Validation of visible/near‐IR atmospheric absorption and solar emission spectroscopic models at 1 cm−1 resolution

Abstract

A Fourier transform infrared (FTIR) spectrometer, operating at 1 cm−1 resolution between 9000 and 24,669 cm−1 (0.405–1.111 μm) has been used to check the spectral composition of databases that form the basis for most atmospheric absorption parameterizations used in climate models, remote sensing, and other radiative transfer simulations. The spectrometer, operating near sea level under clear skies, obtained relative atmospheric transmission measurements of the direct solar beam by means of a heliostat. The spectroscopic data were compared with a line‐by‐line radiative transfer model (LBLRTM) calculation of direct solar beam flux, which used a input data a monochromatic model extraterrestrial solar flux spectrum currently in common use. This intercomparison revealed that the extraterrestrial solar flux spectrum contains 266 solar absorption features that do not appear in the data, resulting in an excess of approximately 1.92 W m−2 in the model's solar constant. The intercomparison also revealed 97 absorption features in the data that do not appear in the HITRAN‐96 database as used by LBLRTM, resulting in a model underestimate of shortwave absorption of ∼0.23 W m−2 for a solar zenith angle of 42°. These small discrepancies revealed by the intercomparison indicate that current extraterrestrial solar irradiance models and spectroscopic databases used by shortwave atmospheric radiative transfer models are nearly entirely complete for purposes of atmospheric energy budget calculation. Thus clear or cloudy sky ‘excess absorption’ is unlikely to be related to an incomplete identification of atmospheric absorbing gases and their spectroscopic features, at 1 cm−1 resolution, for a clean troposphere of normal composition.