• Alexander, M. A., , and C. Deser, 1995: A mechanism for the recurrence of winter midlatitude SST anomalies. J. Phys. Oceanogr., 25 , 122137.

    • Search Google Scholar
    • Export Citation
  • Barsugli, J. J., , and D. S. Battisti, 1998: The basic effects of atmosphere–ocean thermal coupling on midlatitude variability. J. Atmos. Sci., 55 , 477493.

    • Search Google Scholar
    • Export Citation
  • Battisti, D. S., , U. S. Bhatt, , and M. A. Alexander, 1995: A modeling study of the interannual variability of the North Atlantic Ocean. J. Climate, 8 , 30673083.

    • Search Google Scholar
    • Export Citation
  • Bjerknes, J., 1964: Atlantic air–sea interaction. Advances in Geophysics, Vol. 10 Academic Press, 1–82.

  • Curry, R. G., , and M. S. McCartney, 2001: Ocean gyre circulation changes associated with the North Atlantic Oscillation. J. Phys. Oceanogr., 31 , 33743400.

    • Search Google Scholar
    • Export Citation
  • Czaja, A., , and J. Marshall, 2001: Observations of atmosphere–ocean coupling in the North Atlantic. Quart. J. Roy. Meteor. Soc., 127 , 18931916.

    • Search Google Scholar
    • Export Citation
  • Czaja, A., , and C. Frankignoul, 2002: Observed impact of North Atlantic SST anomalies on the North Atlantic Oscillation. J. Climate, 15 , 606623.

    • Search Google Scholar
    • Export Citation
  • Czaja, A., , A. W. Robertson, , and T. Huck, 2003: The role of Atlantic ocean–atmosphere coupling in affecting North Atlantic Oscillation variability. The North Atlantic Oscillation: Climatic Significance and Environment Impact, Geophys. Monogr., Vol. 134, Amer. Geophys. Union, 147–172.

    • Search Google Scholar
    • Export Citation
  • Delworth, T., 1996: North Atlantic interannual variability in a coupled ocean–atmosphere model. J. Climate, 9 , 23562375.

  • Delworth, T., , and R. J. Greatbatch, 2000: Multidecadal thermohaline circulation variability driven by atmospheric surface flux forcing. J. Climate, 13 , 14811495.

    • Search Google Scholar
    • Export Citation
  • Delworth, T., , and M. E. Mann, 2000: Observed and simulated multidecadal variability in the North Atlantic. Climate Dyn., 16 , 661676.

  • Delworth, T., , S. Manabe, , and R. J. Stouffer, 1993: Interdecadal variations in the thermohaline circulation in a coupled ocean–atmosphere model. J. Climate, 6 , 19932010.

    • Search Google Scholar
    • Export Citation
  • Deser, C., , and M. L. Blackmon, 1993: Surface climate variations over the North Atlantic Ocean during winter: 1900–1993. J. Climate, 6 , 17431753.

    • Search Google Scholar
    • Export Citation
  • Eden, C., , and T. Jung, 2001: North Atlantic interdecadal variability: Oceanic response to the North Atlantic Oscillation (1865–1997). J. Climate, 14 , 676691.

    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., , P. Muller, , and E. Zorita, 1997: A simple model of the decadal response of the ocean to stochastic wind forcing. J. Phys. Oceanogr., 27 , 15331546.

    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., , E. Kestenare, , N. Sennéchael, , G. de Coëtlogon, , and F. D’Andréa, 2001: On decadal-scale ocean–atmosphere interactions in the extended ECHAM1/LSG climate simulation. Climate Dyn., 16 , 333354.

    • Search Google Scholar
    • Export Citation
  • Goodman, J., , and J. Marshall, 1999: A model of decadal middle-latitude atmosphere–ocean coupled models. J. Climate, 12 , 621641.

  • Grötzner, A., , M. Latif, , and T. P. Barnett, 1998: A decadal climate cycle in the North Atlantic Ocean as simulated by the ECHO coupled GCM. J. Climate, 11 , 831847.

    • Search Google Scholar
    • Export Citation
  • Halliwell, G., 1998: Simulation of North Atlantic decadal/multidecadal winter SST anomalies driven by basin-scale atmospheric circulation anomalies. J. Phys. Oceanogr., 28 , 521.

    • Search Google Scholar
    • Export Citation
  • Hansen, D. V., , and H. F. Bezdek, 1996: On the nature of decadal anomalies in North Atlantic sea surface temperature. J. Geophys. Res., 101 , 87498758.

    • Search Google Scholar
    • Export Citation
  • Hoerling, M. P., , J. W. Hurrell, , and T. Xu, 2001: Tropical origins for recent North Atlantic climate change. Science, 292 , 9092.

  • Honda, M., , H. Nakamura, , J. Ukita, , L. Kouska, , and K. Takeuchi, 2001: Interannual seesaw between the Aleutian and Iceland Lows. Part I: Seasonal dependence and life cycle. J. Climate, 14 , 10291042.

    • Search Google Scholar
    • Export Citation
  • Hurrel, J. W., , G. Ottersen, , Y. Kushnir, , and M. Visbeck, 2003: An overview of the North Atlantic Oscillation. The North Atlantic Oscillation: Climatic Significance and Environment Impact, Geophys. Monogr., Vol. 134, Amer. Geophys. Union, 1–35.

    • Search Google Scholar
    • Export Citation
  • Jacob, R. L., 1997: Low frequency variability in a simulated atmosphere ocean system. Ph.D. thesis, University of Wisconsin—Madison, 155 pp.

  • Kelly, K. A., , and S. Dong, 2004: The relationship of western boundary current heat transport and storage to midlatitude ocean-atmosphere interaction. Earth Climate: The Ocean–Atmosphere Interaction, Geophys. Monogr., No. 137, Amer. Geophys. Union, 347–364.

    • Search Google Scholar
    • Export Citation
  • Kushnir, Y., 1994: Interdecadal variations in North Atlantic sea surface temperature and associated atmospheric conditions. J. Climate, 7 , 142157.

    • Search Google Scholar
    • Export Citation
  • Kushnir, Y., , W. A. Robinson, , I. Bladé, , N. M. J. Hall, , S. Peng, , and R. T. Sutton, 2002: Atmospheric GCM response to extratropical SST anomalies: Synthesis and evaluation. J. Climate, 15 , 22332256.

    • Search Google Scholar
    • Export Citation
  • Latif, M., , and T. P. Barnett, 1994: Causes of decadal climate variability in the North Pacific/North Atlantic sector. Science, 266 , 634637.

    • Search Google Scholar
    • Export Citation
  • Latif, M., , and T. P. Barnett, 1996: Decadal climate variability over the North Pacific and North America: Dynamics and predictability. J. Climate, 9 , 24072423.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., , and L. Wu, 2004: Atmospheric response to North Pacific SST anomaly: The role of ocean–atmosphere coupling. J. Climate, 17 , 18591882.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., , J. Kutzbach, , and L. Wu, 2000: Modeling climate shift of El Niño in the Holocene. Geophys. Res. Lett., 27 , 22652268.

  • Liu, Z., , B. Otto-Bliesner, , J. Kutzbach, , L. Li, , and C. Shields, 2003: Coupled climate simulation of the evolution of global monsoons in the Holocene. J. Climate, 16 , 24722490.

    • Search Google Scholar
    • Export Citation
  • Mantua, N. J., , S. R. Hare, , Y. Zhang, , J. Wallace, , and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78 , 10691079.

    • Search Google Scholar
    • Export Citation
  • Marshall, J., , H. Johnson, , and J. Goodman, 2001a: A study of the interaction of the North Atlantic Oscillation with the ocean circulation. J. Climate, 14 , 13991421.

    • Search Google Scholar
    • Export Citation
  • Marshall, J., and Coauthors, 2001b: Atlantic climate variability. Int. J. Climatol., 21 , 18631898.

  • Molinari, R. L., , D. A. Mayer, , J. F. Festa, , and H. F. Bezdek, 1997: Multiyear variability in the near-surface temperature structure of the midlatitude western North Atlantic Ocean. J. Geophys. Res, 102 , 32673278.

    • Search Google Scholar
    • Export Citation
  • Okumura, Y., , S. P. Xie, , A. Numaguti, , and Y. Tanimoto, 2001: Tropical Atlantic air–sea interaction and its influence on the NAO. Geophys. Res. Lett., 28 , 15071510.

    • Search Google Scholar
    • Export Citation
  • Peng, S., , A. Mysak, , H. Ritchie, , J. Derome, , and B. Dugas, 1995: The difference between early and middle winter atmospheric response to sea surface temperature anomalies in the northwestern Atlantic. J. Climate, 8 , 137157.

    • Search Google Scholar
    • Export Citation
  • Peng, S., , W. A. Robinson, , and S. Li, 2003: Mechanisms for the NAO responses to the North Atlantic SST tripole. J. Climate, 16 , 19872004.

    • Search Google Scholar
    • Export Citation
  • 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.

    • Search Google Scholar
    • Export Citation
  • Robertson, A. W., , C. R. Mechoso, , and Y. J. Kim, 2000: The influence of Atlantic sea surface temperature anomalies on the North Atlantic Oscillation. J. Climate, 13 , 122138.

    • Search Google Scholar
    • Export Citation
  • Saravanan, R., , and J. C. McWilliams, 1998: Advective ocean–atmosphere interaction: An analytical stochastic model with implications for decadal variability. J. Climate, 11 , 165188.

    • Search Google Scholar
    • Export Citation
  • Sutton, R. T., , and M. R. Allen, 1997: Decadal predictability of North Atlantic sea surface temperature and climate. Nature, 388 , 563567.

    • Search Google Scholar
    • Export Citation
  • Sutton, R. T., , and P. Mathieu, 2002: Response of the atmosphere–ocean mixed layer system to anomalous ocean heat flux convergence. Quart. J. Roy. Meteor. Soc., 128 , 12591275.

    • Search Google Scholar
    • Export Citation
  • Sutton, R. T., , W. A. Norton, , and S. P. Jewson, 2001: The North Atlantic Oscillation—What role for the Ocean? Atmos. Sci. Lett., doi:10.1006/asle.2000.0021.

  • Thompson, D. J., , and J. M. Wallace, 1998: The Arctic oscillation signature in wintertime geopotential height and temperature fields. Geophys. Res. Lett., 25 , 12971300.

    • Search Google Scholar
    • Export Citation
  • Timmerman, A., , M. Latif, , R. Voss, , and A. Grötzner, 1998: Northern Hemispheric interdecadal variability: A coupled air–sea mode. J. Climate, 11 , 19061931.

