2024PASP..136j4503S

Ground-based Precision Radial Velocity (PRV) measurements are inevitably impeded by contamination from telluric absorption features, particularly in the infrared region. Thus, it is crucial to improve modeling of the telluric absorption features down to the spectral noise level. As part of the efforts towards improved PRV measurements, we have taken an existing atmospheric trace gas retrieval algorithm (GFIT) and have successfully adapted it to fit the telluric absorption features in stellar spectra down to the spectral noise level (typically ∼1%). We have established a stellar spectral fitting processing pipeline, Stellar-GFIT, to analyze a series of stellar spectra observed by two spectrographs, PARVI (1.1-1.76 μm) commissioned at the Palomar Observatory (Palomar Mountain, CA) and iSHELL (1-5 μm) deployed on the IRTF (Mauna Kea, HI). For this, we have (1) implemented a Gaussian instrumental line shape function, (2) generated atmospheric models (consisting of temperature, pressure, and volume mixing ratios of all the known trace gases) for the particular observation sites and times, (3) employed the most up-to-date spectroscopic parameters in the target spectral regions, and finally (4) developed a series of spectral fitting intervals of ∼60 cm^–1 width, i.e., micro-windows, customized to the individual orders of each spectrograph. Stellar-GFIT is also capable of handling non-telluric features, such as transitions from a gas cell placed in the starlight beam and stellar features if a model spectrum template is available for the target star. We present spectrum fits from the observations of various target stars and discuss the performance and advantages of our novel approach. One of the major strengths of Stellar-GFIT is an ability to adjust the abundance of atmospheric trace gases simultaneously with determining the stellar doppler shift, mitigating any adverse impacts of short-timescale variations of water vapor.