Draft version January 10, 2023

Typeset using LATEX twocolumn style in AASTeX631

The UNCOVER Survey:

A first-look HST+JWST catalog of 50, 000 galaxies near Abell 2744 and beyond
John R. Weaver ,1 Sam E. Cutler ,1 Richard Pan ,2 Katherine E. Whitaker ,1, 3 Ivo Labbé ,4
Sedona H. Price ,5 Rachel Bezanson ,5 Gabriel Brammer ,6 Danilo Marchesini ,2 Joel Leja ,7, 8, 9
Bingjie Wang (王冰洁) ,7, 8, 9 Lukas J. Furtak ,10 Adi Zitrin ,10 Hakim Atek ,11 Dan Coe ,12, 13, 14
Pratika Dayal ,15 Pieter van Dokkum ,16 Robert Feldmann ,17 Natascha Förster Schreiber ,18
arXiv:2301.02671v1 [astro-ph.GA] 6 Jan 2023

Marijn Franx ,19 Seiji Fujimoto ,20, ∗ Yoshinobu Fudamoto ,21, 22 Karl Glazebrook ,4
Anna de Graaff ,23 Jenny E. Greene,24 Stéphanie Juneau ,25 Susan Kassin ,12 Mariska Kriek ,19
Gourav Khullar ,5 Michael Maseda ,26 Lamiya A. Mowla ,27 Adam Muzzin ,28 Themiya Nanayakkara ,4
Erica J. Nelson ,29 Pascal A. Oesch ,30, 6 Camilla Pacifici ,12 Casey Papovich ,31, 32 David J. Setton ,5
Alice E. Shapley ,33 Renske Smit ,34 Mauro Stefanon ,35, 36 Edward N. Taylor ,4 Andrea Weibel ,30 and
Christina C. Williams 25, 37
1 Department of Astronomy, University of Massachusetts, Amherst, MA 01003, USA
2 Department of Physics and Astronomy, Tufts University, 574 Boston Ave., Medford, MA 02155, USA
3 Cosmic Dawn Center (DAWN), Denmark
4 Centre for Astrophysics and Supercomputing, Swinburne University of Technology, Melbourne, VIC 3122, Australia
5 Department of Physics and Astronomy and PITT PACC, University of Pittsburgh, Pittsburgh, PA 15260, USA
6 Cosmic Dawn Center (DAWN), Niels Bohr Institute, University of Copenhagen, Jagtvej 128, København N, DK-2200, Denmark
7 Department of Astronomy & Astrophysics, The Pennsylvania State University, University Park, PA 16802, USA
8 Institute for Computational & Data Sciences, The Pennsylvania State University, University Park, PA 16802, USA
9 Institute for Gravitation and the Cosmos, The Pennsylvania State University, University Park, PA 16802, USA
10 Physics Department, Ben-Gurion University of the Negev, P.O. Box 653, Beer-Sheva 8410501, Israel
11 Institut d’Astrophysique de Paris, CNRS, Sorbonne Université, 98bis Boulevard Arago, 75014, Paris, France
12 Space Telescope Science Institute (STScI), 3700 San Martin Drive, Baltimore, MD 21218, USA
13 Association of Universities for Research in Astronomy (AURA), Inc. for the European Space Agency (ESA)
14 Center for Astrophysical Sciences, Department of Physics and Astronomy, The Johns Hopkins University, 3400 N Charles St.

Baltimore, MD 21218, USA

15 Kapteyn Astronomical Institute, University of Groningen, P.O. Box 800, 9700 AV Groningen, The Netherlands
16 Department of Astronomy, Yale University, New Haven, CT 06511, USA
17 Institute for Computational Science, University of Zurich, Winterhurerstrasse 190, CH-8006 Zurich, Switzerland
18 Max-Planck-Institut für extraterrestrische Physik, Gießenbachstraße 1, 85748 Garching, Germany
19 Leiden Observatory, Leiden University, P.O.Box 9513, NL-2300 AA Leiden, The Netherlands
20 Department of Astronomy, The University of Texas at Austin, Austin, TX 78712, USA
21 Waseda Research Institute for Science and Engineering, Faculty of Science and Engineering, Waseda University, 3-4-1 Okubo,

Shinjuku, Tokyo 169-8555, Japan

22 National Astronomical Observatory of Japan, 2-21-1, Osawa, Mitaka, Tokyo, Japan
23 Max-Planck-Institut für Astronomie, Königstuhl 17, D-69117, Heidelberg, Germany
24 Department of Astrophysical Sciences, 4 Ivy Lane, Princeton University, Princeton, NJ 08544, USA
25 NSF’s National Optical-Infrared Astronomy Research Laboratory, 950 N. Cherry Avenue, Tucson, AZ 85719, USA
26 Department of Astronomy, University of Wisconsin-Madison, 475 N. Charter St., Madison, WI 53706 USA
27 Dunlap Institute for Astronomy and Astrophysics, 50 St. George Street, Toronto, Ontario, M5S 3H4, Canada
28 Department of Physics and Astronomy, York University, 4700 Keele Street, Toronto, Ontario, ON MJ3 1P3, Canada
29 Department for Astrophysical and Planetary Science, University of Colorado, Boulder, CO 80309, USA
30 Department of Astronomy, University of Geneva, Chemin Pegasi 51, 1290 Versoix, Switzerland
31 Department of Physics and Astronomy, Texas A&M University, College Station, TX, 77843-4242 USA
32 George P. and Cynthia Woods Mitchell Institute for Fundamental Physics and Astronomy, Texas A&M University, College Station,

TX, 77843-4242 USA

Corresponding author: John R. Weaver

john.weaver.astro@gmail.com
2 J. R. Weaver et al.

33 Physics& Astronomy Department, University of California: Los Angeles, 430 Portola Plaza, Los Angeles, CA 90095, USA
34 AstrophysicsResearch Institute, Liverpool John Moores University, 146 Brownlow Hill, Liverpool L3 5RF, UK
35 Departament d’Astronomia i Astrofisica, Universitat de Valencia, C. Dr. Moliner 50, E-46100 Burjassot, Valencia, Spain
36 Unidad Asociada CSIC “Grupo de Astrofisica Extragalactica y Cosmologi” (Instituto de Fisica de Cantabria - Universitat de Valencia)
37 Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721, USA

ABSTRACT
In November 2022, the James Webb Space Telescope (JWST ) returned deep near-infrared images
of Abell 2744 – a powerful lensing cluster capable of magnifying distant, incipient galaxies beyond
it. Together with the existing Hubble Space Telescope (HST ) imaging, this publicly available dataset
opens a fundamentally new discovery space to understand the remaining mysteries of the formation and
evolution of galaxies across cosmic time. In this work, we detect and measure some 50,000 objects across
the 45 arcmin2 JWST footprint down to a 5 σ limiting magnitude of ∼29.9 mag in 0.3200 apertures.
Photometry is performed using circular apertures on images matched to the point spread function of
the reddest NIRCam band, F444W, and cleaned of bright cluster galaxies and the related intra-cluster
light. To give an impression of the photometric performance, we measure photometric redshifts and
achieve a σNMAD ≈ 0.03 based on known, but relatively small, spectroscopic samples. With this
paper, we publicly release HST and JWST PSF-matched photometric catalogs optimized for bright
and extended sources (0.700 apertures) and compact and faint sources (0.3200 apertures) along with
basic photometric redshifts, rest-frame colors, and individual magnification estimates. These catalogs
will set the stage for efficient and deep spectroscopic follow-up of the first JWST -selected samples in
Summer 2023.

Keywords: Catalogs (205), Abell clusters (9), Photometry (1234), James Webb Space Telescope (2291),
Hubble Space Telescope (761), Astronomical methods (1043)

1. INTRODUCTION
The vast distance scales of our Universe relative to the
human timescale implicitly means that there are very
few astrophysical processes we can observe changing in
real time. Dynamical processes in galaxies transpire over
timescales of millions to billions of years. Thus to under-
stand the formation and evolution of galaxies across cos-
mic time necessitates the study of statistically represen-
tative snapshots. Observational campaigns are forced to
make decisions in survey design, generally prioritizing ei-
ther depth (e.g., Williams et al. 1996; Giavalisco et al.
2004; Beckwith et al. 2006; Bouwens et al. 2011; Illing-
worth et al. 2013, 2016; Lotz et al. 2017), volume (e.g.,
Scoville et al. 2007; Jarvis et al. 2013; Aihara et al. 2018;
Abbott et al. 2018), or a mix therein (e.g., Grogin et al.
2011; Koekemoer et al. 2011), in order to assemble un-
biased galaxy populations. All of these surveys share a
common theme: they require a synthesis of the panchro-
matic flux measurements for detected sources into a pho-
tometric catalog as the first step towards modeling the
stellar populations. Such photometric catalogs serve as
the foundation of any galaxy survey, necessary for iden-
∗ Hubble Fellow
tifying robust galaxy samples and for enabling the vast
majority of subsequent science investigations.
The deepest surveys of our Universe to date come
from single ultra-deep pointings with the Hubble Space
Telescope (HST ), either of “blank fields” (i.e., relatively
dark lines of sight through our own Milky Way galaxy;
Williams et al. 1996; Beckwith et al. 2006; Bouwens
et al. 2011; Illingworth et al. 2013) or by targeting known
clusters of galaxies at intermediate redshifts (Lotz et al.
2017; Coe et al. 2019; Salmon et al. 2020; Sharon et al.
2020). One particular advantage of targeting galaxy
clusters is the added boost from strong gravitational
lensing; the richest clusters magnify background galax-
ies by factors of a few up to dozen typically, depend-
ing on the size and position of the background galaxy
with respect to the lens (e.g., Coe et al. 2019). Strong
lens clusters unveil some of the most distant (e.g., Zheng
et al. 2012; Coe et al. 2013; Zitrin et al. 2014; Strait et al.
2021; Bradley et al. 2022; Furtak et al. 2022a; Hsiao et al.
2022; Roberts-Borsani et al. 2022; Williams et al. 2022;
Adams et al. 2023; Atek et al. 2023) and lowest-mass
(e.g., Livermore et al. 2017; Bouwens et al. 2017, 2022;
Atek et al. 2018; Bhatawdekar et al. 2019; Kikuchihara
et al. 2020; Furtak et al. 2021) galaxies known, even sin-
gle candidate stars in some cases (Welch et al. 2022).
 The UNCOVER Photometric Catalog 3

However, this boost from cosmic telescopes comes with

a cost in terms of contamination from cluster galax-
ies/intracluster light and the complex and non-linear
distortions to galaxy morphologies (e.g., Shipley et al.
2018; Bhatawdekar et al. 2019; Pagul et al. 2021; Fox
et al. 2022). In addition, the source-plane area that is
8
being probed behind a lens is smaller. As such, there is 7 5 depth
a trade-off between detecting a higher number of galax- 6 27.0
5 Daper = 0.32"
ies that are boosted in flux but for a smaller area probed
4 27.5
relative to an unlensed field; an effect known as the mag-
nification bias (e.g., Broadhurst et al. 1995). However, 3
since the luminosity function of high-redshift galaxies is 28.0

