The Star-forming Main Sequence and Bursty Star-formation Histories at $z>1.4$ in JADES and AURORA

TL;DR

The study investigates bursty star formation in galaxies across $1.4 < z < 9$ by measuring the SFMS with Hα- and UV-based SFR indicators from JWST JADES DR3 and the AURORA program, applying independent dust corrections. Using a large spectroscopic sample and non-parametric SFH fits from PROSPECTOR, it quantifies the intrinsic SFMS scatter $σ_{ m int}$ as a function of stellar mass and redshift, finding a mass dependence for Hα but not UV and a decline in scatter toward higher redshift, with $σ_{ m int}$ for Hα evolving from $0.36^{+0.02}_{-0.02}$ dex to $0.22^{+0.08}_{-0.07}$ dex and UV from $0.28^{+0.02}_{-0.02}$ dex to $0.20^{+0.08}_{-0.07}$ dex. The redshift evolution of specific SFR is characterized by $γ=1.89^{+0.16}_{-0.15}$ for Hα and $γ=1.36^{+0.13}_{-0.13}$ for UV, and 41–60% of the sample shows evidence against a constant SFH, indicating prevalent burstiness; there is tentative evidence for shorter burst timescales at higher redshift based on Hα-to-UV luminosity ratios and inferred SFHs. The results align with theoretical expectations of bursty SFHs in the early Universe and provide stringent empirical constraints for models of galaxy growth and feedback across cosmic time.

Abstract

We analyze JWST spectroscopic and HST+JWST photometric observations of 659 star-forming galaxies at $1.4 < z < 9$ from DR3 of the JADES survey and the AURORA Cycle 1 program. We measure the star-forming main sequence (SFMS) for galaxies above $10^{8.5}\rm\ M_\odot$ where the sample is largely representative, estimating star-formation rates (SFRs) using the H$α$ line flux and rest-frame far UV (1600Å) continuum measurements, each independently corrected for dust attenuation. We find that the intrinsic, measurement-error-subtracted scatter in the SFMS ($σ_{\rm int}$) increases with decreasing stellar mass for the H$α$-based SFMS, and we find no mass dependence of $σ_{\rm int}$ in the UV-based SFMS. Additionally, we find that $σ_{\rm int}$ decreases with increasing redshift, from $0.36^{+0.02}_{-0.02}$ dex to $0.22^{+0.08}_{-0.07}$ dex (H$α$ SFMS), and from $0.28^{+0.02}_{-0.02}$ dex to $0.20^{+0.08}_{-0.07}$ dex (UV SFMS) between $z\sim2$ and $z\sim 6.5$. We also measure the redshift evolution of the specific SFR and find that, assuming $\rm sSFR\propto (1+z)^γ$, $γ=1.89^{+0.16}_{-0.15}$ for the H$α$-based SFMS, and $γ=1.36^{+0.13}_{-0.13}$ for the UV-based SFMS. Analyzing the observed H$α$/UV luminosity ratios and star-formation histories from the {\sc prospector} fitting code, we find that 41--60\% of the sample is inconsistent with having a constant star-formation history. Finally, we find tentative evidence for shorter SFR burst timescales with increasing redshift based on the distribution of $\rm L_{Hα}/νL_{ν,1600}$ vs. $Δ\rm \log(L_{Hα})$. Taken together, these results are consistent with theoretical predictions of bursty star formation in the early Universe and provide valuable constraints for theoretical models of galaxy evolution.

Figures (3)

• Figure 1: Stellar mass and UV magnitude distributions vs. redshift for the JADES and AURORA spectrscopic sample.
• Figure 2: UV magnitude (at 1600Å) vs. stellar mass for galaxy populations in the range $1.4 < z < 9$. The gray 2D histogram from $1.4 < z < 2.7$ represents measurements of galaxies in GOODS-N and GOODS-S from the 3D-HST program 2014ApJS..214...24S, while the gray 2D histograms at $z>2.7$ are from 2024MNRAS.535.2998S. The points in blue show the full spectroscopic sample that we analyze in this study, and the points in red show the galaxies which we use in our analysis of the SFMS. The blue, red, and gray triangles are the median $M_{UV}$ of the full spectroscopic sample, the SFMS sample, and the 3D-HST and 2024MNRAS.535.2998S samples, respectively, in bins of stellar mass. The error bars on the triangles show the 1$\sigma$ dispersion in $M_{UV}$ for each of the bin. We plot a vertical dashed black line at $10^{8.5}\rm\ M_\odot$ to visualize the representativeness limit. The blue diagonal dashed line shows the best-fit $M_{UV}$ vs. stellar mass relation from 2024MNRAS.535.2998S.
• Figure 3: Comparison of the SFRs of 625 galaxies evaluated using SED-based vs. empirical estimates. Top panel: Comparison of the H$\alpha$-based SFR vs. the SFR from prospector SED fitting averaged over the most recent 10 Myr. Bottom panel: Comparison of the UV-based SFR vs. the SFR from prospector SED fitting averaged over the most recent 100 Myr. In both panels, the 1:1 line is shown as a black dashed line.