JWST CAPERS: Spectroscopic evaluation of very high redshift galaxy candidates

arXiv:2610.12322 · astro-ph.GA · Submitted 2026-10-08 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: I'm Vera, and with me are Jocelyn and Subrahmanyan, guest researcher.

Jocelyn: Today's paper: "JWST CAPERS: Spectroscopic evaluation of very high redshift galaxy candidates".

Vera: The gist JWST/NIRCam imaging has enabled efficient selection of a large number of galaxy candidates at very high redshifts.

Jocelyn: First, who's behind it and why it matters.

Paper summary: Vera: So, looking at the full paper, "JWST CAPERS: Spectroscopic evaluation of very high redshift galaxy candidates," the authors are really focusing on quantifying those different selection criteria and proving which one is superior.

Jocelyn: They argue that for selecting galaxies at z greater than eight point five, the use of a UV LF prior on the posterior redshift distribution from SED fitting is significantly more effective than just using a flat redshift prior or relying only on signal-to-noise ratios.

Subrahmanyan: From a theoretical standpoint, this method suggests that incorporating knowledge about how the ultraviolet luminosity function evolves with redshift directly into the selection process provides a much stronger constraint on the resulting sample of high-redshift galaxies.

Vera: It means if we want to study galaxy evolution out to z = fourteen we should probably move away from just using conservative SNR cuts or simple break strength requirements and start thinking about incorporating these more sophisticated priors <ref:2610.12322#pg1>.

Jocelyn: The paper shows that this approach improves the success rate from about thirty-five percent up to seventy-seven percent, which is a substantial increase in the number of high-redshift galaxies we can actually confirm spectroscopically.

Subrahmanyan: That jump in confirmation efficiency is what makes this method important for building robust statistical analyses of galaxy populations at these extreme epochs.

Vera: And it also shows that even when you look at the sources they did confirm, there are some systematic biases in how those UV slopes were measured using photometric redshifts, which they found was driven by an average overestimation of the photometric redshifts.

Jocelyn: So, the main implication for us is that while JWST is finding a lot of candidates at high redshift, we need to be careful about how we filter them because our initial assumptions in selection can introduce biases in what we measure later on.

Subrahmanyan: The paper concludes that the derived photometric redshifts of these high-z sources are actually quite robust and invariant to precise models of the evolving UV LF when considering current observational constraints at z = nine to fourteen <ref:2610.12322#pg1>.

Vera: So, in short, "JWST CAPERS: Spectroscopic evaluation of very high redshift galaxy candidates" gives us a concrete tool—the UV LF prior—that we should use for future selection efforts targeting galaxies at z greater than eight point five.

Conclusion: Vera: So, this paper, "JWST CAPERS: Spectroscopic evaluation of very high redshift galaxy candidates," it’s basically a deep dive into how we find those super distant galaxies with JWST and whether our selection methods are actually working.

Jocelyn: Right, so it takes all these photometric candidates from the imaging and tests them using real spectroscopic data to see if they are really where they claim to be, specifically at redshifts greater than eight point five.

Subrahmanyan: The authors are looking at a bunch of different ways people try to pick these high-z candidates and ranking them based on how well those methods actually confirm the sources.

Vera: They’re comparing things like just using the brightness signal, or checking the shape of that Lyman-alpha break, against something more complex.

Jocelyn: And what they really highlight is that using a UV luminosity function prior on the redshift distribution from a spectral energy distribution fit is way better than those simpler cuts.

Subrahmanyan: That means when we’re trying to count or study these faint, distant galaxies, that statistical information about the overall galaxy population helps us pick the right ones much more effectively.

Vera: They did confirm forty-eight sources at redshift eight point five and above from one hundred eighty-three initial candidates.

Jocelyn: And they found that even though they have a lot of numbers, there are still some systematic biases in how the photometric redshifts were measured, specifically affecting the UV slope measurements.

Subrahmanyan: The paper suggests that this bias is related to how photometric redshifts are interpreted when modeling specific features like Balmer breaks versus emission lines.

Vera: So what does this mean for us? It tells us that we need to be careful with our selection criteria because the way we filter the candidates directly influences the properties we measure later on.

