A DIAGNOSTIC AND MODELING STUDY OF THE MONTHLY MEAN WINTERTIME ANOMALIES APPEARING IN A 100-YEAR GCM EXPERIMENT

The nature of simulated atmospheric variability on monthly time scales has been investigated by analyzing the output from a 100-year integra tion of a spectral GCM with rhomboidal wavenumber 15 truncation. In th is experiment, the seasonally varying, climatological sea surface temp erature was prescribed throughout the world oceans. The principal mode s of variability in the model experiment were identified by applying a rotated empirical orthogonal function (EOF) analysis to the Northern Hemisphere monthly averaged 515-mb geopotential height for the winter season (November through March). The individual leading spatial modes are similar to the observed north-south dipoles over the North Atlanti c and North Pacific, as well as wavelike patterns in the Pacific/North American and Northern Asian sectors. Quasigeostrophic geopotential te ndencies forced by synoptic-scale (2.5-6 day) eddy vorticity and heat fluxes were computed for those months when the individual EOF modes ar e particularly active. The composite patterns of the eddy-induced tend encies were compared with the corresponding monthly mean anomaly patte rns. It is seen that the forcing due to eddy vorticity transports exhi bits a distinctive barotropic character, and reinforces the monthly av eraged geopotential height anomalies throughout the tropospheric colum n. On the other hand, the eddy heat fluxes lead to dissipation of the monthly mean height anomalies in the upper troposphere, and enhancemen t of the height anomalies in the lower troposphere. Hence, the eddy he at fluxes exert a strong impact on the baroclinic component of the cir culation by destroying the concurrent local monthly mean temperature a nd geopotential thickness anomalies. The above relationships based on model data are in agreement with the corresponding observational resul ts. A linear stationary wave model was then used to mimic the individu al EOF modes appearing in the GCM experiment, and to diagnose the rela tive importance of different types of forcing in the generation of suc h modes. As suggested by the tendency calculations, the transient eddy forcing due to heat fluxes was parameterized as a thermal diffusion m echanism in the stationary wave model. When the model was linearized a bout the climatological zonally averaged basic state, it failed to rep roduce the EOF patterns appearing in the GCM experiment. However, when the same model was linearized about the zonally varying GCM climatolo gy, the response to the total forcing (which includes vorticity fluxes by eddies on submonthly time scales, diabatic heating, and nonlineari ty in those months when the individual EOF modes are active) bears a c onsiderable resemblance to the corresponding anomaly patterns in the G CM. By evaluating the individual contributions of each of the three fo rcing mechanisms to the total linear model solution, it is concluded t hat the transient eddy vorticity fluxes exert the strongest influences . The response to nonlinear effects is negligible, while the forcing d ue to diabatic heating is weak and acts in opposition to the anomaly p atterns in the upper troposphere. The forcing associated with vorticit y fluxes by synoptic-scale transient eddies accounts for approximately half of the total vorticity forcing due to all submonthly fluctuation s. Both the tendency calculations and the stationary wave model result s indicate the crucial role of vorticity transports by transient eddie s. The linear model solutions also illustrate the importance of incorp orating the climatological stationary waves in the basic state. These findings hence suggest that the' monthly mean anomalies in this GCM ex periment are intimately linked to barotropic interactions between tran sient fluctuations of different time scales, and between the monthly m ean anomalies and the climatological stationary waves.