Demonstrating Diversity in Star Formation Histories with the CSI Survey

TL;DR

This study uses spectro-photometric data from the CSI Survey to analyze star formation histories of over 22,000 galaxies, revealing diverse evolutionary paths that challenge simple models and suggest early-set SFHs with gradual declines.

Contribution

It provides the first large-scale, individual SFH measurements for thousands of galaxies, demonstrating the diversity of galaxy evolution and limitations of traditional models.

Findings

Wide diversity of SFHs across galaxy masses

No unique link between halo mass and SFH

Support for a 2-parameter lognormal SFH model

Abstract

We present coarse but robust star formation histories (SFHs) derived from spectro-photometric data of the Carnegie-Spitzer-IMACS Survey, for 22,494 galaxies at 0.3<z<0.9 with stellar masses of 10^9 Msun to 10^12 Msun. Our study moves beyond "average" SFHs and distribution functions of specific star formation rates (sSFRs) to individually measured SFHs for tens of thousands of galaxies. By comparing star formation rates (SFRs) with timescales of 10^10, 10^9, and 10^8 years, we find a wide diversity of SFHs: 'old galaxies' that formed most or all of their stars early; galaxies that formed stars with declining or constant SFRs over a Hubble time, and genuinely 'young galaxies' that formed most of their stars since z=1. This sequence is one of decreasing stellar mass, but, remarkably, each type is found over a mass range of a factor of 10. Conversely, galaxies at any given mass follow a wide range of SFHs, leading us to conclude that: (1) halo mass does not uniquely determine SFHs; (2) there is no 'typical' evolutionary track; and (3) "abundance matching" has limitations as a tool for inferring physics. Our observations imply that SFHs are set at an early epoch, and that--for most galaxies--the decline and cessation of star formation occurs over a Hubble-time, without distinct "quenching" events. SFH diversity is inconsistent with models where galaxy mass, at any given epoch, grows simply along relations between SFR and stellar mass, but is consistent with a 2-parameter lognormal form, lending credence to this model from a new and independent perspective.