f (10 nJy)
2

steep enough (e.g., Finkelstein et al. 2015; Mason et al.
2015), the net effect is a gain in the number density of
detections.
Despite these challenges, galaxy clusters afford our
best opportunity to push to the most extreme depths
and thus to the frontiers of galaxy formation. Cam-
paigns such as the Director’s Discretionary Hubble Fron-
tier Fields program have amassed a rich archival data set
of HST imaging (Lotz et al. 2017; Steinhardt et al. 2020)
that has set the stage for JWST imaging and spectro-
scopic programs (e.g., Willott et al. 2017; Treu et al.
2022; Bezanson et al. 2022; Windhorst et al. 2023). One
cluster is particularly compelling to study: in addition to
a spectacular central core (Lotz et al. 2017; Shipley et al.
2018; Pagul et al. 2021; Kokorev et al. 2022), Abell 2744
contains prominent lensing features within two addi-
tional massive cluster sub-structures in the north and
north-west (Furtak et al. 2022b). Abell 2744 thus con-
tains an unusually large area of high magnification when
combining the various structures. Several early JWST
programs target Abell 2744; here we combine publicly
available HST and JWST photometry from the JWST-
GO-2561, JWST-DD-ERS-1324, and JWST-DD-2756
programs.
In this paper, we present the space-based photomet-
ric catalog for the UNCOVER survey (Bezanson et al.
2022), including derived photometric redshifts and mag-
nification corrections from lensing models presented in
Furtak et al. (2022b). In Section 2, we present an
overview of the data processing, including the reduc-
tion and astrometric correction of images from HST
and JWST . Section 3 describes two approaches for han-
dling the added complexity of intracluster light in the
Abell 2744 cluster. Source detection and photometry
are described in Section 4, including a description of the
methodology adopted to homogenize the point spread
function (PSF), the measurement of total fluxes from
aperture photometry, and a quantification of represen-
tative errors. We present the photometric catalog prop-
erties in Section 5, including depths, galaxy number
Mag (AB)
28.5
1 F140W
F160W 29.0
F125W
F814W F105W F150W F410M
F435W 29.5
F606W F090WF115W F200W
F444W
F277W
F356W 30.0
1.0 2.0 4.0
Wavelength ( m)
Figure 1. Effective catalog depths over the Abell 2744
JWST footprint for the 15 available photometric bands and
their transmission curves. Depths are quoted in 0.3200 di-
ameter apertures and correspond to the area-weighted 50th
(median) and 10th percentiles (dashed and solid lines, re-
spectively). Areas in HST imaging without JWST coverage
are not considered. See the text for details.
counts, and comparisons to other surveys. Finally, we
summarize our findings in Section 6. An appendix con-
tains additional relevant information regarding the sta-
bility of the JWST point spread function in time and
across the detector.
All magnitudes in this paper are expressed in the AB
system (Oke 1974), for which a flux fν in 10 × nJy
(10−28 erg cm−2 s−1 Hz−1 ) corresponds to ABν =
28.9 − 2.5 log10 (fν /µJy). When computing photo-
metric redshifts (zphot ) and corresponding rest-frame
colors, we adopt a standard ΛCDM cosmology with
H0 = 70 km s−1 Mpc−1 , ΩM = 0.3, and ΩΛ = 0.7.
2. DATA
2.1. James Webb Space Telescope
The photometric catalogs presented herein include all
public JWST /NIRCam imaging of Abell 2744 avail-
able to date: the Ultradeep NIRSpec and NIRCam
ObserVations before the Epoch of Reionization (UN-
COVER) Treasury survey (PIs Labbé & Bezanson,
JWST-GO-2561; Bezanson et al. 2022), the Early Re-
4 J. R. Weaver et al.
Filter Depth (5 σ AB) Area (arcmin2 )
10 th th
50 th
90 th
10 50th
90th
Total
F435W 29.48 29.36 28.42 0.94 5.57 15.39 18.54
F606W 29.59 28.83 27.37 0.79 17.17 31.42 36.21
F814W 29.44 28.13 27.01 1.50 14.71 26.15 31.26
F090W 29.63 29.32 29.07 1.60 7.42 11.59 12.91
F105W 29.52 27.16 27.04 3.55 12.48 15.48 20.22
F115W 29.52 29.15 28.22 6.34 26.01 41.83 45.12
F125W 29.02 27.16 27.04 3.91 12.01 14.78 20.08
F140W 28.95 28.87 28.52 1.51 3.41 4.71 5.62
F150W 29.55 29.18 28.43 5.38 25.54 41.43 44.71
F160W 29.11 26.80 26.62 2.72 11.59 15.81 20.15
F200W 29.63 29.26 28.51 5.44 25.67 41.49 44.71
F277W 29.88 29.49 28.80 5.52 25.04 40.61 44.98
F356W 29.95 29.56 28.92 5.88 25.32 40.54 45.67
F410M 29.21 28.88 28.44 3.35 15.52 25.48 28.73
F444W 29.83 29.13 28.34 6.85 25.81 40.91 45.11
Table 1. Effective catalog depths are quoted in 0.3200 diame-
ter apertures and correspond to the area-weighted 10th , 50th
(median), and 90th percentiles. Total areas reflect the union
of the LW detection footprint with the coverage available for
each band.
lease Science (ERS) GLASS-JWST program (PI: Treu,
JWST-DD-ERS-1324; Treu et al. 2022), and a Direc-
tors Discretionary (DD) program (JWST-DD-2756, PI:
Chen). As described in Bezanson et al. (2022), our
dataset combines the deep NIRCam imaging with 4-6
hour exposures in 7 filters (F115W, F150W, F200W,
F277W, F356W, F410M, and F444W) from UNCOVER,
with the ultra-deep imaging with 9-14 hour exposures
from GLASS-ERS in 7 filters (F090W, F115W, F150W,
F200W, F277W, F356W, and F444W). The GLASS-
ERS NIRCam pointing is taken in parallel and thus
offset to the cluster outskirts, thereby extending the
combined science area. Additionally, the DDT pro-
gram includes two epochs of NIRCam imaging in 6
filters (F115W, F150W, F200W, F277W, F356W, and
F444W), totaling ∼ 1 hour per filter. All together, im-
ages in 8 unique JWST filters are combined to extend
the coverage of Abell 2744 to include the nearby cluster
sub-structures.
Next, we summarize the key steps of the image reduc-
tion, referring the reader to Section 3.1 of Bezanson et al.
(2022) for further details. Imaging mosaics are produced
from the flux-calibrated NIRCam exposures released in
Stage 2b of the JWST calibration pipeline (v1.8.4) and
combined with calibration set jwst 0995.pmap. The ex-
posures are then processed, aligned, and co-added using
the Grism redshift and line analysis software for space-
based spectroscopy (Grizli1 , v1.6.0.dev99; Brammer
2019; Kokorev et al. 2022). The pipeline has been op-
timized to handle known JWST artifacts (Rigby et al.
2022). Our flat-field calibration image is custom made
from on-sky commissioning data (COM-1063), updat-
ing the official calibration files to correct for smoothly
varying large-scale structure in the flats and to further
optimize pixel-to-pixel variations. The data reduction
pipeline next subtracts a large scale sky background,
performs an astrometric alignment (see Section 2.4), and
drizzles the images to a common pixel grid of 0.0400 per
pixel using astrodrizzle (Gonzaga et al. 2012).
2.2. Hubble Space Telescope
A wide range of imaging of the Abell 2744 clus-
ter and surrounding area exists within the public HST
archive. Briefly, we summarize the programs relevant
herein. Program HST-GO-11689 (PI: Dupke) and HST-
GO-13386 (PI: Rodney) include deep HST/ACS imag-
ing in the cluster center in 3 filters (F435W, F606W, and
F814W), with program HST-DD-13495 (PI: Lotz; Lotz
et al. 2017) acquiring complementary deep HST /WFC3
observations in 4 filters (F105W, F125W, F140W, and
F160W). While each of the above programs are (deep)
individual pointings limited by the ACS and WFC3
field of view, respectively, the data was later expanded
by a factor of 4 with shallower imaging in 2 ACS fil-
ters (F606W and F814W) and 3 WFC3 filters (F105W,
F125W, and F160W) by the BUFFALO survey (Pro-
gram HST-GO-15117, PIs: Steinhardt & Jauzac; Stein-
hardt et al. 2020). Most recently, the deep optical cover-
age was further expanded by Program JWST-DD-17231
(PI: Treu). A summary of the instruments, filters, pro-
gram IDs, and orbit depths can be found in Table 3 in
Bezanson et al. (2022). Taken together, they contribute
7 unique HST filters. These images are reduced follow-
ing the same procedure as described in Section 2.1 onto
the same 0.0400 pixel grid as the JWST images.
2.3. Astrometry
Astrometric registration of the images is performed
by Grizli using F444W. We adopt star positions from
GAIA DR3 (Gaia Collaboration et al. 2016, 2022). Us-
ing their well known proper motions, the positions of
GAIA stars observed in 2015 are projected to November
2022, the epoch during which the JWST imaging was
acquired. The remaining images are then registered con-
sistently to the F444W filter. In order to independently
1 https://github.com/gbrammer/grizli
 The UNCOVER Photometric Catalog
0.06 ra = -0.002 ± 0.013" Proper motion (mas yr 1)
dec = 0.002 ± 0.010" 0 10
0.04
dec (arcsec)
0.02
0.00
0.02
0.04
0.06
0.050 0.025 0.000 0.025 0.050
ra (arcsec)
Figure 2. Astrometric performance of the imaging dataset
based on positions of bright stars in HST /F160W compared
to GAIA DR3. One and four pixel areas are shown by the
solid and dashed squares, respectively. Filled purple and grey
elliptical contours enclose roughly 50% and 90% of bright
stars, respectively. The median deviation (purple cross) and
the standard deviation of the absolute median deviation are
quoted for each axis corresponding to the innermost 50% of
stars.
test the accuracy of the astrometry, we opt to compare
to stars in the F160W filter instead of F444W, where
saturation and central star clipping is less of an issue.
We perform an additional correction to the proper mo-
tions to shift to the median epoch of the wider F160W
BUFFALO HST imaging in July 2019 (Steinhardt et al.
2020). Figure 2 demonstrates our achieved precision of
≈ 0.01200 or a third of a pixel, measured by the stan-
dard deviation of the median absolute deviation for the
innermost 50% of sources (purple shaded region). The
median bias, based on the same sources, is also small at
≈ 0.00200 , or 5% of a single pixel.
2.4. Spectroscopic redshifts
Spectroscopic redshifts (hereafter zspec ) over our sur-
vey footprint are collected using an automated query
function within Grizli. We match to sources detected
from our nominal bCG-subtracted images within a ra-
dius of 0.300 and find 501 sources with zspec values with
confidence flags 3 or 4, corresponding to secure mea-
surements. Of these, 150 are found in the NASA/IPAC
Extragalactic Database2 , 332 are cluster members with
zspec from Richard et al. (2021), and 19 are grism red-
shifts from GLASS (Treu et al. 2015). The correspond-
ing values are stored in the catalog in the z spec column.
2 https://ned.ipac.caltech.edu/
5
3. REMOVAL OF SKY, INTRA-CLUSTER LIGHT,
AND BRIGHT CLUSTER GALAXIES
In order to achieve the science objectives, we need to
mitigate the contamination of foreground light from the
many bright cluster galaxies (bCGs) and the powerful
intra-cluster light (ICL). Otherwise, the photometry of
distant sources seen through this foreground light will be
inaccurate, potentially mischaracterized, or missing al-
together the rare high-z galaxies magnified by the strong
gravitational lensing of these cluster members. We note
that throughout this paper we adopt the term bCG,
which is not synonymous with the traditional brightest
cluster galaxy (BCG).
3.1. Subtracting fitted models
For robust and tested bCG and ICL subtraction, we
adopt the method described in Ferrarese et al. (2006)
and implemented by Shipley et al. (2018) in the HFF-
DeepSpace (HFFDS) photometric catalogs of six lensing
clusters, including Abell 2744.
The bCGs that contribute significantly to the total
cluster luminosity are first identified from the HFFDS
catalogs; we refer the reader to Shipley et al. (2018)
for a more detailed description of the selection process.
We further expand our selection to accommodate the
wider footprint of the present data set (see Figure 2 in
Bezanson et al. 2022). We note that fewer bCGs are
subtracted from the F410M mosaic, which has a smaller
footprint. To minimize computation time, we generate
“cropped mosaics” using the IRAF IMCOPY task. The
boundaries of these mosaics are defined by the outermost
isophotal cluster radii.
We use Source Extractor to create a crude mask
of all background sources (excluding cluster members).
This is done by using the parameters: DETECT THRESH =
1.2, DEBLEND NTHRESH = 10, DEBLEND MINCONT = 0.01,
which identifies more isolated sources. We repeat this
detection on the masked image, providing a more accu-
rate and precise mask, especially near tightly clustered
galaxies. These two masks are combined to isolate the
cluster galaxies and ultimately smoothed with a Gaus-
sian kernel to account for nearby, poorly modeled light.
Using this mask to isolate each cluster member, we run
ELLIPSE to measure and extract the isophotal parame-
ters out to an arbitrary radius. This is then given to
BMODEL to create the galaxy model. This galaxy model
is subtracted from the cropped mosaic. This process
is repeated for each cluster member, yielding an initial
“residual mosaic”.
Although these initial models provide a good approxi-
mation of the bCG and ICL subtracted mosaic, we adopt
an iterative approach, repeating the modeling process
6 J. R. Weaver et al.

