Shadowing the Sun I: Testing Synthetic Granulation Spectra with NEID Solar Eclipse Observations

Abstract

The subtle Doppler radial velocity (RV) signature of an Earth-mass exoplanet in the Habitable Zone around a nearby Sun-like star is difficult to detect due to stellar variability. While multiple techniques aim to mitigate rotationally-modulated stellar activity, granulation, introducing 0.1-1 m s$^{-1}$ RV shifts, currently sets the floor for detecting Earth-mass RV signals for low-activity stars. Observations of the Rossiter-McLaughlin (RM) effect during a solar eclipse, with a peak-to-trough amplitude of $\sim$1.4 km s$^{-1}$, are a rigorous test of granulation and convective blueshift in the solar spectrum. Using NEID solar feed data of the October 2023 solar eclipse and GRASS-E (a geometry-modified version of GRASS), we compare the observed RVs for 22 lines with the simulated RVs from the synthetic granulation-driven spectra. Residuals of the GRASS-E Fe I 5434 angstrom model show an RMS of 13.47 m s$^{-1}$ over the RM signal. We find that a model with an optimized convective blueshift polynomial for Fe I 5434 significantly outperforms GRASS-E with an RMS of only 2.90 m s$^{-1}$, suggesting that GRASS-E is currently limited by sparse center-to-limb observations. By combining a set of spectral lines, we achieve a weighted RMS of 5.2 m s$^{-1}$ over the RM signal ($\sim$0.37\%), a level of accuracy well beyond the reach of current exoplanet transit observations. In all, we demonstrate that solar eclipse observations allow us to recover the center-to-limb variation of the convective blueshift for individual spectral lines.