JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114,, doi:10.1029/2009jd011737, 2009 Impact of the QBO on surface winter climate Andrew G. Marshall 1,2 and Adam A. Scaife 1 Received 9 January 2009; revised 11 June 2009; accepted 24 June 2009; published 19 September 2009. [1] The European winter surface climate response to the Quasi-Biennial Oscillation (QBO) and its dependence on stratospheric resolution is cleanly assessed in idealized seasonal hindcasts with two versions of the Hadley Centre s atmospheric general circulation model. The standard 38-level version extends to an altitude of 39 km while the extended 60-level version has enhanced stratospheric resolution and reaches 84-km altitude. We show that both models generate a realistic stratospheric polar and surface European Arctic Oscillation (AO) response to the QBO for winter hindcasts initialized on 1 December and suggest that the better representation of the QBO in the L60 model will lead to improved forecasts at seasonal-to-multiannual timescales. Citation: Marshall, A. G., and A. A. Scaife (2009), Impact of the QBO on surface winter climate, J. Geophys. Res., 114,, doi:10.1029/2009jd011737. 1. Introduction [2] The dominant mode of variability in the tropical stratosphere is the Quasi-Biennial Oscillation (QBO [Ebdon, 1960; Baldwin et al., 2001]), which manifests itself as downward propagating patterns of easterly and westerly mean zonal winds [Reed et al., 1961] with an irregular period ranging 25 28 months [Ebdon and Veryand, 1961; Veryand and Ebdon, 1961] and a near-constant amplitude of 20 ms 1 from 5 to 40 mb [Baldwin et al., 2001]. The QBO is driven by the vertical transfer of momentum by equatorial waves [Lindzen and Holton, 1968; Holton and Lindzen, 1972] and affects stratospheric flow in the extratropics by modulating the effects of vertically propagating planetary waves [Holton and Tan, 1980]. This influence is particularly strong in the northern hemisphere during winter [e.g., Baldwin and Dunkerton, 1998] when the polar stratosphere is sensitive to the propagation and breaking of largeamplitude planetary (Rossby) waves [McIntyre and Palmer, 1983, 1984], which disrupt the polar vortex and act as a drag on the strong westerly zonal mean flow. Driven largely by land-sea contrasts and surface topography these waves typically propagate upward and equatorward, but are unable to penetrate into the summer hemisphere where the zonal mean flow is easterly [Charney and Drazin, 1961; Andrews et al., 1987]. [3] When the QBO is in a westerly phase the limiting boundary between easterly and westerly zonal-mean flow (the critical latitude of zero wind) is on the summer side of the equator and planetary waves can propagate into the tropics. However, when the QBO is in an easterly phase the boundary occurs on the winter side of the equator and the effective waveguide for propagation becomes narrower, 1 Hadley Centre, Met Office, Exeter, UK. 2 Now at Centre for Australian Weather and Climate Research, Hobart, Tasmania, Australia. Published in 2009 by the American Geophysical Union. thereby increasing the strength of extratropical planetary wave activity [O Sullivan and Salby, 1990]. The entire stratospheric depth is important for this mechanism [Gray et al., 2001; Pascoe et al., 2006]. The modulation of wave activity during an easterly (westerly) QBO phase thus leads to a greater (reduced) drag on the polar vortex and reduced (enhanced) westerly mean flow characteristic of the negative (positive) phase of the Arctic Oscillation (AO [Thompson and Wallace, 1998, 2000]) and North Atlantic Oscillation (NAO [van Loon and Rogers, 1978]). A clear AO/NAO response to the QBO is also seen at the surface [Holton and Tan, 1980], with the tropospheric anomalies lagging the stratospheric anomalies by about 2 3 weeks [Baldwin and Dunkerton, 1999]. Mean surface temperatures are below normal [e.g., Hurrell, 1995] and the frequency of surface winter cold spells is increased [Thompson and Wallace, 2001] throughout much of Europe, Asia and North America during an easterly QBO phase, with some of the largest cold anomalies observed across northern Eurasia [Thompson et al., 2002]. [4] The impact of the QBO