    • Search Google Scholar
    • Export Citation
  • Toth, Z., , and E. Kalnay, 1997: Ensemble forecasting at NMC and the breeding method. Mon. Wea. Rev., 125 , 32973319.

  • Venzke, S., , M. R. Allen, , R. T. Sutton, , and D. P. Rowell, 1999: The atmospheric response over the North Atlantic to decadal changes in sea surface temperatures. J. Climate, 12 , 25622584.

    • Search Google Scholar
    • Export Citation
  • Visbeck, M., , H. Cullen, , G. Krahman, , and N. Naik, 1998: An ocean model’s response to North Atlantic Oscillation–like wind forcing. Geophys. Res. Lett., 25 , 45214524.

    • Search Google Scholar
    • Export Citation
  • Visbeck, M., , E. Chassignet, , R. Curry, , T. Delworth, , B. Dickson, , and G. Krahmann, 2003: The ocean’s response to NAO variability. The North Atlantic Oscillation: Climatic Significance and Environment Impact, Geophys. Monogr., Vol. 134, Amer. Geophys. Union.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., , and M. Kimoto, 1999: Tropical-extratropical connection in the Atlantic atmosphere-ocean variability. Geophys. Res. Lett., 26 , 22472250.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., , and M. Kimoto, 2000: Ocean atmosphere thermal coupling in the North Atlantic: A positive feedback. Quart. J. Roy. Meteor. Soc., 126 , 33433369.

    • Search Google Scholar
    • Export Citation
  • Weaver, A. J., , E. S. Sarachik, , and J. Marotzke, 1991: Freshwater flux forcing of decadal and interdecadal oceanic variability. Nature, 353 , 836838.

    • Search Google Scholar
    • Export Citation
  • Wu, L., , and Z. Liu, 2002: Is tropical Atlantic variability driven by the North Atlantic Oscillation? Geophys. Res. Lett., 29 .1653, doi:10.1029/2002GL014939.

    • Search Google Scholar
    • Export Citation
  • Wu, L., , and Z. Liu, 2003: Decadal variability in the North Pacific: The eastern North Pacific mode. J. Climate, 16 , 31113131.

  • Wu, L., , Z. Liu, , R. Gallimore, , R. Jacob, , D. Lee, , and Y. Zhong, 2003: Pacific decadal variability: The tropical Pacific mode and the North Pacific mode. J. Climate, 16 , 11011120.

    • Search Google Scholar
    • Export Citation
  • Yulaeva, E., , N. Schneider, , D. Pierce, , and T. Barnett, 2001: Modeling of North Pacific climate variability forced by oceanic heat flux anomalies. J. Climate, 14 , 40274066.

    • Search Google Scholar
    • Export Citation
  • Zorita, E., , and C. Frankignoul, 1997: Modes of North Atlantic decadal variability in the ECHAM1/LSG coupled atmosphere–ocean general circulation model. J. Climate, 10 , 183200.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Leading EOFs of the observed North Atlantic (north of 20°N) winter (DJF) SST anomalies and associated atmospheric 500-mb geopotential height: (a) EOF1 (NAM) and (b) EOF2 (NAT) and regression of 500-mb geopotential height on the (c) NAT and (d) NAM indices. Units for EOF and regression are K and m (per standard deviation of the index), respectively. The magnitude of each EOF is reflected in its pattern.

  • View in gallery

    Power spectrum of the (a) NAM and (b) NAT indices in the observations. The power spectrum is calculated using a multitaper method (three tapers).

  • View in gallery

    As in Fig. 1 but for the FOAM model control simulation.

  • View in gallery

    Power spectrum of (a) NAT, (b) NAM, (c) the subtropical high, and (d) the Icelandic low indices in the FOAM control simulation. The subtropical height and Icelandic low indices are defined as the area-averaged winter 500-mb geopotential height over the region (20°– 45°N, 70°W–20°E) and (50°– 75°N, 90°W– 20°E) respectively.

  • View in gallery

    Leading EOFs of the winter SST anomalies in PC-G, where ocean–atmosphere coupling is deactivated over the global oceans using the partial coupling scheme: (a) EOF1 (NAT) and (b) EOF2 (NAM).

  • View in gallery

    Power spectrum of (a) the NAT and (b) the NAM indices in PC-G, power spectrum of (c) the STH and (d) the IL in PC-G, and power change of (e) the NAT and the NAM and (f) the STH and the IL. In (e) each index is multiplied by the square root of the area-integrated variance before calculating the power spectrum to reflect the amplitude of each mode in each experiment. The power change at a specific frequency is calculated as the difference of power between the control (FC) and the partial coupling (PC-G) experiments divided by the power of the power in PC-G.

  • View in gallery

    Regression of (a) SST, (b) anomalous and (c) mean zonal advection, (d) anomalous and (e) mean meridional advection, (f) anomalous and (g) mean vertical advection, (h) VDC, (i) surface heat flux (positive downward), and (j) wind stress and its curl on the NAT index. All data are bandpassed to retain the variability between 6 and 30 yr. Units for SST, heat balance terms, and wind stress and its curl are °C, W m−2, N m−2, and N m−3 (× 10−9) per standard deviation of the index, respectively.

  • View in gallery

    Lagged regression of SST and upper-400-m heat content on the NAT index. All data are bandpassed to retain variability between 6 and 30 yr. Units for SST and HC are °C and m °C−1 per standard deviation of the tripole index, respectively. Contour interval is 0.08 and 4 for SST and heat content, respectively.

  • View in gallery

    Lagged regression of the anomalous meridional advection (υTy), surface heat flux (HF; positive downward), and wind stress curl on the NAT index. All data are bandpassed to retain the variability between 6 and 30 yr. Units are W m−2 for υTy and HF, N m−3 for the wind stress curl. Contour interval is 1 for υTy. Contour levels for HF are (−20, −10, −5, −3, −2, −1, 1, 2, 3, 5, 10, 20), and for the wind stress curl are (−30, −20, −10, −6, −4, −2, 2, 4, 6, 10, 20, 30) × 10−9, respectively.

  • View in gallery

    (a) DJF mean ocean current velocity averaged over upper 200-m (vectors), climatologic zero wind stress curl line (dark lines), and instantaneous regression of anomalous wind stress curl on the NAT index (contours) [all data are bandpass filtered (6–30 yr)]. Units for current and wind stress curl are cm s−1 and N m−2. Contour interval for the wind stress curl is 4. The Hovmoeller diagram of (b) the zonally averaged (70°–20°W) regression of SST and (c) the upper-400-m heat content on the NAT index, respectively. Units for SST and HC are °C and m °C−1. Contour interval is 0.08 and 4 for SST and heat content, respectively.

  • View in gallery

    (a) First EOF of the North Atlantic winter SST anomalies in the TC experiment. (b) Power spectrum of the normalized time series of EOF1. In the TC experiment, the wind stress for the ocean in the coupled model is constrained to the model climatological wind stress at each time step through interpolation of monthly climatological data.

  • View in gallery

    Ensemble mean atmospheric response (Jan–Feb) to the tripole mixed layer anomaly initiated in earlier winter (1 Nov): (a) 500-mb geopotential height, (b) wind stress (vectors) and curl (contours), and (c) SST averaged in Jan and Feb. Units for SST, wind stress (and curl), and geopotential height are °C, N m−2, and m. The contour interval for the wind stress curl is 2 × 10−9.

  • View in gallery

    Coherence (solid line) and phase (dotted line) spectrum between (a) the NAT and the PDO and (b) the NAM and the PDO indices in the control simulation. The dashed lines represent the 95% confidence level.

  • View in gallery

    Power spectrum of (a) the NAT and (b) the NAM indices in PC-PAF and power change of (c) the NAT and the NAM, and (d) the STH and the IL. In (c) each index is multiplied by the square root of the area-integrated variance before calculating the power spectrum to reflect the amplitude of each mode in each experiment. The power change at a specific frequency is calculated as the difference of power between the control (FC) and the partial coupling (PC-PAF) experiments divided by the power of the power in PC-PAF.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 25 25 6
PDF Downloads 17 17 5

North Atlantic Decadal Variability: Air–Sea Coupling, Oceanic Memory, and Potential Northern Hemisphere Resonance

View More View Less
  • 1 Center for Climatic Research, University of Wisconsin—Madison, Madison, Wisconsin
© Get Permissions
Full access

Abstract

In this paper, the causes and mechanisms of North Atlantic decadal variability are explored in a series of coupled ocean–atmosphere simulations. The model captures the major features of the observed North Atlantic decadal variability. The North Atlantic SST anomalies in the model control simulation exhibit a prominent decadal cycle of 12–16 yr, and a coherent propagation from the western subtropical Atlantic to the subpolar region. A series of additional modeling experiments are conducted in which the air–sea coupling is systematically modified in order to evaluate the importance of air–sea coupling for the North Atlantic decadal variability being studied. This shall be referred to as “modeling surgery.” The results suggest the critical role of ocean–atmosphere coupling in sustaining the North Atlantic decadal oscillation at selected time scales. The coupling in the North Atlantic is characterized by a robust North Atlantic Oscillation (NAO)-like atmospheric response to the SST tripole anomaly, which tends to intensify the SST anomaly and, meanwhile, also provide a delayed negative feedback. This delayed negative feedback is predominantly associated with the adjustment of the subtropical gyre in response to the anomalous wind stress curl in the subtropical Atlantic. Atmospheric stochastic forcing can drive SST patterns similar to those in the fully coupled ocean–atmosphere system, but fails to generate any preferred decadal time scales. The simulated North Atlantic decadal variability, therefore, can be viewed as a coupled ocean–atmosphere mode under the influence of stochastic forcing.

This modeling study also suggests some potential resonance between the Pacific and the North Atlantic decadal fluctuations mediated by the atmosphere. The modeling surgery indicates that the Pacific climate, although not a necessary precondition, can impact the North Atlantic climate variability substantially.

