JADES: JWST Deep Extragalactic Survey

• JADES is a JWST Guaranteed Time Observations program employing deep NIRCam imaging and multi-object NIRSpec spectroscopy to study galaxy evolution from reionization to cosmic noon.
• It uses a tiered survey design with deep and medium imaging, coordinated MIRI parallels, and extensive spectroscopic campaigns to analyze stellar populations, AGN activity, and emission-line diagnostics.
• JADES establishes a legacy dataset for precision studies of cosmic structure, galaxy assembly, and early black hole growth through comprehensive photometric and spectroscopic catalogs.

The JWST Advanced Deep Extragalactic Survey (JADES) is a JWST Guaranteed Time Observations program centered on GOODS-South and GOODS-North that combines deep NIRCam imaging, extensive NIRSpec multi-object spectroscopy, and coordinated MIRI parallels to study galaxy evolution from the first few hundred Myr after the Big Bang through cosmic noon. Built on the legacy of the HUDF/XDF, CANDELS, deep Chandra, ALMA, and related datasets, JADES was designed to construct robust photometric and spectroscopic samples, measure stellar populations, dust, abundances, ionization, morphology, and AGN activity, and constrain the role of galaxies in reionization (Eisenstein et al., 2023).

1. Program conception and scientific scope

JADES was formulated as a joint effort of the NIRCam and NIRSpec Guaranteed Time Observations teams. A 2021 overview described it as a 950-hour program targeting GOODS-North and South to chart galaxy evolution at $z>2$, potentially extending to $z>10$, through combined ultra-deep imaging and multiplexed spectroscopy over $\approx 1$–$5\,\mu$m (Bunker, 2021). The Cycle 1 overview described about 770 hours of Cycle 1 guaranteed time, with deep and medium imaging, extensive NIRSpec spectroscopy, and coordinated MIRI parallels, emphasizing galaxy assembly, reionization, chemical enrichment, dust, morphology, and AGN demographics (Eisenstein et al., 2023).

The scientific design is explicitly twofold. First, JADES targets the reionization era and the first few hundred Myr by combining photometric redshift selection, dropout searches, and spectroscopic confirmation. Second, it addresses the broader evolution of galaxies through cosmic noon by measuring rest-frame optical and ultraviolet continua, emission-line diagnostics, and structural properties. This combination places JADES at the intersection of luminosity-function work, stellar-population inference, ISM physics, AGN studies, and environmental studies.

2. Observing architecture and instrumental design

The original survey architecture used a tiered or “wedding-cake” design. In the Cycle 1 overview, the Deep NIRCam tier in GOODS-S covered approximately 45 arcmin$^2$ with an average of 130 hours spread over nine filters, while the Medium imaging covered approximately 175 arcmin$^2$ at an average of ~20 hours over 8–10 filters. The NIRSpec program included 2 deep pointings of 55 hrs, 14 medium pointings of ~12 hrs, and 15 shallower pointings of ~4 hrs, targeting over 5000 faint sources with low-, medium-, and high-resolution dispersers spanning 0.6–5.3 microns (Eisenstein et al., 2023).

By Data Release 5, the imaging footprint had expanded substantially through co-reduction with additional JWST programs. DR5 presents NIRCam mosaics in up to 18 filters over 469 arcmin$^2$, with 250 arcmin$^2$ having at least 8 filters of coverage; the total NIRCam exposure inventory is 1253 hr in the short-wavelength channel and 1070 hr in the long-wavelength channel across the combined GOODS-S and GOODS-N mosaics (Johnson et al., 22 Jan 2026).

