Improved modeling of telluric absorption features through the retrieval of atmospheric trace gases toward EPRV measurements

Since ground-based Extreme Precision Radial Velocity (EPRV) measurements are inevitably impeded by telluric contamination, it is crucial to model the telluric aborption features down to the spectral noise level to accurately measure the stellar Doppler shifts, and hence determine the bulk composition and surface gravity of terrestrial mass planets. In this talk, we present preliminary results from precision modeling of the telluric features by adopting atmospheric trace gas retrieval algorithm. For this, we used spectra of PARVI (1.1 – 1.76 μm) and iSHELL (1 – 2.5 μm) as a testbed. PARVI is a Laser-Frequency Comb (LFC)-based echellograms offering a resolving power, R, of ~65,000 and iSHELL is a silicon immersion grating spectrograph offering R ~ 80,000, both of which are dedicated to Precision Radial Velocity measurements in the near infrared. They have been operational at Palomar Observatory, CA, and at IRTF at Mauna Kea Observatory, respectively.

We have successfully fitted the telluric absorption lines captured in their stellar spectra to typically ~ 1 % or the spectral noise level by adjusting the trace gas abundances (including water vapor) independently and simultaneously during the spectrum fitting. For this, we have proceeded in the multiple steps; (1) we have adapted a state-of-the-art atmospheric spectrum fitting program, GFIT, which has been developed for atmospheric trace gas retrievals and adopted by the TCCON (Total Column Carbon Observing Network; http://www.tccon.caltech.edu/), (2) we utilize atmospheric models of temperature, pressure and composition individually for sites at the time of the observations, making the best use of climatology and meteorological analyses (e.g., NCEP, GEOS-FPIT), and finally (3) we have employed the most up-to-dated spectroscopy in the target spectral regions. We note that the independent adjustment of the trace gas abundances is key to achieving the precision modeling of the telluric absorption to the noise level, especially water vapor whose abundance is subject to change on short time scales. We present the spectrum fitting results of PARVI and iSHELL spectra and discuss the performance and advantage of our novel approach toward EPRV measurements

Government support acknowledged.