on the northern winter polar vortex has been simulated in numerous modeling studies. O Sullivan and Salby [1990] produced the first calculations linking planetary waves to QBO-induced extratropical circulation anomalies. Hamilton [1998] reproduced the observed extratropical stratospheric planetary wave and circulation response in a general circulation model (GCM) with imposed zonal momentum forcing designed to produce tropical QBO-like behavior, and a realistic extratropical stratospheric response to the QBO was further demonstrated by Niwano and Takahashi [1998] in a GCM that also produced a tropospheric response characteristic of the NAO. The observed downward propagation of zonal mean zonal wind anomalies from the stratosphere to the troposphere has been simulated in recent numerical simulations [e.g., Christiansen, 2001], and several modeling studies have also explored the influence of the QBO on Stratospheric Sudden Warming (SSW [Scherhag, 1952]) events [e.g., Dameris and Ebel, 1990; Holton and Austin, 1991], 1of6
however little attention has been devoted to the impact of the QBO on European seasonal forecasts. [5] Seasonal forecast skill over Europe in winter is currently poor [van Oldenborgh, 2005] due to the high levels of unpredictable natural variability in the extratropics. However, predictability may be present from the El Niño/ Southern Oscillation (ENSO). Observational studies show a clear NAO response in European climate to ENSO in late winter [e.g., Moron and Gouirand, 2003; Brönnimann et al., 2004, 2007]. General circulation model simulations that reproduce the observed response have recently identified a link between ENSO and the northern polar vortex via the stratosphere [Manzini et al., 2006] and shown a European surface NAO response that is large enough to be useful for seasonal forecasting [Ineson and Scaife, 2008]. Indeed, the surface signature of the QBO in NH wintertime is roughly comparable in amplitude to that observed in relation to ENSO [Thompson et al., 2002] and hence the dynamical link between the QBO and northern hemisphere surface climate may also be useful for improving seasonal forecast skill. While a few atmosphere general circulation models (AGCM) are able to reproduce the evolution of the QBO [e.g., Takahashi, 1996; Scaife et al., 2000; Giorgetta et al., 2002], many simulate a tropical stratosphere with mean easterly winds instead of QBO-like oscillations [Boer and Hamilton, 2008]. Thus an observed QBO in the initial conditions will generally dissipate within a few weeks of the forecast with the stratospheric zonal winds relaxing toward model climatology [Hamilton and Yuan, 1992]. This may be artificially rectified by forcing the QBO state into the model as it runs, which has recently been shown to have the potential to modestly improve the forecast skill over the North Atlantic [Boer and Hamilton, 2008] however it is preferable to use a GCM that correctly represents the dynamics of the QBO to best capture the large-scale physical processes and attain additional skill in seasonal forecasts. [6] The Met Office Unified Model (UM [Cullen, 1993]) has been shown to simulate realistic QBO behavior driven by a combination of resolved and parameterized waves [e.g., Scaife et al., 2000]. In the present study we assess the European winter response to the QBO in seasonal hindcast experiments using two versions of the UM that have the same vertical resolution in the troposphere but different stratospheric vertical resolutions and vertical domains. Enhanced stratospheric resolution may help to improve the QBO representation in an AGCM and thus improve the prediction of associated European surface winter cold spells. We therefore also assess the impact of stratospheric resolution on the seasonal prediction of associated European surface anomalies. The model hindcast experiments are described along with the methodology in section 2. The results of the analysis are then presented in section 3 and discussed in section 4. 