Component	Core survey design	Released large-scale implementation
NIRCam imaging	Deep tier: $\sim45$ arcmin$^2$ at average 130 hr; Medium tier: $z>10$0 arcmin$z>10$1 at average $z>10$2 hr (Eisenstein et al., 2023)	DR5 NIRCam mosaics in up to 18 filters over 469 arcmin$z>10$3 (Johnson et al., 22 Jan 2026)
NIRSpec spectroscopy	Deep, medium, and shallow MOS tiers across 0.6–5.3 $z>10$4m (Eisenstein et al., 2023)	DR4 complete spectroscopic sample: 5190 targets (Curtis-Lake et al., 1 Oct 2025)
MIRI parallels	$z>10$5 arcmin$z>10$6 with 43 hr at 7.7 $z>10$7m plus wider medium-depth coverage (Eisenstein et al., 2023)	DR5 MIRI areas: 36.4 arcmin$z>10$8 in F770W, 25.8 arcmin$z>10$9 in F1280W, 22 arcmin$\approx 1$0 in F1500W (Alberts et al., 22 Jan 2026)

The instrumental synergy is central to JADES. NIRCam provides the imaging basis for dropout selection, photometric redshifts, morphology, and medium-band line sensitivity; NIRSpec provides secure redshifts and line/continuum diagnostics; MIRI extends the wavelength baseline into the mid-infrared, constraining rest-frame near-infrared light, dusty systems, and line contamination at very high redshift.

3. Data releases, calibration, and inference frameworks

JADES has evolved into a major public data resource. The DR5 photometric catalogs combine 35 space-based imaging mosaics from JWST/NIRCam, JWST/MIRI, HST/ACS, and HST/WFC3, built from approximately 1250 hours of JADES imaging plus extensive additional public observations. Source detection uses custom signal-to-noise mosaics and deblending optimized for JWST data, while source shapes and apertures are determined with a fast two-dimensional Gaussian regression method. DR5 provides forced circular-aperture photometry, ellipsoidal Kron photometry, curve-of-growth measurements, and template-based photometric redshifts from both native-resolution small apertures and common-PSF Kron photometry (Robertson et al., 22 Jan 2026).

DR5 also adds homogeneous structural and stellar-population products. The morphology release fits single-component Sérsic profiles to all sources detected in JADES NIRCam imaging, reporting over 3 million Sérsic fits for more than $\approx 1$1 sources; in a rest-frame optical sample of 24,692 galaxies at $\approx 1$2, it finds $\approx 1$3 and a relatively constant $\approx 1$4 (Carreira et al., 22 Jan 2026). The stellar-population catalog models roughly half a million sources with Prospector, flexible non-parametric SFHs, nebular emission, dust, metallicities, AGN emission, and an evolving star-forming main sequence prior, yielding posterior distributions for masses, SFRs, SFHs, dust attenuation, metallicities, and AGN contributions (Duan et al., 20 May 2026).

At the image-processing level, DR5 introduced dedicated corrections for low-surface-brightness systematics. The NIRCam wisp-subtraction paper models scattered-light artifacts with non-negative matrix factorization, constructing multi-component, filter- and detector-specific templates that reduce residual RMS and photometric bias in affected regions to levels consistent with clean detector areas (Wu et al., 22 Jan 2026). On the MIRI side, the DR5 coordinated-parallels release adds a custom persistence correction for saturated-source artifacts and emphasizes that pipeline ERR images underestimate uncertainties by factors of approximately 2–3, requiring mosaic-based noise estimation for science analyses (Alberts et al., 22 Jan 2026).

4. Reionization-era galaxies, photometric selection, and early structure

JADES rapidly became a primary dataset for the census of galaxies at cosmic dawn. An early JADES high-redshift catalog assembled 717 candidate galaxies at $\approx 1$5 across 125 square arcminutes of GOODS-S and GOODS-N using 15 JWST/NIRCam filters plus 5 HST/ACS filters; over 93\% of the sources were newly identified from the deep JADES imaging, including 31 new galaxy candidates at $\approx 1$6, and comparison to 42 spectroscopic redshifts yielded no catastrophic outliers with an average offset of $\approx 1$7 (Hainline et al., 2023).

