Variability of stellar granulation and convective blueshift with spectral type and magnetic activity

Context. In solar-type stars, the attenuation of convective blueshift by stellar magnetic activity dominates the RV (radial velocity) variations over the low amplitude signal induced by low mass planets. Models of stars that differ from the Sun will require a good knowledge of the attenuation of the convective blueshift to estimate its impact on the variations. Aims. It is therefore crucial to precisely determine not only the amplitude of the convective blueshift for different types of stars, but also the dependence of this convective blueshift on magnetic activity, as these are key factors in our model producing the RV. Methods. We studied a sample of main sequence stars with spectral types from G0 to K2 and focused on their temporally averaged properties: the activity level and a criterion allowing to characterise the amplitude of the convective blueshift. This criterion is derived from the dependence of the convective blueshift with the intensity at the bottom of a large set of selected spectral lines. Results. We find the differential velocity shifts of spectral lines due to convection to depend on the spectral type, the wavelength (this dependence is correlated with the T-eff and activity level), and on the activity level. This allows us to quantify the dependence of granulation properties on magnetic activity for stars other than the Sun. We are indeed able to derive a significant dependence of the convective blueshift on activity level for all types of stars. The attenuation factor of the convective blueshift appears to be constant over the considered range of spectral types. We derive a convective blueshift which decreases towards lower temperatures, with a trend in close agreement with models for T-eff lower than 5800 K, but with a significantly larger global amplitude. Differences also remain to be examined in detail for larger T-eff. We finally compare the observed RV variation amplitudes with those that could be derived from our convective blueshift using a simple law and find a general agreement on the amplitude. We also show that inclination (viewing angle relative to the stellar equator) plays a major role in the dispersion in RV amplitudes. Conclusions. Our results are consistent with previous results and provide, for the first time, an estimation of the convective blueshift as a function of T-eff, magnetic activity, and wavelength, over a large sample of G and K main sequence stars.