2. Methodology [7] We make use of the 38 and 60 level atmosphere-only configurations of the first generation Hadley Centre Global Environmental Model (HadGAM1 [Martin et al., 2006]) with many of the changes proposed for HadGEM2-A as documented in Collins et al. [2008], and a spatial resolution of 1.25 degrees latitude by 1.875 degrees longitude. The 38 level version (L38) has a model top in the stratosphere at 39.3 km (3 mb) and the 60 level version (L60) has a model top near the mesopause at 84.1 km (0.004 mb). Both models have the same vertical resolution in the troposphere with 28 identical levels below the tropopause, but different vertical resolutions in the stratosphere; the L38 (L60) model has 8 (13) levels in the mid- and lower stratosphere between 10 and 100 mb, and 2 (19) levels above 10 mb. Importantly, both models also have the same gravity wave drag parameters, horizontal resolution and a timestep of 20 minutes, allowing for a clean assessment of the impact of stratospheric resolution. We produce 5-month model hindcasts for a suite of 14 winters over the last 45 years; 1962/63, 1963/64, 1964/65, 1968/69, 1974/75, 1982/83, 1983/84, 1987/88, 1989/90, 1991/92, 1992/93, 1995/96, 1997/98, and 1998/99. These years are chosen so that the QBO anomaly composite (easterly minus westerly winters) is not biased toward cool/warm ENSO phases, so as to assess the extratropical response to the QBO without signal aliasing (assuming a linear superposition). Other climate perturbations that can mask the extratropical winter signal include Stratospheric Sudden Warming (SSW) events, and Atlantic sea surface temperatures during the preceding May which can be a predictor for the winter NAO [Rodwell and Folland, 2002, 2003]. We therefore also ensure that QBO composites approximately average out observed SSW events and Atlantic May SST regimes. Finally, we assume that ocean-atmosphere feedback plays only a minor role in the European climate response to the QBO; this is discussed further in section 4. Fifteen-member ensembles for each winter period are initialized at 6 hourly intervals starting from 12Z on 27 November and ending with 00Z on 1 Dec to produce hindcasts that are integrated from a model start date of 1 December. The initial atmospheric conditions for each hindcast use data from the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis [Uppala et al., 2005]. [8] Each model experiment is forced with time-varying boundary conditions from greenhouse gases including CO 2, CH 4,N 2 O, CFCl 3 and CF 2 Cl 2, and changes in vegetation, sulphur, soot, and biomass emissions. Sea surface temperature and sea-ice extent variations are specified from an analysis of historical observations [Rayner et al., 2003], and atmospheric ozone concentrations were held constant at 1990 levels. Explosive volcanic eruptions are absent from the model simulations to avoid their masking effects on the extratropical circulation [Marshall et al., 2009]. [9] We use both the ECMWF reanalysis (ERA-40 [Uppala et al., 2005]) and the National Centers for Environmental Prediction (NCEP)/NCAR reanalysis 1 data set (NNR1 [Kalnay et al., 1996]) for model comparison and assessment. ERA-40 is a reanalysis of global meteorological quantities spanning a 45-year period from September 1957 to August 2002. The subset of ERA-40 output used in this study is on a 2.5 2.5 global grid (interpolated from T159) with 60 vertical levels interpolated onto 23 standard pressure levels for the years 1979 2001. NNR1 is a global reanalysis spanning the period 1948 2005, produced with a model resolution of T62L28. Both ERA-40 and NNR1 provide global, quality-controlled data sets using frozen data assimilation-forecast systems. The use of such 2of6
Figure 1. QBO index composite for westerly (positive) and easterly (negative) phases over the December April hindcast period for ERA-40 (solid black), L38 (solid dark gray), and L60 (solid light gray), both (a) as an absolute index and (b) as an anomaly relative to climatology. Dashed lines represent 90% confidence intervals using a two-tailed t-test. systems prevents problems of pseudoclimate signals being introduced into the data sets through changes in assimilation techniques and model formulation, although problems due to introduction of observational data over different epochs remain. 