North-Atlantic decadal variability in a coupled atmosphere/ocean/sea-ice model of moderate complexity

Decadal climate variability is studied within the context of a coupled atmosphere/ocean/sea-ice model of moderate complexity (ECBILT). The atmospheric model is a T21 Quasi-Geostrophic three level model with parameterisations for the diabatic physical processes. The ocean model is more or less conventional except that it has a coarse grid, a flat bottom and the winddriven gyre circulation is assumed to be in equilibrium with the time mean windforcing. The sea-ice model is a zero-layer thermodynamic model. The model is computationally efficient. It uses 0.2 hr CPU time for a one year integration on a workstation (SGI Power Indigo).
Short descriptions of the climate of the model atmosphere and ocean are presented when integrated separately, using prescribed, climatological boundary conditions. The climate of the models and their variability is reasonably realistic, although there are significant differences with the observed climate.
After having reached a statistical equilibrium in coupled mode without applying flux corrections, we performed a subsequent thousand year integration. Compared to the current climate, the surface temperatures are two degrees warmer in the tropics to almost eight degrees in the polar regions.
We have analysed the variability properties of the coupled model in the Atlantic European area by computing EOFs. Both the EOFs for the atmosphere and the ocean show good qualitative agreement with observed variability patterns in the extratropics. In the model atmosphere the dominant modes of variability can be associated with the North Atlantic Oscillation and with the East Atlantic pattern.
We have explored the covariability between the atmosphere and ocean by performing a singular value decomposition of SST anomalies and 800 hPa geopotential height anomalies. The first SVD mode shows a red variance spectrum in SST and a white spectrum in 800 hPa height. The second mode shows a peak in both spectra at a timescale of about 18 years. The time scale of this mode is set by the ocean but the physical mechanisms that are operating are not yet unambiguously identified. In our analysis we have concentrated on the second SVD mode. The correlation coefficient between the two timeseries of this mode is 0.7. It shows very good correspondence with the coupled mode found by Grotzner et al (1996) in the ECHO coupled integration.
Additional experiments show that it is essential for the 18 year oscillation to appear that the atmosphere is allowed to respond to SST anomalies, despite the fact that this response is small. The atmosphere responds in two ways to the SST anomalies. First the surface air temperature adjusts to the SST values. Second, the statistics of occurrence of the atmospheric circulation regimes associated with the NAO anomaly pattern changes slightly. The anomalous surface temperature and wind response act to reduce the damping of the SST anomaly. The negative feedback of anomalous surface wind stresses appears to be small.
A decadal mode of variability at a time scale of about 18 years has also been identified in the North-Pacific. There is some indication of a phase relationship with the Atlantic mode. This may be an indication that the dynamical response over the North-Atlantic is influenced by events in the North-Pacific. The mechanisms for this coupling are not yet understood.