Clim Dyn DOI 10.1007/s00382-015-2654-6

Troposphere–stratosphere response to large‑scale North Atlantic Ocean variability in an atmosphere/ocean coupled model N.‑E. Omrani1,2 · Jürgen Bader3,4 · N. S. Keenlyside1,4 · Elisa Manzini3 

Received: 28 May 2014 / Accepted: 11 May 2015 © Springer-Verlag Berlin Heidelberg 2015

Abstract  The instrumental records indicate that the basin-wide wintertime North Atlantic warm conditions are accompanied by a pattern resembling negative North Atlantic oscillation (NAO), and cold conditions with pattern resembling the positive NAO. This relation is well reproduced in a control simulation by the stratosphere resolving atmosphere–ocean coupled Max-Planck-Institute Earth System Model (MPI-ESM). Further analyses of the MPI-ESM model simulation shows that the largescale warm North Atlantic conditions are associated with a stratospheric precursory signal that propagates down into the troposphere, preceding the wintertime negative NAO. Additional experiments using only the atmospheric component of MPI-ESM (ECHAM6) indicate that these stratospheric and tropospheric changes are forced by the warm North Atlantic conditions. The basin-wide warming excites a wave-induced stratospheric vortex weakening, stratosphere/troposphere coupling and a high-latitude tropospheric warming. The induced high-latitude tropospheric warming is associated with reduction of the growth rate of low-level baroclinic waves over the North Atlantic region,

contributing to the negative NAO pattern. For the cold North Atlantic conditions, the strengthening of the westerlies in the coupled model is confined to the troposphere and lower stratosphere. Comparing the coupled and uncoupled model shows that in the cold phase the tropospheric changes seen in the coupled model are not well reproduced by the standalone atmospheric configuration. Our experiments provide further evidence that North Atlantic Ocean variability (NAV) impacts the coupled stratosphere/ troposphere system. As NAV has been shown to be predictable on seasonal-to-decadal timescales, these results have important implications for the predictability of the extratropical atmospheric circulation on these time-scales. Keywords  North Atlantic oscillation (NAO) · Northern annular mode (NAM) · Stratosphere/troposphere coupling · North Atlantic variability (NAV) · Atlantic multidecadal variability (AMV) · Atlantic multidecadal oscillation (AMO) · Ocean–atmosphere interaction

1 Introduction Electronic supplementary material  The online version of this article (doi:10.1007/s00382-015-2654-6) contains supplementary material, which is available to authorized users. * N.‑E. Omrani [email protected] 1

Geophysical Institute, University of Bergen and Bjerknes Centre for Climate Research, Bergen, Norway

2

GEOMAR, Helmholtz Centre for Ocean Research Kiel, Kiel, Germany

3

Max Planck Institute for Meteorology, Hamburg, Germany

4

Uni Climate, Uni Research and The Bjerknes Centre for Climate Research, Bergen, Norway



The observational record shows evidence of a basin-wide variability in the North Atlantic Sea-Surface Temperature (SST) (Deser and Blackmon 1993; Kushnir 1994). Even if the measurement record is still short, the basin-wide SST-fluctuations in the Northern Atlantic are commonly termed the Atlantic multidecadal oscillation or Atlantic multi-decadal variability, referring to its observed pronounced multidecadal variations. However, the simulated basin-wide SST fluctuations have a wide range of time scales (Ba et al. 2014; Kavvada et al. 2013; Mecking et al. 2013; Medhaug and Furevik 2011; Ruiz-Barradas et al. 2013; Zanchettin et al. 2012, 2014) calling into question

13

N.-E. Omrani et al.

the multidecadal oscillatory nature of large-scale North Atlantic SST variability. Unlike the North Atlantic SST-tripole, which is driven mainly by the thermodynamical heat fluxes associated with the atmospheric NAO (Czaja et al. 2003; Kushnir et al. 2002; Visbeck et al. 2003), the basin-wide temperature changes in the North Atlantic can not be explained by the direct thermodynamic atmospheric forcing (Kushnir 1994). The mechanisms for these SST variations remains highly controversial, with large uncertainties in climate models and observations being limited (Keenlyside et al. 2014; Latif and Keenlyside 2011). Although the contribution of external forcing has been suggested (Booth et al. 2012; Mann and Emanuel 2006; Ottera et al. 2010), several studies using standalone ocean (Eden and Jung 2001; Eden and Willebrand 2001) and ocean–atmosphere coupled models (Delworth et al. 1993; Jungclaus et al. 2005; Knight et al. 2005; Vellinga and Wu 2004) suggest that the large-scale Atlantic temperature changes are driven by changes in the poleward ocean-heat transport. This transport is associated with the Atlantic Meridional Overturning Circulation (AMOC). There is some evidence that the NAO (Eden and Jung 2001; Visbeck et al. 2003) and changes in the stratospheric vortex (via the NAO) (Manzini et al. 2012; Reichler et al. 2012) play an important role in low frequency fluctuations of the AMOC. The impact of the NAO on the AMOC occurs through wind-induced changes in heat fluxes and transport of salinity anomalies into high-latitude sinking regions (Delworth et al. 1993; Eden and Jung 2001; Jungclaus et al. 2005; Visbeck et al. 2003), and resulting changes in deep-water formation. The basin-wide SST-changes in the Atlantic are associated with large-scale atmospheric circulation changes in both winter and summer (Deser and Blackmon 1993; Kushnir 1994; Omrani et al. 2014; Zhang and Delworth 2006). In the wintertime the observed large-scale temperature changes in the Atlantic basin are associated with a pattern that resembles the NAO (Omrani et al. 2014). This relation has been reproduced independently in two different standalone AGCM studies (Omrani et al. 2014; Peings and Magnusdottir 2014a). In the first, Omrani et al. (2014) consider the observed 1950s North Atlantic warm conditions and show that the associated winter atmospheric circulation can be largely simulated by the stratosphere resolving standalone atmospheric model MAECHAM5. The observed large-scale warm North Atlantic conditions drive early winter changes in the extratropical stratosphere that propagate into the troposphere resulting in negative NAO pattern in the mid to late winter. Both tropical and extratropical SST play an important role in the atmospheric response. Although the downward propagation of the stratospheric signal into the troposphere is well accepted (Baldwin and Dunkerton 1999; Haynes 2005), there is still