F444W Input Image Modeled & Subtracted Model

N
F (10 nJy)
30 arcsec E
30 arcsec 102 0 102 30 arcsec

F444W Input Image Modeled & Subtracted Model

N
F (10 nJy)
30 arcsec E
30 arcsec 102 0 102 30 arcsec

Figure 3. Zoom-in around the two primary cluster cores of Abell 2744 in F444W. The bright cluster galaxies and intra-cluster
light visible in the input image (left) is removed by subtracting fitted models (middle), with the models themselves on the right.

on residual mosaics. For crowded regions with multi-
ple bCGs we allow for Source Extractor to recreate
and improve the original mask as a replacement for the
master mask.
We see a convergence after 10 iterations for a total of
11 galaxy model iterations (i.e., the original galaxy mod-
eling plus 10 more iterations), similar to Shipley et al.
(2018). Out of the 11 total iterations, we average over
the 4 “best” galaxy models, omitting the first iteration.
We use the IRAF task IMCOMBINE with the following pa-
rameters: combine = “average”, reject = “minmax”,
nlow = “4”, and nhigh = “2” (e.g., see Section 3.1.3
of Shipley et al. 2018). This averaged model effectively
filters out outlier values. Subtracting the average galaxy
model from the cropped mosaic returns the final resid-
ual mosaic. Figure 3 shows the effect of this careful bCG
and ICL subtraction relative to the original mosaic near
the primary cluster core.
With this final subtracted mosaic, we revert our initial
cropping by using the IRAF IMCOPY task. We copy our
subtracted mosaic onto the original mosaic. We then
perform a sky subtraction to remove excess light near
the edges of the galaxy models using a Gaussian interpo-
lation. The background is measured using the Source
Extractor AUTO setting with the following parame-
ters: mesh size of 192 for 0.0200 pixel scale and 96 for
0.0400 pixel scale, limiting magnitude of 15, and a max-
imum threshold of 0.01. The background subtraction
does not significantly change the residual mosaic, but
near the borders where the differences are well-defined
due to our initial cropping this step smooths the previ-
ously defined edges.
3.2. Comparison with median filtering
For comparison, we alternatively mitigate the effect of
bCGs and ICL by performing a median filtering of the
images. Relative to model fitting, median filtering is a
fast operation taking only seconds per image without
multi-processing required. For that reason it has been
a popular approach for rapid catalog building in clus-
ter fields, and serves as a vital tool for testing source
detection and photometry immediately following obser-
vations.
 The UNCOVER Photometric Catalog
We begin by co-adding pixels in the 0.0200 pixel
scale JWST short-wavelength (SW) mosaics (F090W,
F105W, F150W, F200W) to match the 0.0400 pixel scale
of the JWST LW and HST mosaics. Next we make
cutouts of the three cluster cores, each subtending a
1.4 × 1.4 arcmin2 FOV. Each cutout is downsampled by
a factor of 10 along each axis, convolved with a median
filter to remove high frequency signal on scales larger
than ∼ 800 , and then linearly interpolated back to its
original pixel scale. The resulting pixel-matched model
cutouts of the ICL And bCG wings are then subtracted
from their respective cluster cores. The sky variations
across the remainder of the image are subtracted us-
ing SEP, a pythonic version of Source Extractor
by Barbary (2018), assuming a mesh size of 32 pixels
and filter size of 8 pixels, with all images at 0.0400 scale.
This procedure is then repeated independently for ev-
ery mosaic. We find that while the effects of the ICL
are mitigated, the bCG cores remain and their wings
are over-subtracted. Therefore, while the photometry of
objects near the bCGs measured from these median fil-
tered images are generally sub-optimal, catalogs can be
produced quickly and are useful for consistency checks.
Photometric catalogs measured on the bCG and ICL
subtracted images as described above in Section 3.1 are
therefore considered the default.
4. SOURCE DETECTION AND PHOTOMETRY
4.1. Source Detection
Sources are detected on a sky-subtracted noise-
equalized co-added image based on our deepest JWST
imaging in the three long-wavelength (LW) broadband
filters: F277W, F356W, and F444W. Detection is per-
formed with SEP, adopting the configuration listed
in Table 2. While aperture photometry is performed
on point spread function (PSF) matched images (Sec-
tion 4.2), we combine LW images at their native resolu-
tion to maximize sensitivity to identify faint sources. We
detect 50 365 and 51 262 sources in the bCG-subtracted
LW co-add and median filtered co-add, respectively.
The fewer number of sources in the bCG-subtracted ex-
traction is likely driven by comparably better noise prop-
erties and hence fewer false positives near the cluster
cores, despite the addition of some residual bCG light
not fit by our model. In Figure 4, we show a RGB
image of the bCG-subtracted detection bands with el-
lipses marking all faint (> 18 AB), non-flagged (see Sec-
tion 4.6) sources detected from the bCG-subtracted LW
co-add. The effective catalog depth in 0.3200 apertures
of the noise-equalized LW bCG-subtracted co-add de-
tection image is shown in Figure 5 (see Section 5.1 for
further details).
7
Table 2. SEP parameters used for source detection
on the noise-equalized F277W+F356W+F444W co-
added image. Not supplying a weight map implicitly
tells SEP to use THRESH TYPE = ABSOLUTE, suitable for
noise-equalized detection images.
Name Value
KERNEL 1.5 pixel FWHM Gaussian
MINAREA 10 pixels
THRESH 1.5 σ
DEBLEND NTHRESH 16
DEBLEND CONT 0.003
CLEAN Y
CLEAN PARAM 1.6
4.2. Point Spread Function Matching
Before extracting aperture photometry, the PSF of
each image is matched (or “homogenized”). This ap-
proach allows for consistent photometric measurements
within the same aperture size across all bands, which
leads to a better recovery of source colors, zphot , and
physical parameters. We adopt our longest wavelength
NIRCam band, F444W, as our target PSF. This choice is
motivated by F444W being our broadest NIRCam PSF,
meaning that the corresponding images are matched to
the lowest resolution. Additionally, F444W will probe
the reddest rest-frame light (e.g. 1 − 2µ m stellar bulk)
at z & 1, making it an ideal band with which to derive
total fluxes from our aperture photometry. Preserving
the original F444W image properties will maximize con-
sistency within the photometry.
Following methodology described in Skelton et al.
(2014) and Whitaker et al. (2019), we generate empiri-
cal PSFs in all HST bands using stars identified within
the FOV. Point sources are known to inhabit a locus
within a size-magnitude plane (where size is approxi-
mated by the ratio of flux from 200 to 0.3200 apertures
for JWST and 300 to 0.7000 apertures for HST ). This ap-
proach works well for HST , enabling us to identify clean
lists of stars for each filter separately: unsaturated point
sources are selected to be between 14 < m < 24 mag
and 1 < f (300 ) / f (0.7000 ) < 1.2 in the size-magnitude
plane. Next, 400 FOV stamps are made for each star,
with nearby objects and high S/N pixels masked out.
These point sources are then visually inspected and me-
dian combined to create empirical HST PSFs. Finally,
we renormalize each empirical PSF such that its energy
8 J. R. Weaver et al.

- - : NW core

: N core

: primary cluster

Figure 4. Color composite image of the JWST footprint of Abell 2744, with three cluster cores highlighted. Bright cluster
galaxies and intra-cluster light have been subtracted. Elliptical apertures consistent with our aperture-corrected photometry
are shown in green for reliable objects that are not flagged (see Section 4.6) and are fainter than 18th magnitude.

enclosed within a 200 radius aligns with typical calibra-
tion levels3 .
Generating an empirical PSF from the
JWST /NIRCam imaging is difficult, as bright stars
are challenging to identify due to a combination of sat-
urated central pixels and the fact that they are rare due
to a moderately small FOV. Thankfully the wavefront
sensor on JWST provides accurate PSF estimates sam-
pling with roughly a two-day cadence. Using WebbPSF
(Perrin et al. 2012, 2014), we extract JWST PSFs cor-
responding to the sensor report nearest to the middle
of our planned observations, 5 November 2022. For
each JWST filter, a PSF stamp is computed in a 1000
FOV with normalization = exit pupil so that en-
ergy outside the stamp is accounted for correctly. The
stamps are then rotated with linear interpolation to the
median position angle of UNCOVER (41◦ ) and finally
cropped to a 400 FOV. We note that this is a relatively
simple treatment given that due to a guide star failure
UNCOVER was observed at two different PAs, not to
3
https://www.stsci.edu/hst/instrumentation/acs/data-analysis/
aperture-corrections;
https://www.stsci.edu/hst/instrumentation/wfc3/
data-analysis/photometric-calibration/ir-encircled-energy
mention the PAs of the other two programs in the field.
A more complex treatment of the PSF including PA
variations is reserved for future work.
Kernels are produced using Pypher (Boucaud et al.
2016). Pypher generates PSF-matching kernels using
an algorithm based on Wiener filtering (Wiener 1949)
with a tunable regularization parameter, which we set
to 10−4 . All filters are matched to the reference fil-
ter (F444W). Figure 6 shows the PSF growth curves
of every filter relative to the F444W growth curve,
both before (top) and after (bottom) convolving with
a matching kernel. Matched PSFs have almost iden-
tical growth curves to the reference filter, with devia-
tions below the 1% level. The HST NIR filters (F105W,
F125W, F140W, F160W) have the most variation, but
this is only significant for aperture diameters smaller
than 0.3200 . This discrepancy is unavoidable given the
systematic differences in the shapes of the PSF between
HST and JWST , where the latter has significantly more
sub-structure and “snowflake”-like patterns that com-
plicate the PSF matching. This sub-structure will not
significantly impact our photometric measurements, as
it is azimuthally averaged out within the circular aper-
tures.
 The UNCOVER Photometric Catalog 9

LW: F277W + F356W + F444W

-30°26'50.3"
3 arcmin

50.4"
E
Declination

50.5"

50.6"

28.5 29.0 29.5 30.0

5 Depth (ABmag)
in 0.32" apertures
50.7"
0h14m41.19s 41.18s 41.17s 41.16s
Right Ascension
Figure 5. Schematic of depth variation across the noise-equalized LW coadd detection image aggregated from the DDT, GLASS,
and UNCOVER programs. Combining F277W, F356W, and F444W, the 5 σ depths in our LW detection band measured in
0.3200 apertures span 27 − 29 mag. Poisson contributions of the brightest objects feature prominently in our JWST weight maps.