3. Extratropical Response to the QBO [10] We assess L38 and L60 winter hindcasts for the westerly QBO events (QBOW) of 1963/64, 1982/83, 1992/ 93 and 1998/99, and for the easterly QBO events (QBOE) of 1962/63, 1974/75, 1983/84, 1989/90, 1991/92, 1995/96 and 1997/98. A commonly used index for measuring QBO activity is the monthly mean 30 mb zonal mean zonal wind at the equator [Hamilton, 1984]. A composite of the QBO index for QBOW and QBOE years is shown over the December April hindcast period in Figure 1, both as an absolute index and as an anomaly relative to climatology, for ERA-40 and the models. Years used for the climatology are 1963/64, 1964/65, 1968/69, 1982/83, 1987/88, 1991/92, 1992/93, 1995/96, 1997/98, and 1998/99, which give a net observed QBO index over winter (December February; DJF) of less than 0.1 ms 1. Dashed lines in Figure 1 represent 90% confidence intervals using a two-tailed t-test [Student, 1908]. The L60 model reproduces and maintains the strength of the observed QBOW signal throughout boreal winter and into early spring while the L38 model steadily drifts away from the observed signal such that its anomalous westerly strength is around half that observed by April; this translates as a drift to an easterly QBO state in the absolute signal by March and reflects a strong easterly bias in the L38 model climatology. This easterly drift is also apparent in the QBOE curves in Figure 1, for which the strength of the absolute L38 signal agrees well with that observed while the strength of the anomalous L38 signal appears significantly weaker than that observed over the entire hindcast period. The anomalous QBOE signal for the L60 model also drifts over the hindcast period but to a lesser degree, with strength around 75% of that observed by April compared with 50% for the L38 model. In general, the L60 model shows better agreement with the observed QBO throughout the winter and early spring than the L38 model which suffers from an easterly mean bias. [11] The northern winter stratospheric response to the QBO is presented in Figure 2, which shows 50 mb geopotential height anomaly DJF composites of QBOE minus QBOW years for NNR1 and the models. Stippling in Figure 2 indicates statistically significant anomalies at the 90% confidence level using a two-tailed t-test. Both models reproduce a statistically significant intensification (weakening) of the polar vortex during the westerly (easterly) phase of the QBO, as is observed, with the large number of model realizations allowing for a high level of significance. The net QBOE-QBOW polar anomaly for each model peaks around 120 140 m and is characteristic of the negative phase of the AO, although more spatially extensive and up to 50% weaker in magnitude than the NNR1 signal. Note that the NNR1 signal in Figure 2 broadly agrees with that shown by Holton and Tan [1980], Dunkerton and Baldwin [1991], and Hamilton [1998], although the magnitude of the polar anomaly and the sign of weak midlatitude anomalies varies among these studies due to different selections of years for the composites. For example, the peak observed QBOE-QBOW composite anomaly presented by Hamilton [1998] is around 150 m, similar to that for each model shown in Figure 2. [12] The corresponding DJF surface temperature response to QBOE in Figure 3 shows significant cooling over northern Europe for both models, as is observed, characteristic of that associated with the negative phase of the NAO. East of about 15 E the negative temperature anomalies for L38 are around 1.5 times those for L60 and closer to those observed, however anomalous differences between the models are not statistically significant at the 90% confidence level. These results show that the observed European winter surface response to the QBO is broadly reproduced by both models at lead times of 1 3 months with the caveat that the QBO is better sustained in the L60 model than in the L38 model over the 5-month hindcast period. 4. Summary and Perspective [13] In this study we show that the northern winter AO response to the QBO is captured in seasonal hindcasts using the 38-level and 60-level configurations of the Hadley Center s Unified Model. Both models reproduce a weaken- 3of6