13

debate about the mechanism of these propagations (Haynes 2005). Resolving the stratosphere also leads to weaker background winds, which enhance the upward wave propagation and indirectly regulate the tropospheric response to NAV (Omrani et al. 2014). In the second study (Peings and Magnusdottir 2014a), a low-top model (CAM5) was also able to simulate the tropospheric response to basinwide North Atlantic SST anomalies. This is in contrast to results from other low-top multi-model study, which fail to reproduce the observed winter anomalies for warm North Atlantic conditions (Hodson et al. 2010). The difference in the response of the stratosphere in low-top models may also be due to the implementation of sponge-layers and background wind (Omrani et al. 2014). Stratosphere–troposphere interaction adds to previous understanding of the direct and local atmospheric response to the extratropical Atlantic SST, obtained in studies primarily addressing the SST-tripole (see references above). This direct response is primarily baroclinic and can be predicted using linear models (Hoskins and Karoly 1981; Kushnir et al. 2002). This baroclinic response is associated with a synoptic-scale eddy-forcing that transforms it into a large-scale low frequency barotropic NAO-signal (Deser et al. 2007; Peng et al. 2003). The growth rate of the most unstable part of such synoptic eddies is in turn proportional to the low-level meridional temperature gradient or vertical wind shear known as baroclinicity (Czaja et al. 2003; Hoskins and Valdes 1990). The high-latitude North Atlantic warming associated with negative NAO can warm the overlying troposphere by turbulent heat flux and reduce the lower tropospheric horizontal temperature gradient and thus the baroclinicity, which is reflected in a negative NAO-like pattern (Czaja et al. 2003; Deser et al. 2007). In case of cooling the reverse process takes place. The change in baroclinicity and the associated NAO-response can also result from tropical Atlantic SST-changes, which lead through changes in tropical convection and the Hadley-circulation to changes in baroclinicity, planetary-scales transient eddies and the resulting anomalous stationary wave (Terray and Cassou 2002). Not only the North Atlantic SST-anomalies but also oceanic forcing outside the North Atlantic can force the Northern hemispheric coupled stratosphere/troposphere system. Warm (cold) ENSO, for example, drives a weakening (strengthening) of the polar vortex, leading to surface anomalies projecting on negative (positive) NAM or NAO (Ineson and Scaife 2009; Manzini et al. 2006). Furthermore, the relatively small temperature changes in the Indian Ocean associated with ENSO act more efficiently than and in an opposite way to the SST-changes in the tropical Pacific (Fletcher and Kushner 2011). A similar impact of the Indian Ocean on the NAO (or NAM) has been shown in multi-decadal time scales (Bader and Latif

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an…

2003; Hoerling et al. 2001). In addition to the tropical Indo-Pacific Ocean, the extra-tropical Pacific SST (Hurwitz et al. 2012) and Arctic sea-ice (Peings and Magnusdottir 2014b) can also induce changes in the stratosphere/troposphere coupled system. The mechanisms of the impact of Indo-Pacific SST on the NAM include Rossby-wave induced tropospheric changes in Northern Pacific regions that can interfere linearly with background climatological waves and lead to changes in the upward wave propagation from the troposphere into the stratosphere and thus a stratospheric response (Fletcher and Kushner 2011; Garfinkel et al. 2010). As discussed above, our previous results (Omrani et al. 2014) suggest stratosphere–troposphere coupling is important for the atmospheric response to extra-tropical SST anomalies. However, this work was limited to the case of the observed 1950s warm North Atlantic conditions and used only the uncoupled model configuration. The short instrumental record makes robust and consistent observational studies of the relation between NAV and atmospheric circulation difficult. Here we extend on our previous study by investigating this relationship with a coupled model allowing the role of active two-way ocean–atmosphere and stratosphere–troposphere interactions to be investigated. Coupled low-top models simulate only a weak NAV–NAO coupling (Gastineau and Frankignoul 2012). The following questions are addressed: (a) Can a high-top coupled ocean–atmosphere model with simulated SSTs reproduce the observed relationship between large-scale Atlantic SST-changes and the NAO, for both warm and cold North Atlantic conditions? (b) To what extent does the Atlantic SST drive the atmospheric changes seen in the coupled model? (c) Is the stratosphere similarly involved in both warm and cold phases of NAV? Considering the big uncertainties in the origins (Mann and Emanuel 2006; Ottera et al. 2010; Zanchettin et al. 2012), mechanisms (Delworth et al. 1993; Eden and Jung 2001; Eden and Willebrand 2001; Jungclaus et al. 2005; Knight et al. 2005; Visbeck et al. 2003) and time-scales (Ba et al. 2014; Kavvada et al. 2013; Mecking et al. 2013; Ruiz-Barradas et al. 2013; Zanchettin et al. 2012, 2014) of the large-scale NAV, we will address these questions without any time-scale preferences nor assumptions about the mechanisms and the origins of the simulated large-scale North Atlantic SST variability. After describing the data, models and experiments used (Sect.  2), we show the observed and simulated Sea Level Pressure (SLP) anomalies associated with large-scale NAV (Sect.  3). Then we consider the seasonal evolution of the

vertical structure of the atmospheric response to warm North Atlantic conditions (Sect. 4). The cold case is treated analogously (Sect. 5), followed by the summary and conclusion in Sect. 6.

2 Data and methods We use the Hadley center SST and SLP (Rayner et al. 2003) from 1870 to 2009 to identify the observed link between the large-scale North Atlantic SST-variability and tropospheric circulation. To describe the observed largescale North Atlantic SST-variability, we defined North Atlantic Variability Index NAVI as the average of the detrended Atlantic SST anomalies over the region 0–60°N and 75–7.5°W for winter (JFM) months (Sutton and Hodson 2005) without any additional time filtering. As mentioned before our focus is on the sensitivity of the extra tropical circulation to North Atlantic variability independent from its timescale and forcing, and we adopt the name NAVI (Zanchettin et al. 2012) instead of Atlantic Multidecadal Oscillation Index (Sutton and Hodson 2005) to avoid reference to multi-decadal fluctuation. The NAVI is defined in same way in the coupled model. To understand the relationship between Atlantic SST changes and the NAO a control integration of the latest version of the Max-Planck-Institute Earth System Model (MPI-ESM) is analysed and a set of SST-sensitivity experiments with its atmospheric component are performed. The MPI-ESM consists of the atmosphere general circulation model (AGCM) ECHAM6 that is coupled to the global ocean/sea-ice model (MPI-OM) (Giorgetta et al. 2013; Stevens et al. 2013). One difference between ECHAM6 and its older version ECHAM5 is that ECHAM6 is run by default as a high-top model; the model top extends now to 0.01 hPa (~80 km) in the standard setup. ECHAM6 also incorporates a completely new aerosol and ozone climatology and makes use of a new shortwave radiation scheme that is more accurate in offline tests and has less cloud absorption. Substantial differences between ECHAM5 and ECHAM6 concern the land processes, the computation of the surface albedo, and the triggering condition for convection. All experiments with the MPI-ESM and ECHAM6 were performed at T63 (~1.9°) horizontal atmospheric resolution and with 95 vertical levels in the atmosphere, which is able to simulate the Quasi-Biannual Oscillation (QBO) internally. The ocean is run with a tripolar ocean grid (TP04L40). The horizontal ocean resolution at the equator is ~0.4° and the ocean consists of 40 vertical levels. A 500-year-long coupled ocean–atmosphere control integration with prescribed and fixed boundary conditions consistent with preindustrial conditions found ca. 1850 is analysed. All greenhouse gases, aerosols, ozone and the

