Examine Climate Models by Using Infrared Spectrum

We examine global climate models by comparing the satellite-observed high resolution global infrared spectra with the model-simulated counterpart. Because the topof-the-atmosphere outgoing Earth thermal emission at different frequencies is sensitive to different geophysical variables (temperature, water vapor and other greenhouse gas concentrations, clouds, etc.) at various levels, a comparison of observed and simulated spectra is as challenging as examining a variety of model-simulated geophysical variables simultaneously. Thus, the ability of climate models to replicate the observed spectra provides a strict metric for model validation. This approach adopts a forward radiative transfer calculation and thus avoids the uncertainties associated with retrieving geophysical variables from the radiance measurement. After ensuring consistency in the sampling of the observed and simulated spectra and a proper representation of cloud vertical distribution, the infrared spectra simulated from a state-of-the-art Geophysical Fluid Dynamics Lab climate model are shown to yield a good agreement with those measured by the Atmospheric Infrared Sounder onboard Aqua satellite, in terms of global and regional means. The radiance discrepancies in the water vapor and carbon dioxide _2 bands are consistent with the model biases in atmospheric water vapor and temperature. Furthermore, an examination of the seasonal covariations of the infrared spectra and surface temperatures demonstrates the need and a method to use spectrally resolved radiance to understand and constrain model sensitivity. This study shows that infrared spectrum forms an advantageous means for climate model validation. The study also provides an example of the climatic applications using spectroscopic observations and the molecular absorption database.