Investigating the impact of using CO2-collisional parameters in atmospheric water vapor retrievals

The study of the atmospheres of the nearest planets to Earth, Venus and Mars, is of major importance for the understanding of the solar system formation and evolution, as well as the history of Earth. That is why, since the 1960s, numerous space missions have been dedicated to such studies, through the sending of space probes or landers, having ever-increasing spectral resolution and improved signal-to-noise ratio. This is also true for ground-based observation facilities, which provide atmospheric spectra of indisputable quality. However, the constant improvement of observation instruments requires the use of suitable spectroscopic data, at the risk of making an error during atmospheric retrieval process. In the case of the H₂O molecule, found in small amounts in the atmospheres of Mars and Venus, many radiative transfer models use the air-collisional parameters, instead of the CO₂-collisional parameters for the water vapor lines, due to a lack of spectroscopic data. The aim of this study is therefore to contribute to the development of a list, as complete as possible, of CO₂-collisional parameters for H₂O, following our previous article [1], and therefore to assess its impact on the study of atmospheric spectra.

To this end, new H₂O broadened by CO₂ spectra were recorded at room temperature using the GSMA high-resolution Fourier transform infrared spectrometer in several spectral regions of atmospheric interest. Isolated H₂O lines were then selected, and their CO₂-collisional parameters determined using a multi-spectral fitting procedure. Different line shape profiles were used: Voigt,  Rautian, quadratic speed-dependent Voigt and quadratic speed-dependent Rautian. The three latter consider fine physical effects leading to better reproduce the experimental line shape profile [2]. The collisional half-width parameters obtained with Voigt profile were used to determine the interaction potential for the H₂O-CO₂ molecular system. The semi-classical Modified Complex Robert-Bonamy formalism was then used to calculate the half-widths and line shifts for a large number of lines.

In the context of the future European space mission to Venus, and future observations by the VenSpec-H onboard infrared spectrometer, the calculated H₂O broadened by CO₂ linelist was then used to simulate atmospheric spectra of Venus, in two different transparency windows, using the Asimut radiative transfer model [3]. The first spectra window is located around 1.17 µm and will be used to perform observation on the nightside of Venus, under the thick cloud layer, in the low troposphere, between 0 and 15 km of altitude. The infrared signal is weak as it is emitted by the surface and the different hot atmospheric layers. The second spectral region simulated is around 2.34 µm. It will be useful to observe both the nightside, from 30 to 45 km, i.e. under the cloud layer, and the dayside, from 65 to 80 km, i.e. above the cloud layer. The dayside will provide a stronger infrared signal because of the aerosols in the Venusian clouds, which reflect a large part of solar infrared radiation passing through the mesosphere back to space, and thus to the instrument in orbit. The simulated atmospheric spectra were used to retrieve water vapor atmospheric concentration and to investigate the error made using the air-collisional parameters for H₂O lines in a CO₂-rich environment. Samples of 1000 spectra were used to obtain a 1% statistical error on retrieved parameters. Gaussian random noise was added to the samples corresponding to a chosen signal-to-noise ratio. In order to observe the dependence of the error made in function of the signal-to-noise ratio selected, the initial guess of water vapor concentration given and the standard deviation of the fit, these parameters were modified for each spectra sample. The results showed that using air-collisional parameters can lead to significant difference in the retrieved H₂O concentration in the Venus atmosphere. It is therefore necessary to continue measuring the CO₂-collisional parameters for H₂O lines, for the benefit of the planetary community. To conclude, this study is also planned for the second isotopologue of water vapor, HDO, also present in trace form in the atmospheres of our planetary neighbors.

 

[1] L. Régalia et al., Laboratory measurements and calculations of line shape parameters of the H₂O–CO₂ collision system, JQSRT 231 (2019)

[2] É. Ducreux et al., Measurements of H₂O broadened by CO₂ line-shape parameters: beyond the Voigt profile, JQSRT Accepted (2024)

[3] A.C. Vandaele et al., Modeling and retrieval of atmospheric spectra using Asimut, Conference Proc. of the First ‘Atmospheric Science Conference’, ESA SP-628 (2006)