13

N.-E. Omrani et al.

solar insolation at the top of the atmosphere do not vary from year to year. A composite analysis based on the winter (JFM) NAVI was used in observations and coupled model to identify the atmospheric signature of the basinwide North Atlantic SST-changes. The composite criteria are the same in all our analyses and are based on the winter (JFM) NAVI exceeding plus or minus one standard deviation. Three additional SST-sensitivity experiments using the standalone atmospheric component ECHAM6 were performed to isolate the feedback from the Atlantic basin: (1) climatological control, (2) warm North Atlantic and (3) cold North Atlantic. The climatological control simulation was driven by monthly varying climatological SSTs and sea ice concentrations (SICs) computed from the coupled ocean–atmosphere control simulation. All other boundary conditions like e.g., greenhouse gases are kept constant analogous to the coupled simulation throughout the integration. The warm and cold North Atlantic Ocean simulations were driven by a monthly varying SST and SIC climatology that in the Atlantic (40°S–66°N) corresponds respectively to the warm and cold NAVI composites computed from the coupled ocean–atmosphere integration, and elsewhere corresponds to that of the control. To keep preconditioning boundary conditions similar to the coupled model, the warm and cold SST and sea-ice anomalies between June and December are taken 1 year before the corresponding NAVI-composite years (at lag zero) and the anomalies from January to July are taken from the same years (lag zero) as the NAVI-composite years. All analyzed standalone simulations are 40 years long. Treating each winter as independent offers sufficient realizations to test significance, which we test with a two-sided Student T test. To understand the vertical evolution of the atmospheric changes associated with basin-wide North Atlantic temperature variations, we project at each level the geopotential height from the composite analysis in the coupled model and from the standalone model-response onto the corresponding winter (JFM) NAM and NAO patterns. For the coupled model the NAO and NAM patterns were computed from the coupled control simulation, while for the standalone AGCM experiments they were computed from the AGCM control run. In the case of NAM the EOF analysis was performed on the seasonal (JFM) hemispheric geopotential height north of 20°N at each levels. For the NAO, the EOF analysis was performed in the same way but by taking the North Atlantic region from 20° to 80°N and 90°W to 40°E. Because of the local structure of the tropospheric response and hemispheric structure of stratospheric response to large scales NAV (Omrani et al. 2014), we projected the responses on both NAM and NAO, in order to understand pattern and evolution of the atmospheric signature of the basin-wide NAV. In the stratosphere we will however focus only on the projection on the NAM, since

13

the NAO is usually defined in the troposphere. In the troposphere we consider both NAO and NAM patterns. Positive (negative) phase of the NAO and NAM in the troposphere are associated with westerly (easterly) wind anomalies in high latitudes and easterly (westerly) wind anomalies in low-to-mid latitude indicating poleward shift and strengthening of eddy-driven mid-latitude jet. A positive (negative) phases of the NAM in the stratosphere describes an anomalous strong (weak) stratospheric polar vortex and highlatitude westerlies reflecting poleward (equatorward) shift of stratospheric polar jet. The strength of the projection of the geopotential on the NAO (NAM) is used as measure of the similarity of the atmospheric NAV-signature to this pattern, and thus also a measure of the strength of associated wind anomalies. The zonally averaged quasistationary eddy meridional heat flux (computed using the deviation from the zonal mean) is used as proxy for the upward wave propagation (Newman et al. 2001; Nishii et al. 2009), since it is proportional to the vertical component of the quasi-geostrophic Eliassen-Palm-flux vector (Andrews et al. 1987).

3 Observed and simulated atmospheric changes The observed warm North Atlantic conditions defined using the NAVI show positive SST anomalies over the entire North Atlantic, with the largest anomalies in the subpolar, tropical and in the eastern North Atlantic (Fig. 1a). The SST anomalies are nearly opposite for cold conditions (Fig. 1b). Even if there are substantial uncertainties in large parts of the observational record (due to the lack of measurements, especially in the open ocean and before the second half of the last century), the observed SLP patterns associated with the NAV show some resemblance to the NAO (Fig. 1c, d). There are however some discrepancies from the NAO pattern in the warm case and over the open ocean. This could be due to the lack of measurement stations in the observational record before 1950s. For example, the surface circulation change associated with the 1950s warming, where more station data became available, showed better resemblance to the negative NAO (Omrani et al. 2014). The SLP pattern during cold conditions resembles the positive phase of the NAO more closely (Fig. 1d). The long-term coupled model simulation with the high-top model provides a robust and physically consistent way to test the relationship between the NAO and large-scale NAV in more detail. The relationship between NAV and NAO suggested by the observational record is confirmed in the long-term stratosphere-resolving ocean/ atmosphere-coupled simulation. It is identified without any external forcing, in a much longer period than that of the available instrumental record (Fig. 2), providing support for a robust NAV-NAO relationship. In the

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an… Fig. 1  Observed large-scale SST and SLP changes associated with positive and negative NAV index. Observed warm and cold winter (JFM) composite relative to climatology based on one standard deviation of the NAV-index for a, c SST (in °K) and b, d SLP (in Pa). Significant differences (according to twosided T tests) at the 95 %-level are hatched with vertical and at the 90 %-level with horizontal white lines. The composite analysis was performed over the period 1870–2009

observations and the coupled model the positive and negative composites of the SST are in first order linear. The SLP-patterns in the coupled model are also almost linear, while the linearity in the observations is less pronounced (Fig. 2). In the next sections, standalone atmospheric model experiments are used to investigate the role of Atlantic SST (Fig.  2a, b) in driving the tropospheric and stratospheric circulation anomalies identified in the coupled model simulation. The focus will be on the seasonal evolution of the vertical structure of the response. Warm and cold conditions are considered separately.

