Dynamics, Statistics and Predictability of Planetary Flow Regimes

Abstract

We consider regimes of low-frequency variability in large-scale atmospheric dynamics. The model used for the study of these regimes is the fully-nonlinear, equivalent-barotropic vorticity equation on the sphere, with simplified forcing, dissipation and topography. It is found that certain limited regions in the system’s phase space are visited recurrently and for extended periods by model solutions. Flow patterns associated in physical space with these regions correspond to synoptically defined zonal and blocked Northern Hemisphere midlatitude flows.

Based on recent ideas from dynamical system theory, it is shown that the system’s macrodynamics can be described by two or more planetary flow regimes, the expected residence time in each regime, and the transition probabilities from one regime to another. Connections are made with the classical concepts of statistical-synoptic long-range forecasting (LRF). In particular, transitions between regimes in the model’s nonlinear, deterministic dynamics turn out to occur along the directions of maximum variance, i.e., along the first few principal components.

These model-derived ideas are further tested by applying them to a time series of atmospheric data from the Southern Hemisphere. The effect of anomalous surface conditions on planetary flow regimes and their predictability is discussed and perspectives for practical LRF are foreshadowed.