Ca II H+K fluxes from S-indices of large samples: a reliable and consistent conversion based on PHOENIX model atmospheres

Context. Historic stellar activity data based on chromospheric line emission using O.C. Wilson's S-index reach back to the 1960ies and represent a very valuable data resource both in terms of quantity and time-coverage. However, these data are not flux-calibrated and are therefore difficult to compare with modern spectroscopy and to relate to quantitative physics. Aims. In order to make use of the rich archives of Mount Wilson and many other S-index measurements of thousands of main sequence stars, subgiants and giants in terms of physical Ca II H+K line chromospheric surface fluxes and the related R-index, we seek a new, simple but reliable conversion method of the S-indices. A first application aims to obtain the (empirical) basal chromospheric surface flux to better characterise stars with minimal activity levels. Methods. We collect 6024 S-indices from six large catalogues from a total of 2530 stars with well-defined parallaxes (as given by the HIPPARCOS catalogue) in order to distinguish between main sequence stars (2133), subgiants (252) and giants (145), based on their positions in the Hertzsprung-Russell diagram. We use the spectra of a grid of PHOENIX model atmospheres to obtain the photospheric contributions to the S-index. To convert the latter into absolute Ca II H+K chromospheric line flux, we first derive new, colour-dependent photospheric flux relations for, each, main sequence, subgiant and giant stars, and then obtain the chromospheric flux component. In this process, the PHOENIX models also provide a very reliable scale for the physical surface flux. Results. For very large samples of main sequence stars, giants and subgiants, we obtain the chromospheric Ca II H+K line surface fluxes in the colour range of 0.44 < B - V < 1.6 and the related R-indices. We determine and parametrize the lower envelopes, which we find to well coincide with historic work on the basal chromospheric flux. There is good agreement in the apparently simpler cases of inactive giants and subgiants, and distinguishing different luminosity classes proves important. Main sequence stars, surprisingly, show a remarkable lack of inactive chromospheres in the B - V range of 1.1 to 1.5. Finally, we intoduce a new, pure and universal activity indicator: a derivative of the R-index based on the non-basal, purely activity-related Ca II H+K line surface flux, which puts different luminosity classes on the same scale. Conclusions. The here presented conversion method can be used to directly compare historical S-indices with modern chromospheric Ca II H+K line flux measurements, in order to derive activity records over long periods of time or to establish the long-term variability of marginally active stars, for example. The numerical simplicity of this conversion allows for its application to very large stellar samples.