Convective motions in the photosphere are nowadays believed to be the main contributors to observed line asymmetries and shifts in the optical spectra of the Sun and solar-type stars. Several studies have been devoted to the analysis and classification of these observational features in the Sun (e.g. Dravins et al. 1981, hereafter DLN; Balthasar 1984). Some researchers, extending this work to other stars, have seen how the difficulties grow, not only due to the lack of photons but also to the impossibility of accurately removing two major effects: gravitational shifts and the radial velocity of the star, which in turn do not allow the setting up of an absolute velocity scale (e.g. Gray 1982; Dravins 1987a,b). However, absolute line shifts should be included in spectral syntheses and inversion codes that take velocity patterns into account. Line asymmetries and shifts establish a footprint of the dynamics of convection, thereby imposing a major constraint to the theoretical modelling.
The increasing resolving power of the spectrographs and the
improvements of detectors have made systematic high resolution
spectroscopic observations possible over wide spectral ranges. Recent
studies (e.g. Allende Prieto et al. 1995) have extended very
high-quality spectroscopic observations beyond the domain of the
brightest stars. In the solar case, optical, IR and UV atlases are
available from modern observations with very high signal-to-noise ratio
and spectral resolution. For the optical spectrum of the Sun seen as a
star, the Solar Flux Atlas from 296 to 1300 nm (Kurucz et al. 1984,
referred to here as the FTS flux spectrum) has been the most
extensively used. Partially from the same data set (flux), partially from a second set of Kitt Peak FTS spectra (disc-centre intensity), H. Neckel prepared
the Spectral Atlas of Solar Absolute Disk-Averaged and Disk-Center
Intensity from 3290 to 12510 Å (Brault & Neckel 1987; for details
see Neckel 1994). We shall refer to the included
disc-centre spectrum as the FTS disc-centre spectrum. These atlases
achieve signal-to-noise ratios of about 2500 and a resolving power
.
The set of lines with wavelengths accurate enough
to be useful in this context has been significantly enlarged by the
work of Nave et al. (1994), who measured and identified 9501 FeI lines using Fourier transform spectrometers at NSO, Imperial
College (London, UK), and NIST
, from 0.17 to 5
m. In the optical range, they claim an uncertainty below 1 mÅ for many lines, and two orders of magnitude larger for the worst cases.
We have used these relatively new atlases, as well as the more traditional Liège Atlas (Delbouille et al. 1973), to measure central wavelengths for a large subsample of the neutral iron lines included in the line list by Thèvenin (1989, 1990). His compilation joins those lines classified by Moore et al. (1966) as one-blended or unblended, and it was taken as a starting point since the determination of accurate line centres does not require such a clean profile as is needed to measure line asymmetries. This information is combined with the rest air wavelengths from Nave et al. (1994) to deduce the displacements of the solar lines.