Modelling telluric line spectra in the optical and infrared with an application to VLT/X-Shooter spectra

TL;DR

This paper presents a new method using line-by-line radiative transfer modeling to accurately remove telluric lines from optical and infrared spectra, improving over traditional standard star comparison techniques.

Contribution

The authors developed a novel approach that models telluric absorption directly from spectral data, enabling more precise correction without relying on standard star observations.

Findings

Effective removal of telluric lines with minimal residuals

Increased statistical errors are manageable, typically around 10-20%

Method achieves percent-level accuracy in telluric correction

Abstract

Earth's atmosphere imprints a large number of telluric absorption and emission lines on astronomical spectra, especially in the near infrared, that need to be removed before analysing the affected wavelength regions. These lines are typically removed by comparison to A- or B-type stars used as telluric standards that themselves have strong hydrogen lines, which complicates the removal of telluric lines. We have developed a method to circumvent that problem. For our IDL software package tellrem we used a recent approach to model telluric absorption features with the line-by-line radiative transfer model (LBLRTM). The broad wavelength coverage of the X-Shooter at VLT allows us to expand their technique by determining the abundances of the most important telluric molecules H2O, O2, CO2, and CH4 from sufficiently isolated line groups. For individual observations we construct a telluric absorption model for most of the spectral range that is used to remove the telluric absorption from the object spectrum. We remove telluric absorption from both continuum regions and emission lines without systematic residuals for most of the processable spectral range; however, our method increases the statistical errors. The errors of the corrected spectrum typically increase by 10% for S/N~10 and by a factor of two for high-quality data (S/N~100), i.e. the method is accurate on the percent level. Modelling telluric absorption can be an alternative to the observation of standard stars for removing telluric contamination.