Cool Stars 14 - Submitted Abstract # 13
This version created on 05 October 2006

How accurately can we determine stellar parameters:  The case of Teff
in F and G type stars


Frank Grupp, University Observatory Munich
Lyudmila Mashonkina, Institute of Astronomy of Russian Academy of Science 
Thomas Gehren, University Observatory Munich

The accuracy when testing stellar evolution models depends on the
evolutionary model on the one hand and on the accuracy of stellar
parameters on the other.  This talk will concentrate on the matter of
stellar effective temperature determination as one of the most
important stellar parameters and show how much different methods
depend on models and methods used.  Based on my opacity sampling model
atmosphere code MAFAGS-OS (Astron.  Astrophys.  426, 309-322 & Astron.
Astrophys.  420, 289-305 ) I will show the uncertainties in
temperature determination arising from the following concepts:  
   Balmer line profiles 
   Infrared Flux Method 
   Theoretical colours 
Beside the effect of transiting from the most commonly used ODF
models to the more advanced opacity sampling models, the effect
of NLTE and different resonance broadening theories on Hydrogen
Balmer lines, as well as the influence of new atomic bound-free
cross section data will be shown.  The 'sad' conclusion is, that
uncertainties related to models and methods alone are larger
than 100K.  Especially the IRFM is not as model independent as
sometimes suggested.  The stated error budget is related to pure
model input error and assumes 'perfect' observational data. 
With real data the situation gets even worse. The influence of
such uncertainties on abundance determinations is shortly
discussed at the end of the talk.  Finally, the need for better
verification and calibration of the spectroscopic and
photometric methods of parameter determination is shown.  Ideas
of how to achieve this goal are presented.

----------------------------------