Figure 2. 50-mb geopotential height anomaly composites of QBOE minus QBOW winters (DJF) for (a) NNR1, (b) L38, and (c) L60. Stippling indicates statistically significant anomalies at the 90% confidence level using a two-tailed t-test. Contour interval is 20 m with gray-shaded areas indicating negative anomalies. ing of the polar vortex, and enhanced cooling over northern Europe at seasonal timescales comparable to that observed, for composited QBOE-QBOW years over the last half century. While the QBO in the L60 model agrees well with that observed throughout the winter and early spring, the L38 model suffers from an easterly mean bias that causes the stratospheric zonal winds to drift toward model climatology over the hindcast period. This is consistent with model drift issues encountered in previous studies [Hamilton and Yuan, 1992; Boer and Hamilton, 2008] but does not affect the extratropical winter response to the QBO in the model. This is likely to be because the winter response (Figures 2 and 3) is calculated over the first three months of the hindcast period (DJF) when the L38 model drift is relatively small and the absolute zonal mean zonal wind composite for QBOW years is still westerly (Figure 1). With the largest effect of the QBO on the northern winter polar vortex occurring until midwinter both in observations and in the models, it will be interesting to test the impact of this easterly mean bias on the extratropical winter signal with an earlier initialization of several months, for which we would expect the absolute QBOW index to have drifted into an easterly state by winter (from Figure 1). As in the work of Boer and Hamilton [2008], in the absence of a QBO signal the L38 model will not be able to simulate an extratropical QBO response (by definition). Since the L60 4 of 6
Figure 3. 1.5-m temperature anomaly composites of QBOE minus QBOW winters (DJF) for (a) NNR1, (b) L38, and (c) L60. Stippling indicates statistically significant anomalies at the 90% confidence level using a two-tailed t-test. Contour interval is 0.5 K with gray-shaded areas indicating negative anomalies. model better captures and sustains the QBO signal beyond 3 months, we therefore also expect a more realistic extratropical winter response to the QBO in the L60 model and thus improved seasonal prediction of QBO-induced surface anomalies over Europe at these longer lead times. The benefit of enhanced resolution to seasonal forecasting is also supported in a recent study that shows earlier capture of SSW events and better seasonal prediction of winter European temperature anomalies at these longer lead times in the L60 model, compared with the L38 model (A. G. Marshall and A. A. Scaife, Improved predictability of Stratospheric Sudden Warming events in an AGCM with enhanced stratospheric resolution, submitted to Journal of Geophysical Research, 2009). [14] Our assumption that ocean-atmosphere feedback plays only a minor role in the European climate response to the QBO (section 2) is justified by the fact that the basic teleconnection between the QBO and the surface NAO via the extratropical stratosphere does not require any oceanatmosphere interaction. This is further supported by the similar-to-observed extratropical surface winter response to the QBO in both models seen in Figure 3; unless there is a fortuitous cancellation of errors then air-sea interaction is seen to play a minor role. [15] The results of this work have implications for forecasting at multiannual timescales given the 25 28 month periodicity of the QBO; realistic simulation of the QBO and the northern extratropical response in the L60 model should be useful for predicting European surface climate anomalies up to several years ahead due to the multiannual predictability of the QBO. The importance of this result is underscored by the fact that the QBO is one of the few drivers of extratropical climate variability with multiannual predictability that could provide European winter forecast skill at these timescales. [16] Acknowledgments. We thank Sarah Ineson at the Met Office Hadley Centre for helpful discussions during the course of this work. We also thank the three anonymous reviewers for suggested improvements to the manuscript. ERA-40 