Jocelyn: Exactly, it’s a reminder that getting a high confirmation rate isn't just about finding more candidates; it’s about making sure those candidates are actually clean for what they represent.

Subrahmanyan: It shows that incorporating astrophysical knowledge, like how the UV luminosity function should behave at different redshifts, is a critical piece of the puzzle for using these powerful telescopes.

Vera: We’ve seen how good JWST is at finding these things, and now we see exactly how to use those findings to build a more reliable census of the very early universe.

Jocelyn: It sets a clear direction for future work on high-redshift surveys, pushing us toward these more informed selection methods.

Callum T. Donnan, Mark Dickinson, Lorenzo Napolitano, Pablo Arrabal Haro, Anthony J. Taylor, Hollis B. Akins, Ryan Begley, Davide Bevacqua Adam C. Carnall Marco Castellano Fergus Cullen James S. Dunlop Steven L. Finkelstein Carter Flayhart Adriano Fontana Mauro Giavalisco Michaela Hirschmann Ray A Lucas Derek J McLeod Ross J McLure Casey Papovich Laura Pentericci Borja Perez-D´ıaz Thomas M Stanton L Y Aaron Yung

NSF’s National Optical-Infrared Astronomy Research Laboratory · INAF - Observatorio Astronomico di Roma · Center for Space Sciences and Technology, UMBC · Astrophysics Science Division, NASA Goddard Space Flight Center · Department of Astronomy, The University of Texas at Austin · Cosmic Frontier Center, The University of Texas at Austin · Department of Astrophysical Sciences, Princeton University · Armagh Observatory and Planetarium College Hill Armagh Observatory and Planetarium Institute for Astronomy University of Edinburgh Royal Observatory Edinburgh INAF - Osservatorio Astronomico di Roma Laboratory for Multiwavelength Astrophysics School of Physics and Astronomy Rochester Institute of Technology Department of Physics and Astronomy Texas A&M University George P. and Cynthia Woods Mitchell Institute for Fundamental Physics and Astronomy National Tsing Hua University

astro-ph.GA

Submitted: 2026-10-08

Updated: 2026-10-08

Comments: 24 pages, 16 figures, submitted to ApJL

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 77/100

The gist: The gist JWST/NIRCam imaging has enabled efficient selection of a large number of galaxy candidates at very high redshifts.

Key concepts

UV Luminosity Function (UV LF) Prior
This method uses a statistical model of how galaxy luminosities change with redshift to predict the most likely redshift ($p(z)$) for a source. Instead of relying on simple filters, this prior guides the selection process. It significantly improved the success rate for finding high-redshift galaxies from 35% to 77%, making it a superior selection tool compared to standard cuts.
Lyman-$ ext{ m }α$ Break
This is a sharp drop in light observed in the ultraviolet spectrum of a galaxy caused by neutral hydrogen absorption. It occurs when the rest-frame UV light from distant galaxies passes through intervening neutral gas. A stronger break indicates a higher redshift, making it an important feature for confirming high-$z$ candidates.
$δψ^2$ ($δψ^2$) between SED Models
This term measures the statistical difference between two models of a galaxy's spectral energy distribution (SED): one assuming a low redshift and one assuming a high redshift. A larger $δψ^2$ suggests that the high-redshift model is much better at fitting the observed data, indicating stronger evidence for a very high redshift.
Photometric Redshift Overestimation
This occurs when photometric methods incorrectly estimate a source's redshift. In this study, sources were found to be systematically overestimated by an average of $δz ext{ } ext{ } = 0.37$ for confirmed high-$z$ candidates. This bias causes them to appear bluer in UV slope measurements than they truly are when using spectroscopic redshifts.

Terminology

Summary

The gist JWST/NIRCam imaging has enabled efficient selection of a large number of galaxy candidates at very high redshifts.

