Short-term spectroscopic variability in the pre-main sequence Herbig Ae star AB Aurigae during the MUSICOS 96 campaign

We present results of the spectroscopic monitoring of AB Aur obtained during the MUSICOS 96 campaign. The analysis is mainly focussed on the He I D3 line, on the Hα line, and on a set of photospheric lines. The star was monitored irregularly for more than 200 hours. We confirm the high level of variability of spectral lines in AB Aur. We find that the photospheric lines have a profile differing significantly from a classical rotational profile. The dominant features of this abnormal photospheric profile are a blue component, in absorption, whose velocity is modulated with a 34hr period, and a red component, stable in velocity but of variable intensity, with a possible periodicity near 43 hrs. The He I D3 line exhibits two well-defined components: a blue component, always in emission with a velocity modulated with a 45hr period, and a red component of variable intensity, alternatively in emission and in absorption, occurring at a fixed velocity, with a variable intensity possibly modulated with a 45 hr period. The Hα line, showing a P Cygni profile, also exhibits pseudo-periodic variations of its blue absorption component, but its variability appears more complicated than that of the other lines studied here. We suggest that the blue component of the photospheric lines is modulated by the star's rotation, with a period of 34 hrs, due to a highly inhomogeneous photosphere, involving significant radial flows. Our model also involves downflows onto the stellar pole to account for the red components of the photospheric lines and of the He I D3 line. We propose two different interpretations of the behavior of the blue component of the He I D3 line. In the first one, this component is formed in a wind originating from the star's equatorial regions. In this interpretation, the rotation period of the equatorial regions of the star is 45 hrs, implying a 25% surface differential rotation, with the pole rotating faster than the equator. The second interpretation involves a wind originating from a region of a circumstellar disk, at a distance of 1.6 stellar radii from the star's center, with a rotation period of 45 hrs. We are not able to decide which one of these two interpretations is more likely, on the basis of the data presented here.