CAOSP abstracts, Volume: 38, No.: 2, year: 2008

Abstract: Abundance analyses are usually performed by empirical methods. The existing semi-automatic software packages mostly are strongly dependent on an initial or fixed set of fundamental parameters and the set of lines provided by the researcher. Furthermore the implemented iterative process of fitting certain lines or regions of a star's spectrum often ignores the parameters' physical meaning. Thus, the accuracy of the derived results and moreover the comparibility of different investigations are in many cases left to the researchers judgement. However, several attempts to study how uncertainties in single fundamental or line parameters translate to the precision of abundance determination have been successfully conducted. We started to investigate the combined effect of variations in effective temperature, surface gravity, microturbulence, line blending, metallicity (i.e. abundance pattern) together with the number of atomic lines used. Our aim is to be able to make quantitative statements on this general uncertainty in abundance analysis and to define a way to automate determination of fundamental parameters and element abundances whilst: providing a traceable measure of accuracy for the results; being not restricted to a certain instrument, resolution or object; minimizing the effort of preparation.