Alternatively, we have similar success matching PSFs
with Photutils (Bradley et al. 2022), which uses ra-
tios of Fourier transforms to generate a matching kernel.
Photutils also requires the selection of one of several
window functions, used to filter high-frequency noise
from the Fourier ratios. The best window function for
a given reference PSF and the ideal values for the tun-
ing parameters used to scale the window function are
not straightforward and usually require testing differ-
ent combinations. While Photutils has similar PSF-
matching success to Pypher, it is less convenient due to
the additional parameters that need tuning. Moreover,
we find Photutils can only generate kernels comparable
to Pypher when the PSF is oversampled by a factor of
3 or more. Finally, we find that PSF-matching meth-
ods that utilize linear combinations of shapelets (e.g.
Gauss-Hermite or Gauss-Laguerre polynomials) to gen-
erate kernels (e.g., Skelton et al. 2014) should be avoided
when matching to JWST filters. The intrinsic symme-
try of these functions cannot effectively match the sig-
nificant spikes and extended structure in JWST PSFs.
However, shapelet-based algorithms perform the best for
PSF-matching to an HST reference filter (e.g., F160W).
4.3. Aperture Photometry
Photometry is measured in both 0.3200 and 0.7000 di-
ameter circular apertures with SEP. These “color” aper-
ture measurements are then corrected to total fluxes by
scaling them against total fluxes in F444W estimated
from elliptical Kron-like apertures (Kron 1980). To be
consistent with classical Source Extractor, photom-
etry is extracted on the F444W image in elliptical aper-
tures whose semi-major and semi-minor axes (estimated
from the detection image in pixels) are grown by a factor
10
0.32"
0.70" aperture
1.4
1.2
Fi /FF444W ( < r)
1.0
0.8
0.6
1.02
1.01
Fi /FF444W ( < r)
1.00
0.99
0.98
0.0 0.5 1.0 1.5 2.0 2.5
Aperture Diameter [arcsec]
Figure 6. PSF growth curves for each filter before (top) and
after (bottom) matching to F444W. After matching, all fil-
ters have deviations below the 1% level at the smallest aper-
ture diameter used (0.3200 ). Growth curves are shown rela-
tive to the F444W growth curve; a value of 1 indicates perfect
matching with F444W. Dashed lines indicate the ±1% devi-
ations from exact matching (solid black line). Dotted lines
indicate the location of 0.3200 and 0.7000 aperture diameters.
of 2.5× the Kron “radius” (a unitless factor), with 3.5 as
the mimumum allowed factor. Then, for sources whose
circularized Kron radius is less than the circular aper-
ture radius, the circular apertures are used instead of the
Kron-scaled ellipse. The resulting measurements from
this procedure are commonly referred to as “AUTO”
flux densities.
We additionally correct each measurement for light
missed at large radii (especially important for JWST )
by dividing the flux measurement of each source by the
fraction of the total light from the PSF curve of growth
within each respective circularized Kron radius. We
stress that this correction for the JWST bands in par-
ticular is on the order of 10 − 20%, significantly larger
than for HST due to the significant amount of light that
is characteristically scattered to large radii in JWST
PSFs.
Photometric uncertainties are derived by means of an
independent estimate of the background noise. For each
J. R. Weaver et al.
band we place 10 000 circular apertures in regions out-
f435w f115w f200w
f606w
f814w
f125w
f140w
f277w
f356w
side detected sources, within our detection footprint,
f090w f150w f410m and with good coverage in that band. We opt to use the
f105w f160w f444w segmentation maps to mark detected sources, making
the placement of the apertures dependent not only on
the footprint of the detection image, but also its union
with the footprint for that particular filter. For a given
filter, apertures are placed on the corresponding noise-
equalize image to account for the variation in depth
across the field (e.g., see Figure 5). Outlier measure-
ments greater than 5σ are removed, and the width of
the flux distribution is estimated by the standard devia-
tion, and is unitless. To obtain physical noise estimates
consistent with our photometry, the width is multiplied
by the effective local noise around each source estimated
by the inverse of the square root of the median of a 9 × 9
pixel box on the 40 mas weight map. Total flux errors are
then computed by multiplying the resulting per-object
noise by the ratio of total to the 0.32 or 0.700 aperture
flux from F444W.
Figure 7 illustrates the growth of photometric un-
certainty with magnitude. While the growth of pho-
3.0 tometric uncertainty in HST bands show dual locii
corresponding to the shallower BUFFALO and deeper
HFF observations, that of JWST bands follow many
locii which are blended together due to the multi-modal
depths produced by the overlapping DDT, UNCOVER,
and GLASS programs (see also Section 5.1).
Photomery is corrected for line-of-sight attenuation
through the Galaxy, adopting the dust maps of Schlafly
& Finkbeiner (2011) and the attenuation law of Fitz-
patrick & Massa (2007). Given the small footprint, we
opt to apply the median E(B − V) = 0.01. The dust col-
umn density in the direction of Abell 2744 is favorably
low, resulting in attenuation corrections in each band
on the order of 1% or less. We report flux densities and
their uncertainties in Fν units of 10 nJy corresponding
to an AB magnitude zeropoint of 28.9.
4.4. Source Magnification
Abell 2744 is one of the most powerful lensing clusters
known (Merten et al. 2011; Jauzac et al. 2015, e.g.). Ac-
cording to revised estimates of all three cluster cores by
Furtak et al. (2022b), Abell 2744 magnifies most objects
in our survey footprint by at least 2×, with objects seen
nearest to the cluster cores magnified 10 − 100× (see
Figure 1 of Furtak et al.). Magnification estimates for
sources in each catalog are computed following the Fur-
tak et al. parametric strong lensing model assuming
zphot derived from EAzY in Section 5.3 (and from zspec
where available). Future work presented in B. Wang
et al. (in preparation) will provide updated magnifica-
 The UNCOVER Photometric Catalog 11

1.0
0.4 F435W F606W F814W
0.2
0.0
0.8
0.4 F090W F105W F115W
0.2
Mag Uncertainty (AB)

0.0
0.6
0.4 F125W F140W F150W
0.2
0.0
0.4
0.4 F160W F200W F277W
0.2
0.0
0.2
0.4 F356W F410M F444W
0.2
0.00.0 20 22 24 260.2 28 30 20 0.422 24 26 28 0.6 30 20 22 0.8
24 26 28 30 1.0
Mag (AB)

Figure 7. Photometric uncertainty as a function of magnitude shown by log10 -scaled 2D histograms for total fluxes derived in
0.3200 diameter apertures. Catalog depths for each filter measured in the same aperture size corresponding to the 50th and 10th
percentile areas are indicated by the dashed and solid lines, respectively.

tion estimates using zphot from Prospector-β (John-
son et al. 2021, B. Wang et al. submitted). Note that
while magnification estimates are provided in the cat-
alogs, measurements (e.g. fluxes) are not corrected for
magnification.
4.5. Identifying Stars
Constructing reliable photometric catalogs of extra-
galactic sources requires identifying foreground stars
and other spurious sources which would otherwise con-
taminate galaxy samples. While no one identifier is com-
plete and strictly pure, stars in HST phototmetric sur-
veys have been traditionally identified by the fact that
they are typically bright and unresolved. However, due
to the staggering efficiency of JWST /NIRCam, stars
of similar brightness often saturate the detector pixels
making their identification surprisingly difficult.
To overcome this new obstacle, we first identify
stars using traditional methods as described in Skel-
ton et al. (2014). Figure 8 presents the flux ratio
in a larger 0.7000 aperture relative to a 0.3200 aper-
ture as a function of magnitude for JWST /F200W
(top left) and HST /F160W (top right). Stars are se-
lected to have a flux ratio less than 1.5 for magnitudes
brighter than 24 AB (HST /F160W, yellow) AND 25 AB
(JWST /F200W, blue). While we can identify point-
like objects a magnitude fainter for the larger JWST -
only footprint at the native F200W resolution relative to
HST /F160W, saturation sets in sooner and thus justifies
the combination of the two selections. It is also possible
to directly identify saturated sources in JWST imagese
as they have clipped centers where the weight maps have
a value of zero. We identify bright saturated stars by re-
quiring F200W < 18 AB as well as zero weight in F200W
at the source centroid. To be as thorough as possible,
we also leverage GAIA DR3 (Gaia Collaboration et al.
2016, 2022, green) to identify stars with non-zero proper
motion. These last two criteria are reliable, easily ac-
cessible, and enable robust star identification when HST
coverage is missing and/or where objects are brighter
than the saturation limit of NIRCam. There are 5 ex-
12 J. R. Weaver et al.
ample postage stamps of F200W-selected unsaturated
stars shown at the top of Figure 8 (outlined in blue).
However, a point-like indicator is unsuitable for the
faintest stars that are likely to intermix with the gen-
eral galaxy population. Some literature studies have
opted to identify stars by comparing their fit quality be-
tween galaxy and stellar templates, either alone or in
combination with resolution criteria, e.g. Weaver et al.
(2022a). To enable these comparisons, we use EAzY to
quantify the goodness-of-fit of our SEDs to theoretical
PHOENIX BT-Settl stellar templates (Allard et al. 2012)
and include the corresponding χ2 estimates in our cat-
alog. However, we do not use χ2 to flag additional faint
stars. Such a comparison is prone to incorrectly reject-
ing a non-zero number of potentially interesting and lit-
tle known objects for which we have poor templates, e.g.
high-z galaxies. Consequently, galaxy candidates fainter
than ∼ 25 mag are liable to be contaminated by difficult
to identify foreground stars and so require additional
scrutiny.
4.6. Getting Started: The Use Flag
A “use” flag is particularly useful when familiarizing
oneself with any given photometric catalog. Ideally, one
simple selection yields a clean sample of galaxies across
cosmic time for further analysis. Following Skelton et al.
(2014), the USE PHOT flag in our photometric catalog re-
quires stars to be removed (see Section 4.5) and sets a
minimum signal-to-noise threshold of 3 for a color aper-
ture in the longest wavelength F444W filter. However,
with the addition of the novel JWST photometry, there
are further selections required to sufficiently clean up
the catalog as described below. For this first genera-
tion, we opt to be more conservative at the expense of
completeness for certain more exotic parameter spaces
(e.g., extreme high redshifts).
There also appear to be a number of bad pixels in
the LW channels of NIRCam where the similar dither
patterns between all LW bands make them broader
than a single pixel and in the same location in all LW
bands. These bad pixels are typically moderately bright
(∼ 22 − 24 mag) and not found in SW bands, thus nat-
urally mimicking bright z & 10 galaxies with strong
Lyman breaks. Thankfully in our data, and likely in
most data sets with reasonable dither patterns, they ap-
pear as bright objects in LW bands (e.g., F444W) with
sizes smaller than the PSF. Furthermore, while our bCG
modeling and subtraction enables searches for objects
within the immediate vicinity of the bCGs, it also pro-
duces undesirable residual features in some cases which
are detected in our catalogs. We find that both sets of
artifacts can be identified in F444W having a flux ratio
less than 1.1 with magnitudes brighter than 26 AB (see
bottom right panel of Figure 8). The catalog includes a
column FLAG ARTIFACT to signal objects satisfying this
selection.
The areas immediately surrounding the subtracted
bCGs are in all cases contaminated by residual struc-
tures. We further flag all detections within a conserva-
tive 400 radius of a known bCG centroid computed as
part of the modeling procedure. The catalog includes
a column FLAG NEARBCG to signal objects satisfying this
selection. While caution that some residual features out-
side this radius remain unflagged, it is not possible to
perfect flag genuine residual features without also catch-
ing real sources of interest. A more sophisticated treat-
ment will be explored in future work.
Beyond this, we identify sources with unphysically
large Kron radii that result in regions where residuals
are strong. For these objects, the aperture photometry
is robust but the correction to total flux density is biased
by the erroneously large Kron radius. This impacts <1%
of the objects within the catalog. The catalog includes
a column FLAG BADKRON to signal objects satisfying this
selection.
Finally, there exist significant issues with the flat field
in the LW images of ultra-deep GLASS pointing north-
west of the main cluster that leads to several thousand
spurious detections. Thankfully, the SW images do not
suffer from this problematic flat fielding and so we flag
any source in the GLASS region that is undetected in
the F200W images (SNR F200W < 3) or for which we
do not have F200W coverage (e.g. SW chip gaps). Al-
though simple, these two criteria are remarkably effec-
tive at identifying these spurious detections that could
otherwise be easily mistaken for high-z galaxies in this
region. However, we caution that they may also catch
rare but genuine high-z galaxies in the same region. The
catalog includes a column FLAG BADFLAT to signal ob-
jects satisfying this selection.
Together, we build a single USE PHOT flag which
removes undesirable objects, including bright stars
(FLAG STAR), sources within 300 radius of bCGs
(FLAG NEARBCG), bad LW pixels and model residual fea-
tures (FLAG ARTIFACT), unphysically large Kron radii
(FLAG BADKRON), and LW flat field artifacts in the NW
region (FLAG BADFLAT). To ensure the reliability of total
photometry based on F444W kron radii, this flag also
excludes any object with a SNR< 3 in the F444W color
aperture (FLAG LOWSNR). By selecting all objects where
USE PHOT=1, this reduces our total sample to 23 765
galaxy candidates with reliable photometry.
5. CATALOG PROPERTIES
 The UNCOVER Photometric Catalog 13

Figure 8. The selection of stars and spurious objects (bad pixels or model residuals) are determined from the flux ratio in
large to small apertures above a limiting magnitude threshold. Stars that are unsaturated in JWST are selected from the
native resolution F200W image (blue), and co-added with HST /F160W selected stars (yellow) and known GAIA stars (green).
The LW images suffer from additional bad pixels and sources with contaminated photometry due to bad model residuals that
are flagged (red). Together, the bottom left panel shows that these flagged sources are well separated from the general galaxy
population, with smaller values of F200W flux radius for a given magnitude. Examples of point source postage stamps and bad
pixels are shown at the top (left and right, respectively).
14 J. R. Weaver et al.