Corresponding author address: Lixin Wu, Center for Climatic Research, University of Wisconsin—Madison, 1225 West Dayton Street, Madison, WI 53706. Email: lixinwu@facstaff.wisc.edu

Abstract

In this paper, the causes and mechanisms of North Atlantic decadal variability are explored in a series of coupled ocean–atmosphere simulations. The model captures the major features of the observed North Atlantic decadal variability. The North Atlantic SST anomalies in the model control simulation exhibit a prominent decadal cycle of 12–16 yr, and a coherent propagation from the western subtropical Atlantic to the subpolar region. A series of additional modeling experiments are conducted in which the air–sea coupling is systematically modified in order to evaluate the importance of air–sea coupling for the North Atlantic decadal variability being studied. This shall be referred to as “modeling surgery.” The results suggest the critical role of ocean–atmosphere coupling in sustaining the North Atlantic decadal oscillation at selected time scales. The coupling in the North Atlantic is characterized by a robust North Atlantic Oscillation (NAO)-like atmospheric response to the SST tripole anomaly, which tends to intensify the SST anomaly and, meanwhile, also provide a delayed negative feedback. This delayed negative feedback is predominantly associated with the adjustment of the subtropical gyre in response to the anomalous wind stress curl in the subtropical Atlantic. Atmospheric stochastic forcing can drive SST patterns similar to those in the fully coupled ocean–atmosphere system, but fails to generate any preferred decadal time scales. The simulated North Atlantic decadal variability, therefore, can be viewed as a coupled ocean–atmosphere mode under the influence of stochastic forcing.

This modeling study also suggests some potential resonance between the Pacific and the North Atlantic decadal fluctuations mediated by the atmosphere. The modeling surgery indicates that the Pacific climate, although not a necessary precondition, can impact the North Atlantic climate variability substantially.

Corresponding author address: Lixin Wu, Center for Climatic Research, University of Wisconsin—Madison, 1225 West Dayton Street, Madison, WI 53706. Email: lixinwu@facstaff.wisc.edu

1. Introduction

Observational studies show that the North Atlantic sea surface temperature exhibits substantial interannual to decadal variability (e.g., Bjerknes 1964; Deser and Blackmon 1993; Kushnir 1994; Hansen and Bezdek 1996; Sutton and Allen 1997; Monlinari et al. 1997; Czaja and Marshall 2001). Over the North Atlantic, SST anomalies are characterized by two distinctive patterns: a tripole, with anomalies extending from the southeast coast of the United States into the central basin, surrounded by anomalies with opposite polarities in the midlatitudes south of Greenland and the Tropics south of 30°N, and a monopole with uniform polarity over the entire North Atlantic (e.g., Deser and Blackmon 1993). These SST variations are associated with a redistribution of atmospheric mass between the Arctic and the subtropical Atlantic, namely, the North Atlantic Oscillation (NAO), and a change of the heat and moisture transport, as well as storm tracks, etc., over the North Atlantic (see reviews by Marshall et al. 2001a; Hurrell et al. 2003; Czaja et al. 2003).

The causes and mechanisms of North Atlantic decadal variability, however, have not been fully understood. Current debates concern the role of ocean–atmosphere coupling and oceanic processes in the generation of North Atlantic decadal variability. Both observations and modeling studies suggest that the prominent and robust NAO variability intrinsic to the atmosphere drives large-scale interannual SST variability through changes of surface heat flux as well as the Ekman current (e.g., Bjerknes 1964; Battisti et al. 1995; Delworth 1996; Marshall et al. 2001a; Visbeck et al. 2003). However, on decadal to multidecadal time scales, observational evidence suggests that SST variations may not be directly associated with local air–sea interaction, and nonlocal oceanic dynamics tend to be more important (e.g., Kushnir 1994; Hansen and Bezdek 1996; Latif and Barnett 1996; Sutton and Allen 1997; Halliwell 1998; Visbeck et al. 1998). It is conceivable that both changes of thermocline gyres and the Atlantic themohaline circulation can contribute to decadal SST variations (e.g., Goodman and Marshall 1999; Marshall et al. 2001b).

Several mechanisms on decadal variability in the North Atlantic have been proposed. At longer time scales of a few decades, it has been argued that the basin-scale SST anomaly is primarily associated with the thermohaline-driven meridional overturning cell (MOC) of the North Atlantic Ocean (e.g., Kushnir 1994; Delworth et al. 1993). Theoretical and OGCM modeling studies indicate the possibility of self-sustained oscillation of the MOC (e.g., Weaver et al. 1991). It still remains uncertain whether the multidecadal variability is purely an oceanic process or whether it involves atmospheric stochastic forcing and/or active air–sea coupling (e.g., Delworth and Greatbatch 2000; Timmerman et al. 1998). At decadal time scales, it has been argued that SST variations may arise from intrinsic atmospheric variability in conjunction with the mean flow advection or intrinsic Rossby wave dynamics in the ocean (Frankignoul et al. 1997; Saravanan and McWilliams 1998; Wu and Liu 2003). Observations indicate a propagation of coherent SST fluctuations from the southeast coast of the United States, across the Atlantic, to the northwest of Scotland in the United Kingdom at decadal time scales (Hansen and Bezdek 1996; Sutton and Allen 1997), seemingly following the mean gyre circulation. However, as some coupled GCMs demonstrated, SST variations, including propagation, may also be associated with anomalous advection of the mean temperature gradient (e.g., Grötzner et al. 1998) as a reflection of the adjustment of the ocean to the anomalous wind stress forcing (e.g., Curry and McCartney 2001; Frankignoul et al. 2001; Eden and Jung 2001; Marshall et al. 2001b; Visbeck et al. 2003). On the other hand, similar mechanisms put forward by Latif and Barnett (1994) for the North Pacific have been suggested for the decadal variability in the North Atlantic (Grötzner et al. 1998), although this mechanism is not apparent in other coupled GCMs (e.g., Zorita and Frankignoul 1997; Frankignoul et al. 2001).

The inconsistent mechanisms for North Atlantic decadal variability in different coupled GCMs may reflect the inconsistency of the atmospheric response to midlatitude SST anomalies in different AGCMs (see a review by Kushnir et al. 2002). A positive feedback between the SST tripole and the NAO has been found in statistical analysis of observations (Czaja and Frankignoul 2002) and some modeling studies (e.g., Venzke et al. 1999; Sutton et al. 2001; Watanabe and Kimoto 2000; Peng et al. 2003). Such a feedback is weaker than the intrinsic NAO variability and is usually underestimated in AGCM studies with fixed SST anomalies. Recent studies suggest that the atmospheric response is determined not only by the SST, but also by the heat flux (e.g., Yulaeva et al. 2001; Sutton and Mathieu 2002; Liu and Wu 2004). The correct response, however, is generated only in the fully coupled ocean–atmosphere system that produces the correct combination of SST and heat flux naturally (Liu and Wu 2004).

The North Atlantic climate variability may also be affected by tropical SST variability. Recent modeling studies suggest that the North Atlantic climate change since 1950 is linked to a progressive warming of tropical Pacific and Indian Ocean sea surface temperature (Hoerling et al. 2001). Other AGCM studies also indicate that SST in the tropical Atlantic may affect the NAO variability through the atmospheric bridge of Rossby waves in response to the shifting of the intertropical convergence zone (ITCZ) (e.g., Robertson et al. 2000; Watanabe and Kimoto 1999; Okumura et al. 2001).

In this paper, we will explore the role of ocean–atmosphere coupling and oceanic dynamics in the generation of North Atlantic decadal variability by conducting a series of modeling surgical experiments in which the air–sea coupling is systematically modified (Wu et al. 2003). The model-simulated tripole exhibits significant decadal variability with a period of approximately 12–16 yr, consistent with the observations (Deser and Blackmon 1993). The surgical experiments clearly demonstrate the critical role of ocean–atmosphere coupling in sustaining the decadal oscillation at these selected time scales. The upper-ocean heat budget further reveals the importance of ocean dynamics in creating the SST anomalies and providing the slow memory for the decadal oscillation.

The paper is constructed as follows. A brief description of the coupled model and observational data are given in section 2. Section 3 describes the observed and model-simulated North Atlantic variability. The role of ocean–atmosphere coupling and oceanic dynamics in the generation of North Atlantic variability are studied in section 4 and 5, respectively. In section 6, the remote impact of the Pacific climate is investigated. The paper is concluded by a summary and further discussion.

2. Model description and data

The model we used is the Fast Ocean–Atmosphere Model, version 1.5 (FOAM1.5; Jacob 1997). The AGCM is the R15 version of the National Center for Atmospheric Research (NCAR) community climate model, version 2 (CCM2), but with 19 vertical levels, and with the atmospheric physics replaced by those of the community climate model, version 3 (CCM3). The OGCM is developed following the Geophysical Fluid Dynamics Laboratory (GFDL) Modular Ocean Model (MOM), with a horizontal resolution of 1.4° latitude × 2.8° longitude × 24 vertical levels. The model has a thermodynamical sea ice component model. Without flux adjustment, the model captures most of the major features of the observed climatology (Jacob 1997; Liu et al. 2003), as in most state-of-the-art climate models, and reasonable climate variability, such as ENSO (Liu et al. 2000), Pacific decadal variability (Wu et al. 2003), and tropical Atlantic variability (Wu and Liu 2002). The considerable errors in the model’s climatological SST are in the Gulf Stream region, with errors up to ±2°C. The simulated Gulf Stream is also too diffusive due to the coarse resolution of the ocean model. The model also generates a stable meridional overturning circulation in the Atlantic, although the magnitude is slightly stronger than that of the observed (not shown). Here, we will focus on the cold season [December–February (DJF)], and we will address the warm season in a separate study. We did not include November and March in the cold season due to the format of data archive of the experiments.

The observational data used in this study are from the Hadley Centre Global Sea Ice and Sea Surface Temperature dataset (HadISST: Rayner et al. 2003) from 1870 to 2003, and the geopotential height from National Centers for Environmental Prediction (NCEP) reanalysis from 1948. To be consistent with the model, we used the DJF SST and atmospheric geopotential height. For the observation, no significant difference is seen when November and March data were included.

3. Observed and model-simulated North Atlantic variability

a. Observed variability

Over the North Atlantic, the first and second EOF of the observed winter SST anomalies (north of 20°N, after detrending) exhibits two distinctive patterns (Figs. 1a,b). The first EOF, which accounts for 23% of the total variance, has uniform polarity over the entire North Atlantic (Fig. 1a). This mode will be referred to as the North Atlantic monopole (NAM). The second EOF, which accounts for 18% of the total variance, exhibits a tripole pattern with anomalies extending from the southeast coast of the United States into the central basin, surrounded by anomalies in the midlatitudes south of Greenland and the Tropics south of 25°N, with opposite polarities (Fig. 1b). This mode will be referred to as the North Atlantic tripole (NAT).

