Testing Climate Model Fidelity Using Lag Autocovariance of Infrared Radiances

A number of studies have shown that climate model sensitivity to changes in radiative forcing can in some cases be diagnosed using statistics based on lag autocorrelations of model variables. This offers the promise of ranking models by their fidelity to real world data using a metric directly relevant to the models' ability to predict climate sensitivty. Here we examine the ability of these statistics to determine model climate sensitivity using infrared brightness temperature at a number of frequencies, and examine how the error properties of realistic simulated satellite data sets propagate to the error in determining climate sensitivity and climate model fidelity.