Riffing with MOOG: the Impact of Departures from the LTE Approximation on the Analysis of Blue Abundance Probes in Evolved Stars

Developed by Chris Sneden in the early 1970’s, the line transfer code MOOG is a Fortran-based program that performs a wide range of line analysis and spectrum synthesis tasks including determinations of equivalent widths, stellar fluxes, and chemical abundances. The standard version of MOOG relies upon the assumptions of local thermodynamic equilibrium (LTE) and one-dimensional geometry. Consequently, MOOG treats scattering as pure absorption and employs a simplified source function (that directly scales with the Planck function). The LTE approximation appears to be sufficient for the derivation of abundances in select stars (e.g. warm, relatively metal-rich dwarfs). However, the use of the simplified source function is inadequate for abundance determinations in evolved stars of low temperature and metallicity as it fails to account for the main source of continuous opacity in the bluest wavelength regions (Rayleigh scattering). Over a two-year period, I undertook a major redevelopment of MOOG to include an expanded form of the source function that incorporates a mean intensity (and accordingly, a scattering contribution). I then altered the solution of the radiative transfer equation in the MOOG program to implement a short characteristics approach with an accelerated lambda iteration scheme (MOOG-SCAT). I will discuss these various modifications made to MOOG as well as continued efforts to enhance its functionality. I will report on recent work to produce an open source, readily available version of MOOG-SCAT on Github. Finally, I will attempt to describe the impact of the workhorse MOOG code on chemical composition analyses and the legacy of its creator, my mentor and friend, Chris Sneden.