# Non-parametric Star Formation History Reconstruction with Gaussian Processes I: Counting Major Episodes of Star Formation

**Authors:** Kartheik G. Iyer, Eric Gawiser, Sandra M. Faber, Henry C. Ferguson, Anton M. Koekemoer, Camilla Pacifici, Rachel Somerville  
**arXiv:** [1901.02877](https://arxiv.org/abs/1901.02877)  
**Date:** 2019-01-09  
**Summary page:** https://www.summarizepaper.com/en/paper/1901.02877v1/

## Key points

- Study of star formation histories (SFHs) in galaxies provides insights into physical processes regulating star formation during galaxy growth and quenching
- Authors improve upon Dense Basis SFH reconstruction method by introducing nonparametric description of SFH based on lookback times at which a galaxy assembles certain quantiles of its stellar mass
- Method uses Gaussian Processes to create smooth SFHs that are independent of any functional form, with a flexible number of parameters adjusted to extract maximum possible information from the spectral energy distributions (SEDs) being fit
- Authors apply this method to reconstruct SFHs of 48,791 galaxies with $H<25$ at $0.5 < z < 3.0$ across five CANDELS fields
- Using reconstructed SFHs, authors study how galaxies evolve as they become more massive over cosmic time
- They quantify fraction of galaxies showing multiple major episodes of star formation and find that median time between two peaks is $\sim 0.42_{-0.10}^{+0.15}t_{univ}$ Gyr, where $t_{univ}$ is age of universe at given redshift and remains roughly constant with stellar mass

## Summary

The study of star formation histories (SFHs) in galaxies provides insights into the physical processes that regulate star formation during galaxy growth and quenching. In this paper, the authors improve upon the Dense Basis SFH reconstruction method of Iyer & Gawiser (2017) by introducing a nonparametric description of SFH based on the lookback times at which a galaxy assembles certain quantiles of its stellar mass. The method uses Gaussian Processes to create smooth SFHs that are independent of any functional form, with a flexible number of parameters adjusted to extract maximum possible information from the spectral energy distributions (SEDs) being fit. The authors apply this method to reconstruct the SFHs of 48,791 galaxies with $H<25$ at $0.5 < z < 3.0$ across five CANDELS fields. Using these reconstructed SFHs, they study how galaxies evolve as they become more massive over cosmic time. They quantify the fraction of galaxies that show multiple major episodes of star formation and find that the median time between two peaks is $\sim 0.42_{-0.10}^{+0.15}t_{univ}$ Gyr, where $t_{univ}$ is the age of the universe at a given redshift and remains roughly constant with stellar mass.