4 Response to the warm phase 4.1 In the coupled ocean–atmosphere model The winter NAM-patterns computed from the coupled and uncoupled model (Fig. 3a, b, e, f) show the well-known quasi-hemispheric and zonally symmetric NAM structure (Thompson and Wallace 2000; Thompson et al. 2000). The annular structure is more pronounced in the stratosphere (Fig. 3e, f) than in the troposphere (Fig. 3a, b). The NAO has more localized centres (Fig. 3c, d) that are more pronounced over the North Atlantic and its surrounding

13

N.-E. Omrani et al. Fig. 2  Simulated large-scale SST and SLP changes in the MPI-ESM model associated with positive and negative NAVI. Simulated warm and cold winter (JFM) composite relative to climatology in MPI-ESM model, based on one standard deviation of the NAVI, for a, c SST (in °K) and b, d SLP (in Pa). Significant differences (according to two-sided T tests) at the 95 %-level are hatched vertical with and at the 90 %-level with horizontal white lines

regions (consistent with its definition). In case of the coupled model, the early-to-mid winter stratospheric geopotential height anomalies, associated with warm conditions, project negatively on the NAM (Fig. 4a) showing a quasi-hemispheric weakening of the stratospheric vortex and westerlies in January and February (Fig. 5a, b). The weakening of the westerlies propagates down into the troposphere and persists over the wintertime as an NAO-like pattern (Figs. 4a, 6a, b). The tropospheric changes project better on the NAO than on the NAM (Fig. 4a, c) indicating a more localized structure of the tropospheric signature of the NAV (Fig. 6a, b). These results make no assumptions about the differences between NAO and NAM as internal

13

mode of variability, since these modes are to first order driven by the internal atmospheric dynamics and not by large-scale SST changes (Limpasuvan and Hartmann 1999, 2000). 4.2 In the atmosphere‑only model The standalone atmospheric model ECHAM6 experiment forced by the prescribed warm North Atlantic conditions reproduces most of the wintertime atmospheric signal simulated by the coupled model for the warm phase (Figs. 4, 5, 6). This suggests Atlantic SST mainly force the atmospheric changes associated with the large-scale Atlantic

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an… Fig. 3  Stratospheric and tropospheric quasi-hemispheric patterns associated with NAM and NAO in coupled and standalone model configurations. The winter (JFM) geopotential height patterns (in m) at a 500 hPa associated with corresponding NAM for a coupled and b uncoupled model, and associated with the corresponding NAO for c coupled and d uncoupled model. e and f are like a and b but for 30 hPa level. The EOF analysis was performed in the area 20–80°N and 90°W–40°E for NAO and the whole hemisphere north of 20°N for NAM at 30 hPa. The patterns are computed as the covariance of the first principal component and the corresponding hemispheric geopotential height

warm conditions. The comparison between the coupled and uncoupled experiments shows that the coupling and possibly also the SST and sea-ice forcing outside the Atlantic region used in our simulation enhance the persistence of the atmospheric signal in the troposphere and lowermost stratosphere (Fig. 4). In the uncoupled experiments the stratospheric response is stronger and deeper (Figs. 4a, b, 5) than in the coupled experiment, leading to larger differences in the upper stratosphere. However, the difference in the midstratospheric response between coupled and uncoupled model is significant only in the North Pacific and over the

North American continent around the polar regions, not in the Arctic and North Atlantic regions (Suppl. Fig. 1A&B). The strength of the tropospheric response in the uncoupled model becomes weaker than in the coupled model as the surface is approached (Fig. 4c, d). However, these differences are only significant in the North Pacific and in the American and European continents up and down-stream of the North Atlantic storm-track region, but not in the core of the NAO centers (Supp. Fig. 1C&D). In all our experiments the same external climatological radiative forcing with no year-to-year changes was

13

N.-E. Omrani et al.

Fig. 4  Simulated seasonal evolutions of the vertical structure of atmospheric changes associated with large-scale Atlantic warming conditions. Simulated seasonal evolutions (3 month running mean) of the vertical structure of atmospheric changes associated with largescale Atlantic warming presented as a projection of geopotential height anomalies (in gpm) from the warm composite on winter (JFM)

a northern hemispheric NAM and c localized NAO. The NAM-, NAO-patterns and the associated projections were computed at different levels between 1000 and 0.1 hPa. b and d are like a and c but for the projection of the geopotential height response on the NAM and NAO in the AGCM standalone simulations. The NAM and NAO patterns were computed at different levels from the control simulation

imposed, suggesting wave-induced dynamics as the driver for the stratospheric changes (Andrews et al. 1987). The poleward heat flux in the high latitude lowermost stratosphere is used to describe the upward wave propagation (Newman et al. 2001; Nishii et al. 2009). As shown in Fig. 7a, b, the decrease (or negative tendency) in the stratospheric vortex (Figs. 4, 5) and the associated warming (Fig.  7c, d) are maintained by the upward propagation of planetary waves. The response of the upward wave activities in the standalone simulation is much higher than in the coupled model, which is consistent with the stronger stratospheric circulation response seen in the uncoupled

experiments. The difference in the NAV-induced wave propagation between coupled and uncoupled configurations (Fig. 7a, b) is consistent with the differences in the stratospheric response. The weak NAO-response to the Atlantic SST tripole pattern found in previous studies was interpreted as a direct SST impact on the growth rate of the most unstable baroclinic waves, which is proportional to the low-level meridional temperature gradient and the associated vertical wind shear (Hoskins and Valdes 1990). This impact on the baroclinicity occurs through (1) extra-tropical turbulent heat fluxes (Czaja et al. 2003), which affect the low-level

13

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an… Fig. 5  Simulated stratospheric changes associated with large-scale Atlantic warming conditions. a, b The evolution of the precursory changes in the coupled model based on the warm composite analogous to previous figures. Plotted are the 30 hPa (in gpm) for 3 month means centred on c January and d February, which correspond to the time period preceding the strongest simulated tropospheric response to the model Atlantic SST. c, d is as in a, b, but simulated by the standalone atmospheric model in response to the Atlantic warm conditions taken from the coupled model (see Fig. 2a). Significant differences (according to two-sided T tests) at the 95 %-level are hatched with vertical and at the 90 %-level with horizontal white lines

meridional temperature gradient, and (2) convection over the tropical Atlantic, which impacts the subtropical jet (Terray and Cassou 2002). The NAO-response to the largescale North Atlantic warm conditions cannot be apparently explained by these mechanisms alone, as our previous lowtop model experiments (Omrani et al. 2014), just as several other low-top multi-model results (Hodson et al. 2010), do not reproduce the observed NAO-change. Instead, the high-top model simulations (Omrani et al. 2014) suggest a direct impact of the stratosphere through stratosphere/ troposphere coupling (see also Fig. 4) and an indirect impact through the reduction of background wind (Omrani et al. 2014), which favors the upward wave propagation in both stratosphere and troposphere (Charney and Drazin 1961). The high-latitude tropospheric warming (Fig. 7c, d)

can contribute also to the reduction of tropospheric baroclinicity seen in Fig. 7c, d; this warming is, according to thermal wind balance, a part of the downward propagating stratospheric easterly wind anomalies that characterize the negative NAO or NAM (Fig. 4). The mechanisms of the stratospheric involvement in the impact of large-scale NAV on the tropospheric circulation can be described as follows: the large-scale North Atlantic Warming induces upward wave propagation into the stratosphere (Fig. 7a, b), which leads to stratospheric warming (Fig. 7c, d) and stratospheric vortex weakening (Fig. 4a, b) that propagate into the troposphere. In addition, the associated tropospheric warming in the high latitude North Atlantic regions induces a decrease of the growth rate of baroclinic waves (Fig. 7c, d). The mechanisms proposed here act in addition

13

N.-E. Omrani et al. Fig. 6  Simulated tropospheric changes associated with large-scale Atlantic warming conditions. As in Fig. 4 but for 500 hPa geopotential height (in gpm) and 3 month mean centred on a, c February and b, d March

to the classical direct Atlantic SST-impact on baroclinic wave activity (Czaja et al. 2003; Terray and Cassou 2002), leading to a more efficient activation of negative NAO-like patterns.

