Accelerated Iterative Method for Solving Steady Problems of Linearized Atmospheric Models

A new approach, referred to as the accelerated iterative method (AIM), is developed for obtaining steady atmospheric responses with a zonally varying basic state. The linear dynamical operator is divided into two parts, one associated with the zonally symmetric component and the other with the asymmetric component of the basic state. To ensure an accelerated convergence of the iteration to the true solution, the two parts of the operator are modified by adding and subtracting an identical “accelerating” operator. AIM is shown to be an efficient scheme well suited for computing the higher-resolution, steady atmospheric response of barotropic and more so of baroclinic numerical models linearized about a zonally varying basic state. A preliminary application of AIM to the T42 baroclinic model linearized about the observed winter (December–February) climatology is presented. A series of steady responses forced by the diabatic heating and transient eddy forcing, both estimated from reanalysis data for individual winters during 1960–2002, captures a certain part of the observed interannual variability associated with dominant teleconnection patterns, such as the North Atlantic Oscillation and the Pacific–North American pattern. Thus, AIM should be a useful tool for the diagnostic studies of the low-frequency variability of the atmosphere.