On North Atlantic lntedecadal Variability: A Stochastic View
177
thousand-year integrations: A fully coupled run with an ocean-atmosphere general
circulation model, an atmosphere-slab run, and an atmosphere-only run with fixed
climatogical SSTs. aur results presented in this paper are mostly consistent with
the results of Manabe et al. (1996). In particular, Manabe et al. (1996) conclude
that the stochastic forcing plays a major role in generating low-frequency variability.
Moreover, it would be desirable to understand the relative roles of changes in the
wind-driven and thermohaline circulations in changing the North Atlantic SST.
Sensitivity experiments with our coupled ocean-atmosphere model would help to
get further insight into this matter. One could, for instance, inhibit certain atmospheric feedbacks in coupled integrations, such as the fresh water feedback or the
wind stress feedback and compare the level of SST variability with that in the control run. Such experiments will provide important informations about the mechanisms that lead to the decadal and interdecadal variations in the SST. These
experiments are underway, and we shall report on the their outcomes in forthcoming papers.
Finally, model verification is stiH an important issue. We have used so far relatively coarse-resolution models to investigate the dynamics and predictability of
decadal and interdecadal variability. Such models have considerable problems in
simulating important aspects of the global climate and its variability. Relatively
large flux corrections, are applied to inhibit the coupled models to drift into unrealistic climates. It is unknown how these corrections affect the variability of complex
non-linear systems such as the climate system. Furthermore, we have no good estimate of the space-time structure of the low-frequency variability from observations. Paleoclimatic datasets wiH become more and more important to validate the
coupled models. An example of such a dataset, for instance, is the multi-century
reconstruction of the North Atlantic OsciHation from tree rings (Cook et al.
(1997)).
Acknowlegements
The authors would like to thank Drs. G. Lohmann, E. Maier-Reimer, and U.
Mikolajewicz for many fruitful discussions. This work was supported by the German govemment by its "Decadal Predictability" and "Ocean-CLIVAR" programmes and the European Union through its "DICE" and "SINTEX" projects. The
model runs were conducted at the Deutsches Klimarechenzentrum in Hamburg.
177
thousand-year integrations: A fully coupled run with an ocean-atmosphere general
circulation model, an atmosphere-slab run, and an atmosphere-only run with fixed
climatogical SSTs. aur results presented in this paper are mostly consistent with
the results of Manabe et al. (1996). In particular, Manabe et al. (1996) conclude
that the stochastic forcing plays a major role in generating low-frequency variability.
Moreover, it would be desirable to understand the relative roles of changes in the
wind-driven and thermohaline circulations in changing the North Atlantic SST.
Sensitivity experiments with our coupled ocean-atmosphere model would help to
get further insight into this matter. One could, for instance, inhibit certain atmospheric feedbacks in coupled integrations, such as the fresh water feedback or the
wind stress feedback and compare the level of SST variability with that in the control run. Such experiments will provide important informations about the mechanisms that lead to the decadal and interdecadal variations in the SST. These
experiments are underway, and we shall report on the their outcomes in forthcoming papers.
Finally, model verification is stiH an important issue. We have used so far relatively coarse-resolution models to investigate the dynamics and predictability of
decadal and interdecadal variability. Such models have considerable problems in
simulating important aspects of the global climate and its variability. Relatively
large flux corrections, are applied to inhibit the coupled models to drift into unrealistic climates. It is unknown how these corrections affect the variability of complex
non-linear systems such as the climate system. Furthermore, we have no good estimate of the space-time structure of the low-frequency variability from observations. Paleoclimatic datasets wiH become more and more important to validate the
coupled models. An example of such a dataset, for instance, is the multi-century
reconstruction of the North Atlantic OsciHation from tree rings (Cook et al.
(1997)).
Acknowlegements
The authors would like to thank Drs. G. Lohmann, E. Maier-Reimer, and U.
Mikolajewicz for many fruitful discussions. This work was supported by the German govemment by its "Decadal Predictability" and "Ocean-CLIVAR" programmes and the European Union through its "DICE" and "SINTEX" projects. The
model runs were conducted at the Deutsches Klimarechenzentrum in Hamburg.