1997A&A...326.1143H

We compare hydrogen and calcium line calculations for dM1 (Teff=3500K) stellar chromospheres with high resolution observations of a selected sample of stars with the same spectral type ((R-I)_K=0.875±0.05). We bring evidence that grids of uniform model atmospheres in the plane-parallel and hydrostatic equilibrium approximations can reproduce the average spectral signatures throughout the entire activity range. Observations confirm that when magnetic activity level rises, the Hα line is first weak, then increases in absorption strength, rapidly fills in and eventually goes into emission. We obtain a correlation between the Hα line width and equivalent width that is in good agreement with our model calculations. Simultaneous Hα and CaII line observations allow to remove the degeneracy in Hα equivalent width for low activity (weak absorption) and intermediate activity stars (filled in profiles). We show that the latter group represents a significant proportion of the stellar population. Within the active stars group, we find an exclusion zone in the [0.25Å;-1Å] Hα equivalent width domain, that can be simply explained by the rapid change from the absorption to the emission regimes when the chromospheric pressure increases. In our sample of 154 stars, covering a large luminosity range, we found no ``zero-Hα'' stars but instead a minimum (possibly ``basal'') Hα equivalent width of ∼0.20Å which, with reference to our models, suggests a transition region column mass of log(M)~-5.5. This implies that for an overwhelming majority of M1 type dwarfs the amount of non-thermal energy input in the chromosphere is much higher than in the Sun, and by reference to acoustic heating calculations, that they are also magnetically much more active (per unit area). Our observations provide evidence for gradual and important changes in the integrated physical properties of the chromosphere throughout the activity range. For example, the equivalent widths of the H and K line cores are tightly correlated with their ratio, the later decreasing from low activity (∼1.50 at -0.2Å) to high activity stars (∼1.06 at -15Å). The K lines are also typically 30% broader than the H lines and their widths increase with increasing activity level. This suggests that their optical depths in their region of formation also increase with increasing activity level. Our calcium line calculations reproduce the observed trends. We confirm a near UV and blue excess in active dMe stars that increases with activity level; in average 0.12 magnitudes in U-B (and up to 0.26mag.) and 0.03 magnitudes in B-V. This excess is about three times larger than expected from our calculations for a given atmospheric pressure, and together with discrepancies between models and observations for spectral lines, all converge to imply that high pressure plages with a filling factor of about 30% are present on these stars. Low metallicity halo dwarfs in our sample also exhibit a U-B excess, but in the case of single dMe stars an effect of metallicity is excluded. In our sample, single dMe stars are more luminous than their less active absorption line counterparts. We present an activity-luminosity relationship for the CaII lines; namely, the CaII line fluxes rise as the power of 5.4 of the stellar radius. Hence, we expect the stellar magnetic flux to rise approximately as the power of 7.4 of the radius; an important constraint for the dynamo mechanism. Such a correlation is also found with Hα and L_X, the X-ray luminosity. With our present understanding, these correlations are, at least partly, activity-metallicity relationships. We emphasize the importance of metallicity on stellar activity as a whole, i.e., metal deficient stars are also activity deficient.