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

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.

💡 Research Summary

**
This paper presents a systematic search for galaxy overdensities in the JWST Cycle 2 BEACON pure‑parallel NIRCam imaging survey, exploiting 20 independent fields that each contain at least six broadband filters (F090W, F115W, F150W, F277W, F356W, and F444W). The total surveyed area is ≈ 400 arcmin² with a 10σ depth of m ≈ 27.3–28 AB mag in F444W, enabling detection of low‑mass galaxies out to z ≈ 7. Source detection was performed on a weighted combination of the three longest‑wavelength NIRCam images, and photometry was measured in 0.16″ apertures and corrected to total fluxes. Galactic extinction corrections were applied using the Milky Way reddening law.

Photometric redshifts were derived with eazy‑py, employing a custom FSPS‑based template set (QSF12 v3) and a magnitude prior based on F444W. Crucially, the full redshift probability distribution function, P(z), was retained for every object, allowing the authors to incorporate redshift uncertainties directly into the overdensity analysis. Validation against spectroscopic samples in overlapping legacy fields (COSMOS, EGS, GOODS‑S, UDS) yields a normalized median absolute deviation σ_NMAD = 0.024 and an outlier fraction < 10 % for galaxies with z_phot > 1.5, confirming the reliability of the photometric redshifts in the redshift range of interest.

Physical parameters (stellar mass, star‑formation rate, specific SFR, dust attenuation) were obtained via CIGALE SED fitting, fixing the redshift to the median of the PDF (z₅₀). The authors adopted a delayed‑exponential star‑formation history, a Chabrier IMF, Bruzual & Charlot (2003) stellar population models, a metallicity grid Z = 0.0001–0.05, and included nebular emission based on Inoue (2011) with log U = −2. Dust attenuation follows a modified Calzetti law with E(B−V) ranging from 0 to 1.5.

The core of the analysis is a weighted adaptive kernel density estimation (wKDE) technique that maps three‑dimensional galaxy density (RA, Dec, redshift) by convolving each galaxy’s P(z) with an adaptive kernel whose bandwidth varies with local density. This method fully exploits the redshift PDFs, avoiding the loss of information inherent in single‑point redshift estimates. Overdensities are defined as regions exceeding the background mean by > 4σ. Applying this to the 1.5 < z < 5 interval yields 207 statistically significant overdensities.

For each overdensity the authors estimate: (1) total halo mass via a halo‑occupation‑distribution (HOD) model calibrated to the observed galaxy counts; (2) the quenched‑galaxy fraction, defined as the proportion of galaxies with sSFR < 10⁻¹¹ yr⁻¹; (3) the median sSFR of the member galaxies; and (4) a local galaxy density measured as the surface density of galaxies within a 0.5 Mpc projected radius around the overdensity centre. Halo masses span 10¹³–10¹⁴ M⊙, quenched fractions range from ~10 % to ~70 %, and median sSFR values decline from ≈10⁻⁸ yr⁻¹ at low redshift to ≈10⁻⁹ yr⁻¹ at z ≈ 2, with a notable rise again at higher redshift.

Two principal trends emerge from the correlation analysis. First, more massive halos host higher quenched fractions and lower median sSFRs, consistent with the classic picture of halo‑mass‑driven quenching. However, this trend weakens for z ≳ 2, suggesting that at earlier epochs cold‑flow streams can penetrate shock‑heated massive halos, sustaining star formation despite the large halo mass. This observation aligns with theoretical expectations of a gas‑accretion mode transition (cold → hot) around the “cosmic noon” epoch (z ≈ 2), as described by Dekel & Birnboim (2006) and subsequent simulations.

Second, the local galaxy density also correlates with quenching: at z < 2, higher densities correspond to higher quenched fractions and lower sSFRs, reflecting the well‑known environmental quenching seen in clusters and groups. Intriguingly, this environmental dependence diminishes around z ≈ 2–3 and shows tentative evidence of reversal at z > 3, where denser regions exhibit elevated sSFRs. The authors interpret this reversal as a signature of abundant cold‑gas supply along filaments feeding proto‑clusters at early times, effectively turning dense environments into sites of enhanced star formation rather than suppression.

Importantly, the study finds only a weak correlation between halo mass and local density, indicating that large‑scale halo properties and small‑scale environmental conditions act semi‑independently in shaping galaxy evolution. Comparison with cosmological hydrodynamic simulations confirms that the observed sSFR‑density reversal matches the epoch when cold streams dominate gas accretion, supporting the notion that the transition in gas‑accretion mode drives the observed environmental trends.

Overall, the paper demonstrates the power of JWST pure‑parallel imaging combined with full PDF‑aware density estimation to probe galaxy environments at high redshift with reduced cosmic‑variance bias. The results provide compelling observational evidence that both halo mass and local environment influence galaxy quenching, but that their relative importance evolves with redshift in a manner consistent with a transition from cold‑flow‑dominated to hot‑halo‑dominated gas accretion. These findings advance our understanding of how massive structures and the cosmic web jointly regulate star formation during the peak epoch of cosmic activity.