top of page

Elemental Abundances in Red Giant Stars with APOGEE 

Keywords:

#Stellar       #APOGEE       #Chemical       #Regression

In my Modern Stellar Astrophysics course, we explored the life cycle of stars, from their formation and internal processes to their explosive deaths, using real astronomical survey data. A major component of the course was a midterm project combining stellar evolution theory with statistical analysis of observed stars. For this, I worked on a data-driven study of red giant stars from the APOGEE spectroscopic survey to investigate chemical abundance patterns across the galactic disk.

ms1.png
ms2.png

Course & Project

The goal of our midterm project was to uncover patterns in elemental abundances of red giant stars and investigate how these patterns vary with galactocentric radius. APOGEE’s infrared spectroscopy provides detailed abundance measurements for a wide array of elements in stars across the Milky Way, offering insight into stellar nucleosynthesis and galactic evolution.

Our team applied multivariate regression techniques to explore which individual or combinations of elements were most predictive of overall chemical behavior. The goal was to tie these patterns to physical processes like stellar dredge-up events and chemical enrichment histories.

Methodology

Starting with a catalog of over 700,000 stars, we filtered the APOGEE DR17 dataset to isolate red giants in the galactic disk. We applied quality cuts based on parallax, surface gravity, and metallicity to yield a cleaner sample of ~220,000 stars. For each, we extracted elemental abundances and computed galactocentric radius using stellar coordinates.

We then applied multivariate linear regression to test the predictive power of various element trios. Magnesium (Mg) and Titanium-II (Ti-II) emerged as especially strong predictors, appearing in most of the top-ranked element combinations with average 𝑅^2 values around 0.33. While not extremely strong, this correlation suggested these elements are closely tied to global chemical enrichment trends.

The main challenges of the project lay in the complexity of real astrophysical data: intrinsic scatter, stellar evolution effects, and observational biases. Our models were not able to account for measurement uncertainty or non-linear trends, limiting their interpretive power. However, the project underscored the potential of using elements like Mg and Ti-II as tracers of chemical enrichment and stellar evolution.
 

You can view our midterm paper here:

bottom of page