1996A&A...305..209H

The variation in the continuum intensity (50-5000nm) of late-type M dwarfs is investigated via two grids of model atmospheres with different temperature minima. It is shown that the (E)UV intensity is mostly dependent on the transition region pressure, although the temperature minimum also plays an important role. We also observe a significant frequency redistribution of the photospheric flux when changing the minimum temperature, and a black-body type of emission from the lower chromosphere. We examine the formation of the continuum and point out that, although some differences appear for very low or very high activity levels, in general the global picture is much alike the Sun. We show that the UV continua are very good diagnostics of cool dwarf atmospheres, from the temperature minimum to the transition region. Our calculations give a good overview of the domains where physical parameters and spectral signatures are most likely observed. We compute the UBV(RIJKL)_J broad band fluxes for our models and conclude that the chromospheric contribution should be detectable in the U band and possibly also in the B band. We compare our calculations to recent high resolution observations for selected stars in a narrow spectral range ((R-I)_K=0.875±0.05); we show that Hα line profiles behave as expected, with a tight correlation between the line width and equivalent width. Hα emission line stars show an excess in U-B color but not in B-V. They are also more luminous than their less active absorption line counterparts, which indicates that active dwarfs have not yet reached the main sequence and are intermediate between T Tauri stars and main sequence stars. The anomalously large proportion of active stars towards late spectral types is attributed to the very slow contraction phase for low mass stars. We calculate the fluxes in the Extreme Ultraviolet Explorer Sn/SiO band (500-740A) and show that they are compatible with observed upper limits. We further compare our results for the upper activity range to pre-main sequence models and observations. They strongly support the case for a chromospheric contribution to Hα and the blue/UV excess for those objects (T Tauris, naked T Tauris, YSOs). An important conclusion is that at high pressures, corresponding to active dMe stellar atmospheres, the chromosphere becomes a very efficient radiator at continuum wavelengths. From log(M)~-5 (column mass), radiative losses in the continuum rise exponentially and faster than in Hydrogen spectral lines because of their larger optical depths. As a consequence, HI spectral lines have a small or negligible contribution to the total HI (lines and continua) and white light radiative budget. We show that the hydrogen series dominate the radiative cooling in spectral lines (from 40% to 90%) for Hα emission line stars. The cooling in the (E)UV continuum overwhelms the total radiative budget and is much larger than that in outstanding chromospheric and transition region lines (e.g. CaII, MgII, HI Lyman and Balmer). We propose this as a possible contribution for the apparent saturation observed in some spectral lines, therefore questioning the suggestion of saturation in magnetic activity levels.