5.2. Galaxy Number Counts

27.0
We compute the galaxy number counts for our cat-
27.5 alog in three JWST /NIRCam bands (F150W, F356W,
F444W) and one HST /WFC3 band (F150W), shown in
28.0 Figure 10. Instead of assuming the nominal total sci-
5 Depth (ABmag)

ence areas listed in Table 1. Only sources with reliable

28.5
29.0
29.5
LW Detection
30.0 F435W F115W
F606W F125W
F814W F140W
30.5 F090W F150W
F105W F160W
0 10 20 30 40
Cumulative Area (arcmin2)
Figure 9. Depth as a function of cumulative area for each
filter mosaic as well as the LW detection image. Measure-
ments are taken from 10 000 empty apertures of 0.3200 diam-
eter, per filter. Grey dotted lines mark depths at 27, 28, and
29 mag.
With photometry in hand, we summarize the key
properties of our catalog including effective catalog
depths, galaxy number counts, photometric redshifts,
and a brief comparison of measured JWST photome-
try to that expected from the best-fit SED template so-
lutions. Although we provide total photometry from
both 0.3200 and 0.7000 diameter apertures, in this section
we generally refer to the smaller 0.3200 diameter mea-
surements. A comparison between the two apertures is
shown in Figure 14 of Appendix B.
5.1. Photometric depths
As introduced in Section 1, the Abell 2744 imaging
consists of three overlapping JWST and two overlap-
ping HST programs. Consequently, the depth of any
of these mosaics cannot be fully described by a single
value. While the photometric uncertainties are com-
puted to account for this variation, Figure 9 explicitly
illustrates the effective catalog depth of each filter as a
function of cumulative area corresponding to 0.3200 aper-
tures. The depths of the HST data are roughly bimodal
as a result of the deep HFF observations contrasted with
the wider but shallower BUFFALO coverage. Mean-
while for JWST , the contribution of the shallow DDT
observation from NIRCam’s Module A is readily visible,
but the depths produced by including UNCOVER and
GLASS create many small regions of varying depths.
photometry are shown (i.e. not stars), but keeping low
SNR sources. We stress that these counts should not
be used to precisely quantify completeness nor survey
depth, but are merely an accessible means by which to
validate our catalogs against literature measurements.
In all cases we find good agreement with the num-
F200W ber counts presented in Kokorev et al. (2022), also mea-
F277W
F356W sured from Abell 2744, as well as the field counts of
F410M Weaver et al. (2022a) from COSMOS. We note that our
F444W
1.6µm counts are pre-selected by our LW detection from
JWST . This being said, our 1.6µm counts similarly turn
over around 27 mag which is expected as no additional
HST observations have been taken since Kokorev et al..
Furthermore, our 1.5µm counts (again pre-selected by
our LW detection) agree well with 1.6µm counts, but
continue another 0.5 mag deeper thanks to the deep
JWST imaging. For 3.6 and 4.4µm, our JWST imaging
is significantly deeper than the existing Spitzer /IRAC
coverage in both this field and COSMOS (with hints
that they are plausibly complete to about 25 − 26 mag;
see Weaver et al. 2022b), turning over at about 28 and
29 mag, respectively. Overall, our observed galaxy num-
ber counts serves to demonstrate that we may confi-
dently springboard from well-studied HST surveys to
probe orders of magnitude deeper with JWST .
5.3. Photometric Redshifts
In order to provide an impression as to photometric
performance, we compute zphot using EAzY (Brammer
et al. 2008). We use all bands available for each object
and set a minimum error floor of 5%, an increase from
the default 1% to more realistically reflect the calibra-
tion uncertainties of JWST /NIRCam. Given the con-
siderable uncertainty as to the real high-z galaxy SEDs,
we forgo the usual pre-processing step of iteratively tun-
ing zeropoints to avoid biasing our colors to those of the
SED templates. We also turn off both magnitude and
β-slope priors for similar reasons.
We compute zphot for four combinations of image pro-
cessing: (A) bCG-subtracted or (B) median filtered, and
photometry: total fluxes derived from (1) 0.3200 and (2)
0.7000 apertures. We additionally compute zphot from
two SED template sets: the default FSPS FULL set used
frequently in the literature, and the newer SFHZ CORR
set which features z-dependent priors on allowable star-
formation histories and an observed z ≈ 8 emission-line
 The UNCOVER Photometric Catalog
106
F150W
F160W
2
105
N mag 1 deg
104
UVISTA H
Weaver et al. 2022
F160W
Kokorev et al. 2022
10318 20 22 24 26 28 30 18
Mag (AB)
Figure 10. Galaxy number counts for sources identified from our LW detection image and measured in total magnitudes from
F150W, F160W, F356W, and F444W. Stars and objects with unreliable photometry are removed, but keeping low SNR sources.
Comparable F160W-selected counts by Kokorev et al. (2022) from Abell 2744 are shown by the unfilled grey boxes. Also shown
are the izY JHKs -selected counts from of field galaxies from COSMOS by Weaver et al. (2022a). Errorbars denote 1 σ poisson
uncertainties.
galaxy spectrum from Carnall et al. (2023) to provide re-
alistic line ratios – two considerations especially impor-
tant for identifying robust high-z galaxies candidates.
One popular way of judging photometric accuracy is
to compare zphot and zspec , where available. We use
N = 329 spectroscopic sources (see Section 2.4), remov-
ing N = 172 sources, including bright bCGs which are
modeled and subtracted in imaging set A as well as stars
and other unreliable objects (USE PHOT = 1). zphot per-
formance is then assessed using the normalized median
absolute deviation (NMAD, Hoaglin et al. 1983), defined
as

|∆z − median(∆z)|
σNMAD = 1.48 × median
1 + zspec
following Brammer et al. (2008) as it is less sensitive to
outliers compared to other definitions (e.g. Ilbert et al.
2006). We additionally quantify an outlier fraction as
the fraction of objects with |zphot − zspec | ≥ 0.015 (1 +
zspec ).
In general, the zphot performance in comparison to
zspec is equally good and essentially agnostic to our
bCG/ICL subtraction method and adopted photomet-
ric color aperture size (Figure 11). This, however, is
expected as objects in our spectroscopic sample are
notably brighter than many of the sources found in
this new, deep imaging. This has two consequences:
firstly, the assessable zphot performance does not depend
strongly on source magnitude and secondly, the zphot
performance reflects that of bright and easy-to-measure
sources.
Unsurprisingly, we find significantly better zphot per-
formance using the SFHz CORR SED templates com-
15
F356W F444W
IRAC CH1 IRAC CH2
Weaver et al. 2022 Weaver et al. 2022
IRAC CH1 IRAC CH2
Kokorev et al. 2022 Kokorev et al. 2022
20 22 24 26 28 30 18 20 22 24 26 28 30
Mag (AB) Mag (AB)
pared to the default FSPS FULL. Therefore the addi-
tion of z-dependent star-formation histories and realis-
tic line emission serves to greatly enhance the ability of
EAzY to correctly recover the zphot of known spectro-
scopic sources. This, in combination with the aforemen-
tioned advantages imparted by the model-based bCG-
subtraction make for an obvious pairing shown in Fig-
ure 11 where we achieve a σNMAD = 0.0256 after remov-
ing the 7.5% outlying sources. The outliers are mostly
at zphot <0.5 for zspec >2; this trend is expected as our
photometric catalog does not constrain the bluest rest-
frame light of nearby objects (i.e. no U band) and hence
zphot <0.5 may be unreliable. For zphot >0.5, the outlier
 fraction reduces to 5.2%.
, (1) Figure 12 shows a comparison of the distribution of
the total zphot sample to the redshifts of the generally
lower-z, bright spectroscopic sample.
We provide the zphot and several common rest-frame
fluxes (e.g. UVJ) for all objects in the catalog. How-
ever, more sophisticated methods utilizing extensive
physically-based priors and advanced sampling tech-
niques will enable even more robust zphot and physical
parameters (e.g. stellar mass). A forthcoming paper by
B. Wang et al. (in preparation) will detail how we have
applied Prospector-β to this end.
5.4. Validation of JW ST photometry
Given that JWST is a relatively new facility, much
is still to be learned about its performance. Conse-
quently, photometric extractions of JWST imaging in
the literature are only just becoming available and so
the traditional demonstration of comparing our JWST
photometry to the published literature is not possible.
16 J. R. Weaver et al.
7 Thankfully, computing zphot for every object provides
6 N=265 (20, 7.5%)
5 NMAD=0.0284 (0.0256)
access to the observed-frame fluxes expected by the cor-
responding best-fit SED template4 . While this is ar-
4 guably a circular exercise, the SED template combina-
tions allowed by EAzY still only span a finite volume in
3 color-color space. By attempting to maximize the likeli-
hood of the fit (corresponding to Z PHOT), EAzY is effec-
2 tively maximizing its agreement with the error-weighted
zphot

colors that we provide it; yet large-scale biases in indi-

vidual filters should stand out given our well-sampled 15
1 filter SEDs.
Figure 13 shows the photometric agreement ∆ Mag
between our measured fluxes and those predicted by the
best-fit model from EAzY. While here we assume the
newer SFHz CORR template set, we find similar agree-
ment using the FSPS FULL set. The largest offset appears
0 in F814W at ≈ 0.114 mag, and is relatively systematic.
z/1 + z

+3 We note that in a handful of other filters the difference

0
3 trends upwards at the 50th percentile depth but flattens
0 1 2 3 4 5 6 7
zspec
Figure 11. Performance of zphot assessed by comparison
with known zspec , described in Section 2.4. Outlier sources
with zphot wrong by more than 0.15 ∆z/(1+zspec ) are colored
red. Lower panel shows residuals in normalized units of σ
away from zspec . N = 328 sources are compared finding 20
outliers (7.5%), an overall tightness σNMAD = 0.0284 and
an outlier-removed σNMAD = 0.0256. zphot are derived from
0.3200 aperture total fluxes extracted from bCG-subtracted
images with EAzY assuming the SFHz CORR SED template
set.
towards the 10th percentile depth. This behavior corre-
lates with the variation in depth and so can be readily
explained by differences in the photometric performance
of similarly bright objects with a range (or bimodality)
of signal-to-noise. We also investigate ∆ Mag as a func-
tion of z, finding similar levels of agreement even at
z & 6.
Although this is a relatively basic comparison, it
nonetheless demonstrates that our photometry appears
reasonable and that the SEDs of most objects are de-
scribable by EAzY. Comparisons to literature HST
photometric catalogs are included in Appendix B.

Age (Gyr)
6. SUMMARY

13 6 3 2 1.5 1 0.8 0.6 0.5 0.4 0.3 We present a first generation of high angular resolu-
tion space-based photometric catalogs of the strong lens
103 zphot cluster Abell 2744, including a combination of archival
zspec HST imaging in 7 bands with public JWST imaging in 8
bands from 3 programs (UNCOVER, GLASS/ERS, and
Count

102 a DDT program). With an ultra-deep noise-equalized

F277W+F356W+F444W detection image at the na-
101 tive JWST resolution (Figure 5), we present 0.4-4.4µm
panchromatic coverage for roughly 50k sources within
100 the extended cluster region, including two newly de-
0.0 2.5 5.0 7.5 10.0 12.5 15.0 tailed structures to the north and northwest of the clus-
z ter heart. After removing stars, artefacts, and other spu-
rious sources, we present reliable photometry for 24,265
Figure 12. Distribution of zphot (teal) and zspec (purple) galaxy candidates.
shown up to z = 15. zphot are derived from 0.3200 aperture to-
tal fluxes extracted from bCG-subtracted images with EAzY
4 Note that since EAzY uses linear combination of basis templates,
assuming the SFHz CORR SED template set.
the resulting best-fit model is not limited to physically allowed
solutions.
 The UNCOVER Photometric Catalog 17

1.0
0.5 f435w f606w f814w
0.0

0.5 = 0.025 = 0.039 = 0.116

0.8
0.5 f090w f105w f115w
0.0
Mag (Best-fit - AperPy)

0.5 = 0.011 = 0.050 = 0.006

0.6
0.5 f125w f140w f150w
0.0

0.5 = 0.039 = 0.016 = 0.016

0.4
0.5 f160w f200w f277w
0.0

0.5 = 0.032 = 0.003 = 0.032

0.2
0.5 f356w f410m f444w
0.0

0.5 = 0.032 = 0.030 = 0.001

0.00.0 20 22 24 0.2
26 28 30 20 0.422 24 26 28
0.6 30 20 22 0.824 26 28 301.0
AperPy Mag (AB) AperPy Mag (AB) AperPy Mag (AB)

Figure 13. Photometry measured by AperPy in this work compared against predicted fluxes from the best-fit SED from
EAzY in all 15 available bands. In each panel, the difference in AB magnitude (∆ Mag) as a function of magnitude measured
by AperPy is summarized by the log10 -scaled 2D grey density histogram. Colored curves indicate binned medians with two-sided
envelopes enclosing 68% of sources per bin. The overall median offset is labeled on each panel computed on all magnitudes up
to the magnitude limit of the deepest 10% of the corresponding image indicated by the colored solid vertical line; that of the
median depth is indicated by the colored dashed vertical line. The typical photometric error derived by AperPy is indicated by
the dotted grey curves.

