Every photon that reaches Earth carries the fingerprint of its stellar origin, encoded in spectral lines that reveal temperature, composition, and motion. But extracting these cosmic secrets from noisy, instrument-affected data requires sophisticated modeling that can separate genuine stellar signals from terrestrial and instrumental contamination. Enter SMART - a framework that turns this challenge into a solved problem through forward-modeling wizardry.
SMART harnesses Markov Chain Monte Carlo algorithms to model high-resolution near-infrared spectra from premier instruments including Keck/NIRSPEC, SDSS/APOGEE, and IGRINS. Rather than simply fitting curves, it forward-models the entire observation process - from stellar atmosphere through Earth’s atmosphere to detector response. The toolkit includes advanced defringing algorithms based on Rojo & Harrington’s wavelet methods, telluric correction capabilities, and robust uncertainty quantification through Bayesian inference.
Developed by the stellar astrophysics community at UCSD, Northwestern, and partner institutions, SMART represents years of hard-won experience in precision spectroscopy. Whether you’re hunting exoplanets through radial velocity measurements, characterizing brown dwarf atmospheres, or mapping stellar populations across the galaxy, this toolkit transforms complex spectroscopic analysis from months of custom coding into reproducible, publication-ready science.
⭐ Stars: 9
💻 Language: Python
🔗 Repository: chihchunhsu/smart