5 Response to the cold phase Unlike the warm conditions, where a significant precursory vortex weakening is seen in the whole stratosphere (Figs.  4, 5), the vortex strengthening associated with the cold conditions of the coupled model are confined to the lowermost stratosphere (Fig. 8a, c). The circulation changes also peak in January instead of February or March, in the troposphere. The stratospheric changes in

13

the coupled model are much smaller for the cold North Atlantic conditions than for the warm ones and they show a more wave-like instead of a quasi-annular hemispheric structure (Figs. 5a, b, 9a, b). Whereas the tropospheric geopotential height from the composite shows a high symmetry between warm and cold North Atlantic conditions (Figs.  2, 6a, b, 10a, b), the symmetry in the stratosphere is extremely weak (Figs. 5a, b, 9a, b). The tropospheric changes for the cold conditions project more positively on the local NAO than on the quasi-hemispheric NAM (Figs. 8a, c, 10a, b). The standalone sensitivity experiment with ECHAM6 forced by prescribed large-scale cooling does not reproduce the results of the coupled model (Figs. 8, 9, 10). The highest discrepancies between the coupled and

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an…

Fig. 7  Simulated dynamical changes associated with large-scale Atlantic warming conditions. The seasonal evolution (3 month running mean) of the warm composite relative to the climatology (a) of zonally averaged eddy heat flux (in °K m/s) in the lower stratosphere (100 hPa) and c of the air temperature at 50 and 850 hPa levels averaged between 80°W and 20°E and 50–80°N as well as of the baro-

clinicity averaged over 40°W–10°E and 50–55°N. b, d as a, c but for the response to Atlantic warm conditions in the AGCM standalone experiments. The poleward heat flux in a, b is used as an indicator of upward wave propagation. The baroclinicity (c, d) is proportional to the growth rate of the most unstable part of synoptic wave activity

uncoupled mode are seen in the stratosphere. In the troposphere the response is significant but it doesn’t agree well with the coupled model, especially in February.

coupled MPI-ESM without year-to-year variations in external forcing. This suggests that the observed relationship between NAV and atmospheric circulation changes over the north Atlantic region may be robust. The seasonal evolution of the vertical structure of atmospheric changes associated with warm North Atlantic conditions revealed a downward propagation of easterly wind anomalies from the stratosphere into the troposphere. These results are consistent with the stratospheric role for the NAO response to the observed 1950s warm conditions shown in our previous study (Omrani et al. 2014). The sensitivity experiments using the standalone atmospheric model component of MPI-ESM (ECHAM6) show that the downward propagation of the signal in the coupled model

6 Summary and conclusion We have shown that the observed basin-wide warm North Atlantic conditions are associated with a winter SLP pattern resembling the negative NAO, and that the cold conditions are associated with a pattern resembling the positive NAO. A similar relation is found in a long-term control simulation using the stratosphere resolving atmosphere–ocean

13

N.-E. Omrani et al.

Fig. 8  Simulated seasonal evolution of the vertical structure of atmospheric changes associated with large-scale Atlantic cooling conditions. As Fig. 4 but for the cold phase. The simulated seasonal evolution (3 month running mean) of the vertical structure of atmospheric changes associated with large-scale Atlantic cooling as a pro-

jection of geopotential height anomalies from the cold composite on winter (JFM) a northern hemispheric NAM and b localized NAO. c and d are like a and b but for the projection of the geopotential height response on the NAM and NAO in the AGCM standalone simulations

and the associated atmospheric changes are mainly forced by the large-scale North Atlantic warm conditions. The differences between coupled and uncoupled models for the warm conditions were only significant in limited regions outside of the core of NAM and NAO centers. However, these differences increase from the middle to upper stratosphere. The mechanism of the atmospheric response to warm conditions involves wave-maintained stratosphere/troposphere coupling, high latitude tropospheric warming, and reduction of low-level baroclinicity. This mechanism does not exclude but instead adds to the mechanisms known from previous studies, which showed that the reduction

of low-level baroclinicity can also be induced directly by high latitude turbulent heat fluxes from the ocean and the warming in the tropical North Atlantic. The coupled model shows high symmetry between warm and cold phases in the troposphere. This symmetry is, however, very weak in the stratosphere in both pattern and magnitude. In contrast to the warm phase, the atmospheric changes associated with the cold phase from the coupled model are confined to the troposphere and lower stratosphere. The standalone atmospheric model ECHAM6 forced by the cold SST-anomalies restricted to the Atlantic is not able to reproduce the results from the coupled model for cold North Atlantic conditions. The comparison of cold

13

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an… Fig. 9  Simulated stratospheric changes associated with large-scale Atlantic cooling conditions. As Fig. 5 but for the cold case. a, b The evolution of the stratospheric precursory changes in the coupled model based on the cold composite. Plotted are the 30 hPa (in gpm) for 3 month means centred on c January and d February. c, d is as in a, b, but simulated by the standalone atmospheric model in response to the Atlantic cold conditions taken from the coupled model (see Fig. 2b). Significant differences (according to two-sided T tests) at the 95 %-level are hatched with vertical and at the 90 %-level with horizontal white lines

and warm conditions in the standalone configuration shows a very high degree of non-linearity in the stratosphere, which may also impact the non-linearity in the tropospheric response and strengthen it through stratosphere/troposphere coupling. The comparison of the standalone and coupled model shows that the active two-way atmosphere/ocean interaction and possibly the impact of other basins, Arctic sea-ice and internal atmospheric variability may modulate the shape and strength of tropospheric response to the NAV and that this modulation is strongest in the cold NAV-phase. The differences of the atmospheric NAV-signature in the coupled and uncoupled model may be an artifact of the differences in the model configurations. Different configurations of the same model can lead to different background circulation and thus to different atmospheric

wave dynamics, which has been shown to impact not only the occurrence of stratospheric warming but also the atmospheric response to different forcings like, greenhouse gases (GHG), ENSO- and NAV-induced SST variability. The significant difference in stratospheric wave dynamics and even in the Major Stratospheric Warming (MSW) between ocean/atmosphere-coupled and standalone atmosphere configurations of the same model has been shown using the Community Earth System Model (CESM) (Hansen et al. 2014). Similar differences in wave dynamics and circulation changes have been shown between configurations with and without Quasi-Biennial Oscillation (QBO) (Hansen et al. 2014) and between high-top and low-top configurations in response to GHG (Karpechko and Manzini 2012), NAV