DR5 expanded that census to 2081 sources selected at $\approx 1$8 over 469 square arcmin, spanning $\approx 1$9 to $5\,\mu$0, including 19 objects at $5\,\mu$1. In that release, the UV-continuum slope relation is parameterized as $5\,\mu$2, with fitted slopes of $5\,\mu$3 at $5\,\mu$4–9, $5\,\mu$5 at $5\,\mu$6–11, and $5\,\mu$7 at $5\,\mu$8. The same catalog reports that approximately 27\% of GOODS-S sources and 28\% of GOODS-N sources at $5\,\mu$9 are morphologically extended, with multiple bright knots or flattened morphologies detected even at $^2$0–14 (Hainline et al., 22 Jan 2026).

JADES has also identified large-scale structure during the first 500 Myr. A prominent candidate overdensity at $^2$1 in GOODS-S contains 18 galaxies with consistent photometric redshifts and clear F115W dropouts within $^2$2 comoving Mpc in projection. The inferred overdensity parameter is $^2$3, corresponding to an approximately 4× enhancement over the mean field density and a $^2$4 Poisson significance. This structure hosts one-third of all comparably bright sources and contributes approximately 48\% of the total star formation in the GOODS-S slice at $^2$5; a radial trend in photometrically inferred Ly$^2$6 transmission is consistent with an emerging ionized bubble (Wu et al., 22 Jan 2026).

The spectroscopic program reinforces these photometric results statistically. DR4 presents spectra for 5190 targets and derives robust redshifts for 3297 galaxies, including 396 at $^2$7, while rest-UV luminosity functions at $^2$8 agree well with previous photometric and spectroscopic determinations and show modest interloper fractions in UV-selected bins (Curtis-Lake et al., 1 Oct 2025).

5. Spectroscopic diagnostics, galaxy demographics, and black-hole growth

The spectroscopic conception of JADES was to use NIRSpec over 0.6–5 μm at $^2$9, 1000, and 2700 to secure redshifts and measure rest-frame optical and ultraviolet lines including H$^2$0, H$^2$1, [O III], [O II], [N II], [S II], Ly$^2$2, and He II $^2$3, thereby constraining dust, star formation, metallicity, abundances, ionization, excitation, and ionizing photon production (Bunker, 2021). DR4 operationalizes that design with the complete NIRSpec sample, defining “gold” UV- and F444W-selected subsamples for uniform analyses and demonstrating high success rates for both UV-bright reionization-era targets and F444W-selected galaxies at $^2$4 (Curtis-Lake et al., 1 Oct 2025).

One major spectroscopic result is the identification of a substantial population of faint AGN through stacking. By stacking $^2$5 NIRSpec grating spectra at $^2$6, JADES detected a broad component of H$^2$7 without a corresponding broad [O III] component in multiple stacks, implying emission from the Broad Line Region rather than outflows. The inferred black holes have masses of a few times $^2$8 and accrete at $^2$9–0.1. The resulting black-hole mass function at $^2$0 rises steeply toward low masses, and the paper argues that the measurements are consistent with models in which black holes grow through short bursts of super-Eddington accretion (Geris et al., 27 Jun 2025).

This spectroscopic dimension is central to JADES’s role in early-universe studies. It ties photometric samples to secure rest-frame optical diagnostics, exposes populations that are not recoverable from imaging alone, and enables direct tests of whether high-redshift galaxies and AGN follow local scaling relations or depart from them.

6. Time-domain, Galactic foreground, and dusty-galaxy science

JADES is not restricted to static extragalactic imaging. The JADES Transient Survey used two nearly identical NIRCam epochs in the GOODS-S Deep Field separated by roughly one year to discover and classify 79 supernovae, including 38 at $^2$1, 23 at $^2$2, 8 at $^2$3, and 7 at $^2$4. The measured detection rate is $^2$5–2 SNe per arcmin$^2$6 per year, and multi-epoch NIRCam photometry was sufficient to classify a Type Ia SN at $^2$7, a Type IIP SN at $^2$8, and a Type Ic-BL SN at $^2$9. The same study showed that two apparent $^2$0 galaxy candidates from a single epoch were actually fading transients, demonstrating that moderate- and high-redshift supernovae can contaminate extreme dropout selections (DeCoursey et al., 2024).

