278
to obtain more insight into the oceanic adjustment, we conducted a series
of" ocean-only" integrations in which we forced a coarse-resolution version
(3.5 0 x 3.5°) of our oceanic general circulation model (OGCM) HOPE by
prescribed low-frequency periodic wind stress variations. The spatial wind
stress pattern by which we forced our OGCM was derived from the standalone integrations with our atmosphere model (section 3.3) and is shown
in Fig. 5, and the time evolution was assumed to be sinusoidal.
We discuss here two cases in which the forcing period amounts to five
years and twenty years, respectively. As can be seen clearly in Fig. 6
showing the time evolution of the sea level anomalies along two particular
latitudes bands, the ocean is not in equilibrium with the wind stress forcing
at a period of five years. Consistent with planetary wave propagation, slow
westward phase propagation can be seen between 20° Nand 30° N (Fig.
6a), while the response further to the north is more stationary in character
(Fig. 6b). Similar results are obtained for a forcing period of ten years
(not shown). The ocean approaches an equilibrium response for forcing
periods longer than about twenty years. This is demonstrated in Fig. 7
which shows the case for a forcing period of exactly twenty years. No phase
propagation is seen at this period, neither in the subtropics (Fig. 7a) nor in
the mid-latitudes (Fig. 7b). Thus, our ocean-only experiments with lowfrequency periodic wind stress forcing support fully the picture that the
ocean has a sufficiently long memory to enable decadal oscillations. The
precise period of the decadal oscillations, however, will not only depend
on the wave transit times but also on the nature of the ocean-atmosphere
interactions.
4 Simple coupled model
In summary, both our CGCM results and the uncoupled atmospheric and
oceanic experiments support the picture that the decadal mode arises from
unstable ocean-atmosphere interactions over the North Pacific, so that the
mode is best described as an inherently coupled air-sea mode. The memory
of the coupled system resides in the ocean, while the atmosphere responds
passively to the slowly varying boundary conditions. It is interesting to
note that this paradigm for the generation of the decadal variability over
the North Pacific is similar to that for the ENSO phenomenon in the tropical Pacific (e.g. Schopf and Suarez (1988), Neelin et al. (1994)).
It is therefore tempting to construct a simple ocean-atmosphere model
to obtain more insight into the oceanic adjustment, we conducted a series
of" ocean-only" integrations in which we forced a coarse-resolution version
(3.5 0 x 3.5°) of our oceanic general circulation model (OGCM) HOPE by
prescribed low-frequency periodic wind stress variations. The spatial wind
stress pattern by which we forced our OGCM was derived from the standalone integrations with our atmosphere model (section 3.3) and is shown
in Fig. 5, and the time evolution was assumed to be sinusoidal.
We discuss here two cases in which the forcing period amounts to five
years and twenty years, respectively. As can be seen clearly in Fig. 6
showing the time evolution of the sea level anomalies along two particular
latitudes bands, the ocean is not in equilibrium with the wind stress forcing
at a period of five years. Consistent with planetary wave propagation, slow
westward phase propagation can be seen between 20° Nand 30° N (Fig.
6a), while the response further to the north is more stationary in character
(Fig. 6b). Similar results are obtained for a forcing period of ten years
(not shown). The ocean approaches an equilibrium response for forcing
periods longer than about twenty years. This is demonstrated in Fig. 7
which shows the case for a forcing period of exactly twenty years. No phase
propagation is seen at this period, neither in the subtropics (Fig. 7a) nor in
the mid-latitudes (Fig. 7b). Thus, our ocean-only experiments with lowfrequency periodic wind stress forcing support fully the picture that the
ocean has a sufficiently long memory to enable decadal oscillations. The
precise period of the decadal oscillations, however, will not only depend
on the wave transit times but also on the nature of the ocean-atmosphere
interactions.
4 Simple coupled model
In summary, both our CGCM results and the uncoupled atmospheric and
oceanic experiments support the picture that the decadal mode arises from
unstable ocean-atmosphere interactions over the North Pacific, so that the
mode is best described as an inherently coupled air-sea mode. The memory
of the coupled system resides in the ocean, while the atmosphere responds
passively to the slowly varying boundary conditions. It is interesting to
note that this paradigm for the generation of the decadal variability over
the North Pacific is similar to that for the ENSO phenomenon in the tropical Pacific (e.g. Schopf and Suarez (1988), Neelin et al. (1994)).
It is therefore tempting to construct a simple ocean-atmosphere model