13

N.-E. Omrani et al. Fig. 10  Simulated tropospheric changes associated with large-scale Atlantic cooling conditions. As in Fig. 9 but for 500 hPa geopotential height (in gpm) and 3 month mean centred on a, c February and b, d March

(Omrani et al. 2014) and ENSO (Cagnazzo and Manzini 2009). The differences between the climatologies of our uncoupled control and coupled model simulations show conditions that are similar to the easterly QBO-phase (EQBO) in the tropical stratosphere (Suppl. Fig. 2). Such background wind promotes, analogous to the EQBO-phase, weakening of the polar vortex and westerlies according to the HoltonTan effect (Holton and Tan 1980). This weakening is maintained by upward and poleward wave propagation. These differences in the poleward propagation of planetary waves induced by background wind may explain the differences in the upper stratospheric response of uncoupled and coupled models in the warm case (Suppl. Fig. 2), but not in the cold case where no large stratospheric anomalies are simulated.

13

As discussed in the introduction, the changes in the coupled stratosphere/troposphere system can also be forced by SST-changes outside the Atlantic basin. The observed SST-changes associated with the NAV are significant in the Indo-Pacific basins and show much higher changes in the warm than in the cold phase (Fig. 11). However, these anomalies remain much weaker compared to the typical ENSO-anomalies. The Indo-Pacific regions undergo strong internal variability on different timescales, which, in addition to a short observational time record, make it difficult to assess if the NAV can lead to such changes in the Indo-Pacific SST. The long-term coupled model results show, however, much weaker SST changes in the IndoPacific regions during large-scale NAV than the observations (Fig. 11). This may contribute to differences between

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an… Fig. 11  Observed global SST changes associated with positive and negative NAV index. Observed warm (a) and cold (b) winter (JFM) composite relative to climatology based on one standard deviation of the NAV-index for SST (in°K). Significant differences (according to two-sided T tests) at the 95 %-level are hatched with vertical and at the 90 %-level with horizontal white lines. The composite analysis was performed over the period 1870–2009

model and observation. In particular, the well-known strengthening (weakening) of the Aleutian low in response to warm (cold) ENSO-like conditions during warm (cold) NAV-phases can provide an additional precursory wave changes that acts to strengthen (weaken) the stratospheric vortex. Such changes should be weak in our coupled model simulation (Figs. 6, 10) due to the weaker model NAV-related SST-changes in the tropical Indo-Pacific regions (compared to the observation, Figs. 11, 12). These are weakest in the cold case, and thus they are unlikely to explain the differences between our coupled and uncoupled simulations. However the tropospheric response to

the cold NAV seen in the Pacific could be important for the non-linear behaviour in the stratospheric NAV-signature in our model simulations. The response—in both coupled and uncoupled simulation—shows a lowering of the pressure in the sub-polar North Pacific (Figs. 2d, 10), which is stronger and has larger spatial extent in the uncoupled simulation. Such changes can interfere positively with climatological waves and force the Major Sudden Stratospheric Warming (MSSW) (Garfinkel et al. 2012) and thus stratospheric vortex weakening that may dampen the possible strengthening of stratospheric vortex due to the cold NAV. As shown in Fig. 13, the tropospheric precursors of

13

N.-E. Omrani et al. Fig. 12  Simulated global SST changes in the MPI-ESM model associated with positive and negative NAVI. Simulated warm and cold winter (JFM) composite relative to climatology in MPI-ESM model, based on one standard deviation of the NAVI, for a, b SST (in °K). Significant differences (according to twosided T tests) at the 95 %-level are hatched vertical with and at the 90 %-level with horizontal white lines

MSSW in both coupled and uncoupled model show lowering of pressure in sub-polar western pacific supporting this hypothesis. This provides plausible explanation of the non-linearity in the stratospheric response to the NAV in our model. Another difference between coupled and uncoupled simulations is the lack of synoptic temporal SST-variation in the standalone simulations, which were driven by the monthly climatological SST. The impact of daily SSTvariation, which is present in the coupled model on our results is unclear and need additional studies. This could be important especially for the tropospheric response, where

13

synoptic or baroclinic eddies are important for the large scale circulation patterns (Lorenz and Hartmann 2003). Finally the constructive and/or destructive interference between the SST-forced atmospheric response and atmospheric changes induced by SST-response to internal atmospheric variability (known as weather-noise) could also play important role in modulating the structure and strength of the total response to SST in the coupled model (Chen and Schneider 2014; Chen et al. 2013). The atmospheric response to the weather-noise-induced SST-changes is missing in the stand-alone simulation, since the SST is prescribed. Thus, the differences in the response between

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an… Fig. 13  Precursory tropospheric changes associated with major sudden stratospheric warming (MSSW). Composite of 500 hPa geopotential height anomalies preceding major sudden stratospheric warming by 5–20 days, MSSW for a the coupled model simulation and b the uncoupled control simulation

the coupled and uncoupled models may also be due to such constructive or destructive interference (which occurs only in the coupled model). This may be more important in the cold case. Our work is a sensitivity study of the coupled troposphere/stratosphere system to the large-scale NAV without any assumption on the time scale, mechanisms and forcing of Atlantic variability. This study identifies a possible impact of the NAV on stratosphere/tropospherecoupled system that is supported by the observational record (Omrani et al. 2014; Peings and Magnusdottir 2014a). Because of the long oceanic memory, this link has the potential to strongly enhance the predictability of the NAO and NAM on inter-annual to decadal timescales. However the mechanisms for NAV and its interaction with the whole stratosphere/troposphere coupled system on different time scales remains poorly understood and controversial, as does the role of internal and external factors. Acknowledgments  We are grateful to Sandro Wellyanto Lubis, Hisashi Nakamura, Marco Giorgetta and Mojib Latif for many fruitful discussion. Computing resources at the Deutsche Klimarechenzentrum, and the Norddeutscher Verbund für Hoch—und Höchstleistungsrechnen are also acknowledged. We are also grateful to our reviewers for the very constructive comments. The work was supported by the Deutsche Forschungsgemeinschaft under the Emmy Noether—Programm (Grant KE 1471/2-1); also by the European Union SUMO (ERC Grant # 266722) and STEPS (PCIG10-GA-2011-304243) projects; the DecCen project funded by the research council of Norway; by the Centre for Climate Dynamics at the Bjerknes centre, Norway; by the Max-Planck-Society, and by the Federal Ministry of Education and Research in Germany (BMBF) through the research programme ‘‘MiKlip’’ (FKZ: 01LP1158A).