The survey also produces Galactic science in deep extragalactic fields. A NIRSpec/PRISM spectrum confirmed JADES-GS-BD-9 as a T5–T6 dwarf with $^2$1–900 K, metallicity $^2$2, distance 1.8–2.3 kpc, and proper motion $^2$3 mas yr$^2$4, implying a tangential velocity consistent with the thick disk edge or halo (Hainline et al., 2024). A broader JADES analysis then identified 41 brown dwarf and brown dwarf candidates across GOODS-S and GOODS-N, with 31 consistent with T-dwarf temperatures out to 5–6 kpc and 10 consistent with Y-dwarf temperatures out to 1–2 kpc; this work further argued that Y dwarfs are plausible contaminants in searches for ultra-high-redshift galaxies (Hainline et al., 30 Sep 2025).

JADES also has direct relevance for dusty galaxy evolution. A GOODS-S analysis of dusty star-forming galaxies and proto-spheroids adopted an 87.5 arcmin$^2$5 JADES-matched footprint and found that photometric redshifts estimated from simulated JWST photometry showed $^2$6 accuracy and that JWST can provide stellar mass estimates up to 0.1 dex for the majority of proto-spheroids at $^2$7. In that framework, JADES reaches lower-mass dusty systems during cosmic noon than pre-JWST surveys, while Spitzer and Herschel data remain necessary for robust dust and obscured-SFR constraints (Mitra et al., 10 Jun 2025).

7. Legacy status, limitations, and future development

JADES now functions as a layered public archive as much as a single observing campaign. DR5 NIRCam imaging releases both combined mosaics and per-epoch or per-program subregion mosaics, explicitly to support variability studies, transient searches, and proper-motion work (Johnson et al., 22 Jan 2026). The MIRI coordinated-parallels release adds forced photometry tied to NIRCam detections and provides rest-frame near-infrared constraints on early galaxies, dusty systems, and quiescent candidates (Alberts et al., 22 Jan 2026). The morphology, photometric, and stellar-population catalogs together establish a uniform basis for statistical studies of galaxy sizes, stellar masses, SFHs, dust, AGN fractions, and high-redshift selection (Robertson et al., 22 Jan 2026).

Several caveats recur across JADES science analyses. High-redshift overdensity membership based on photometric redshifts requires spectroscopic confirmation; Ly$^2$8-based bubble inferences are explicitly described as tentative; field contamination, projection effects, and cosmic variance remain important uncertainties in reionization-era structure studies (Wu et al., 22 Jan 2026). Single-epoch dropout searches can be contaminated by supernovae (DeCoursey et al., 2024), and ultra-cool brown dwarfs can mimic the colors of the most extreme high-redshift candidates (Hainline et al., 30 Sep 2025). On the instrumental side, low-surface-brightness artifacts such as wisps and persistence required dedicated mitigation in DR5, and the MIRI release emphasizes that science-grade uncertainties must be derived from mosaic statistics rather than pipeline ERR maps alone (Wu et al., 22 Jan 2026).

Future directions are already embedded in the release papers. DR5 morphology papers state that multi-component surface-brightness profiles are planned (Carreira et al., 22 Jan 2026); high-redshift photometric papers emphasize improved template sets and Ly$^2$9 damping-wing prescriptions (Hainline et al., 22 Jan 2026); overdensity work calls for JWST/NIRSpec, MIRI/LRS, and ALMA [O III] 88 μm follow-up to secure systemic redshifts and map ionized regions (Wu et al., 22 Jan 2026). Taken together, these developments suggest that JADES has transitioned from an initial deep-field survey into a broader, legacy-quality framework for precision studies of cosmic dawn, galaxy assembly, AGN growth, time-domain astrophysics, and the Galactic foreground.