,
N. E.
, and
Coauthors
,
1998
:
The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis
.
Proc. Roy. Soc. London
,
454
,
903
993
.
Inoue
,
T.
, and
J.
Matsumoto
,
2004
:
A comparison of summer sea level pressure over east Eurasia between NCEP-NCAR Reanalysis and ERA-40 for the period 1960-99
.
J. Meteor. Soc. Japan
,
82
,
951
958
.
Jones
,
P. D.
,
T.
Jónsson
, and
D.
Wheeler
,
1997
:
Extension to the North Atlantic Oscillation using early instrumental pressure observations from Gibraltar and South-West Iceland
.
Int. J. Climatol.
,
17
,
1433
1450
.
Kerr
,
R.
,
2000
:
A North Atlantic climate pacemaker for the centuries
.
Science
,
288
,
1984
1986
.
Knight
,
J. R.
,
R. J.
Allan
,
C. K.
Folland
,
M.
Vellinga
, and
M. E.
Mann
,
2005
:
A signature of persistent natural thermohaline circulation cycles in observed climate
.
Geophys. Res. Lett.
,
32
, L20708,
doi:10.1029/2005GL024233
.
Kushnir
,
Y.
,
1994
:
Interdecadal variations in North Atlantic sea surface temperature and associated atmospheric conditions
.
J. Climate
,
7
,
141
157
.
Latif
,
M.
, and
T. P.
Barnett
,
1994
:
Causes of decadal climate variability over the North Pacific and North American
.
Science
,
266
,
634
637
.
Lo
,
T.-T.
, and
H.-H.
Hsu
,
2010
: Change in dominant decadal modes and the late 1980s abrupt warming in the extratropical Northern Hemisphere. Atmos. Sci. Lett.,11, 210–215,
doi:10.1002/asl.275
.
Lu
,
J.
,
G.
Chen
, and
D. M. W.
Frierson
,
2008
:
Response of the zonal mean atmospheric circulation to El Niño versus global warming
.
J. Climate
,
21
,
5835
5851
.
Lu
,
R.
,
B.
Dong
, and
H.
Ding
,
2006
:
Impact of the Atlantic multidecadal oscillation on the Asian summer
.
Geophys. Res. Lett.
,
33
, L24701,
doi:10.1029/2006GL027655
.
Mantua
,
N. J.
,
S. R.
Hare
,
Y.
Zhang
,
J. M.
Wallace
, and
R. C.
Francis
,
1997
:
A Pacific interdecadal climate oscillation with impacts on salmon production
.
Bull. Amer. Meteor. Soc.
,
78
,
1069
1079
.
Mitchell
,
T. D.
,
T. R.
Carter
,
P. D.
Jones
,
M.
Hulme
, and
M.
New
,
2004
: A comprehensive set of high-resolution grids of monthly climate for Europe and the globe: The observed record (1901–2000) and 16 scenarios (2001–2100). Tyndall Centre Working Paper 55, 30 pp.
Omrani
,
N.-E.
,
N. S.
Keenlyside
,
J.
Bader
, and
E.
Manzini
,
2013
: Stratosphere key for wintertime atmospheric response to warm Atlantic decadal conditions. Climate Dyn.,
doi:10.1007/s00382-013-1860-3, in press
.
Pan
,
Y. H.
, and
A. H.
Oort
,
1983
:
Global climate variations connected with sea surface temperature anomalies in the eastern equatorial Pacific Ocean for the 1958–70 period
.
Mon. Wea. Rev.
,
111
,
1244
1258
.
Parker
,
D.
,
C.
Folland
,
A.
Scaife
,
J.
Knight
,
A.
Colman
,
P.
Baines
, and
B.
Dong
,
2007
:
Decadal to multidecadal variability and the climate change background
.
J. Geophys. Res.
,
112
, D18115,
doi:10.1029/2007JD008411
.
Rayner
,
N. A.
,
D. E.
Parker
,
E. B.
Horton
,
C. K.
Folland
,
L. V.
Alexander
,
D. P.
Rowell
,
E. C.
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:10.1029/2002JD002670
.
Roeckner
,
E.
, and
Coauthors
,
2003
: The atmospheric general circulation model ECHAM5. Part I: Model description. Max Planck Institute for Meteorology Rep. 349, 127 pp.
Roeckner
,
E.
, and
Coauthors
,
2006
:
Sensitivity of simulated climate to horizontal and vertical resolution in the ECHAM5 atmosphere model
.
J. Climate
,
19
,
3771
3791
.
