Improving spectroscopic line parameters by means of atmospheric spectra: Theory and example for water vapor and solar absorption spectra

Current atmospheric remote sensing experiments measuring high resolution spectra apply instruments of similar quality as the laboratory experiments that are especially equipped to derive spectroscopic parameters. We propose a method that uses high quality atmospheric spectra to improve the spectroscopic parameterization. It is an optimal estimation method that applies the uncertainty ranges given in current spectroscopic databases as a priori covariance and the residuals between the simulated and measured atmospheric spectra as new measurements. We test the method by updating the current HITRAN parameters of 49 mid-infrared water vapor lines (situated in 15 spectral windows between 795 and ) by the information present in high quality ground-based Fourier transform infra-red (FTIR) spectra. We show that the application of a speed dependent Voigt line shape model is important. The updated water vapor parameterization, compared to the original one, leads to lower residuals, larger measurement information content, and better agreement between remotely sensed and coincident in situ water vapor profiles. Using the new line parameterization, a state-of-the-art ground-based FTIR system is able to monitor upper tropospheric water vapor amounts and middle/upper tropospheric HDO/H₂O ratios.