BEACON: JWST NIRCam Pure-parallel Imaging Survey. IV. A Systematic Search for Galaxy Overdensities and Evidence for Gas Accretion Mode Transition

TL;DR

BEACON presents a systematic search for galaxy overdensities at $1.5<z<5$ using JWST/NIRCam pure-parallel data across 20 fields. By applying a weighted-adaptive kernel density estimator to full photometric redshift PDFs, the study identifies $207$ significant overdensities and derives their halo masses via an updated stellar-to-halo mass relation, revealing assembly in halos above $\sim10^{12}\,M_\odot$. The analysis uncovers two quenching pathways tied to halo mass and redshift: a hot, massive-halo regime suppressing star formation by $z\sim2$, and a cold-in-hot regime at higher redshift where cold streams sustain activity. Local-density effects are significant for $z<2$, with higher density correlating with larger stellar masses and lower sSFR, while trends weaken or reverse at $z>2$, consistent with evolving gas supply and accretion modes. Overall, the results illuminate the complex interplay between individual galaxies and their large-scale environments, marking environmental quenching as it begins to emerge in the cosmic noon epoch, and motivate spectroscopic follow-up to confirm the structures and test gas-accretion models.

Abstract

We systematically search for galaxy overdensities using 20 independent fields with a minimum of six filters (F090W, F115W, F150W, F277W, F356W, and F444W) from BEACON, the JWST Cycle 2 NIRCam pure-parallel imaging survey. We apply an adaptive kernel-density estimation method that incorporates the full photometric redshift probability distribution function of each galaxy to map galaxy overdensities, and identify 207 significant ($>4\,σ$) overdensities at $1.5<z<5$. We measure the quenched-galaxy fraction, the median specific star formation rate (sSFR), the total halo mass, and the local galaxy density of each system. By investigating the correlation among these observables, we find that galaxy quenching proceeds in two paths:($i$) Overdensities within more massive halos exhibit higher quenched fractions and lower averaged sSFRs. This trend weakens at $z\gtrsim2$, consistent with cold gas streams penetrating shock-heated massive halos and sustaining star formation activity at early times. ($ii$) We also find a dependence of the same parameters on local densities at $z<2$, where the quenched fraction increases and the sSFR decreases toward higher densities. The environmental trend in sSFR weakens at $z\sim2$--$3$ and shows tentative evidence for a reversal at $z>3$, potentially due to a larger cold gas supply in earlier times. Our study reveals a complex interplay between individual galaxies and large-scale environmental properties, marking the onset of environmental effects on galaxy quenching in massive halos at cosmic noon.

Figures (12)

• Figure 1: Comparison between eazy-derived photometric redshifts and spectroscopic redshifts compiled from various previous studies. Red dotted line corresponds to the outlier limit between photometric redshift and spectroscopic redshift. Grey dots correspond to the outliers.
• Figure 2: Mass completeness limit for the 20 selected BEACON fields. The 2D hexagon bins show the stellar mass distributions at $0<z<8$. Red open circles show the mass completeness limits calculated in the redshift bins with a step of 0.7, with the red line indicating its quadratic fitted function.
• Figure 3: The flowchart of the overdensity map creation, following the direction of arrows. (1) Weighting of each galaxy is calculated by its photometric redshift PDF. (2) Each field is divided into redshift slices following the definition from Section \ref{['sec:zslice']} (3) For each slice (red box), calculate surface density using Equation \ref{['eq:1']} around each galaxy (red dot) based on all surrounding galaxies. Different dot sizes represent different weights. (4) The surface density of each galaxy, where larger size and opacity corresponds to larger local surface density. (5) Dividing the mosaic into defined pixel grid. (6) Calculation on each pixel grid $(x,y)$ following Equation \ref{['eq:sigmaxy']}. (7) Iterative calculation of global bandwidth. (8) Optimal global bandwidth selection based on LCV. (9) Map of local bandwidth of each galaxy calculated with $h_i = h\times\lambda_i$. (10) Edge correction factor map. (11) The final overdensity map with contour step of $\Delta \delta =1$. Each data point corresponds to galaxies that are located inside the redshift slice. Star markers show galaxies that are spatially included in the overdense region.
• Figure 4: Example of a two-dimensional overdensity map of the beacon_0055–3749 field at $z = 2.20$. The map is plotted in the $x$–$y$ plane, converted into comoving coordinates in the unit of cMpc. White contours indicate one-step increments of the overdensity factor ($\delta$). White dots mark galaxies that have a probability of being included within the map, accounting for their photometric redshift uncertainties. The star symbols denote galaxies located within the $>1\sigma$ overdensity contour, representing the probable members of the overdense region. The inset panel in the bottom-left corner shows the photometric redshift distribution of galaxies within the redshift slice.
• Figure 5: The distribution of halo mass of our overdensities as a function of redshift. Error bar represents 1$\sigma$ error of the halo mass estimation from bootstrapping. Solid circle: data points from this study, color-coded by their measured quiescent galaxy fraction (top) and median of specific star formation rate (bottom). Solid lines: dividing lines of the halo accretion modes to cold, hot, and cold in hot modes as predicted by 2006MNRAS.368....2D. Dashed lines: median halo mass growth history for each present-day halo with a masses of $\log{(M_h(z=0))}=12.0, 13.0, 14.0, \text{and } 15.0$, derived by 2013ApJ...770...57B. Small gray dot: local clusters detected with Sunyaev-Zel'dovich (SZ) effects with ROSAT All-Sky Survey (X-ray) or Planck Satellite (FIR) 2019AA...626A...7T, South Pole Telescope 2019ApJ...878...55B, and Atacama Cosmology Telescope 2025arXiv250721459A. Grey hex-plots: groups and protoclusters identified by 2022ApJ...933....9L using halo-based group finder. Pentagon (green): protocluster candidates from 2023ApJ...943..153B using a similar method to this study. Square (purple): high redshift protoclusters compiled from various studies. See Section \ref{['sec:halomass']} for detailed references.