2012A&A...543A..80F

BA-type supergiants show a high potential as versatile indicators for modern astronomy. This paper constitutes the first in a series that aims at a systematic spectroscopic study of Galactic BA-type supergiants. Various problems will be addressed, including in particular observational constraints on the evolution of massive stars and a determination of abundance gradients in the Milky Way. The focus here is on the determination of accurate and precise atmospheric parameters for a sample of Galactic BA-type supergiants as prerequisite for all further analysis. Some first applications include a recalibration of functional relationships between spectral-type, intrinsic colours, bolometric corrections and effective temperature, and an exploration of the reddening-free Johnson Q and Stroemgren [c₁] and β-indices as photometric indicators for effective temperatures and gravities of BA-type supergiants. An extensive grid of theoretical spectra is computed based on a hybrid non-LTE approach, covering the relevant parameter space in effective temperature, surface gravity, helium abundance, microturbulence and elemental abundances. The atmospheric parameters are derived spectroscopically by line-profile fits of our theoretical models to high-resolution and high-S/N spectra obtained at various observatories. Ionization equilibria of multiple metals and the Stark-broadened hydrogen and the neutral helium lines constitute our primary indicators for the parameter determination, supplemented by (spectro-)photometry from the UV to the near-IR. We obtain accurate atmospheric parameters for 35 sample supergiants from a homogeneous analysis. Data on effective temperatures, surface gravities, helium abundances, microturbulence, macroturbulence and rotational velocities are presented. The interstellar reddening and the ratio of total-to-selective extinction towards the stars are determined. Our empirical spectral-type-T_eff scale is steeper than reference relations from the literature, the stars are significantly bluer than usually assumed, and bolometric corrections differ significantly from established literature values. Photometric T_eff-determinations based on the reddening-free Q-index are found to be of limited use for studies of BA-type supergiants because of large errors of typically ±5%(1σ statistical)±3%(1σ systematic), compared to a spectroscopically achieved precision of 1-2% (combined statistical and systematic uncertainty with our methodology). The reddening-free [c₁]-index and β on the other hand are found to provide useful starting values for high-precision/accuracy analyses, with uncertainties of ±1%±2.5% in T_eff, and ±0.04±0.13dex in logg(1σ-statistical, 1σ-systematic, respectively).