ar X iv : a st ro - p h / 04 09 21 2 v 1 9 S ep 2 00 4 1 WAVELENGTH SHIFTS IN SOLAR - TYPE SPECTRA

Spectral-line displacements away from the wavelengths naively expected from the Doppler shift caused by stellar radial motion may originate as convective shifts (correlated velocity and brightness patterns in the photosphere), as gravitational redshifts, or perhaps be induced by wave motions. Absolute lineshifts, in the past studied only for the Sun, are now accessible also for other stars thanks to astrometric determination of stellar radial motion, and spectrometers with accurate wavelength calibration. Comparisons between spectroscopic apparent radial velocities and astrometrically determined radial motions reveal greater spectral blueshifts in F-type stars than in the Sun (as theoretically expected from their more vigorous convection), further increasing in A-type stars (possibly due to atmospheric shockwaves). Work is in progress to survey the spectra of the Sun and several solar–type stars for " unblended " photospheric lines of most atomic species with accurate laboratory wavelengths available. One aim is to understand the ultimate information content of stellar spectra, and in what detail it will be feasible to verify models of stellar atmospheric hydrodynamics. These may predict bisectors and shifts for widely different classes of lines, but there will not result any comparison with observations if such lines do not exist in real spectra, or are too blended for meaningful measurement. An important near–future development to enable a further analysis of stellar surface structure will be the study of wavelength variations across spatially resolved stellar disks, e.g., the center–to–limb wavelength changes along a stellar diameter, and their spatially resolved time variability. 1. Really understanding stellar atmospheres Improved understanding of physical processes in stellar atmospheres comes from the interplay between theory and observations. The validity of, e.g., hydrodynamic simulations of stellar photospheres can be tested by whether they succeed in reproducing observed shapes and asymmetries of photospheric spectral lines. On one hand, the success of such models (at least for solar-type stars) is encouraging in suggesting that many physical processes probably are beginning to get well understood. On the other hand, however, similar spectral–line shapes may sometimes be predicted by models with different parameter combinations, showing the need for further observational constraints. New diagnostic tools are needed to, e.g., segregate 2-dimensional models from 3-dimensional ones, to clarify effects of non-LTE, or of 3-dimensional radiative– transfer effects in line formation. The need for additional diagnostic tools is made more urgent whenever significant predictions are obtained from hydrodynamic models which disagree with, e.g., simple one–dimensional ones. For example, any …