# Reducing activity-induced variations in a radial-velocity time series of the Sun as a star

## TL;DR

This study demonstrates that applying kernel regression to magnetic activity indicators significantly reduces activity-induced variations in the Sun's radial velocity measurements, improving the detection of stellar signals.

## Contribution

It introduces a method using kernel regression on magnetic activity proxies to mitigate activity-related noise in solar radial velocity data.

## Key findings

- Magnetic flux is the best activity proxy over a month.
- Regression residuals have a standard deviation of 1.04 m/s.
- Activity-induced fluctuations are reduced by a factor of 2.8.

## Abstract

The radial velocity of the Sun as a star is affected by its surface convection and magnetic activity. The moments of the cross-correlation function between the solar spectrum and a binary line mask contain information about the stellar radial velocity and line-profile distortions caused by stellar activity. As additional indicators, we consider the disc-averaged magnetic flux and the filling factor of the magnetic regions. Here we show that the activity-induced radial-velocity fluctuations are reduced when we apply a kernel regression to these activity indicators. The disc-averaged magnetic flux proves to be the best activity proxy over a timescale of one month and gives a standard deviation of the regression residuals of 1.04 m/s, more than a factor of 2.8 smaller than the standard deviation of the original radial velocity fluctuations. This result has been achieved thanks to the high-cadence and time continuity of the observations that simultaneously sample both the radial velocity and the activity proxies.