Midlatitude Error Growth in Atmospheric GCMs: The Role of Eddy Growth Rate

Abstract

Several studies have established that atmospheric flows have a finite range of predictability, which may be reasonably considered a consequence of the underlying dynamics. In the midlatitudes, error growth is predominantly associated with baroclinic disturbances. We consider midlatitude error growth in two models: an idealized dry dynamical core and a comprehensive atmospheric general circulation model (GCM). By systematically varying equator to pole temperature gradients in the dynamical core, we show that with increasing Eady growth rates, the time elapsed before errors saturate decreases, shortening the window in which weather predictions may be useful. We also consider the limits of midlatitude predictability in the comprehensive moist GCM in a range of climates. Our results show that the times to error saturation are shorter in warmer climates than colder climates, suggesting that warmer climates are inherently less predictable.