data used in this study was provided by ECMWF, and the NCEP-NCAR reanalyses were provided by the NOAA/CIRES Climate Diagnostics Center. This work was supported by the Joint DECC and MoD Integrated Climate Programme-(DECC) GA01101, (MoD) CBC/ 2B/0417_Annex C5. References Andrews, D. G., J. R. Holton, and C. B. Leovy (1987), Middle Atmosphere Dynamics, 489 pp., Elsevier, New York. Baldwin, M. P., and T. J. Dunkerton (1998), Quasi-biennial modulation of the Southern Hemisphere stratospheric polar vortex, Geophys. Res. Lett., 25, 3343 3346. Baldwin, M. P., and T. J. Dunkerton (1999), Propagation of the Arctic Oscillation from the stratosphere to the troposphere, J. Geophys. Res., 104, 30,937 30,946. Baldwin, M. P., et al. (2001), Quasi-biennial oscillation, Rev. Geophys., 39, 179 229. Boer, G. J., and K. Hamilton (2008), QBO influence on extratropical predictive skill, Clim. Dyn., 31, 987 1000. Brönnimann, S., J. Luterbacher, J. Staehelin, T. M. Svendby, G. Hansen, and T. Svenøe (2004), Extreme climate of the global atmosphere and stratosphere in 1940 42 related to El Nino, Nature, 431, 971 974. Brönnimann, S., E. Xoplaki, C. Casty, A. Pauling, and J. Luterbacher (2007), ENSO influence on Europe during the last centuries, Clim. 5of6
Dyn., 28, 181 197. Charney, J. G., and P. G. Drazin (1961), Propagation of planetary-scale disturbances from the lower into the upper atmosphere, J. Geophys. Res., 66, 83 109. Christiansen, B. (2001), Downward propagation of zonal mean zonal wind anomalies from the stratosphere to the troposphere: Model and reanalysis, J. Geophys. Res., 106, 27,307 27,322. Collins, W. J., et al. (2008), Evaluation of the HadGEM2 model, Met Office Hadley Centre Technical Note 74. Cullen, M. J. P. (1993), The unified forecast/climate model, Meteorol. Mag., 122, 81 94. Dameris, M., and A. Ebel (1990), The quasi-biennial oscillation and major stratospheric warmings: A three-dimensional model study, Ann. Geophys., 8, 79 85. Dunkerton, T. J., and M. P. Baldwin (1991), Quasi-biennial modulation of planetary wave fluxes in the Northern Hemisphere winter, J. Atmos. Sci., 48, 1043 1061. Ebdon, R. A. (1960), Notes on the wind flow at 50 mb in tropical and subtropical regions in January 1957 and in 1958, Q. J. R. Meteorol. Soc., 86, 540 542. Ebdon, R. A., and R. G. Veryand (1961), Fluctuations in equatorial stratospheric winds, Nature, 189, 791 793. Giorgetta, M. A., E. Manzini, and E. Roeckner (2002), Forcing of the Quasi-Biennial Oscillation from a broad spectrum of atmospheric waves, Geophys. Res. Lett., 29(8), 1245, doi:10.1029/2002gl014756. Gray, L. J., S. J. Phipps, T. J. Dunkerton, M. P. Baldwin, E. F. Drysdale, and M. R. Allen (2001), A data study of the influence of the equatorial upper stratosphere on northern-hemisphere stratospheric sudden warmings, Q. J. R. Meteorol. Soc., 127, 1985 2003. Hamilton, K. (1984), Mean wind evolution through the quasi-biennial cycle in the tropical lower stratosphere, J. Atmos. Sci., 41, 2113 2125. Hamilton, K. (1998), Effects of an imposed Quasi-Biennial Oscillation in a comprehensive troposphere-stratosphere-mesosphere general circulation model, J. Atmos. Sci., 55, 2393 2418. Hamilton, K., and L. Yuan (1992), Experiments on tropical stratospheric mean wind variations in a spectral GCM, J. Atmos. Sci., 49, 2464 2483. Holton, J. R., and R. S. Lindzen (1972), An updated theory for the quasibiennial cycle of the tropical stratosphere, J. Atmos. Sci., 29, 1076 1080. Holton, J. R., and H.-C. Tan (1980), The influence of the equatorial quasibiennial oscillation on the global circulation at 50 mb, J. Atmos. Sci., 37, 2200 2208. Holton, J. R., and J. Austin (1991), The influence of the QBO on sudden stratospheric warmings, J. Atmos. Sci., 48, 607 618. Hurrell, J. (1995), Decadal trends in the North Atlantic oscillation: Regional temperature and precipitation, Science, 269, 676 679. Ineson, S., and A. A. Scaife (2008), The role of the stratosphere in the European climate response to El Niño, Nat. Geosci., doi:10.1038/ ngeo381. Kalnay, E., et al. (1996), The NCEP/NCAR 40-Year Reanalysis Project, Bull. Am. Meteorol. Soc., 77, 437 471. Lindzen, R. S., and J. R. Holton (1968), A theory of the quasi-biennial