The NAT and NAM tend to exhibit distinctive temporal behavior, as is shown by the power spectrum of the normalized time series of these two modes (Fig. 2). The NAM tends to be dominated by decadal to multidecadal variability (Fig. 2a), while the NAT is dominated by interannual to decadal variability (Fig. 2b). The NAT exhibits a broad peak in the decadal band roughly between 12 and 16 yr, with the power decreasing toward longer time scales (Fig. 2b). In contrast, the power spectrum of the NAM is dominated by multidecadal variations (Fig. 2a).

Associated with these two SST modes, the atmosphere also reveals coherent large-scale patterns. We take the DJF 500-mb geopotential height from the NCEP reanalysis and then regress on each SST mode index (defined as the normalized EOF time series) within the period from 1948 to 2003. Associated with the NAT, the atmosphere exhibits a north–south dipole in the 500-mb height, with a resemblance to the NAO pattern. A stronger-than-normal Icelandic low (IL) and a stronger-than-normal subtropical high (STH) are associated with warming in the subtropics and cooling in the subpolar and eastern subtropical Atlantic (Fig. 1d). Associated with the warm (cold) NAM SST anomaly, a downstream ridge (trough) is shown with the center located along 45°N (Fig. 1c). The atmospheric Z500 regression patterns associated with the NAT and the NAM resemble the first and second EOF modes of Z500, respectively (not shown). It should be noted that the instantaneous regression here reflects the predominant forcing of the atmosphere on the ocean. Overall, the results of the observational analysis here are consistent with previous observational studies, although different datasets are used (e.g., Deser and Blackmon 1993; Kushnir 1994).

b. Model-simulated North Atlantic variability

The leading two EOFs of the North Atlantic winter SST anomalies (north of 20°N) from a 400-yr FOAM control (CTRL) simulation are shown in Fig. 3. The first two leading EOFs correspond to the NAT and the NAM, which account for 32% and 14% of the total variance, respectively (Figs. 3a,b). Overall, the simulated tripole and monopole resemble that of the observed (Fig. 1a, b), in spite of the shifting of the EOF orders as well as some other notable deficiencies. Compared with the observations, the overall amplitudes of the NAT are stronger, especially over the eastern subtropical and subpolar Atlantic; the center of the western subtropical lobe is somewhat shifted from the eastern coast of the United States to the interior ocean (Fig. 3a versus Fig. 1b). The NAM also shows an eastward shift and weaker amplitudes than that of the observed (Fig. 3b versus Fig. 1a).

The simulated NAT exhibits distinctive decadal variability. The power spectrum of the NAT time series shows a pronounced decadal peak between 12 and 16 yr and a decreasing tendency of power toward a very low frequency, consistent with that of the observation (Fig. 4a versus Fig. 2b). In contrast to the NAT, the NAM shows an increasing tendency toward a very low frequency (Fig. 4b) similar to that of the observed (Fig. 2a), although it also demonstrates significant decadal variability within 12–16 yr.

The patterns of the atmospheric circulation associated with these two SST modes also resemble the observations. In the model, a NAT with warming in the western subtropical and cooling in the subpolar and eastern subtropical Atlantic is associated with a stronger-than-normal Icelandic low and a stronger-than-normal subtropical high (Fig. 3c); a warm NAM is associated with a downstream ridge in the midlatitudes (Fig. 3d). The pattern correlations between the simulated and the observed are statistically significant (Fig. 3c versus Fig. 1d and Fig. 3d versus Fig. 1c). There are, of course, some deficiencies. The most notable deficiency is the eastward shifting of the action centers in the subtropics for each atmospheric pattern.

The coherence between these basin-scale SST anomalies and the atmospheric circulations can be seen in the temporal characteristics of the atmospheric variability. We calculate the power spectrum of the IL and the STH, respectively (Figs. 4c,d). Both the IL and STH exhibit substantial interannual and decadal variability and bear this similarity for up to 20 yr. Both the IL and the STH exhibit the same decadal peaks as the NAT and the NAM, although the statistical significance is weaker. The IL power keeps decreasing toward lower frequencies, in contrast to the STH which retains power toward multidecadal time scales. To further examine the covariability between the ocean and atmosphere over the North Atlantic, a cross-spectrum analysis is performed. The coherence spectrum shows that the variation of the NAT is significantly associated with the IL, while the NAM is significantly associated with the STH (not shown). The tendency of the IL and STH power spectrum appears to be consistent with analysis of the observed sea level pressure data (Czaja and Marshall 2001).

In summary, the model reasonably reproduces the observed North Atlantic decadal variability. Both observations and the model control simulation indicate a potential ocean–atmosphere coupling over the North Atlantic at decadal time scales. In the following, we will further explore the role of the ocean–atmosphere coupling in the North Atlantic decadal variability by a series of surgical experiments. Unless specified, all of the experiments are integrated for 400 yr from the same initial condition as the control simulation.

4. The role of ocean–atmosphere coupling

Because the intrinsic atmospheric decorrelation time scale is about 1 week or so, the coherent variations between the ocean and atmosphere at decadal time scales strongly suggest the potential role of ocean–atmosphere coupling in the generation of North Atlantic decadal variability. To explore the role of ocean–atmosphere coupling over the Atlantic, a partial coupling (PC) experiment, denoted as PC-G, is performed. In this experiment, over the entire globe, the atmospheric model sees a prescribed model climatological annual cycle of SST and drives the ocean with the flux that is calculated using the predicted SST at each time step; therefore, the ocean–atmosphere coupling is deactivated (Wu et al. 2003). In PC-G, variability in the atmosphere and ocean is generated predominantly by the atmospheric internal stochastic forcing and/or oceanic internal processes. It should be noted that the sea ice is still coupled to the atmosphere in PC-G and, therefore, can affect the atmospheric variability by atmosphere–sea ice feedback. The partial coupling approach may underestimate the atmospheric internal variability due to an amplified thermodynamic damping caused by the fixed SST boundary condition (Barsugli and Battisti 1998), but, nevertheless, the fixed SST forcing of an AGCM provides a useful reference for studying the role of the coupled ocean–atmosphere interaction in the generation of climate variability, and is frequently used in Atmospheric Model Intercomparison Project (AMIP) experiments.

Without ocean–atmosphere coupling, both the NAT and the NAM remain robust, and similar to those in the CTRL (Figs. 5a,b). The pattern correlation with the CTRL exceeds 0.95 for both the NAT and the NAM. This indicates that the ocean–atmosphere coupling is not necessary for the generation of these two modes, which is consistent with previous modeling studies (e.g., Battisti et al. 1995; Delworth 1996). The conclusion, however, does not exclude the impact of the ocean–atmosphere coupling on these two modes, especially at decadal time scales.

Compared with the CTRL, the magnitudes of the NAT in PC-G are reduced over both the subpolar and the eastern subtropical Atlantic but are somewhat enhanced over the western subtropical Atlantic (Fig. 5a versus Fig. 3a). The magnitudes are reduced by 20% over the subploar Atlantic, but more substantially so (by about 40%) over the eastern subtropical Atlantic. A substantial reduction of the magnitudes is also seen for the NAM, which appears to be reduced by over 30% in comparison with the CTRL (Fig. 5b versus Fig. 3b).

Perhaps the most striking change in the absence of the ocean–atmosphere coupling occurs on the decadal evolution of both the NAT and NAM. Without ocean–atmosphere coupling, the decadal oscillation seen in the CTRL disappears for both the NAT and the NAM (Figs. 6a,b). Their spectrum does not show any prominent peaks at decadal time scales, in sharp contrast to those in CTRL (Figs. 4a,b). Furthermore, both the STH and the IL experience a substantial reduction of variance at decadal time scales (Figs. 6c,d). This suggests the critical role of ocean–atmosphere coupling in sustaining the North Atlantic decadal oscillation at these selected time scales, although the reduction in PC-G may be partly attributed to the amplified thermodynamic damping due to the fixed SST boundary condition for the atmosphere.

The role of ocean–atmosphere coupling in selectively sustaining the North Atlantic decadal oscillation is further demonstrated by the power change at different frequencies. For both the NAT and NAM, the most substantial power enhancement by the ocean–atmosphere coupling appears at decadal time scales. The NAT attains the maximum intensification between 12 and 16 yr by a factor of 8–10, in contrast to the interannual and multidecadal time scales where no substantial enhancement occurs (Fig. 6e). A similar power enhancement at decadal time scales is also seen for the NAM (Fig. 6e). The role of ocean–atmosphere coupling in sustaining the North Atlantic decadal oscillation is also seen in the atmosphere (Fig. 6f). The coupling intensifies both the IL and the STH at decadal time scales. The intensification peaks around 12–16 yr by a factor of 2.5 and 4.5 for the IL and STH, respectively. The coupling also intensifies the atmospheric variability at some interannual time scales, but is less significant than the decadal time scales.

It is noted that the NAM in PC-G retains remarkable multidecadal variability (Fig. 6b), although it is weaker than that in CTRL (Fig. 6e). This tends to suggest that the ocean–atmosphere coupling is not necessary for the generation of the North Atlantic multidecadal variability. Potential mechanisms of the North Atlantic multidecadal variability may involve the North Atlantic thermohaline circulation (e.g., Kushnir 1994; Delworth and Mann 2000) and/or atmospheric stochastic forcing through spatial resonance (e.g., Saravanan and McWilliams 1998). Due to the short integration (400 yr), we will not further explore the mechanisms of the simulated multidecadal variability.

In summary, the PC-G experiment clearly demonstrates the important role of ocean–atmosphere coupling in sustaining the North Atlantic decadal oscillation at the selected time scales. The atmospheric stochastic forcing can generate both the NAT and the NAM, but fails to give rise to any preferred decadal time scales. A further study on the nature of ocean–atmosphere coupling will be given in a later section.

In the next section, we will explore the mechanisms of the simulated North Atlantic decadal variability. We will primarily focus on the NAT, because at the decadal time scale the simulated NAM appears to be in quadrature with the NAT, although in the observations the NAM tends to be more distinctive from the NAT (not shown).

