Faculty of Science, Medicine and Health - Papers: part A

Improving atmospheric CO₂ retrievals using line mixing and speed-dependence when fitting high-resolution ground-based solar spectra

Joseph Mendonca, University of Toronto
Kimberly Strong, University of TorontoFollow
Geoffrey Toon, California Institute of Technology
Debra Wunch, California Institute of TechnologyFollow
K Sung, California Institute of Technology
Nicholas M. Deutscher, University of WollongongFollow
David W. T Griffith, University of WollongongFollow
J E. Franklin, Dalhousie University

RIS ID

107057

Publication Details

Mendonca, J., Strong, K., Toon, G. C., Wunch, D., Sung, K., Deutscher, N. M., Griffith, D. W. T. & Franklin, J. E. (2016). Improving atmospheric CO₂ retrievals using line mixing and speed-dependence when fitting high-resolution ground-based solar spectra. Journal of Molecular Spectroscopy, 323 15-22.

Abstract

A quadratic speed-dependent Voigt spectral line shape with line mixing (qSDV + LM) has been included in atmospheric trace-gas retrievals to improve the accuracy of the calculated CO2 absorption coefficients. CO2 laboratory spectra were used to validate absorption coefficient calculations for three bands: the strong 20013 ← 00001 band centered at 4850 cm−1, and the weak 30013 ← 00001 and 30012 ← 00001 bands centered at 6220 cm−1 and 6340 cm−1 respectively, and referred to below as bands 1 and 2. Several different line lists were tested. Laboratory spectra were best reproduced for the strong CO2 band when using HITRAN 2008 spectroscopic data with air-broadened widths divided by 0.985, self-broadened widths divided by 0.978, line mixing coefficients calculated using the exponential power gap (EPG) law, and a speed-dependent parameter of 0.11 used for all lines. For the weak CO2 bands, laboratory spectra were best reproduced using spectroscopic parameters from the studies by Devi et al. in 2007 coupled with line mixing coefficients calculated using the EPG law. A total of 132,598 high-resolution ground-based solar absorption spectra were fitted using qSDV + LM to calculate CO2 absorption coefficients and compared to fits that used the Voigt line shape. For the strong CO2 band, the average root mean square (RMS) residual is 0.49 ± 0.22% when using qSDV + LM to calculate the absorption coefficients. This is an improvement over the results with the Voigt line shape, which had an average RMS residual of 0.60 ± 0.21%. When using the qSDV + LM to fit the two weak CO2 bands, the average RMS residual is 0.47 ± 0.19% and 0.51 ± 0.20% for bands 1 and 2, respectively. These values are identical to those obtained with the Voigt line shape. Finally, we find that using the qSDV + LM decreases the airmass dependence of the column averaged dry air mole fraction of CO2 retrieved from the strong and both weak CO2 bands when compared to the retrievals obtained using the Voigt line shape.