oscillation, J. Atmos. Sci., 25, 1095 1107. Manzini, E., M. A. Giorgetta, M. Esch, L. Kornblueh, and E. Roeckner (2006), The influence of sea surface temperatures on the northern winter stratosphere: Ensemble simulations with the MAECHAM5 model, J. Clim, 19, 3863 3881. Marshall, A. G., A. A. Scaife, and S. Ineson (2009), Enhanced seasonal prediction of European winter warming following volcanic eruptions, J. Clim., doi:10.1175/2009jcli3145.1, in press. Martin, G. M., M. A. Ringer, V. D. Pope, A. Jones, C. Dearden, and T. J. Hinton (2006), The physical properties of the atmosphere in the New Hadley Centre Global Environmental Model (HadGEM1): Part I. Model description and global climatology, J. Clim., 19, 1274 1301. McIntyre, M. E., and T. N. Palmer (1983), Breaking planetary waves in the stratosphere, Nature, 305, 593 600. McIntyre, M. E., and T. N. Palmer (1984), The surf zone in the stratosphere, J. Atmos. Terr. Phys., 46, 825 849. Moron, V., and I. Gouirand (2003), Seasonal modulation of the El Niño- Southern Oscillation relationship with sea level pressure anomalies over the North Atlantic in October-March 1873 1996, Int. J. Clim., 23, 143 155. Niwano, M., and M. Takahashi (1998), Notes and correspondence: The influence of the equatorial QBO on the Northern Hemisphere winter circulation of a GCM, J. Meteorol. Soc. Jpn., 76, 453 461. O Sullivan, D., and M. L. Salby (1990), Coupling of the quasi-biennial oscillation and the extratropical circulation in the stratosphere through planetary wave transport, J. Atmos. Sci., 47, 650 673. Pascoe, C. L., L. J. Gray, and A. A. Scaife (2006), A GCM study of the influence of equatorial winds on the timing of sudden stratospheric warmings, Geophys. Res. Lett., 33, L06825, doi:10.1029/2005gl024715. 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(D14), 4407, doi:10.1029/ 2002JD002670. Reed, R. J., W. J. Campbell, L. A. Rasmussen, and R. G. Rogers (1961), Evidence of a downward propagating annual wind reversal in the equatorial stratosphere, J. Geophys. Res., 66, 813 818. Rodwell, M. J., and C. K. Folland (2002), Atlantic air-sea interaction and seasonal predictability, Q. J. R. Meteorol. Soc., 128, 1413 1443. Rodwell, M. J., and C. K. Folland (2003), Atlantic air-sea interaction and model validation, Ann. Geophys., 46, 47 56. Scaife, A., N. Butchart, C. D. Warner, D. Stainforth, W. Norton, and J. Austin (2000), Realistic quasi-biennial oscillations in a simulation of the global climate, Geophys. Res. Lett., 27, 3481 3484. Scherhag, R. (1952), Die explosionsartigen Stratospärenerwärmungen des Spätwinters 1951 1952, Ber. Dtsch. Wetterdienst (US Zone), 6, 51 63. Student (1908), The probable error of a mean, Biometrika, 6, 1 25. Takahashi, M. (1996), Simulation of the stratospheric quasi-biennial oscillation using a general circulation model, Geophys. Res. Lett., 23, 661 664. 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, 1297 1300. Thompson, D. W. J., and J. M. Wallace (2000), Annular model in the extratropical circulation: Part I. Month-to-month variability, J. Clim., 13, 1000 1016. Thompson, D. W. J., and J. M. Wallace (2001), Regional climate impacts of the Northern Hemisphere annular mode, Science, 293, 85 89. Thompson, D. W. J., M. P. Baldwin, and J. M. Wallace (2002), Stratospheric connection to Northern Hemisphere wintertime weather: Implications for prediction, J. Clim., 15, 1421 1428. Uppala, S. M., et al. (2005), The ERA-40 re-analysis, Quart. J. R. Meteorol. Soc., 131, 2961 3012. van Loon, H., and J. C. Rogers (1978), The seesaw in winter temperatures between Greenland and northern Europe: Part I. General description, Mon. Weather Rev., 106, 296 310. van Oldenborgh, G. J. (2005), Comments on Predictability of winter climate over the North Atlantic European region during ENSO events, J. Clim., 18, 2770 2772. Veryand, R. G., and R. A. Ebdon (1961), Fluctuations in tropical stratospheric winds, Meteorol. Mag., 90, 125 143. A. G. Marshall, Centre for Australian Weather and Climate Research, CSIRO Marine and Atmospheric Research, Castray Esplanade, GPO Box 1538, Hobart, Tas 7000, Australia. (andrew.marshall@csiro.au) A. A. Scaife, Hadley Centre, Met Office, FitzRoy Road, Exeter, Devon EX1 3PB, UK. 6of6