The Seasonal Cycle of Interannual Variability and the Dynamical Imprint of the Seasonally Varying Mean State

Abstract

Various aspects of the seasonal cycle of interannual variability of the observed 300-hPa streamfunction are documented and related to dynamical influences of the seasonality of the mean circulation. The stochastically excited nondivergent barotropic vorticity equation linearized about upper-tropospheric climatological mean states from each month of the year is used to identify characteristics of interannual variability that the seasonal cycle of the mean state should modulate. The result is interannual variability with (a) extratropical centers of variance that are much stronger in winter than summer and that are confined to midlatitudes during the warm season, (b) an annual cycle of preferred scales in midlatitudes with largest scales occurring during winter and a semiannual cycle of scales in the subtropics, and (c) streamfunction tendencies from interannual fluxes that adjust to the seasonally varying climatological eddies in such a way as to damp them. Because these same properties are also shown to exist in nature, it is concluded that the linear framework is a useful means of understanding the seasonality of interannual disturbances and that seasonality of the mean state leaves a pronounced imprint on interannual variability. Analysis of an ensemble of general circulation model integrations indicates the signatures of seasonality produced in the stochastically driven linear framework are more useful for understanding intrinsic interannual variability than variability caused by seasonally varying sea surface temperature anomalies. Furthermore, it is found that the intrinsic variability of the GCM has properties very much like those in nature, another indication that organization resulting from anomalous forcing structure is not required for production of many aspects of the observed seasonality of interannual variability.