References Andrews DG, Holton JR, Leovy CB (1987) Middle atmosphere dynamics. Academic Press, London Ba J et al (2014) A multi-model comparison of Atlantic multidecadal variability. Clim Dyn. doi:10.1007/s00382-014-2056-1 Bader J, Latif M (2003) The impact of decadal-scale Indian Ocean sea surface temperature anomalies on Sahelian rainfall and the North Atlantic Oscillation. Geophys Res Lett 30(22):2169. doi:10.102 9/2003gl018426 Baldwin MP, Dunkerton TJ (1999) Downward propagation of the Arctic Oscillation from the stratosphere to the troposphere. J Geophys Res 104:30937–30946 Booth BBB, Dunstone NJ, Halloran PR, Andrews T, Bellouin N (2012) Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability (vol 484, pg 228, 2012). Nature 485(7399):534. doi:10.1038/Nature11138 Cagnazzo C, Manzini E (2009) Impact of the stratosphere on the winter tropospheric teleconnections between ENSO and the North Atlantic and European Region. J Clim 22(5):1223–1238. doi:10. 1175/2008jcli2549.1 Charney JG, Drazin PG (1961) Propagation of Planetary-Scale disturbances from the lower into the upper atmosphere. J Geophys Res 66(1):83–109. doi:10.1029/JZ066i001p00083 Chen H, Schneider EK (2014) Comparison of the SST-forced responses between coupled and uncoupled climate simulations. J Clim 27(2):740–756. doi:10.1175/Jcli-D-13-00092.1 Chen H, Schneider EK, Kirtman BP, Colfescu I (2013) Evaluation of weather noise and its role in climate model simulations. J Clim 26(11):3766–3784. doi:10.1175/Jcli-D-12-00292.1 Czaja A, Robertson AW, Huck T (2003) The role of atlantic ocean– atmosphere coupling in affecting North Atlantic Oscillation variability. In: Hurrell JW, Kushnir Y, Ottersen G, Visbeck M (eds) The North Atlantic Oscillation: climate significance and environmental impact. Am Geophys Union, Washington, DC, pp 147–172 Delworth T, Manabe S, Stouffer RJ (1993) Interdecadal variations of the thermohaline circulation in a coupled ocean–atmosphere model. J Clim 6:1993–2011

13

N.-E. Omrani et al. Deser C, Blackmon ML (1993) Surface climate variations over the North-Atlantic Ocean during Winter—1900–1989. J Clim 6(9):1743–1753 Deser C, Thomas RA, Peng S (2007) The transient atmospheric circulation response to North Atlantic SST and Sea Ice anomalies. J Clim 20(18):4751–4767 Eden C, Jung T (2001) North Atlantic interdecadal variability: oceanic response to the North Atlantic Oscillation (1865–1997). J Clim 14(5):676–691 Eden C, Willebrand J (2001) Mechanism of interannual to decadal variability of the North Atlantic circulation. J Clim 14(10):2266–2280 Fletcher CG, Kushner PJ (2011) The role of linear interference in the annular mode response to tropical SST forcing. J Clim 24(3):778–794. doi:10.1175/2010jcli3735.1 Garfinkel CI, Hartmann DL, Sassi F (2010) Tropospheric precursors of anomalous northern hemisphere stratospheric polar vortices. J Clim 23(12):3282–3299. doi:10.1175/2010jcli3010.1 Garfinkel CI, Butler AH, Waugh DW, Hurwitz MM, Polvani LM (2012) Why might stratospheric sudden warmings occur with similar frequency in El Nino and La Nina winters? J Geophys Res-Atmos 117:D19106. doi:10.1029/2012jd017777 Gastineau G, Frankignoul C (2012) Cold-season atmospheric response to the natural variability of the Atlantic meridional overturning circulation. Clim Dyn 39(1–2):37–57. doi:10.1007/ S00382-011-1109-Y Giorgetta MA et al (2013) Climate and carbon cycle changes from 1850 to 2100 in MPI-ESM simulations for the coupled model intercomparison project phase 5. J Adv Model Earth Syst 5(3):572–597. doi:10.1002/Jame.20038 Hansen F, Matthes K, Petrick C, Wang W (2014) The influence of natural and anthropogenic factors on major stratospheric sudden warmings. J Geophys Res Atmos 119(13):8117. doi:10.1002/20 13jd021397 Haynes PH (2005) Stratospheric dynamics. Annu Rev Fluid Mech 37:263–293 Hodson DLR, Sutton RT, Cassou C, Keenlyside N, Okumura Y, Zhou TJ (2010) Climate impacts of recent multidecadal changes in Atlantic Ocean sea surface temperature: a multimodel comparison. Clim Dynam 34(7–8):1041–1058. doi:10.1007/S00382-009-0571-2 Hoerling MP, Hurrell JW, Xu TY (2001) Tropical origins for recent North Atlantic climate change. Science 292(5514):90–92 Holton JR, Tan HC (1980) The influence of the equatorial quasi-biennial oscillation on the global circulation at 50 Mb. J Atmos Sci 37(10):2200–2208. doi:10.1175/1520-0469(1980)037<2200:Tio teq>2.0.Co;2 Hoskins BJ, Karoly DJ (1981) The steady linear response of a spherical atmosphere to thermal and orographic forcing. J Atmos Sci 38(6):1179–1196 Hoskins BJ, Valdes PJ (1990) On the existence of storm-tracks. J Atmos Sci 47(15):1854–1864. doi:10.1175/1520-0469(1990)047< 1854:oteost>2.0.co;2 Hurwitz MM, Newman PA, Garfinkel CI (2012) On the influence of North Pacific sea surface temperature on the Arctic winter climate. J Geophys Res Atmos 117:D19110. doi:10.1029/201 2jd017819 Ineson S, Scaife AA (2009) The role of the stratosphere in the European climate response to El Nino. Nat Geosci 2(1):32–36. doi:10.1038/Ngeo381 Jungclaus JH, Haak H, Latif M, Mikolajewicz U (2005) Arctic– North Atlantic interactions and multidecadal variability of the meridional overturning circulation. J Clim 18(19):4013–4031. doi:10.1175/Jcli3462.1 Karpechko AY, Manzini E (2012) Stratospheric influence on tropospheric climate change in the Northern Hemisphere. J Geophys Res Atmos. doi:10.1029/2011jd017036