5. Mechanisms of North Atlantic decadal oscillation

The partial coupling experiment suggests the critical role of ocean–atmosphere coupling in sustaining the North Atlantic decadal oscillation, but it is not clear how the decadal SST anomalies are generated and sustained by the coupling with the atmosphere. Observational evidence suggests that SST variations at decadal time scales may not be directly associated with local air–sea interaction, and nonlocal oceanic dynamics tend to be more important (e.g., Kushnir 1994; Hansen and Bezdek 1996; Sutton and Allen 1997). In the following, we will explore the role of oceanic dynamics in the generation of SST decadal oscillation and how ocean–atmosphere coupling is involved.

a. The role of ocean dynamics

To explore the role of oceanic dynamics in the generation of SST decadal oscillation, we will implement a three-step approach. First, an analysis of the upper-ocean heat budget is conducted to identify key processes contributing to the SST decadal anomalies. Second, a lagged regression analysis is performed to study how oceanic dynamics are involved in setting the time scales and switching the phase of the oscillation. Finally, we performed some diagnostic experiments to further assess the key processes identified in the first two steps. To focus on decadal time scales, all of the data are bandpassed before the analysis in order to retain the variability between 6 and 30 yr.

1) Upper-ocean heat budget

To see how SST anomalies are created over the North Atlantic, we analyze the ocean’s upper-100-m heat budget in the control simulation. Various terms in the heat balance include surface heat flux (HF), anomalous advection (uTx, υTy, and wTz, mean advection (UT ′x, VT ′y, and WT ′z), horizontal diffusion (HD), and vertical diffusion and convection (VDC). The anomalous DJF current (u′, υ′, and w′) is defined as the deviation from the model’s climatological mean DJF current (U, V, and W). All of these terms are averaged over the upper 100 m and are then regressed on the NAT index. The instantaneous regression of all these terms is shown in Fig. 7 (except for the small horizontal diffusion terms). In addition, we also calculated the regression of SST and wind stress and its curl. The heat budget analysis reveals that the NAT is primarily associated with the surface heat flux (Fig. 7i), anomalous meridional advection (Fig. 7d), and vertical mixing and convection (Fig. 7h). Specifically, in the subpolar Atlantic, cold (warm) anomalies are created predominantly by the surface heat flux (Fig. 7i), in response to the anomalous westerlies (easterlies) (Fig. 7j), and are damped by vertical convection and mixing (Fig. 7h). The anomalous westerlies (easterlies) also bring cold (warm) water from higher (lower) latitudes to enhance the cold (warm) SST anomalies (Fig. 7d). In the western subtropical Atlantic, warm (cold) SST anomalies are primarily created by anomalous meridional advection (Fig. 7d) and are damped by vertical convection and mixing (Fig. 7h). The former tends to be caused by the strengthening (weakening) of the subtropical gyre in response to the interior anomalous negative (positive) wind stress curl (Fig. 7j). The surface heat flux (Fig. 7i) tends to enhance the SST anomalies near the east coast of the United States and downstream of the Gulf Stream, but dampen the SST anomalies over the center of the western subtropical lobe. The northward mean advection, albeit weaker, tends to reduce the warming (cooling) near the southern and northern boundaries of the western subtropical lobe by moving cold (warm) anomalies from the eastern subtropical Atlantic and transporting warm (cold) anomalies to the subpolar Atlantic (Fig. 7e). In the eastern subtropical Atlantic, cold (warm) anomalies are created by the surface heat flux (Fig. 7i) and anomalous upwelling (downwelling) (Fig. 7f). The former is associated with the anomalous northeast (southeast) trades, which enhance (reduce) latent heat loss, while the latter tends to be associated the positive (negative) wind stress curl as a result of the equatorward diminishing of the anomalous northeast (southwest) trades (Fig. 7j). In this region, the northward mean flow tends to diminish SST anomalies by moving SST anomalies toward the extratropics (Fig. 7e).

In summary, the analysis of the upper-ocean heat budget reveals a dominant role of the anomalous meridional advection, heat flux, and anomalous vertical advection in the generation of decadal SST anomalies over the western subtropical Atlantic, subpolar Atlantic, and eastern subtropical Atlantic, respectively. In the next section, we will examine how the key processes identified in the above analysis are involved in setting the time scales and switching the phase of the North Atlantic decadal oscillation.

2) Life cycle of North Atlantic decadal variability

To demonstrate the life cycle of the North Atlantic decadal oscillation, we calculated the lagged regressions of the SST and upper-400-m heat content (HC) anomalies on the NAT index, respectively (Fig. 8). Lag-0 regression corresponds to a mature NAT (Fig. 8a1). The heat content anomalies show a similar pattern as that of the SST with the thermocline deepening in the western subtropics and shoaling in the subpolar and eastern subtropical Atlantic (Fig. 8b1). After that, the warm subtropical anomalies tend to move toward the subpolar region (Figs. 8a2–a), resulting in a NAM-like pattern with uniform polarity over the subtropical and subpolar Atlantic (Fig. 8a4). The poleward propagation of subtropical warm anomalies is primarily associated with the anomalous meridional advection (Figs. 9a1–a), and is more clearly seen in the HC anomalies (Figs. 8b2–b). The northward anomalous advection tends to be forced by the interior anomalous negative wind stress curl, which creates an intergyre gyre across the subtropical and the subpolar gyres (Figs. 9c1–c) (Marshall et al. 2001b). Consequently, the initial negative SST gradient across the subtropical/subpolar gyre boundary (Fig. 8a1) is relaxed by this anticyclonic intergyre gyre (Fig. 8a4).

The relaxation of the SST gradient by the intergyre gyre is further reinforced by the subtropical gyre adjustment, which ultimately reverses the SST gradient across the subtropical/subpolar gyre boundary. In the subtropical Atlantic, the negative HC anomalies show a clear westward propagation (Figs. 8b1–b) with a cross-basin time of about 2–6 yr within 15°–30°N. The westward propagation seems to be consistent with the first baroclinic Rossby wave, which is forced by the Ekman upwelling as a response to the anomalous negative wind curl in the eastern subtropics. This upwelling Rossby wave will spin down the subtropical gyre, reduce the northward heat transport, and essentially create cold anomalies in the subtropics. At lag +3 yr, a cold anomaly, albeit weak, first appears in the southwestern Atlantic (Fig. 8a4) accompanied by a negative HC anomaly (Fig. 8b4). This cold anomaly is primarily caused by the reduction of the meridional heat advection (Fig. 9a4) associated with the weakening of the subtropical gyre as a response to the positive wind curl in eastern subtropics (15°–30°N). After that, both SST and HC anomalies tend to grow and propagate northward (Figs. 8a5–a, 8b5–b), due to the further reduction of the meridional heat advection (Figs. 9a5–a). Associated with the SST phase switching, the wind stress curl also changes its sign and intensifies as SST grows (Figs. 9c6–c), indicating a positive feedback at work. Such a feedback can also be inferred from the heat flux. Along the propagation of the cold subtropical anomalies (Figs. 9a4–a), the heat flux tends to diminish the SST anomalies, suggesting forcing of the ocean on the atmosphere (Figs. 9b5–b). This region is the entrance of the storm track; therefore, such a perturbation may distort the storm track and exert a significant atmospheric response (Sutton and Allen 1997). The NAT reappears after about 6 yr but with a reduced amplitude (Fig. 8b7), suggesting that the decadal tripole mode is a damped coupled ocean–atmosphere mode (e.g., Czaja and Marshall 2001).

It is noted that the model-simulated North Atlantic SST anomalies demonstrate a coherent propagation along the Gulf Stream from the subtropical western Atlantic to the subpolar Atlantic (e.g., Sutton and Allen 1997). Consistent with observations, the SST anomalies originate in the southwestern subtropical Atlantic and propagate subsequently to the north (Figs. 9a4–a) and then across the Atlantic (Figs. 9a1–a), within about a decade, seemingly following the path of the simulated Gulf Stream (Fig. 10a). This SST propagation tends to be in phase with the thermocline (Figs. 8b1–b). The northward propagation of SST and HC is further summarized in Figs. 10b and 10c. The speed for both SST and HC is about 1.6 cm s−1, although the northward propagation tendency for HC (Fig. 10c) is clearer than that of the SST (Fig. 10b). The propagation, based on the heat budget analysis, is primarily associated with dynamic adjustment of the subtropical gyre to the interior anomalous wind stress curl.

In summary, the mechanisms of the decadal oscillation in our model simulation are similar to other coupled modeling studies (Grötzner et al. 1998). The phase switch of the decadal oscillation tends to be associated with the anomalous wind stress curl in the subtropical Atlantic. At the mature phase (Fig. 7a), both westerlies and northeast trades are enhanced, resulting in a negative wind curl straddling the subtropical and subpolar gyre (35° and 55°N) and a positive wind curl in the subtropics (10°–35°N) (Fig. 10a). The former tends to create warm SST anomalies in the midlatitudes by shifting the mean gyre boundary to the north and, thus, intensifying the northern part of the western subtropical lobe, resulting in a northward propagation (e.g., Marshall et al. 2001b), while the latter tends to spin down the subtropical gyre and create cold SST anomalies in the western subtropics. The adjustment of the subtropical gyre to the anomalous wind stress in the subtropics essentially provides a delayed negative feedback to switch the phase of the decadal oscillation. In the subtropics (10°–35°N), the cross-basin time scale of the first baroclinic wave ranges roughly from 2 (in the south) to 6 (in the north) yr in the model, essentially setting the time scale of the oscillation.

In the next section, we will further explore the role of wind-driven oceanic adjustment in sustaining the North Atlantic decadal oscillation by carrying a new sensitivity experiment.

3) The role of wind-driven adjustment

To explicitly show the important role of the wind-driven adjustment of the oceanic gyre circulation, we perform an experiment in which the wind stress for the ocean is constrained to the model mean climatologic annual cycle, and, thus, the ocean and the atmosphere are only thermodynamically coupled (TC). Compared with the control simulation, the dynamic adjustment of the ocean circulation to the variable wind stress is completely eliminated in the TC experiment. The TC experiment is still different from an atmospheric model coupled to a mixed layer ocean because the mean oceanic advection remains.

Among the leading EOFs of the winter SST anomalies in the TC experiment, only the first EOF shares some similarity with the NAT (Fig. 11a). This mode explains about the same percentage of the total variance as the NAT in the control simulation, but both the spatial pattern and temporal characteristics differ from the NAT substantially. The most remarkable difference appears over the western subtropics off of the U.S. East Coast and the subpolar region, where the SST variance is substantially reduced (Fig. 11a versus Fig. 3a). In the western subtropics off of the U.S. East Coast, the reduction of the SST variance is associated with the suppression of the anomalous meridional advection in the TC experiment, as inferred from the heat budget analysis (Fig. 7d). Over the subpolar region, the reduction of the SST variance tends to be caused by the southward extension of the climatologic ice margin in the TC experiment (not shown). This climatologic drift may be attributed to the physical deficiencies in the thermodynamic ice model component of the coupled system.