Selection Methodology Evaluation

The study evaluates the selection methodology of galaxy candidates at z ≥ 8.5 using 183 candidates with secure spectroscopic redshifts obtained from the CAPERS JWST/NIRSpec PRISM program The success of high-z confirmation is strongly correlated with the photometric signalto-noise ratio (SNR) in the detection filter, the strength of the Lyman-α break, and the ∆χ2 between the best-fitting low-redshift and high-redshift spectral energy distribution model However, the most efficient selection method is the use of a UV LF prior on the inferred p(z) of a source, raising the success rate from ≃ 35% to ≃ 77% while achieving greater completeness than conservative SNR and ∆χ2 cuts

Spectroscopic Confirmation Results

Out of the 183 candidates we spectroscopically confirm 48 sources at z ≥ 8.5 via detections of emission lines and/or the Lyman-α break (particularly important for those with no emission line detection) The highest-redshift source in CAPERS as UDS-111893 at zspec = 12.98, with no sources confirmed at z > 13 The spectroscopic redshift measurement is performed by deriving an SNR-weighted centroid for sources with emission lines detected at SNR > 3, or by modeling the Lyman-α break for sources without emission lines

Effectiveness of Selection Criteria

The paper tests several selection techniques on the VHZ sample to quantify their effectiveness The success rate of selecting sources at z ≥ 8.5 rises with increasing “VHZ count”, defined as the number of contributed subsamples that selected the target as a VHZ candidate, rising from ∼ 7% to ∼ 100% from VHZ count = 1 to 5 The most efficient selection technique tested is the use of a UV LF prior on the posterior redshift distribution from SED fitting, which improves the success rate of selecting galaxies at z > 8 significantly from ≃ 35% to ≃ 77% while only suffering a moderate drop in completeness (≃ 79% to ≃ 69%)

Nature of Redshift Failures

A number of VHZ candidates that are formally considered redshift failures (dz > 0.15) are still confirmed at z ≥ 8 These failures often occur because photometric redshift fitting interprets the strong flux in F410M and F444W incorrectly as a Balmer break rather than strong [O III] λ4959, 5007 and Hβ emission Low-z contaminants at 8.5 ≤ zphot ≤ 13 are characterized by bluer median UV slopes compared to the total VHZ candidate sample

Observational Properties of Confirmed Sources

The observational properties of the 48 sources we successfully confirm reveal a systematic bias in photometric UVslope measurements towards moderately bluer UV slopes due to an average overestimation of photometric redshifts This systematic overestimation drives a systematic bias towards bluer UV slopes when measured using photometric redshifts, than in reality becomes apparent with the use of the spectroscopic redshifts

Conclusion and Recommendation

Overall, this work quantifies a number of different selection criteria with the most efficient being the use of a UV LF prior on the derived p(z) and therefore we recommend this for future selection of galaxies at z ≥ 8.5 The use of a UV LF prior over the typically assumed flat redshift prior is significantly the most effective selection technique tested in this work

How it works

The study utilizes JWST/NIRCam imaging across three legacy extragalactic fields—COSMOS, UDS and EGS—to target very high redshift galaxy candidates Spectroscopic redshifts are measured using NIRSpec PRISM mode, employing either SNR-weighted centroids for sources with detected emission lines or modeling the Lyman-α break for those without lines The analysis tests various photometric selection criteria, including SNR cuts, minimum break strength requirements (e.g., > 1.5 mag AB), and the ∆χ2 between low-redshift and high-redshift solutions SED fitting templates are tested using sets like “tweak fsps QSF v12 v3” to determine the best photometric redshifts, showing that including sufficiently blue spectral templates is important for achieving high completeness when selecting galaxies at z ≥ 8.5

Key Findings on Selection Criteria

The success of high-redshift spectroscopic confirmation is strongly correlated with the signal-tonoise in the detection filter, the strength of the Lyman-α break and the ∆χ2 between the low and high redshift solutions A strong ∆χ2 requirement is more effective at maintaining completeness while successfully removing low-redshift contaminants than simply strong SNR requirements The use of a UV LF prior on the posterior redshift distribution from SED fitting is highly efficient at selecting high-redshift galaxies while maintaining high completeness outperforming even the most conservative SNR and ∆χ2 requirements

Systematic Bias in Properties

The analysis of the 48 confirmed sources reveals a systematic overestimation in photometric redshifts of ⟨∆z⟩ = 0.37 among the spectroscopically confirmed sources at zspec ≥ 8.5 with dz < 0.15 This average photometric redshift overestimation drives a systematic bias towards bluer UV slopes when measured using photometric redshifts, than in reality becomes apparent with the use of the spectroscopic redshifts The median UV-slope changes from β ≃ −2.5 to β ≃ −2.3 when moving from photometric to spectroscopic measurements

