Stochastic Forcing and Prediction of Low-Frequency Planetary Scale Flow

Abstract

A dynamical model incorporating observed field data is used to estimate the potential importance of linear and nonlinear vorticity advection to climate forecast models. Forecasts of 30-day averages benefit from inclusion of the linear advection term, but the nonlinear advection appears only marginally helpful. For intermediate averaging times (e.g., 10 days), both advection terms appear to be important. Analysis of the nonlinear terms suggests that they could be most adequately parameterized as a noise process that is ?white? in wavenumber space and ?red? in the time domain.