In addition to the spatial difference, the temporal evolution of the first EOF mode in TC is also strikingly different from the NAT. Without dynamic coupling, the decadal oscillation as seen in the control simulation is virtually suppressed (Fig. 11b), indicating the role of the oceanic gyre adjustment in setting the slow memory for the North Atlantic decadal oscillation. No significant decadal variability was found in other EOF modes (not shown). Due to the drift of the climatological ice margin in the TC experiment, results from this experiment remain inconclusive. Nevertheless, the TC experiment tends to be supportive for the previous heat budget analysis.

b. Atmospheric response to North Atlantic SST anomaly

The previous partial coupling experiment suggests the important role of ocean–atmosphere coupling in sustaining North Atlantic decadal variability. The key element of this coupled process is how changes of the atmospheric circulation in response to SST anomalies feed back on the ocean.

A traditional modeling approach to assess the feedback of the ocean to the atmosphere is to use an AGCM with a prescribed SST anomaly as the lower boundary condition (e.g., Peng et al. 1995; Venzke et al. 1999; Sutton et al. 2001; Peng et al. 2003) or an AGCM coupled to a mixed layer ocean forced by a prescribed flux (e.g., Sutton and Mathieu, 2002; Yulaeva et al. 2001). These two approaches sometimes yield contradictory atmospheric responses. Recently, we find that the correct response of the atmosphere to SST anomalies can be generated only in the fully coupled ocean–atmosphere system (Liu and Wu 2004).

In Liu and Wu (2004), the response of the atmosphere to North Pacific SST anomaly is studied by initiating a mixed-layer temperature anomaly in the North Pacific in the fully coupled ocean–atmosphere model. Here, we use the same approach to study the response of the atmosphere to North Atlantic SST anomaly. To dynamically estimate the atmospheric response to a winter SST anomaly in the coupled model, we performed a 40-member fully coupled ensemble of experiments, each starting from a November condition, that consists of a basic state and an anomaly. The basic state is a 1 November atmosphere–ocean state of a year of the control simulation, which is selected 8 yr apart. The anomaly is a warm mixed layer temperature anomaly that is of the same shape as the NAT and NAM, respectively (Figs. 3a,b), but is normalized to have an amplitude of 1°C and uniformly extended from the surface to a 200-m depth. This mixed layer anomaly can be thought of as generated by oceanic dynamics. Each experiment is integrated for 14 months. The difference between the ensemble mean of these 40 experiments and the ensemble mean of the control experiments is the response of the atmosphere to the mixed layer anomaly. Liu and Wu (2004) have shown that this approach is the most natural way to yield responses of the atmosphere to an extratropical Pacific SST anomaly that are consistent with the control simulation as well as observations. This approach is equivalent to the “breeding” method (Toth and Kalnay 1997) that favors the slowest decaying coupled ocean–atmosphere mode.

The atmospheric response to the North Atlantic tripole exhibits a north–south dipole in 500-m height with a stronger-than-normal Icelandic low and stronger-than-normal subtropical high (Fig. 12a). The amplitude is about 15 and 35 m K−1 for the IL and STH, respectively. The response resembles the model-simulated NAO pattern, although the subtropical lobe exhibits an additional center over the North American continent. The response is equivalent baratropic, with the surface circulation characterized by anomalous westerlies in the midlatitudes and northwest trades in the subtropics (Fig. 12b). This surface wind pattern results in a positive wind curl in the subtropics and a negative wind curl in the midlatitudes. The response is relatively stronger in later winter (February–March), compared with earlier winter (November–January), although the initial tripole has been distorted with a much-reduced eastern subtropical Atlantic lobe in late winter (Fig. 12c). This tends to suggest that the atmospheric response is predominantly determined by the extratropical SST anomalies.

The response in our coarse-resolution model (R15) is more or less consistent with other higher-resolution AMIP-type studies (e.g., Peng et al. 2003). The atmospheric response to the NAT bears a similar pattern and polarity as the atmospheric forcing of the NAT (Fig. 12b versus Fig. 7j), indicating a positive feedback at work.

The positive feedback also exists for the NAM. The response of the atmosphere to the NAM exhibits a robust warm-ridge structure in the midlatitdues (not shown). The response is also equivalent baratropic with an amplitude of 30 m K−1 in the 500-mb geopotential height.

The atmospheric response to both the NAT and the NAM, although significant, is still much weaker than the atmospheric intrinsic stochastic variability. This determines that the extratropical ocean–atmospheric interaction is dominated by the forcing of the atmosphere on the ocean, especially at interannual time scales. At decadal time scales, the coupled feedback turns out to be relatively important because of the shorter decorrelation time scales intrinsic to the atmosphere. While the detailed mechanism of the atmospheric response to the extratropical SST anomalies may be complicated (e.g., Kushnir et al. 2002) and beyond the scope of this paper, it is likely that the eddy-induced vorticity flux tends to play an important role in generating such an equivalent barotropic response (e.g., Peng et al. 2003; Liu and Wu 2004). Nevertheless, our “breeding” experiment suggests a positive feedback at work for the North Atlantic ocean–atmosphere interaction, which further substantiate the results of our partial coupling experiment.

6. A further discussion: Remote impact of the Pacific climate

The North Atlantic climate is subjected to multiple forcing mechanisms, including both local and remote resources. One of the dominant remote forcings is from the Pacific, including both the tropical and North Pacific climate. Recent modeling studies suggested that the North Atlantic climate change since 1950 is linked to a progressive warming of tropical Pacific and Indian Ocean sea surface temperature (Hoerling et al. 2001). Observations and the coupled model also indicate the climatic linkage of the North Pacific and the North Atlantic in both the atmosphere (Thompson and Wallace 1998; Honda et al. 2001) and the ocean (Timmerman et al. 1998; Kelly and Dong 2004).

The linkage between the Pacific and the North Atlantic climate is also evident in our model control simulation. To focus on decadal time scales, we use the Pacific decadal oscillation (PDO) index (defined as the principal component of the first EOF of the winter SST anomalies north of 20°N; Mantua et al. 1997) to represent the Pacific decadal climate. The coherence spectrum shows that the simulated PDO and the NAT vary coherently at time scales of 12–16 yr without any significant lags (Fig. 13a). This suggests a potential resonance of these two variability modes, mediated by the atmospheric teleconnection. The PDO also correlates with the NAM at decadal time scales with a lead of 2–3 yr (Fig. 13b). Therefore, there is a possibility that the Pacific decadal oscillation triggers the North Atlantic decadal variability.

To truly assess the role of the Pacific climate in the generation of the North Atlantic decadal variability, we performed a partial coupling surgical experiment, PC-PAF, in which the impact of the Pacific climate variability is eliminated. In the experiment, the atmospheric model sees the prescribed model control annual cycle of the SST over the entire Pacific, but is coupled to the ocean elsewhere.

The PC-PAF experiment clearly shows that the North Atlantic decadal variability originates predominantly locally within the North Atlantic ocean–atmosphere interaction. Without the Pacific impact, both the NAT and NAM patterns remain similar to these in the control simulation (not shown). Furthermore, these two modes also remain significant to decadal variability, similar to the control simulation (Figs. 14a,b versus Figs. 4a,b). The NAT exhibits a prominent peak around 12 yr (Fig. 14a), while the NAM exhibits substantial decadal (16–20 yr) and multidecadal variability (Fig. 14b).

The Pacific climate, although not a dominant forcing, can nevertheless impact the North Atlantic climate. The Pacific significantly enhances the NAT decadal variability broadly between 10 and 20 yr, with a peak around 14 yr (Fig. 14c). A similar enhancement is also seen for the IL (Fig. 14d). The decadal variability of the NAM tends to be less affected by the Pacific, although the variability between 8 and 12 yr is somewhat enhanced. Perhaps the most notable change for the NAM is the suppression of the interannual variability by the Pacific. This is also seen for the STH (Fig. 14d). In the presence of the Pacific climate variability, the power of both the NAM and STH between 2 and 6 yr is reduced by over 30% of that in the absence of the Pacific impact. The details of the teleconnective process between these two basins remain a topic for further study.

Nevertheless, our PC-PAF experiment suggests that the North Atlantic decadal variability arises predominantly from the ocean–atmosphere interaction over the Atlantic basin, although the Pacific climate can exert significant impact.

7. Summary and discussion

In this paper, the causes and mechanisms of North Atlantic decadal variability is explored in a series of coupled ocean–atmosphere simulations. The model control simulation captures the major features of the observed North Atlantic decadal variability. The simulated North Atlantic SST anomalies exhibit a prominent decadal cycle of 12–16 yr and a coherent propagation from the western subtropical Atlantic to the subpolar region. The coupled modeling surgery studies suggest a critical role of the ocean–atmosphere coupling in sustaining the North Atlantic decadal oscillation. The coupling in the North Atlantic is characterized by a robust NAO-like atmospheric response to the SST tripole anomaly, which tends to intensify the latter and, meanwhile, also provide a delayed negative feedback. This delayed negative feedback is predominantly associated with the adjustment of the subtropical gyre in response to the anomalous wind stress curl in the eastern subtropical Atlantic. Although atmospheric stochastic forcing can drive SST patterns similar to those in the fully coupled ocean–atmosphere system, it fails to generate any preferred decadal time scales. Therefore, the North Atlantic decadal variability can be viewed as a damped coupled ocean–atmosphere mode under the influence of stochastic forcing.

Our modeling study also suggests that the Pacific climate variability can affect the North Atlantic decadal fluctuations. The model control simulation indicates a potential resonance between the Pacific decadal oscillation and the North Atlantic decadal variability. Eliminating the impact of the Pacific climate variability can reduce the variance of the North Atlantic decadal fluctuations but cannot disrupt the North Atlantic decadal cycle. This suggests that the North Atlantic decadal oscillation arises predominantly from ocean–atmospheric interaction locally in the Atlantic.

Several questions remain to be further studied. The model shows a coherent SST propagation from the western subtropics to the subpolar Atlantic at about a decade. This is consistent with observations (e.g., Sutton and Allen 1997), but is dominated by anomalous advection instead of the mean advection. The latter also contributes to the model SST anomalies in the southwestern subtropics, but is weaker compared to the former. In reality, the mean advection may play a more important role, which tends to be underestimated in all GCMs (e.g., Grötzner et al. 1998).

