Keywords - PCA of continuous 2D curves, SFH, Bayesian Analysis, Population of galaxies, photometric data, spectroscopic data
References - This analysis makes extensive use of Sparre et al 2015 method of performing PCA on 2D SFH data. This also builds up on my recent publication to understand the similarity in SFHs from two Bayesian SED fitting tools each optimized for these ~3000 LEGA-C galaxies Kaushal et al 2023.
We can obtain two types of observational data of galaxies -
Photometric data (small)
Spectroscopic data (large)
Bayesian modeling of these data together will give us information about their Star-formation Histories (SFHs) as well as other internal properties like dust content, metal content, ages of stars, star-formation rates etc.
My recent publication focused on comparing the SFHs of these galaxies at 3 physical timestamps, namely - t10, t50 and t90 - corresponding to the times when a galaxy forms 10%, 50% and 90% of its stellar mass Kaushal et al 2023.
In this notebook, we explore a step further and systematically study major star formation modes of galaxies of the LEGA-C Survey at - not just at few time-stamps, but across their full lifetime/history.For that, we adopt PCA technique described in detail in Sparre et al 2015 (inspired by Cohn & Voort 2015).SFH of a galaxy is seen as a vector in an N-dimensional space.As >90% of the information could be represented in first 3 components, we will study PC0, PC1 and PC2 only. Note that they describe scatter around the mean SFH and diagonalize the scatter matrix.
They characterize the most important modes of star-formation histories in the population.
The PC0-mode accounts for galaxies forming stars early or late, depending on whether the coefficient q0 is positive (earlier than sample mean) or negative (later than sample mean) . PC1 and PC2 determine the more detailed evolution of the SFH and cross the zero-point two and three times, respectively.
Comments