With this paper, we release the UNCOVER photo-
metric catalogs derived from small (0.3200 diameter) and
larger (0.7000 diameter) circular apertures that include
PSF homogenization of images to the JWST /F444W
resolution that are cleaned of contaminating light from
bCGs and ICL. The photometry is corrected to total
based on the ratio of flux within a Kron-like aperture
relative to a circular aperture, plus an additional correc-
tion of order 5-20% for missing light beyond the Kron ra-
dius as determined from the PFS curve of growth in the
F444W filter. Details including how to access the cat-
alogs, column descriptions, and general use recommen-
dations can be found in Appendix C or directly through
the UNCOVER survey webpage: https://jwst-uncover.
github.io/DR1.html#PhotometricCatalogs.
The UNCOVER photometric catalogs are among the
deepest catalogs publicly available, reaching effective 5σ
depths greater than 29,AB in all 15 bands for the 0.3200
diameter apertures. These depths do not account for ex-
tra magnification factors from strong gravitational lens-
ing for those background galaxies in optimal configura-
tions. When combining the survey depths with strong
lensing, UNCOVER is the deepest view into our uni-
verse to date.
18 J. R. Weaver et al.
ACKNOWLEDGMENTS
The authors would like to acknowledge the generos-
ity of Vasily Kokorev, as well as Diego Paris, Adriano
Fontana, Emiliano Merlin, and Tommaso Treu for mak-
ing their photometry available for comparison. We all
thank you for the many fruitful discussions that im-
proved our catalogs.
This work is based in part on observations made with
the NASA/ESA/CSA James Webb Space Telescope.
The data were obtained from the Mikulski Archive for
Space Telescopes at the Space Telescope Science Insti-
tute, which is operated by the Association of Univer-
sities for Research in Astronomy, Inc., under NASA
contract NAS 5-03127 for JWST . These observations
are associated with JWST-GO-2561, JWST-ERS-1324,
and JWST-DD-2756. Support for program JWST-GO-
2561 was provided by NASA through a grant from the
Space Telescope Science Institute, which is operated by
the Associations of Universities for Research in Astron-
omy, Incorporated, under NASA contract NAS5-26555.
This research is also based on observations made with
the NASA/ESA Hubble Space Telescope obtained from
the Space Telescope Science Institute, which is oper-
ated by the Association of Universities for Research in
Astronomy, Inc., under NASA contract NAS 5–26555.
These observations are associated with programs HST-
GO-11689, HST-GO-13386, HST-GO/DD-13495, HST-
GO-13389, HST-GO-15117, and HST-GO/DD-17231.
This work has made use of data from the Euro-
pean Space Agency (ESA) mission Gaia (https://www.
cosmos.esa.int/gaia), processed by the Gaia Data Pro-
cessing and Analysis Consortium (DPAC, https://www.
cosmos.esa.int/web/gaia/dpac/consortium). Funding
for the DPAC has been provided by national institu-
tions, in particular the institutions participating in the
Gaia Multilateral Agreement.
Cloud-based data processing and file storage for this
work is provided by the AWS Cloud Credits for Re-
search program. This research has made use of the
NASA/IPAC Extragalactic Database (NED), which is
funded by the National Aeronautics and Space Adminis-
APPENDIX
A. PHOTOMETRIC COMPARISONS
While the bulk of this work focused on the photom-
etry extracted on our bCG- and ICL-subtracted im-
ages, both methods introduce considerable uncertainty
into the extracted photometry that is difficult to quan-
tify. Figure 15 summarizes the differences in photome-
try extracted from the nominal bCG-subtracted images
tration and operated by the California Institute of Tech-
nology.
The Cosmic Dawn Center is funded by the Dan-
ish National Research Foundation (DNRF) under grant
#140. RB acknowledges support from the Research
Corporation for Scientific Advancement (RCSA) Cot-
trell Scholar Award ID No: 27587. LF and AZ
acknowledge support by Grant No. 2020750 from
the United States-Israel Binational Science Foundation
(BSF) and Grant No. 2109066 from the United States
National Science Foundation (NSF), and by the Min-
istry of Science & Technology, Israel. PD acknowl-
edges support from the NWO grant 016.VIDI.189.162
(“ODIN”) and from the European Commission’s
and University of Groningen’s CO-FUND Rosalind
Franklin program. HA acknowledges support from
CNES (Centre National d’Etudes Spatiales). RS ac-
knowledges an STFC Ernest Rutherford Fellowship
(ST/S004831/1). MS acknowledges support from the
CIDEGENT/2021/059 grant, from project PID2019-
109592GB-I00/AEI/10.13039/501100011033 from the
Spanish Ministerio de Ciencia e Innovación - Agen-
cia Estatal de Investigación, and from Proyecto AS-
FAE/2022/025 del Ministerio de Ciencia y Innovación
en el marco del Plan de Recuperación, Transformación
y Resiliencia del Gobierno de España
Facilities: JWST (NIRCam, NIRSpec, and
NIRISS), HST (ACS and WFC3), GAIA
Software: astropy (Astropy Collaboration et al.
2013, 2018, 2022), Source Extractor (Bertin &
Arnouts 1996), SEP (Barbary 2016a), extinction
(Barbary 2016b), SFDMap (Schlegel et al. 1998;
Schlafly & Finkbeiner 2011, github.com/kbarbary/
sfdmap), WebbPSF (Perrin et al. 2012, 2014), EAzY
(Brammer et al. 2008), Pypher (Boucaud et al. 2016),
photutils (Bradley et al. 2022), astrodrizzle (Gon-
zaga et al. 2012), Grizli (github.com/gbrammer/grizli),
numpy (van der Walt et al. 2011), matplotlib (Hunter
2007)
and from the median-filtered images for all 15 avail-
able bands. Generally there is good agreement, espe-
cially on the bright end from HST /WFC3 and JWST .
However, there appears to be a small bias in the two
HST /ACS bands: F606W and F814W on the order of
0.1 mag for bright sources whereby the median-filtered
photometry is fainter. This is expected as median fil-
 The UNCOVER Photometric Catalog
tering is more likely to oversubtract compared to fitted
models. The different choices of background subtrac-
tion may also play a role. Faint sources are typically
0.1-0.2 mag fainter in the median-filtered photometry as
well, likely for the same reason. Finally we note that
F444W is expectantly identical since F444W total fluxes
are derived from the same Kron-like apertures.
This work forwards two reference catalog with pho-
tometry extracted on robust bCG-subtracted images in
corrected 0.32 and 0.700 apertures. Figure 14 demon-
strates the differences between the two for all 15 avail-
able bands. We find agreement below 0.03 mag for all
bands, with some having essentially no bias. There are
some minor trends on the faint end likely driven by low
signal-to-noise sources on the order of 0.1 mag or less.
Although comparisons with predicted model fluxes
from EAzY are useful (see Figure 13), we can also lever-
age the existing HST catalogs to make comparisons for
ACS and WFC3 derived photometry. Comparisons for
all 7 common bands for Shipley et al. (2018) and Koko-
rev et al. (2022) are shown in Figures 16 and 17, respec-
tively. We choose to show the total fluxes derived from
0.700 apertures to be consistent with their choice of aper-
ture size, and adopt photometry based on the nominal
bCG-subtracted images.
Compared to Shipley et al., we generally find good
agreement, although we note that the HST /ACS
F606W and F814W fluxes in our catalog are systemat-
ically brighter by ∼ 0.1 mag. Despite the similar meth-
ods in both bCG-subtraction and aperture size, we find
that these offsets are minimized when using the (less pre-
ferred) median-filtered images, pointing to differences in
the background subtraction that will be explored in fu-
ture work.
Compared to Kokorev et al., we again generally find
good agreement. While the bright end has almost no
bias, our photometry of fainter sources is generally 0.2-
0.3 mag brighter. However, as detailed in Kokorev et al.
(2022), the authors elected to forgo PSF homogenization
which makes this comparison particularly hazardous for
faint, point-like sources. As seen before, the bimodailty
in depth for some of the HST bands drives up ∆ Mag
for sources in the shallow areas.
At the time of this writing there are no published cat-
alogs of JWST photometry in Abell 2744. However,
we compare with a separate photometric catalog under
development by the GLASS team (Paris et al. 2023).
Both catalogs are based on roughly the same public
datasets from HST and JWST , though Paris et al. per-
forms an independent data reduction and analysis to
ours. Several notable differences between the two cata-
logs include (1) independent image reduction pipelines,
19
(2) source detection (the GLASS catalog is F444W de-
tected, whereas we use a noise-equalized LW detection),
(3) the treatment of bCG/ICL (modeled out herein,
and not in GLASS), and (4) different software and ap-
proaches to the PSF homogenization, aperture photom-
etry, and total flux corrections. Despite these differences
in approach, we note that the aperture photometry for
the JWST filters is in remarkable agreement overall,
with essentially no systematic offsets, and the catalog
as a whole agrees within 0.1 AB.
B. JWST/NIRCAM PSF STABILITY
While JWST is already proving to be a revolution-
ary facility to advance nearly all areas of astrophysics,
the typical behavior of the telescope resolution, both in
time and across the detectors, remains a concern. Unlike
the generally stable PSF enjoyed by HST , the moving
hexagonal mirrors of JWST means that the PSF can
be highly variable. Thankfully, the Wavefront Sensor
(WSS) samples the observatory mirrors on an approx-
imate two-day cadence with sufficient detail to recon-
struct the PSF at any detector location in any band.
This is most easily accessed by WebbPSF, which in-
cludes near-live reports from the WSS to enable PSF
reconstruction on the fly, as well as tools to visualize
the typical PSF behaviour with time.
Nardiello et al. (2022) has already presented a detailed
analysis of the PSF variability of NIRCam finding ∼9%
variation across their FOV and 3-4% over multi-epoch
exposures based on peak-to-peak variations measured
from PSF residuals. However, their results are con-
cerned with the shape of the PSF (especially it’s core)
which is directly relevant to their PSF-fitting photome-
try of dense star clusters. The power of aperture pho-
tometry, crucially, is that it is sensitive only to the en-
ergy enclosed by a given aperture and not about pre-
cisely where that energy is located within the aperture.
By now studying the variation of the enclosed energy
we can not only better understand the photometric im-
pact of these variations, but do so in the context of our
particular observations from UNCOVER.
Our photometry is derived from images PSF-matched
to F444W and so their accuracy strongly depends on the
F444W PSF characteristics. As described in Section 4.2,
we adopt single JWST PSFs generated by WebbPSF
corresponding to 5 November, 2022 near the expected
mid-point of the planned UNCOVER visits5 . In the
following analysis we generate two regular grids of 144
PSF samples for each NIRCam LW detector: NRCA5
5 Due to a guide star acquisition failure on 31 October, visit 1.1
was repeated successfully on 15 November.
20 J. R. Weaver et al.