Future Directions

The paper suggests that stringent ∆χ2 requirements may be more efficient than stringent SNR or Lyman-α break strength requirements The use of a UV LF prior on the derived p(z) is recommended for future selection of galaxies at z ≥ 8.5 This method is shown to be significantly the most effective selection technique tested in this work The final conclusion is that the derived photometric redshifts of high-z sources are invariant to precise model of the evolving UV LF when considering current observational constraints on the UV LF at z = 9 − 14

Appendix Details

In Appendix A, Lyman-α break modeling for redshift measurement is described in detail, testing the need for additional damping through a local neutral hydrogen reservoir The results of this fitting show generally good accuracy but with a slight average overestimation of the spectroscopic redshift of ∆z ≃ 0.03 A possible cause is a failure to allow sufficient damping of the Lyman-α break in our model which motivates the inclusion of a further free parameter which parameterizes the column density of a local hydrogen reservoir The results of fitting with this additional prior causes an underestimation in spectroscopic redshift of ∆z ≃ 0.18 with increased scatter suggesting that for a number of sources, the Lyman-α break is being modeled a local column density too high and therefore worsening the average best-fitting Lyman-α break redshift

Notable Sources

The CAPERS confirmed VHZ sample contains notable sources such as UDS-111893, which is the highest redshift source in CAPERS at zspec = 12.

Improvements for AI systems

  1. This improved system can perform spectroscopic quantification of JWST/NIRCam photometric selection at z ≥ 8.5 by utilizing a UV LF prior on the inferred p(z) of a source, which raising the success rate from ≃ 35% to ≃ 77% while achieving greater completeness than conservative SNR and ∆χ2 cuts.

  2. The system can accurately assess selection methodology by quantifying success rates as a function of various criteria, such as the strength of the Lyman-α break or the ∆χ2 between the best-fitting low-redshift and high-redshift spectral energy distribution model, providing insights into which criteria are more efficient at maintaining completeness while successfully removing low-redshift contaminants.

  3. The system can classify redshift failures by analyzing observational properties, noting that it can identify sources where photometric redshift fitting interprets the strong flux in F410M and F444W incorrectly as a Balmer break rather than strong [O III] λ4959, 5007 and Hβ emission.

  4. The system can provide calibrated observational properties for confirmed high-redshift galaxies, such as the rest-frame UV absolute magnitude of the galaxies at zspec ≥ 8.5 and rest-frame UV continuum slopes, β, noting that these measurements reveal a systematic bias in photometric UVslope measurements towards moderately bluer UV slopes due to an average overestimation of photometric redshifts.

Abstract

JWST/NIRCam imaging has enabled efficient selection of a large number of galaxy candidates at very high redshifts. Accurate and robust selection is crucial for statistical analyses, such as measurement of the ultraviolet luminosity function (UV LF). A variety of selection methods and criteria have been proposed in the literature. We evaluate the selection methodology of galaxy candidates at z at least8.5 using 183 candidates with secure spectroscopic redshifts obtained from the CAPERS JWST/NIRSpec PRISM program. The success of high- z confirmation is strongly correlated with the photometric signal-to-noise ratio (SNR) in the detection filter, the strength of the Lyman- α break, and the Δχ squared between the best-fitting low-redshift and high-redshift spectral energy distribution model. However, the most efficient selection method is the use of a UV LF prior on the inferred p(z) of a source, raising the success rate from 35% to 77% while achieving greater completeness than conservative SNR and Δχ squared cuts. Out of the 183 candidates we spectroscopically confirm 48 sources at z at least8.5 via detections of emission lines and/or the Lyman- α break (particularly important for those with no emission line detection). We report the highest-redshift source in CAPERS as UDS-111893 at z spec=12.98, with no sources confirmed at z>13. This work provides spectroscopic quantification of JWST/NIRCam photometric selection at z at least8.5, and demonstrates the effectiveness of a UV LF prior for high-redshift photometric selection.

Sources

Related papers