CO2 Spectroscopy Evaluation Using Atmospheric Solar Absorption Spectra

We evaluated the improvements in successive versions (1996 - 2004) of HITRAN (1) and other molecular line parameter data set (2) to correctly simulate infrared (IR) and near-infrared (NIR) CO 2 transmittance spectra. Understanding the global sources and sinks of CO 2 requires highly accurate measurements (ó 0.3%) and makes extreme de- mands on the spectroscopy. We evaluated the line parameter data sets by fitting solar absorption spectra measured by the JPL MkIV FTIR spectrometer (3) and the Caltech FT spectrometer (4). The JPL spectra cover the entire 650 - 5650 cm-1 spectral region simultaneously and were measured during recent balloon ìights under conditions that were closely monitored by in situ sensors(5). The Caltech spectra, measured during sum- mer 2004 from Park Falls, WI, cover the entire 4,000 - 11,000 cm-1 spectral region simultaneously. The measurements were made under conditions that were also closely monitored by in situ sensors (continuous CO 2 monitors, weekly ìask sampling and air- craft profiles).(6) This work benefits the HITRAN user community by documenting and highlighting inadequacies in the data set and quantifying their likely e’ect on retrieved CO 2 col- umn. The consistency between retrieved column-averaged mole fraction of CO 2 ( XCO 2 ) and measured surface CO 2 is an important demonstration of the adequacy of the spec- troscopy. The work provides an objective basis for assessing the adequacy and self- consistency of the existing HITRAN data set and for quantifying the improvements (or otherwise) produced by new CO 2 line parameters. It also helps to prioritize needs for additional laboratory studies.(7