1.0
0.5 f435w f606w f814w
0.0

0.5 = 0.030 = 0.047 = 0.046

0.8
0.5 f090w f105w f115w
0.0

0.5 = 0.019 = 0.018 = 0.008

Mag (0.70" - 0.32")

0.6
0.5 f125w f140w f150w
0.0

0.5 = 0.006 = 0.006 = 0.008

0.4
0.5 f160w f200w f277w
0.0

0.5 = 0.021 = 0.008 = 0.024

0.2
0.5 f356w f410m f444w
0.0

0.5 = 0.006 = 0.004 = 0.000

0.00.0 20 22 24 0.2
26 28 30 20 0.422 24 26 28
0.6 30 20 22 0.824 26 28 301.0
0.32" Mag (AB) 0.32" Mag (AB) 0.32" Mag (AB)

Figure 14. Photometry measured by AperPy on the nominal bCG-subtracted images in corrected 0.3200 versus 0.7000 apertures.
Format follows that of Fig. 14.

and NRCB5. We set normalization = exit pupil so
that the 400 FOV stamps are normalized such that the
energy at large radii is accounted for correctly. Here we
focus on the UNCOVER dataset only leaving the char-
acterization of the PSF behaviour during the GLASS
and DDT visits to future work. Although shown below
only for F444W, we have repeated the analysis for all
NIRCam bands and find similar results.
Figure 18 quantifies the wavefront error (WFE) from
1 October to 30 November, 2022, during which the UN-
COVER program was executed. While there were sig-
nificant anomalies in the WFE at the end of October,
all four UNCOVER visits were executed with nominal
PSF behaviour with a WFE ≈ 65 − 75 nm, well within
the target control range6 . Of even greater significance
6
https://jwst-docs.stsci.edu/jwst-observatory-hardware/
jwst-wavefront-sensing-and-control
to this work, the relative enclosed energy ∆EE during
that same window was effectively static meaning that
any uncertainty due to the time evolution of the F444W
PSF when measuring photometry can be safely ignored.
Figure 19 provides insight into the spatial variability
of the PSF across NRCA5 and NRCB5 on 5 Novem-
ber. Instead of examining FWHM as provided by
WebbPSF, we measure enclosed energy (EE) as it is
more pertinent to assessing uncertanties in aperture
photometry. While the EE relative to that of the av-
erage PSF h EE i differs between NCRA5 and NCRB5,
we find that the spatial variation is generally smooth
and increases with decreasing aperture size such that
the variability in EE is  0.1% in 0.7000 apertures and
∼ 1% at 0.3200 .
We caution, however, that these results do not nec-
essarily apply to model-based photometric techniques
(e.g. GALSIM, Rowe et al. 2015) because although the
 The UNCOVER Photometric Catalog 21

1.0
0.5 f435w f606w f814w
0.0

0.5 = 0.106 = 0.131 = 0.127

0.8
0.5 f090w f105w f115w
Mag (Median-Filtered - bCG-Subtracted)

0.0

0.5 = 0.039 = 0.059 = 0.038

0.6
0.5 f125w f140w f150w
0.0

0.5 = 0.058 = 0.082 = 0.050

0.4
0.5 f160w f200w f277w
0.0

0.5 = 0.056 = 0.058 = 0.065

0.2
0.5 f356w f410m f444w
0.0

0.5 = 0.081 = 0.085 = 0.137

0.00.0 20 22 24 0.2
26 28 30 20 0.422 24 26 28
0.6 30 20 22 0.824 26 28 301.0
bCG-Subtracted Mag (AB) bCG-Subtracted Mag (AB) bCG-Subtracted Mag (AB)

Figure 15. Photometry measured by AperPy on the nominal bCG-subtracted images based on corrected 0.3200 apertures
compared to that of the median filtered images in all 14 bands. Format follows that of Fig. 13.

summed energy at fixed radius is relatively constant, the
exact shape of the PSF itself may vary enough that the
spatial dependence may need to be taking into account,
as stressed by Nardiello et al..
Together, these two pieces of evidence point to a
favourably small contribution to the error budget from
the JWST PSF behavior for UNCOVER – both in time
evolution and spatial variation – such that we can con-
fidently neglect these uncertainties in our photometric
catalogs.
C. CATALOG COLUMN DESCRIPTIONS
This paper is accompanied by two photometric cat-
alogs. They are both derived from the same bCG-
subtracted images which have been PSF-matched to
that of F444W, and share the same LW-selected ob-
jects. Fluxes are reported in total as described in Sec-
tion 4.3, accounting for any additional light outside the
Kron aperture. The catalogs differ in their choice of
aperture size. While the columns included in the cat-
alogs are described in Table 3, we also include a dedi-
cated README file with the catalogs in the case of future
changes.
We strongly recommend using photometry derived
from smaller 0.3200 diameter apertures for high-z science
cases, whereas the photometry derived from the larger
0.7000 apertures is more suitable for brighter, extended
objects at lower-z. The photometry is reported in Fν
in units of 10 nJy corresponding to a zeropoint of 28.9
AB. Although zphot , rest-frame colors, and magnifica-
tion estimates are reported for all objects from each set
of photometry, clean samples of galaxies can be readily
identified with USE PHOT=1 (see Section 4.6).
The photometric catalogs and all related documen-
tation and high-level science data products as de-
scribed herein are accessible via the UNCOVER survey
22 J. R. Weaver et al.

1.0
0.5 f435w f606w f814w
Mag (Shipley+18 - AperPy) 0.0
0.8
0.5 = 0.059 = 0.116 = 0.153
0.5 f105w
0.6 f125w f140w
0.0
0.4
0.5 = 0.095 = 0.078 = 0.066
20 22 24 26 28 30 20 22 24 26 28 30
0.5 f160w AperPy Mag (AB) AperPy Mag (AB)
0.2
0.0
0.5 = 0.100
0.00.0 20 22 24 0.2
26 28 30 0.4 0.6 0.8 1.0
AperPy Mag (AB)

Figure 16. Photometry measured by AperPy on the nominal bCG-subtracted images based on corrected 0.700 apertures
compared to that of Shipley et al. (2018) in the 7 common HST bands. Format follows that of Fig. 13.

1.0
0.5 f435w f606w f814w
Mag (Kokorev+22 - AperPy)

0.0
0.8
0.5 = 0.124 = 0.143 = 0.109
0.5 f105w
0.6 f125w f140w
0.0
0.4
0.5 = 0.102 = 0.093 = 0.101
20 22 24 26 28 30 20 22 24 26 28 30
0.5 f160w AperPy Mag (AB) AperPy Mag (AB)
0.2
0.0
0.5 = 0.114
0.00.0 20 22 24 0.2
26 28 30 0.4 0.6 0.8 1.0
AperPy Mag (AB)

Figure 17. Photometry measured by AperPy on the nominal bCG-subtracted images based on corrected 0.700 apertures
compared to that of Kokorev et al. (2022) in the 7 common HST bands. Format follows that of Fig. 13.

webpage: https://jwst-uncover.github.io/DR1.html#
PhotometricCatalogs.

REFERENCES
Abbott, T. M. C., Abdalla, F. B., Allam, S., et al. 2018,
ApJS, 239, 18, doi: 10.3847/1538-4365/aae9f0
Adams, N. J., Conselice, C. J., Ferreira, L., et al. 2023,
MNRAS, 518, 4755, doi: 10.1093/mnras/stac3347
Aihara, H., Arimoto, N., Armstrong, R., et al. 2018, PASJ,
70, S4, doi: 10.1093/pasj/psx066
Allard, F., Homeier, D., & Freytag, B. 2012, Philosophical
Transactions of the Royal Society of London Series A,
370, 2765, doi: 10.1098/rsta.2011.0269
Antwi-Danso, J., Papovich, C., Leja, J., et al. 2022, arXiv
e-prints, arXiv:2207.07170.
https://arxiv.org/abs/2207.07170
Astropy Collaboration, Robitaille, T. P., Tollerud, E. J.,
et al. 2013, A&A, 558, A33,
doi: 10.1051/0004-6361/201322068
Astropy Collaboration, Price-Whelan, A. M., Sipőcz, B. M.,
et al. 2018, AJ, 156, 123, doi: 10.3847/1538-3881/aabc4f
 The UNCOVER Photometric Catalog 23

Observatory WFE at NIRCam NRCA3

120 Telescope WFE

visit 1.2
visit 1.3
visit 1.4

visit 1.1
Wavefront Error
100
(nm)

Wavefront control
target range
80

60

0.04
±3% change (stability requirement)
0.02
(F200W)
EE

0.00

0.02
EE within 0.31 arcsec
0.04 EE within 0.08 arcsec
2022-10-10 2022-10-24 2022-11-07 2022-11-21
Figure 18. Evolution of the wavefront error (WFE) for JWST generated from WebbPSF from 1 October to 30 November,
2022. Top: the measured WFE for NIRCam Module A3 (NRCA3, yellow) is shown along with that of the general telescope
facility (blue). Bottom: the enclosed energy (EE) of the F200W PSF in 0.31 and 0.0800 apertures (green and purple, respectively)
relative to the median. Relevant control and stability criteria are indicated. Behaviour during the four UNCOVER visits is
effectively unchanged, with relative EE . 0.1%.

Astropy Collaboration, Price-Whelan, A. M., Lim, P. L.,
et al. 2022, apj, 935, 167, doi: 10.3847/1538-4357/ac7c74
Atek, H., Richard, J., Kneib, J.-P., & Schaerer, D. 2018,
MNRAS, 479, 5184, doi: 10.1093/mnras/sty1820
Atek, H., Shuntov, M., Furtak, L. J., et al. 2023, MNRAS,
519, 1201, doi: 10.1093/mnras/stac3144
Barbary, K. 2016a, Journal of Open Source Software, 1, 58,
doi: 10.21105/joss.00058
—. 2016b, extinction v0.3.0, Zenodo,
doi: 10.5281/zenodo.804967
Barbary, K. 2018, SEP: Source Extraction and
Photometry, Astrophysics Source Code Library, record
ascl:1811.004. http://ascl.net/1811.004
Beckwith, S. V. W., Stiavelli, M., Koekemoer, A. M., et al.
2006, AJ, 132, 1729, doi: 10.1086/507302
Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393,
doi: 10.1051/aas:1996164
Bezanson, R., Labbe, I., Whitaker, K. E., et al. 2022, arXiv
e-prints, arXiv:2212.04026.
https://arxiv.org/abs/2212.04026
Bhatawdekar, R., Conselice, C. J., Margalef-Bentabol, B.,
& Duncan, K. 2019, MNRAS, 486, 3805,
doi: 10.1093/mnras/stz866
Boucaud, A., Bocchio, M., Abergel, A., et al. 2016, A&A,
596, A63, doi: 10.1051/0004-6361/201629080
Bouwens, R. J., Illingworth, G., Ellis, R. S., Oesch, P., &
Stefanon, M. 2022, ApJ, 940, 55,
doi: 10.3847/1538-4357/ac86d1
Bouwens, R. J., Oesch, P. A., Illingworth, G. D., Ellis,
R. S., & Stefanon, M. 2017, ApJ, 843, 129,
doi: 10.3847/1538-4357/aa70a4
Bouwens, R. J., Illingworth, G. D., Oesch, P. A., et al.
2011, ApJ, 737, 90, doi: 10.1088/0004-637X/737/2/90
Bradley, L., Sipőcz, B., Robitaille, T., et al. 2022,
astropy/photutils: 1.5.0, 1.5.0, Zenodo,
doi: 10.5281/zenodo.6825092
24 J. R. Weaver et al.

Figure 19. Behavior of JWST /NIRCam as measured by the wavefront sensor on 5 November, 2022, from which our JWST
PSFs are derived, based on a grid of 144 PSF realizations generated with WebbPSF. Top: the spatial dependence of the energy
enclosed in 0.3200 apertures for NIRCam Module A5 (NRCA5) and B5 (NRCB5) in F444W relative to that of the average PSF.
Middle: enclosed energies (EE) as a function of aperture size colored by distance from the respective module center. Bottom:
enclosed energies as a function of aperture size relative to that of the average PSF for each module, colored as before. PSF
variability across NIRCam for F444W is minimal with a relative EE . 0.1% in even small 0.3200 apertures.
 The UNCOVER Photometric Catalog 25

Table 3. Catalog columns

Column name Description

id Unique identifier
x X centroid in image coordinates
y Y centroid in image coordinates
ra RA J2000 (degrees)
dec Dec J2000 (degrees)
faper F444W F444W flux within a 0.32/0.70 arcsecond aperture
eaper F444W 1 σ F444W error within a 0.32/0.70 arcsecond aperture
fX Total flux for each filter X (Fν , zeropoint = 28.9)
eX 1 σ error for each filter X (Fν , zeropoint = 28.9)
wX Weight relative to the maximum within image X (see text)
tot cor Aperture-to-total correction including both aperture-to-Kron and Kron-to-total
z spec Spectroscopic redshift, where available
kron radius Kron radius factor (Source Extractor-like, unitless)
kron radius circ Circularized Kron radius (arcsec)
a image Semi-major axis (A IMAGE, pixels)
b image Semi-minor axis (B IMAGE, pixels)
theta J2000 Position angle of the major axis (counter-clockwise, measured from East)
use phot 1 for reliable sources (good = 1; all flag XXX = 0)
flag lowsnr 1 if source has aperture SNR< 3 in F444W
flag star 1 if source is identified as a star (see Section 4.5)
flag artifact 1 if source is identified as an artifact (see Section 4.6)
flag nearbcg 1 if within 300 of a known bCG (see Section 4.6)
flag badkron 1 if source has erroneously large kron radius (see Section 4.6)
flag badflat 1 if source is an artifact from bad flats in the NW region (see Section 4.6)
z phot Best-fit photometric redshift via maximum likelihood
z025/160/500/840/975 Redshift posterior percentiles, e.g. z025 → 2.5%
z phot chi2 Total χ2 at z phot
nusefilt Number of filters used in the fit
restU Rest-frame U -band flux (Fν , zeropoint = 28.9)
restU err Rest-frame U -band flux uncertainty (Fν , zeropoint = 28.9)
restV Rest-frame V -band flux (Fν , zeropoint = 28.9)
restV err Rest-frame V -band flux uncertainty (Fν , zeropoint = 28.9)
restJ Rest-frame J-band flux (Fν , zeropoint = 28.9)
restJ err Rest-frame J-band flux uncertainty (Fν , zeropoint = 28.9)
restus Rest-frame synth. u-band flux (Fν , zeropoint = 28.9)
restus err Rest-frame synth. u-band flux uncertainty (Fν , zeropoint = 28.9)
restgs Rest-frame synth. g-band flux (Fν , zeropoint = 28.9)
restgs err Rest-frame synth. g-band flux uncertainty (Fν , zeropoint = 28.9)
restis Rest-frame synth. i-band flux (Fν , zeropoint = 28.9)
restis err Rest-frame synth. i-band flux uncertainty (Fν , zeropoint = 28.9)
star min chi2 χ2 of best-fit stellar template
mu Magnification (best-fit; = 1 for foreground objects)
mu025/160/840/975 Magnification uncertainty percentiles

