Activity sensitive spectral lines of M dwarfs in the CARMENES visible and near infrared spectral range: impact on radial velocity determinations and stellar parameters determination

In this contribution we summarize our project devoted to identify activity-sensitive spectral lines in the CARMENES visible and near-infrared spectral range of M dwarfs. The aim is to contribute to solve the problem of stellar activity in RV measurements to search for exoplanets around
these stars and in the determination of precise stellar parameters. To identify lines with a significant chromospheric contribution, apart from well known
activity indicators (Na I D 1 D 2 He I D 3 Hα, and Ca II IRT lines, He I 10830 Å, Paγ and Paβ lines), we have used the spectral subtraction technique
using our Python code iSTARMOD Labarga Montes 2020 choosing as reference the spectrum of the star with lower activity. We confirm the new
activity sensitive lines by analysing the correlation with the other well known activity indicators in the same spectra and their temporal evolution in two
particular active stars EV Lac (EV Lacertae, M 3.5) and YZ CMi (YZ Canis Minoris, M 4.5). They are specially active stars with strong flares and strong
magnetic fields. In addition, we analyse line by line the template spectrum (co added of all the individual spectra available) of these two
stars applying also the spectral subtraction using in this case as reference star an inactive M dwarf star of similar spectral type to search for
magnetically sensitive lines, that is lines with detectable Zeeman broadening. After this analysis on YZ CMi and EV Lac, we have found 84 and 97
chromospheric activity-sensitive lines and 160 and 170 magnetically-sensitive spectral lines respectively. We are now studying the impact of the
elimination of the activity sensitive spectral lines identified in this way on the RV determination using cross correlation functions with weighted binary
masks as in Lafarga et al. (2020) and on the stellar parameters determination by spectral synthesis as in Marfil et al. (2021).