In the last decade it has been shown that the visible spectrum of Classical T Tauri stars (CTTs) is composed of a "normal" stellar spectrum and a smoothly varying continuum with emission lines superposed (Walker 1987; Hartigan et al. 1989, HHKHS hereafter). This continuum excess (excess hereafter) makes the photospheric absorption lines to appear veiled, i.e. less deep than those of a star with the same spectral type. The excess is bluer than the normal stellar continuum and seems to be related with some mechanism of the accretion process: emission from an active photosphere (Calvet et al. 1984), from a boundary layer (Lynden-Bell & Pringle 1974; Bertout et al. 1988), or from the base of magnetospheric accretion columns (Hartmann et al. 1994).
In order to model a CTTs, it is necessary to know the amount of excess, or equivalently, the amount of veiling (defined as the ratio between the excess and the local continuum of the underlying normal star). Several methods have been developped to estimate the veiling. The most popular one is that proposed by HHKHS (see also Basri & Batalha 1990) which consists in comparing, within a small wavelength interval of a few tens of angstroms (where the veiling and the extinction towards the object can be assumed constant), high resolution spectra of a CTTs with a standard star of the same spectral type. More recently, Gullbring et al. (1998, GHCC hereafter) introduced a new method to derive both the visual extinction and the excess spectral shape from spectrophotometrically calibrated data by veiling analysis in a limited number of optimized photospheric absorption lines.
This article intends to study within a rigorous framework, veiling, visual extinction and excess extraction from the spectra of CTTs. Sect. 2 investigates further the method of HHKHS for veiling estimate. We introduce a non linear least square fit which allows to deal with non constant noise, and we derive a formal expression for the resulting error of the veiling. We also discuss the main sources of bias and the influence of the spectral resolution on the veiling estimate. In Sect. 3, we generalize the method of GHCC for the determination of the visual extinction and excess spectral shape by introducing a "continuous" method which is tested successfully on simulated data. The originality of this new approach is that it uses all the avalaible information contained in the spectra. We derive a formal expression for the visual extinction error and we conclude this work by a brief discussion of the results.
© European Southern Observatory (ESO) 1999
Online publication: February 23, 1999