Here, we only focused on the cold season. Recent observational studies suggest that the summer SST anomalies may exert a significant atmospheric response in the following winter (e.g., Czaja and Frankignoul 2002), thus, bridging the winter to winter atmospheric variability. Our study here tends to suggest that such a seasonal bridge may also be necessary for the North Atlantic decadal oscillation. The strongest atmospheric response to the winter SST tripole appears in later winter or earlier spring, as is shown in our “breeding experiments” and other higher-resolution AGCM studies (e.g., Peng et al. 2003). This will potentially force a summer SST anomaly, which can affect the atmospheric circulation in the following winter in such a way that can enhance the reemerged SST anomalies (Alexander and Deser 1995). Further studies are needed to clarify this seasonal bridging process.

The North Atlantic variability may also be affected by the tropical Atlantic, as is shown in other AGCM studies (e.g., Watanabe and Kimoto 1999; Robertson et al. 2000; Okumura et al. 2001). The shifting of the ITCZ by the change of the tropical Atlantic SST may impact the North Atlantic through the atmospheric bridge. Our “breeding experiment” tends to suggest that the tropical Atlantic may not be necessary for the response of the atmosphere to the tripole SST. Nevertheless, this effect needs to be further clarified by additional partial coupling experiments.

Finally, our modeling studies suggest a potential predictability of the North Atlantic decadal SST anomalies. SST in the western subtropical Atlantic is predominantly created by the anomalous meridional heat transport that is dictated by the anomalous wind stress in the subtropical Atlantic, and is further amplified by the opposite anomalous wind stress curl in the midlatitudes. The interplay between the wind stress curl in the subtropics and the midlatitudes in the evolution of extratropical SST anomalies provides a potential predictability for the North Atlantic decadal variability. Further surgical experiments similar to the TC experiment, but constraining dynamic coupling regionally, are needed to clarify these dynamic effects.

Acknowledgments

This work is supported by NASA and DOE. Discussions with Drs. Thomas Delworth, John Marshall, and David Battisti were helpful. Comments from both reviewers were helpful in improving the paper. We thank Dr. Robert Jacob for the continuous development of FOAM. Computer time allocations from NCAR are appreciated.

REFERENCES

  • Alexander, M. A., , and C. Deser, 1995: A mechanism for the recurrence of winter midlatitude SST anomalies. J. Phys. Oceanogr., 25 , 122137.

    • Search Google Scholar
    • Export Citation
  • Barsugli, J. J., , and D. S. Battisti, 1998: The basic effects of atmosphere–ocean thermal coupling on midlatitude variability. J. Atmos. Sci., 55 , 477493.

    • Search Google Scholar
    • Export Citation
  • Battisti, D. S., , U. S. Bhatt, , and M. A. Alexander, 1995: A modeling study of the interannual variability of the North Atlantic Ocean. J. Climate, 8 , 30673083.

    • Search Google Scholar
    • Export Citation
  • Bjerknes, J., 1964: Atlantic air–sea interaction. Advances in Geophysics, Vol. 10 Academic Press, 1–82.

  • Curry, R. G., , and M. S. McCartney, 2001: Ocean gyre circulation changes associated with the North Atlantic Oscillation. J. Phys. Oceanogr., 31 , 33743400.

    • Search Google Scholar
    • Export Citation
  • Czaja, A., , and J. Marshall, 2001: Observations of atmosphere–ocean coupling in the North Atlantic. Quart. J. Roy. Meteor. Soc., 127 , 18931916.

    • Search Google Scholar
    • Export Citation
  • Czaja, A., , and C. Frankignoul, 2002: Observed impact of North Atlantic SST anomalies on the North Atlantic Oscillation. J. Climate, 15 , 606623.

    • Search Google Scholar
    • Export Citation
  • Czaja, A., , A. W. Robertson, , and T. Huck, 2003: The role of Atlantic ocean–atmosphere coupling in affecting North Atlantic Oscillation variability. The North Atlantic Oscillation: Climatic Significance and Environment Impact, Geophys. Monogr., Vol. 134, Amer. Geophys. Union, 147–172.

    • Search Google Scholar
    • Export Citation
  • Delworth, T., 1996: North Atlantic interannual variability in a coupled ocean–atmosphere model. J. Climate, 9 , 23562375.

  • Delworth, T., , and R. J. Greatbatch, 2000: Multidecadal thermohaline circulation variability driven by atmospheric surface flux forcing. J. Climate, 13 , 14811495.

    • Search Google Scholar
    • Export Citation
  • Delworth, T., , and M. E. Mann, 2000: Observed and simulated multidecadal variability in the North Atlantic. Climate Dyn., 16 , 661676.

  • Delworth, T., , S. Manabe, , and R. J. Stouffer, 1993: Interdecadal variations in the thermohaline circulation in a coupled ocean–atmosphere model. J. Climate, 6 , 19932010.

    • Search Google Scholar
    • Export Citation
  • Deser, C., , and M. L. Blackmon, 1993: Surface climate variations over the North Atlantic Ocean during winter: 1900–1993. J. Climate, 6 , 17431753.

    • Search Google Scholar
    • Export Citation
  • Eden, C., , and T. Jung, 2001: North Atlantic interdecadal variability: Oceanic response to the North Atlantic Oscillation (1865–1997). J. Climate, 14 , 676691.

    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., , P. Muller, , and E. Zorita, 1997: A simple model of the decadal response of the ocean to stochastic wind forcing. J. Phys. Oceanogr., 27 , 15331546.

    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., , E. Kestenare, , N. Sennéchael, , G. de Coëtlogon, , and F. D’Andréa, 2001: On decadal-scale ocean–atmosphere interactions in the extended ECHAM1/LSG climate simulation. Climate Dyn., 16 , 333354.

    • Search Google Scholar
    • Export Citation
  • Goodman, J., , and J. Marshall, 1999: A model of decadal middle-latitude atmosphere–ocean coupled models. J. Climate, 12 , 621641.

  • Grötzner, A., , M. Latif, , and T. P. Barnett, 1998: A decadal climate cycle in the North Atlantic Ocean as simulated by the ECHO coupled GCM. J. Climate, 11 , 831847.

    • Search Google Scholar
    • Export Citation
  • Halliwell, G., 1998: Simulation of North Atlantic decadal/multidecadal winter SST anomalies driven by basin-scale atmospheric circulation anomalies. J. Phys. Oceanogr., 28 , 521.

    • Search Google Scholar
    • Export Citation
  • Hansen, D. V., , and H. F. Bezdek, 1996: On the nature of decadal anomalies in North Atlantic sea surface temperature. J. Geophys. Res., 101 , 87498758.

    • Search Google Scholar
    • Export Citation
  • Hoerling, M. P., , J. W. Hurrell, , and T. Xu, 2001: Tropical origins for recent North Atlantic climate change. Science, 292 , 9092.

  • Honda, M., , H. Nakamura, , J. Ukita, , L. Kouska, , and K. Takeuchi, 2001: Interannual seesaw between the Aleutian and Iceland Lows. Part I: Seasonal dependence and life cycle. J. Climate, 14 , 10291042.

    • Search Google Scholar
    • Export Citation
  • Hurrel, J. W., , G. Ottersen, , Y. Kushnir, , and M. Visbeck, 2003: An overview of the North Atlantic Oscillation. The North Atlantic Oscillation: Climatic Significance and Environment Impact, Geophys. Monogr., Vol. 134, Amer. Geophys. Union, 1–35.

    • Search Google Scholar
    • Export Citation
  • Jacob, R. L., 1997: Low frequency variability in a simulated atmosphere ocean system. Ph.D. thesis, University of Wisconsin—Madison, 155 pp.

  • Kelly, K. A., , and S. Dong, 2004: The relationship of western boundary current heat transport and storage to midlatitude ocean-atmosphere interaction. Earth Climate: The Ocean–Atmosphere Interaction, Geophys. Monogr., No. 137, Amer. Geophys. Union, 347–364.

    • Search Google Scholar
    • Export Citation
  • Kushnir, Y., 1994: Interdecadal variations in North Atlantic sea surface temperature and associated atmospheric conditions. J. Climate, 7 , 142157.

    • Search Google Scholar
    • Export Citation
  • Kushnir, Y., , W. A. Robinson, , I. Bladé, , N. M. J. Hall, , S. Peng, , and R. T. Sutton, 2002: Atmospheric GCM response to extratropical SST anomalies: Synthesis and evaluation. J. Climate, 15 , 22332256.

    • Search Google Scholar
    • Export Citation
  • Latif, M., , and T. P. Barnett, 1994: Causes of decadal climate variability in the North Pacific/North Atlantic sector. Science, 266 , 634637.

    • Search Google Scholar
    • Export Citation
  • Latif, M., , and T. P. Barnett, 1996: Decadal climate variability over the North Pacific and North America: Dynamics and predictability. J. Climate, 9 , 24072423.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., , and L. Wu, 2004: Atmospheric response to North Pacific SST anomaly: The role of ocean–atmosphere coupling. J. Climate, 17 , 18591882.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., , J. Kutzbach, , and L. Wu, 2000: Modeling climate shift of El Niño in the Holocene. Geophys. Res. Lett., 27 , 22652268.

  • Liu, Z., , B. Otto-Bliesner, , J. Kutzbach, , L. Li, , and C. Shields, 2003: Coupled climate simulation of the evolution of global monsoons in the Holocene. J. Climate, 16 , 24722490.

    • Search Google Scholar
    • Export Citation
  • Mantua, N. J., , S. R. Hare, , Y. Zhang, , J. Wallace, , and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78 , 10691079.

    • Search Google Scholar
    • Export Citation
  • Marshall, J., , H. Johnson, , and J. Goodman, 2001a: A study of the interaction of the North Atlantic Oscillation with the ocean circulation. J. Climate, 14 , 13991421.

    • Search Google Scholar
    • Export Citation
  • Marshall, J., and Coauthors, 2001b: Atlantic climate variability. Int. J. Climatol., 21 , 18631898.

  • Molinari, R. L., , D. A. Mayer, , J. F. Festa, , and H. F. Bezdek, 1997: Multiyear variability in the near-surface temperature structure of the midlatitude western North Atlantic Ocean. J. Geophys. Res, 102 , 32673278.

    • Search Google Scholar
    • Export Citation
  • Okumura, Y., , S. P. Xie, , A. Numaguti, , and Y. Tanimoto, 2001: Tropical Atlantic air–sea interaction and its influence on the NAO. Geophys. Res. Lett., 28 , 15071510.