13

Kavvada A, Ruiz-Barradas A, Nigam S (2013) AMO’s structure and climate footprint in observations and IPCC AR5 climate simulations. Clim Dynam 41(5–6):1345–1364. doi:10.1007/ S00382-013-1712-1 Keenlyside NS, Ba J, Mecking J, Omrani N-O, Latif M, Zhang R, Msadek R (2014) North Atlantic multi-decadal variability— mechanisms and predictability. In: Chang C-P, Ghil M, Latif M, Wallace M (eds) Climate change: multidecadal and beyond. World Scientific Publishing, Singapore Knight JR, Allan RJ, Folland CK, Vellinga M, Mann ME (2005) A signature of persistent natural thermohaline circulation cycles in observed climate. Geophys Res Lett 32(20):L20708. doi:10.102 9/2005gl024233 Kushnir Y (1994) Interdecadal variations in North Atlantic sea surface temperature and associated atmospheric conditions. J Clim 7:141–157 Kushnir Y, Robinson WA, Blade I, Hall NMJ, Peng S, Sutton R (2002) Atmospheric GCM response to extratropical SST anomalies: synthesis and evaluation. J Clim 15(16):2233–2256 Latif M, Keenlyside NS (2011) A perspective on decadal climate variability and predictability. Deep Sea Res II 58:1880–1894 Limpasuvan V, Hartmann DL (1999) Eddies and the annular modes of climate variability. Geophys Res Lett 26(20):3133–3136 Limpasuvan V, Hartmann D (2000) Wave-maintained annulare modes of climate variability. J Clim 13:4414–4429 Lorenz DJ, Hartmann DL (2003) Eddy-zonal flow feedback in the Northern Hemisphere winter. J Clim 16:1212–1227 Mann ME, Emanuel KA (2006) Atlantic hurricane trends linked to climate change. EOS Trans AGU 87(24):233. doi:10.1029/200 6EO240001 Manzini E, Giorgetta MA, Esch M, Kornblueh L, Roeckner E (2006) The influence of sea surface temperatures on the northern winter stratosphere: ensemble simulations with the MAECHAM5 model. J Clim 19(16):3863–3881 Manzini E, Cagnazzo C, Fogli PG, Bellucci A, Muller WA (2012) Stratosphere–troposphere coupling at inter-decadal time scales: implications for the North Atlantic Ocean. Geophys Res Lett 39:L05801. doi:10.1029/2011gl050771 Mecking J, Keenlyside N, Greatbatch R (2013) Stochastically-forced multidecadal variability in the North Atlantic: a model study. Clim Dyn. doi:10.1007/s00382-013-1930-6 Medhaug I, Furevik T (2011) North Atlantic 20th century multidecadal variability in coupled climate models: sea surface temperature and ocean overturning circulation. Ocean Sci Discuss 8(1):353–396. doi:10.5194/osd-8-353-2011 Newman PA, Nash ER, Rosenfield JE (2001) What controls the temperature of the Arctic stratosphere during the spring? J Geophys Res Atmos 106(D17):19999–20010. doi:10.1029/200 0jd000061 Nishii K, Nakamura H, Miyasaka T (2009) Modulations in the planetary wave field induced by upward-propagating Rossby wave packets prior to stratospheric sudden warming events: a casestudy. Q J R Meteor Soc 135(638):39–52. doi:10.1002/Qj.359 Omrani NE, Keenlyside NS, Bader J, Manzini E (2014) Stratosphere key for wintertime atmospheric response to warm Atlantic decadal conditions. Clim Dynam 42(3–4):649–663. doi:10.1007/ S00382-013-1860-3 Ottera OH, Bentsen M, Drange H, Suo LL (2010) External forcing as a metronome for Atlantic multidecadal variability. Nat Geosci 3(10):688–694 Peings Y, Magnusdottir G (2014a) Forcing of the wintertime atmospheric circulation by the multidecadal fluctuations of the North Atlantic Ocean. Environ Res Lett 9(3):034018. doi:10.1088/1748-9326/9/3/034018 Peings Y, Magnusdottir G (2014b) Response of the wintertime Northern Hemisphere atmospheric circulation to current and projected

Troposphere–stratosphere response to large-scale North Atlantic Ocean variability in an… Arctic Sea ice decline: a numerical study with CAM5. J Clim 27(1):244–264. doi:10.1175/Jcli-D-13-00272.1 Peng S, Robinson WR, Li S (2003) Mechanisms for the NAO responses to the North Atlantic SST tripole. J Climate 16:1987 Rayner NA, Parker DE, Horton EB, Folland CK, Alexander LV, Rowell DP, Kent EC, Kaplan A (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:10.1 029/2002JD002670 Reichler T, Kim J, Manzini E, Kroger J (2012) A stratospheric connection to Atlantic climate variability. Nat Geosci 5(11):783– 787. doi:10.1038/Ngeo1586 Ruiz-Barradas A, Nigam S, Kavvada A (2013) The Atlantic multidecadal oscillation in twentieth century climate simulations: uneven progress from CMIP3 to CMIP5. Clim Dyn 41(11–12):3301– 3315. doi:10.1007/S00382-013-1810-0 Stevens B et al (2013) Atmospheric component of the MPI-M earth system model: ECHAM6. J Adv Model Earth Syst 5(2):146– 172. doi:10.1002/Jame.20015 Sutton RT, Hodson DLR (2005) Atlantic Ocean forcing of North American and European summer climate. Science 309:115–118 Terray L, Cassou C (2002) Tropical Atlantic sea surface temperature forcing of quasi-decadal climate variability over the North Atlantic-European region. J Clim 15(22):3170–3187

Thompson DWJ, Wallace JM (2000) Annular modes in the extratropical circulation. Part I: month-to-month variability. J Clim 13(5):1000–1016 Thompson DWJ, Wallace JM, Hegerl GC (2000) Annular modes in the extratropical circulation. Part II: trends. J Clim 13(5):1018–1036 Vellinga M, Wu PL (2004) Low-latitude freshwater influence on centennial variability of the Atlantic thermohaline circulation. J Clim 17(23):4498–4511. doi:10.1175/3219.1 Visbeck M, Chassignet E, Curry R, Delworth T, Dickson B, Krahmann G (2003) The ocean’s response to North Atlantic Oscillation variability, in “The North Atlantic Oscillation. In: Hurrell JW, Kushnir Y, Ottersen G, Visbeck M (eds) The North Atlantic Oscillation: climate significance and environmental impact. American Geophysical Union, Washington, DC, pp 113–146 Zanchettin D, Rubino A, Matei D, Bothe O, Jungclaus JH (2012) Multidecadal-to-centennial SST variability in the MPI–ESM simulation ensemble for the last millennium. Clim Dyn. doi:10.1007/s00382-012-1361-9 Zanchettin D, Bothe O, Muller W, Bader J, Jungclaus JH (2014) Different flavors of the Atlantic multidecadal variability. Clim Dyn 42(1–2):381–399. doi:10.1007/S00382-013-1669-0 Zhang R, Delworth TL (2006) Impact of Atlantic multidecadal oscillations on India/Sahel rainfall and Atlantic hurricanes. Geophys Res Lett 33(17):L17712. doi:10.1029/2006gl026267

13