PHD Discussions Logo

Ask, Learn and Accelerate in your PhD Research

Question Icon Post Your Answer

Question Icon

3 years ago in Asteroseismology , Astrophysics By Kabir

How can I identify oscillation frequencies from a power spectrum?

I'm analyzing Kepler light curves for main-sequence stars to do asteroseismology. After computing the power spectrum, I'm left with a forest of peaks. My advisor says many are noise. What statistical or algorithmic approaches can I use to distinguish genuine p-mode oscillation peaks from random fluctuations or instrumental artifacts with confidence?

 

All Answers (1 Answers In All)

By Ravneet Singh Answered 1 year ago

The key is to move beyond visual inspection. First, I always model the background noise a combination of granulation, activity, and white noise and subtract it from the spectrum. What remains should have peaks that stand above the local noise floor. Then, I calculate a signal-to-noise ratio (S/N) for each peak; in solar-like oscillators, a common threshold is S/N > 4-6. Crucially, look for patterns: real p-modes have near-regular spacing (the large frequency separation). If you see a single, isolated high peak without this pattern or harmonics, be skeptical it's often an artifact or a granulation remnant.

 

Your Answer