Methods for estimating climate anomaly forcing patterns

Inverse methods for determining the anomalous mean forcing functions responsible for climate change are investigated. First, an iterative method is considered, and it is shown to successfully reproduce forcing functions for various idealized and observed climate states using quasigeostrophic simulations. Second, a new inverse method that is more computationally efficient is presented. This method closes the mean-field equations by representing the second-order statistical moments, the transient eddy heat and momentum (or potential vorticity) fluxes, as linear functions of the mean field. The coefficients of the linear parameterization are determined by least squares regression. It is shown that the new method also successfully reproduces the anomalous forcing functions responsible for climatic changes in quasigeostrophic simulations.