395
3.2 SST Anomalies in the Northwest Atlantic
Coupled climate models are also beginning to exhibit natural variability.
Delworth et at. (1993) integrate the GFDL climate model for 600 years and
find natural variability whose spectral properties are remarkably similar to
observations (Fig. 5a). Superimposed on a red-noise spectrum a number of
spectral peaks are visible. Interdecadal oscillations of 40-60 yr are evident
in the maximum meridional overturning in the North Atlantic reaching
amplitudes of about 2 Sv (Fig. 5b). When the thermohaline circulation is
weak, decreased advection of lower-latitude warm and saline waters into
the central regions of the North Atlantic generates a pool of anomalously
cold and fresh water. The thermal anomaly dominates and hence generates a geostrophically controlled cyclonic circulation at the surface (baroclinic vortex). The western half of this anomalous circulation enhances the
mean northward current of warm and saline waters located in the center
of the Atlantic which is part of the large-scale conveyor belt circulation.
The strengthened conveyor then carries more saline and warm low-latitude
waters into this region. Again, the thermal contribution is stronger and
creates an anomalous warm pool which is associated with anti-cyclonic circulation. The latter weakens the mean flow again, and the cycle begins
anew.
The oscillation is distinctly irregular, a common feature of nonlinear dynamical systems. Although the first 200 years of the integration show a
quasi-periodic cycle of 40-50 years, the periods are longer in the following
400 years. This is indicative of some preconditioning of the ocean independent of the feedback mechanism. One plausible possibility is the volume,
i.e. the heat and salt content, of the anomalous pool. The bigger the pool,
the longer it takes a certain mass transport anomaly to erode the SST
anomaly. While the amplitude of the anomaly governs the strength of this
anomalous mass flux via the pressure gradients, the spatial extent itself
does not influence it but determines the time scale on which the anomaly
can be removed. For an estimate of orders of magnitude we assume a typical extent of this pool (65°W-45°W, 35°N-50oN, (Delworth et at. 1993))
and a depth ofthe anomaly of roughly 300m (Greatbatch and Zhang 1995).
The corresponding volume of 8.2· 10 14 m 3 is renewed once in 26 years by a
flux anomaly of 1 Sv. Changing pool sizes are thus likely to be responsible for the changes of the period lengths during different segments of the
600-year run of Delworth et at. (1993). It is intriguing that a similar in-
3.2 SST Anomalies in the Northwest Atlantic
Coupled climate models are also beginning to exhibit natural variability.
Delworth et at. (1993) integrate the GFDL climate model for 600 years and
find natural variability whose spectral properties are remarkably similar to
observations (Fig. 5a). Superimposed on a red-noise spectrum a number of
spectral peaks are visible. Interdecadal oscillations of 40-60 yr are evident
in the maximum meridional overturning in the North Atlantic reaching
amplitudes of about 2 Sv (Fig. 5b). When the thermohaline circulation is
weak, decreased advection of lower-latitude warm and saline waters into
the central regions of the North Atlantic generates a pool of anomalously
cold and fresh water. The thermal anomaly dominates and hence generates a geostrophically controlled cyclonic circulation at the surface (baroclinic vortex). The western half of this anomalous circulation enhances the
mean northward current of warm and saline waters located in the center
of the Atlantic which is part of the large-scale conveyor belt circulation.
The strengthened conveyor then carries more saline and warm low-latitude
waters into this region. Again, the thermal contribution is stronger and
creates an anomalous warm pool which is associated with anti-cyclonic circulation. The latter weakens the mean flow again, and the cycle begins
anew.
The oscillation is distinctly irregular, a common feature of nonlinear dynamical systems. Although the first 200 years of the integration show a
quasi-periodic cycle of 40-50 years, the periods are longer in the following
400 years. This is indicative of some preconditioning of the ocean independent of the feedback mechanism. One plausible possibility is the volume,
i.e. the heat and salt content, of the anomalous pool. The bigger the pool,
the longer it takes a certain mass transport anomaly to erode the SST
anomaly. While the amplitude of the anomaly governs the strength of this
anomalous mass flux via the pressure gradients, the spatial extent itself
does not influence it but determines the time scale on which the anomaly
can be removed. For an estimate of orders of magnitude we assume a typical extent of this pool (65°W-45°W, 35°N-50oN, (Delworth et at. 1993))
and a depth ofthe anomaly of roughly 300m (Greatbatch and Zhang 1995).
The corresponding volume of 8.2· 10 14 m 3 is renewed once in 26 years by a
flux anomaly of 1 Sv. Changing pool sizes are thus likely to be responsible for the changes of the period lengths during different segments of the
600-year run of Delworth et at. (1993). It is intriguing that a similar in-