Stellar classification from simulated DIVA spectra
Abstract
The Double Interferometer for Visual Astrometry (DIVA) will obtain, apart from astrometry, intermediate-resolution spectra covering the entire visual range for over 10 million stars. The classification of such spectra according to T_eff and log g has been simulated. Using Kurucz spectral energy distributions and optional binning of the signal on the CCD detectors, simulated DIVA data have been produced in the spectral range of ~= 3130 to ~= 9990 Å. Binning of 4 or 8 pixels leads to 66 or 32 spectral bands. The band width (resulting from the point spread function of the DIVA optics) varies between ~= 160 and ~= 380 Å for 66 and between ~= 220 and ~= 400 Å for 32 bands. The Kurucz models (in ~= 20 Å steps) were scaled to V = 10, V = 12, or V = 14. The observational scatter was calculated for reasonable values for the DIVA CCD detector dark current and read-out noise. For hot stars of V ~= 12 the accuracy achieved for classification is ~= 0.5 in log g and ~= 10 percent in T_eff. For stars of intermediate temperature these accuracies are 0.3 in log g and 5 percent in temperature. For the coolest stars the accuracies are better than in the other temperature regimes. However, between 6000 and 4800 K the gravity determination turns out to be difficult. The results for 32 bands are, in general, as good as or better than those for 66 bands.
- Publication:
-
Astronomische Gesellschaft Abstract Series
- Pub Date:
- 1999
- Bibcode:
- 1999AGAb...15..146E
- Keywords:
-
- Spectral Classification: Stars