Asymptotic Giant Branch (AGB) stars are the near-end state of low-to-intermediate mass stars (we have focused on stars between 0.6 M⊙ and 2.0 M⊙). Through combined effects of opacity driven pulsations and dust formation, these stars develop a strong pulsation and dust driven stellar wind andconsequently experience significant mass loss, ultimately losing their stellar envelopes and becomingwhite dwarfs. Surveys of nearby galaxies tell us the observable properties of these mass-losing stars, andultimately we should be able to replicate the properties of these populations using atmospheric models. We have been investigating the effects of modifying parameters of atmospheric models generated using George Bowen’s atmospheric pulsation code on the properties of the resulting grid of stellar of models. Initial results suggested extreme modifications to both the driving amplitude and dust properties were necessary to produce stars with the observed mass-loss rates at observed luminosities and pulsation periods. However, continued investigation has shown that modification of radius-mass-luminosity and pulsation-mass-radius relations to better match the results of dynamic 3D models of these stars significantly improves agreement between the properties of the modeled atmospheres and observations. Further, these modified constraints lead to self-consistent determinations of which stars should be losing mass from multiple methods. This implies we will need 3D models to correctly determine the relations between stellar mass, luminosity, radius, and pulsation period on the AGB.
Prager et al. (Tue,) studied this question.