ADASSX

Recent surveys, such as the Sloan Digital Sky Survey (SDSS), have discovered thousands of white dwarfs, low-mass non-fusing stellar remnants. Because these white dwarfs (WDs) radiate away their energy as they cool, changing their spectral features, their spectral characterization and evolution serve as "cosmic clocks", used to date stellar populations and stellar formation history. Emerging studies have applied machine learning techniques to analyze this population; however, these methods are often hindered due to the large number of low-resolution WD spectra and the computational intensity of atmospherically modeled synthetic spectra. This work introduces WD-SYNSPEC, a parameterized framework to generate noisy synthetic WD spectra in the optical range based on spectral type, effective temperature, and surface gravity. Able to simulate multiple spectral types, WD-SYNSPEC helps solve both the forward and inverse modeling problems in WD spectroscopy by (1) creating a sample of unlabeled spectra and (2) creating a population of spectra on which existing spectral fitting methods can be applied. WD-SYNSPEC functions by applying a blackbody spectrum, leveraging a new framework based on the introduction (and modeling) of a broad-spectrum correction factor to account for type-based shifts in the blackbody spectrum, and modeling spectral lines as least-squares optimized Voigt profiles. WD-SYNSPEC uses both theoretical modeling and empirical observations to generate morphologically accurate spectra, and adds dataset variation via the introduction of Gaussian noise. Although WD-SYNSPEC is not a true model for fitting WD atmospheres, and does not consider factors such as convection or non-equilibrium chemistries, it makes up for this in its extensibility and ability to generate populations of synthetic spectra rapidly due to its low time complexity. These methods demonstrate WD-SYNSPEC’s effectiveness and possible use within a future, more complex deep-learning-based modeling and classification pipeline.