Rossby
,
C.-G.
, and
H. C.
Willett
,
1948
:
The circulation of the upper troposphere and lower stratosphere
.
Science
,
108
,
643
652
.
Scaife
,
A. A.
,
J. R.
Knight
,
G. K.
Vallis
, and
C. K.
Folland
,
2005
:
A stratospheric influence on the winter NAO and North Atlantic surface climate
.
Geophys. Res. Lett.
,
32
, L18715,
doi:10.1029/2005GL023226
.
Schlesinger
,
M. E.
, and
N.
Ramankutty
,
1994
:
An oscillation in the global climate system of period 65–70 years
.
Nature
,
367
,
723
726
.
Sutton
,
R. T.
, and
L. R.
Hodson
,
2007
:
Climate response to basin-scale warming and cooling of the North Atlantic Ocean
.
J. Climate
,
20
,
891
907
.
Takaya
,
K.
, and
H.
Nakamura
,
2001
:
A formulation of a phase-independent wave-activity flux for stationary and migratory quasigeostrophic eddies on a zonally varying basic flow
.
J. Atmos. Sci.
,
58
,
608
627
.
Thompson
,
D. W. J.
, and
J. M.
Wallace
,
1998
:
The Arctic Oscillation signature in the wintertime geopotential height and temperature fields
.
Geophys. Res. Lett.
,
25
(9),
1297
1300
.
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
.
Thompson
,
D. W. J.
,
J. M.
Wallace
, and
G. C.
Hegerl
,
2000
:
Annular modes in the extratropical circulation. Part II: Trends
.
J. Climate
,
13
,
1018
1036
.
Ting
,
M.
,
Y.
Kushnir
,
R.
Seager
, and
C.
Li
,
2011
:
Robust features of Atlantic multi-decadal variability and its climate impacts
.
Geophys. Res. Lett.
,
38
, L17705,
doi:10.1029/2011GL048712
.
Uppala
,
S. M.
, and
Coauthors
,
2005
:
The ERA-40 Re-Analysis
.
Quart. J. Roy. Meteor. Soc.
,
131
,
2961
3012
.
Wallace
,
J. M.
, and
H.-H.
Hsu
,
1985
:
Another look at the index cycle
.
Tellus
,
37A
,
478
486
.
Wei
,
W.
, and
G.
Lohmann
,
2012
:
Simulated Atlantic multidecadal oscillation during the Holocene
.
J. Climate
,
25
,
6989
7002
.
Wu
,
Z.
,
N. E.
Huang
,
S. R.
Long
, and
C. K.
Peng
,
2007
:
On the trend, detrending and variability of nonlinear and non-stationary time series
.
Proc. Natl. Acad. Sci. USA
,
104
,
14 889
14 894
.
Wu
,
Z.
,
N. E.
Huang
,
J. M.
Wallace
,
B. V.
Smoliak
, and
X.
Chen
,
2011
:
On the time-varying trend in global-mean surface temperature
.
Climate Dyn.
,
37
,
759
773
,
doi:10.1007/s00382-011-1128-8
.
Yulaeva
,
E.
, and
J. M.
Wallace
,
1994
:
The signature of ENSO in global temperature and precipitation fields derived from the microwave sounding unit
.
J. Climate
,
7
,
1719
1736
.
Zhang
,
R.
, and
T. L.
Delworth
,
2006
:
Impact of Atlantic multidecadal oscillation on India/Sahel rainfall and Atlantic hurricanes
.
Geophys. Res. Lett.
,
33
, L17712,
doi:10.1029/2006GL026267
.
Zhang
,
R.
, and
T. L.
Delworth
,
2007
:
Impact of the Atlantic multidecadal oscillation on the North Pacific climate variability
.
Geophys. Res. Lett.
,
34
, L23078,
doi:10.1029/2007GL031601
.
Zhang
,
R.
,
T. L.
Delworth
, and
I. M.
Held
,
2007
:
Can the Atlantic Ocean drive the observed multidecadal variability in Northern Hemisphere mean temperature?
Geophys. Res. Lett.
,
34
, L02709,
doi:10.1029/2006GL028683
.
Zhang
,
Y.
,
J. M.
Wallace
, and
D. S.
Battisti
,
1997
:
ENSO-like interdecadal variability: 1900–93
.
J. Climate
,
10
,
1004
1020
.
Zhou
,
J.
, and
K.
Tung
,
2013
:
Deducing multidecadal anthropogenic global warming trends using multiple regression analysis
.
J. Atmos. Sci.
, 70, 3–8.