X = filter name, as defined in Section 2. Synthetic rest-frame ugi filters are detailed in Antwi-Danso et al. (2022).
26 J. R. Weaver et al.
Bradley, L. D., Coe, D., Brammer, G., et al. 2022, arXiv
e-prints, arXiv:2210.01777.
https://arxiv.org/abs/2210.01777
Brammer, G. 2019, Grizli: Grism redshift and line analysis
software, Astrophysics Source Code Library, record
ascl:1905.001. http://ascl.net/1905.001
Brammer, G. B., van Dokkum, P. G., & Coppi, P. 2008,
ApJ, 686, 1503, doi: 10.1086/591786
Broadhurst, T. J., Taylor, A. N., & Peacock, J. A. 1995,
ApJ, 438, 49, doi: 10.1086/175053
Carnall, A. C., Begley, R., McLeod, D. J., et al. 2023,
MNRAS, 518, L45, doi: 10.1093/mnrasl/slac136
Coe, D., Zitrin, A., Carrasco, M., et al. 2013, ApJ, 762, 32,
doi: 10.1088/0004-637X/762/1/32
Coe, D., Salmon, B., Bradač, M., et al. 2019, ApJ, 884, 85,
doi: 10.3847/1538-4357/ab412b
Ferrarese, L., Côté, P., Jordán, A., et al. 2006, ApJS, 164,
334, doi: 10.1086/501350
Finkelstein, S. L., Ryan, Russell E., J., Papovich, C., et al.
2015, ApJ, 810, 71, doi: 10.1088/0004-637X/810/1/71
Fitzpatrick, E. L., & Massa, D. 2007, ApJ, 663, 320,
doi: 10.1086/518158
Fox, C., Mahler, G., Sharon, K., & Remolina González,
J. D. 2022, ApJ, 928, 87, doi: 10.3847/1538-4357/ac5024
Furtak, L. J., Atek, H., Lehnert, M. D., Chevallard, J., &
Charlot, S. 2021, MNRAS, 501, 1568,
doi: 10.1093/mnras/staa3760
Furtak, L. J., Shuntov, M., Atek, H., et al. 2022a, MNRAS,
doi: 10.1093/mnras/stac3717
Furtak, L. J., Zitrin, A., Weaver, J. R., et al. 2022b, arXiv
e-prints, arXiv:2212.04381.
https://arxiv.org/abs/2212.04381
Gaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al.
2016, A&A, 595, A1, doi: 10.1051/0004-6361/201629272
Gaia Collaboration, Vallenari, A., Brown, A. G. A., et al.
2022, arXiv e-prints, arXiv:2208.00211.
https://arxiv.org/abs/2208.00211
Giavalisco, M., Ferguson, H. C., Koekemoer, A. M., et al.
2004, ApJL, 600, L93, doi: 10.1086/379232
Gonzaga, S., Hack, W., Fruchter, A., & Mack, J. 2012, The
DrizzlePac Handbook
Grogin, N. A., Kocevski, D. D., Faber, S. M., et al. 2011,
ApJS, 197, 35, doi: 10.1088/0067-0049/197/2/35
Hoaglin, D. C., Mosteller, F., & Tukey, J. W. 1983,
Understanding robust and exploratory data analysis,
Wiley Series in Probability and Mathematical Statistics,
(John Wiley,)
Hsiao, T. Y.-Y., Coe, D., Abdurro’uf, et al. 2022, arXiv
e-prints, arXiv:2210.14123.
https://arxiv.org/abs/2210.14123
Hunter, J. D. 2007, Computing in Science Engineering, 9,
90, doi: 10.1109/MCSE.2007.55
Ilbert, O., Arnouts, S., McCracken, H. J., et al. 2006, A&A,
457, 841, doi: 10.1051/0004-6361:20065138
Illingworth, G., Magee, D., Bouwens, R., et al. 2016, arXiv
e-prints, arXiv:1606.00841.
https://arxiv.org/abs/1606.00841
Illingworth, G. D., Magee, D., Oesch, P. A., et al. 2013,
ApJS, 209, 6, doi: 10.1088/0067-0049/209/1/6
Jarvis, M. J., Bonfield, D. G., Bruce, V. A., et al. 2013,
MNRAS, 428, 1281, doi: 10.1093/mnras/sts118
Jauzac, M., Richard, J., Jullo, E., et al. 2015, MNRAS,
452, 1437, doi: 10.1093/mnras/stv1402
Johnson, B. D., Leja, J., Conroy, C., & Speagle, J. S. 2021,
ApJS, 254, 22, doi: 10.3847/1538-4365/abef67
Kikuchihara, S., Ouchi, M., Ono, Y., et al. 2020, ApJ, 893,
60, doi: 10.3847/1538-4357/ab7dbe
Koekemoer, A. M., Faber, S. M., Ferguson, H. C., et al.
2011, ApJS, 197, 36, doi: 10.1088/0067-0049/197/2/36
Kokorev, V., Brammer, G., Fujimoto, S., et al. 2022, arXiv
e-prints, arXiv:2207.07125.
https://arxiv.org/abs/2207.07125
Kron, R. G. 1980, ApJS, 43, 305, doi: 10.1086/190669
Livermore, R. C., Finkelstein, S. L., & Lotz, J. M. 2017,
ApJ, 835, 113, doi: 10.3847/1538-4357/835/2/113
Lotz, J. M., Koekemoer, A., Coe, D., et al. 2017, ApJ, 837,
97, doi: 10.3847/1538-4357/837/1/97
Mason, C. A., Trenti, M., & Treu, T. 2015, ApJ, 813, 21,
doi: 10.1088/0004-637X/813/1/21
Merten, J., Coe, D., Dupke, R., et al. 2011, MNRAS, 417,
333, doi: 10.1111/j.1365-2966.2011.19266.x
Nardiello, D., Bedin, L. R., Burgasser, A., et al. 2022,
MNRAS, 517, 484, doi: 10.1093/mnras/stac2659
Oke, J. B. 1974, ApJS, 27, 21, doi: 10.1086/190287
Pagul, A., Sánchez, F. J., Davidzon, I., & Mobasher, B.
2021, ApJS, 256, 27, doi: 10.3847/1538-4365/abea9d
Paris, D., Merlin, E., Fontana, A., et al. 2023,
doi: 10.48550/ARXIV.2301.02179
Perrin, M. D., Sivaramakrishnan, A., Lajoie, C.-P., et al.
2014, in Society of Photo-Optical Instrumentation
Engineers (SPIE) Conference Series, Vol. 9143, Space
Telescopes and Instrumentation 2014: Optical, Infrared,
and Millimeter Wave, ed. J. Oschmann, Jacobus M.,
M. Clampin, G. G. Fazio, & H. A. MacEwen, 91433X,
doi: 10.1117/12.2056689
 The UNCOVER Photometric Catalog
Perrin, M. D., Soummer, R., Elliott, E. M., Lallo, M. D., &
Sivaramakrishnan, A. 2012, in Society of Photo-Optical
Instrumentation Engineers (SPIE) Conference Series,
Vol. 8442, Space Telescopes and Instrumentation 2012:
Optical, Infrared, and Millimeter Wave, ed. M. C.
Clampin, G. G. Fazio, H. A. MacEwen, & J. Oschmann,
Jacobus M., 84423D, doi: 10.1117/12.925230
Richard, J., Claeyssens, A., Lagattuta, D., et al. 2021,
A&A, 646, A83, doi: 10.1051/0004-6361/202039462
Rigby, J., Perrin, M., McElwain, M., et al. 2022, arXiv
e-prints, arXiv:2207.05632.
https://arxiv.org/abs/2207.05632
Roberts-Borsani, G., Morishita, T., Treu, T., et al. 2022,
ApJL, 938, L13, doi: 10.3847/2041-8213/ac8e6e
Rowe, B. T. P., Jarvis, M., Mandelbaum, R., et al. 2015,
Astronomy and Computing, 10, 121,
doi: 10.1016/j.ascom.2015.02.002
Salmon, B., Coe, D., Bradley, L., et al. 2020, ApJ, 889,
189, doi: 10.3847/1538-4357/ab5a8b
Schlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 737, 103,
doi: 10.1088/0004-637X/737/2/103
Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ,
500, 525, doi: 10.1086/305772
Scoville, N., Aussel, H., Brusa, M., et al. 2007, ApJS, 172,
1, doi: 10.1086/516585
Sharon, K., Bayliss, M. B., Dahle, H., et al. 2020, ApJS,
247, 12, doi: 10.3847/1538-4365/ab5f13
Shipley, H. V., Lange-Vagle, D., Marchesini, D., et al. 2018,
ApJS, 235, 14, doi: 10.3847/1538-4365/aaacce
Skelton, R. E., Whitaker, K. E., Momcheva, I. G., et al.
2014, ApJS, 214, 24, doi: 10.1088/0067-0049/214/2/24
Steinhardt, C. L., Jauzac, M., Acebron, A., et al. 2020,
ApJS, 247, 64, doi: 10.3847/1538-4365/ab75ed
27
Strait, V., Bradač, M., Coe, D., et al. 2021, ApJ, 910, 135,
doi: 10.3847/1538-4357/abe533
Treu, T., Schmidt, K. B., Brammer, G. B., et al. 2015,
ApJ, 812, 114, doi: 10.1088/0004-637X/812/2/114
Treu, T., Roberts-Borsani, G., Bradac, M., et al. 2022,
ApJ, 935, 110, doi: 10.3847/1538-4357/ac8158
van der Walt, S., Colbert, S. C., & Varoquaux, G. 2011,
Computing in Science Engineering, 13, 22,
doi: 10.1109/MCSE.2011.37
Weaver, J. R., Kauffmann, O. B., Ilbert, O., et al. 2022a,
ApJS, 258, 11, doi: 10.3847/1538-4365/ac3078
Weaver, J. R., Davidzon, I., Toft, S., et al. 2022b, arXiv
e-prints, arXiv:2212.02512.
https://arxiv.org/abs/2212.02512
Welch, B., Coe, D., Zackrisson, E., et al. 2022, ApJL, 940,
L1, doi: 10.3847/2041-8213/ac9d39
Whitaker, K. E., Ashas, M., Illingworth, G., et al. 2019,
ApJS, 244, 16, doi: 10.3847/1538-4365/ab3853
Wiener, N. 1949, Extrapolation, Interpolation, and
Smoothing of Stationary Time Series: With Engineering
Applications (The MIT Press),
doi: 10.7551/mitpress/2946.001.0001
Williams, H., Kelly, P. L., Chen, W., et al. 2022, arXiv
e-prints, arXiv:2210.15699.
https://arxiv.org/abs/2210.15699
Williams, R. E., Blacker, B., Dickinson, M., et al. 1996, AJ,
112, 1335, doi: 10.1086/118105
Willott, C. J., Abraham, R. G., Albert, L., et al. 2017,
CANUCS: The CAnadian NIRISS Unbiased Cluster
Survey, JWST Proposal. Cycle 1, ID. #1208
Windhorst, R. A., Cohen, S. H., Jansen, R. A., et al. 2023,
AJ, 165, 13, doi: 10.3847/1538-3881/aca163
Zheng, W., Postman, M., Zitrin, A., et al. 2012, Nature,
489, 406, doi: 10.1038/nature11446
Zitrin, A., Zheng, W., Broadhurst, T., et al. 2014, ApJL,
793, L12, doi: 10.1088/2041-8205/793/1/L12