Fine-tuning stellar population models

  • L. Gustav Strömbäck

Student thesis: Doctoral Thesis


In this work we present new, high-to-intermediate spectral resolution evolutionary population synthesis models, complementing and extending the widely used Maraston models. The new models are based on four popular libraries of empirical stellar spectra, which has necessitated some modifications to the original code, while keeping much of the original ingredients – such as stellar energetics, treatment of the thermally pulsating asymptotic giant branch, and mass loss recipe – intact. In addition, we have computed models at very high resolution (R = 20000) based on the theoretical MARCS library, that extends far into the infrared region of the electromagnetic spectrum. A library-dependent, but model independent comparison is made, where both photometric and spectroscopic similarities and discrepancies are highlighted. We find that stellar population models employing empirical stellar spectra exhibit considerably bluer (B-V) colours compared to models adopting theoretical spectra synthesised from the Kurucz model atmospheres (such as the BaSeL library), but that some differences arise between the empirical libraries due to, in particular, the adopted temperature scale. Furthermore, the results obtained with the theoretical MARCS library are fully consistent with the empirical libraries in this respect. The same effect can be found also in other EPS models that are based on empirical stars. We show that this discovery, whose origin can be traced mainly to cool stars (Teff < 5000 K), leads to improved photometric agreement with both galaxy and Milky Way globular cluster data. Spectral energy distributions of the latter are also used for testing the models concerning their ability to reproduce, through full SED-fitting, the cluster ages and metallicities as derived through independent fitting in colour-magnitude diagrams. In general, the agreement is very promising, although the higher resolution of the new models cannot alleviate the age-metallicity degeneracy in the optical in any significant way. A comparison with models of absorption line indices with variable abundance ratios is also made, both for the full SED-fitting procedure and when measuring indices directly on the SED. We obtain satisfying agreement in the first case, but in the second case only when a subsample of indices are used, the combination of which is little sensitive to abundance ratio effects. As a side track we exploit for the first time for population synthesis purposes the vast stellar database of the Sloan Digital Sky Survey, but find that the metallicity and age range of the stars is currently too narrow for computing sensible population models. Finally, we re-invigorate the Ca K line for the purpose of absorption feature diagnostics, presenting a new version of the index which will help in separating solutions for a key case of the age-metallicity degeneracy.
Date of AwardJan 2012
Original languageEnglish
SupervisorClaudia Maraston (Supervisor), Daniel Benjamin Thomas (Supervisor) & Bob Nichol (Supervisor)

Cite this