    • Search Google Scholar
    • Export Citation
  • Peng, S., , A. Mysak, , H. Ritchie, , J. Derome, , and B. Dugas, 1995: The difference between early and middle winter atmospheric response to sea surface temperature anomalies in the northwestern Atlantic. J. Climate, 8 , 137157.

    • Search Google Scholar
    • Export Citation
  • Peng, S., , W. A. Robinson, , and S. Li, 2003: Mechanisms for the NAO responses to the North Atlantic SST tripole. J. Climate, 16 , 19872004.

    • Search Google Scholar
    • Export Citation
  • 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.

    • Search Google Scholar
    • Export Citation
  • Robertson, A. W., , C. R. Mechoso, , and Y. J. Kim, 2000: The influence of Atlantic sea surface temperature anomalies on the North Atlantic Oscillation. J. Climate, 13 , 122138.

    • Search Google Scholar
    • Export Citation
  • Saravanan, R., , and J. C. McWilliams, 1998: Advective ocean–atmosphere interaction: An analytical stochastic model with implications for decadal variability. J. Climate, 11 , 165188.

    • Search Google Scholar
    • Export Citation
  • Sutton, R. T., , and M. R. Allen, 1997: Decadal predictability of North Atlantic sea surface temperature and climate. Nature, 388 , 563567.

    • Search Google Scholar
    • Export Citation
  • Sutton, R. T., , and P. Mathieu, 2002: Response of the atmosphere–ocean mixed layer system to anomalous ocean heat flux convergence. Quart. J. Roy. Meteor. Soc., 128 , 12591275.

    • Search Google Scholar
    • Export Citation
  • Sutton, R. T., , W. A. Norton, , and S. P. Jewson, 2001: The North Atlantic Oscillation—What role for the Ocean? Atmos. Sci. Lett., doi:10.1006/asle.2000.0021.

  • Thompson, D. J., , and J. M. Wallace, 1998: The Arctic oscillation signature in wintertime geopotential height and temperature fields. Geophys. Res. Lett., 25 , 12971300.

    • Search Google Scholar
    • Export Citation
  • Timmerman, A., , M. Latif, , R. Voss, , and A. Grötzner, 1998: Northern Hemispheric interdecadal variability: A coupled air–sea mode. J. Climate, 11 , 19061931.

    • Search Google Scholar
    • Export Citation
  • Toth, Z., , and E. Kalnay, 1997: Ensemble forecasting at NMC and the breeding method. Mon. Wea. Rev., 125 , 32973319.

  • Venzke, S., , M. R. Allen, , R. T. Sutton, , and D. P. Rowell, 1999: The atmospheric response over the North Atlantic to decadal changes in sea surface temperatures. J. Climate, 12 , 25622584.

    • Search Google Scholar
    • Export Citation
  • Visbeck, M., , H. Cullen, , G. Krahman, , and N. Naik, 1998: An ocean model’s response to North Atlantic Oscillation–like wind forcing. Geophys. Res. Lett., 25 , 45214524.

    • Search Google Scholar
    • Export Citation
  • Visbeck, M., , E. Chassignet, , R. Curry, , T. Delworth, , B. Dickson, , and G. Krahmann, 2003: The ocean’s response to NAO variability. The North Atlantic Oscillation: Climatic Significance and Environment Impact, Geophys. Monogr., Vol. 134, Amer. Geophys. Union.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., , and M. Kimoto, 1999: Tropical-extratropical connection in the Atlantic atmosphere-ocean variability. Geophys. Res. Lett., 26 , 22472250.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., , and M. Kimoto, 2000: Ocean atmosphere thermal coupling in the North Atlantic: A positive feedback. Quart. J. Roy. Meteor. Soc., 126 , 33433369.

    • Search Google Scholar
    • Export Citation
  • Weaver, A. J., , E. S. Sarachik, , and J. Marotzke, 1991: Freshwater flux forcing of decadal and interdecadal oceanic variability. Nature, 353 , 836838.

    • Search Google Scholar
    • Export Citation
  • Wu, L., , and Z. Liu, 2002: Is tropical Atlantic variability driven by the North Atlantic Oscillation? Geophys. Res. Lett., 29 .1653, doi:10.1029/2002GL014939.

    • Search Google Scholar
    • Export Citation
  • Wu, L., , and Z. Liu, 2003: Decadal variability in the North Pacific: The eastern North Pacific mode. J. Climate, 16 , 31113131.

  • Wu, L., , Z. Liu, , R. Gallimore, , R. Jacob, , D. Lee, , and Y. Zhong, 2003: Pacific decadal variability: The tropical Pacific mode and the North Pacific mode. J. Climate, 16 , 11011120.

    • Search Google Scholar
    • Export Citation
  • Yulaeva, E., , N. Schneider, , D. Pierce, , and T. Barnett, 2001: Modeling of North Pacific climate variability forced by oceanic heat flux anomalies. J. Climate, 14 , 40274066.

    • Search Google Scholar
    • Export Citation
  • Zorita, E., , and C. Frankignoul, 1997: Modes of North Atlantic decadal variability in the ECHAM1/LSG coupled atmosphere–ocean general circulation model. J. Climate, 10 , 183200.

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

Leading EOFs of the observed North Atlantic (north of 20°N) winter (DJF) SST anomalies and associated atmospheric 500-mb geopotential height: (a) EOF1 (NAM) and (b) EOF2 (NAT) and regression of 500-mb geopotential height on the (c) NAT and (d) NAM indices. Units for EOF and regression are K and m (per standard deviation of the index), respectively. The magnitude of each EOF is reflected in its pattern.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 2.
Fig. 2.

Power spectrum of the (a) NAM and (b) NAT indices in the observations. The power spectrum is calculated using a multitaper method (three tapers).

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 3.
Fig. 3.

As in Fig. 1 but for the FOAM model control simulation.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 4.
Fig. 4.

Power spectrum of (a) NAT, (b) NAM, (c) the subtropical high, and (d) the Icelandic low indices in the FOAM control simulation. The subtropical height and Icelandic low indices are defined as the area-averaged winter 500-mb geopotential height over the region (20°– 45°N, 70°W–20°E) and (50°– 75°N, 90°W– 20°E) respectively.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 5.
Fig. 5.

Leading EOFs of the winter SST anomalies in PC-G, where ocean–atmosphere coupling is deactivated over the global oceans using the partial coupling scheme: (a) EOF1 (NAT) and (b) EOF2 (NAM).

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 6.
Fig. 6.

Power spectrum of (a) the NAT and (b) the NAM indices in PC-G, power spectrum of (c) the STH and (d) the IL in PC-G, and power change of (e) the NAT and the NAM and (f) the STH and the IL. In (e) each index is multiplied by the square root of the area-integrated variance before calculating the power spectrum to reflect the amplitude of each mode in each experiment. The power change at a specific frequency is calculated as the difference of power between the control (FC) and the partial coupling (PC-G) experiments divided by the power of the power in PC-G.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 7.
Fig. 7.

Regression of (a) SST, (b) anomalous and (c) mean zonal advection, (d) anomalous and (e) mean meridional advection, (f) anomalous and (g) mean vertical advection, (h) VDC, (i) surface heat flux (positive downward), and (j) wind stress and its curl on the NAT index. All data are bandpassed to retain the variability between 6 and 30 yr. Units for SST, heat balance terms, and wind stress and its curl are °C, W m−2, N m−2, and N m−3 (× 10−9) per standard deviation of the index, respectively.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 8.
Fig. 8.

Lagged regression of SST and upper-400-m heat content on the NAT index. All data are bandpassed to retain variability between 6 and 30 yr. Units for SST and HC are °C and m °C−1 per standard deviation of the tripole index, respectively. Contour interval is 0.08 and 4 for SST and heat content, respectively.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 9.
Fig. 9.

Lagged regression of the anomalous meridional advection (υTy), surface heat flux (HF; positive downward), and wind stress curl on the NAT index. All data are bandpassed to retain the variability between 6 and 30 yr. Units are W m−2 for υTy and HF, N m−3 for the wind stress curl. Contour interval is 1 for υTy. Contour levels for HF are (−20, −10, −5, −3, −2, −1, 1, 2, 3, 5, 10, 20), and for the wind stress curl are (−30, −20, −10, −6, −4, −2, 2, 4, 6, 10, 20, 30) × 10−9, respectively.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 10.
Fig. 10.

(a) DJF mean ocean current velocity averaged over upper 200-m (vectors), climatologic zero wind stress curl line (dark lines), and instantaneous regression of anomalous wind stress curl on the NAT index (contours) [all data are bandpass filtered (6–30 yr)]. Units for current and wind stress curl are cm s−1 and N m−2. Contour interval for the wind stress curl is 4. The Hovmoeller diagram of (b) the zonally averaged (70°–20°W) regression of SST and (c) the upper-400-m heat content on the NAT index, respectively. Units for SST and HC are °C and m °C−1. Contour interval is 0.08 and 4 for SST and heat content, respectively.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 11.
Fig. 11.

(a) First EOF of the North Atlantic winter SST anomalies in the TC experiment. (b) Power spectrum of the normalized time series of EOF1. In the TC experiment, the wind stress for the ocean in the coupled model is constrained to the model climatological wind stress at each time step through interpolation of monthly climatological data.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 12.
Fig. 12.

Ensemble mean atmospheric response (Jan–Feb) to the tripole mixed layer anomaly initiated in earlier winter (1 Nov): (a) 500-mb geopotential height, (b) wind stress (vectors) and curl (contours), and (c) SST averaged in Jan and Feb. Units for SST, wind stress (and curl), and geopotential height are °C, N m−2, and m. The contour interval for the wind stress curl is 2 × 10−9.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 13.
Fig. 13.

Coherence (solid line) and phase (dotted line) spectrum between (a) the NAT and the PDO and (b) the NAM and the PDO indices in the control simulation. The dashed lines represent the 95% confidence level.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

Fig. 14.
Fig. 14.

Power spectrum of (a) the NAT and (b) the NAM indices in PC-PAF and power change of (c) the NAT and the NAM, and (d) the STH and the IL. In (c) each index is multiplied by the square root of the area-integrated variance before calculating the power spectrum to reflect the amplitude of each mode in each experiment. The power change at a specific frequency is calculated as the difference of power between the control (FC) and the partial coupling (PC-PAF) experiments divided by the power of the power in PC-PAF.

Citation: Journal of Climate 18, 2; 10.1175/JCLI-3264.1

* Center for Climatic Research Contribution Number 844.

Save