JWST Advanced Deep Extragalactic Survey (JADES)

• JADES is a cutting-edge extragalactic survey that maps galaxy formation and evolution from cosmic dawn to cosmic noon using ultra-deep infrared imaging and spectroscopy.
• It employs a multi-tiered 'wedding-cake' design with instruments like NIRCam, NIRSpec, and MIRI to capture high-resolution data across a redshift range of 1 to 15.
• The survey delivers advanced data products and catalogs that enable quantitative studies of reionization, stellar mass assembly, and morphological evolution in galaxies.

The JWST Advanced Deep Extragalactic Survey (JADES) is a flagship program of the James Webb Space Telescope (JWST) targeting galaxy formation, assembly, and evolution from cosmic dawn through cosmic noon. Leveraging ultra-deep near- and mid-infrared imaging and spectroscopy, JADES systematically samples the properties and demographics of galaxies in the GOODS-South and GOODS-North legacy deep fields. The survey exploits the unique power of JWST’s NIRCam and NIRSpec instruments, complemented by MIRI coordinated parallels, with unprecedented depth, spatial resolution, and rest-frame optical coverage across a redshift interval spanning 1 ≲ z ≲ 15. JADES’s multi-tier “wedding-cake” design, advanced data processing pipelines, and comprehensive public data releases provide a foundational resource for extragalactic astrophysics, enabling quantitative investigations of reionization, stellar mass assembly, dust-obscured populations, AGN activity, morphological evolution, environmental structure, and the emergence of the first cosmic structures.

1. Survey Design, Architecture, and Instrumentation

JADES was allocated ~770 (ultimately ~950) hours of JWST Cycle 1–4 observation, focusing on GOODS–South (GOODS-S) and GOODS–North (GOODS-N), the premier extragalactic reference fields with maximal multiwavelength legacy overlap. The two-tier imaging architecture comprises a “Deep” region (~45 arcmin² in GOODS-S) with ~130 hours/filter over nine NIRCam wide/medium bands, and a “Medium” tier (~175 arcmin² split between both GOODS fields) with ~20 hours/filter in 8–10 bands. NIRCam imaging spans 0.7–5 μm using filters F070W, F090W, F115W, F150W, F200W, F277W, F335M, F356W, F410M, and F444W (Deep) or a similar set in Medium. Point-source depths reach ≳30 AB mag in the Deep core.

Spectroscopy is carried out using NIRSpec’s Micro-Shutter Assembly (MSA) in multi-object mode, employing both the low-dispersion Prism (R = 30–300, 0.6–5.5 μm) and three medium-resolution gratings (G140M, G235M, G395M; R ≈ 500–1500, 0.7–5.1 μm), with high-resolution G395H (R ≈ 2700) deployed for select targets and ultra-deep integrations. The MSA multiplexes up to 200 targets per configuration. Spectroscopic coverage is “wedding-caked” into ultradeep (≥32 h), deep (≥16 h), and medium (∼1–1.8 h) integration subfields, targeting over 5000 galaxies per field (Eisenstein et al., 2023, Rieke et al., 2023, D'Eugenio et al., 2024, Scholtz et al., 1 Oct 2025).

Coordinated MIRI parallel imaging in F770W, F1280W, and F1500W achieves 5σ depths of up to 28.2 AB at 7.7 μm over ~10 arcmin² and ≳26 AB over ~40 arcmin² total, essential for probing dust-obscured systems and rest-frame optical emission at z > 6 (Alberts et al., 22 Jan 2026).

2. Data Processing, Photometry, Morphology, and Catalog Products

JADES employs a rigorously characterized, multi-stage data reduction pipeline, starting from raw JWST/NIRCam, MIRI, and public HST exposures:

• Stage 1 calibration covers dark, linearity, cosmic-ray, and persistence corrections.
• Stage 2 performs flat-fielding, custom “sky flats,” 1/f noise removal, and wisp subtraction using non-negative matrix factorization (NMF), achieving improved background uniformity and faint-source reliability (Wu et al., 22 Jan 2026).
• Stage 3 aligns exposures astrometrically to ≤5 mas (rms, Gaia-tied), mosaicks to a 0.03″ pixel scale, and outputs per-filter, per-program tiles and deep combined mosaics (Johnson et al., 22 Jan 2026).

Source detection is based on SNR-weighted stacks, with advanced deblending (e.g., watershed, 2D Gaussian regression) and size estimation (Robertson et al., 22 Jan 2026). Forced aperture, Kron elliptical, and curve-of-growth photometry are computed per object and band, with uncertainties derived via pixel-level regression incorporating correlated noise and empirical variance scaling.

Morphological analysis is performed via Bayesian single-component Sérsic modeling (pysersic over JAX/NumPyro, NUTS sampler), yielding resolved effective radii, Sérsic index, axis ratio, and orientation parameters for >10⁵ sources per field and filter (~3 million fits total) (Carreira et al., 22 Jan 2026). Bulge–disk decompositions on deep subfields support direct tracing of structural evolution.

The data releases (e.g., DR5) provide:

• Multi-HDU FITS catalogs with photometry, photo-z, morphology, and error characterization for ≳500,000 objects across ∼469 arcmin² in GOODS-S and GOODS-N.
• Astrometrically aligned science mosaics in up to 18 NIRCam and 8 MIRI bands, plus HST/ACS and WFC3 reprojected to the JWST grid.
• Public online database and interactive “FitsMap” genome for mosaic visualization, catalog query, and cross-referencing (Robertson et al., 22 Jan 2026).

3. Spectroscopic Campaigns and Derived Physical Properties

NIRSpec MSA spectroscopy is executed at multiple depths, enabling population-scale rest-optical/UV line diagnostics out to z ≳ 14.2 (Scholtz et al., 1 Oct 2025). Data reduction includes:

• Wavelength and flat-field calibration (empirical path-loss correction, intra-shutter zero-point adjustment).
• Background subtraction (per-exposure, per-shutter, nod pairs, cosmic-ray rejection).
• Independent 1D/2D spectral extraction for point vs. extended sources, with flexible pixel apertures (3, 5, and 15 pixels).
• Comprehensive flux and uncertainty propagation; covariance structures supplied for ultra-deep stacks.

Redshifts are determined from emission lines or continuum breaks, with systematic offset Δz_phot–z_spec/(1+z_spec) ≃ 0.026, and outlier fractions ~7% for robust sub-samples (Duan et al., 20 May 2026). Line fluxes are extracted via pPXF [prism] or MCMC [medium gratings], with typical sensitivity: 5σ ≲ 5×10⁻¹⁹ erg s⁻¹ cm⁻² (ultradeep prism), 3–5×10⁻¹⁸ erg s⁻¹ cm⁻² (medium gratings).

Derived physical parameters use the Prospector Bayesian SED framework with an evolving SF Main Sequence prior and flexible 7-bin non-parametric SFHs. Outputs include:

• Stellar mass (completeness to M★ ∼ 10⁷ M⊙ at z∼1, ∼10⁸ M⊙ at z∼6; typical σ≃0.25–0.5 dex).
• SFRs (recent 10 Myr average), sSFR, stellar and gas-phase metallicities, dust attenuation, AGN fractions.
• Multi-component posterior distributions, fit statistics, and diagnostics for 500,000 objects.
• Data validated against >16,000 spectroscopic redshifts and NIRSpec photometric spectra (Duan et al., 20 May 2026).

4. Galaxy Assembly, Morphological and Environmental Evolution

JADES reveals the statistical properties of galaxies from z ∼ 1 to z > 15, with particular advance in characterizing the structure and diversity of the early galaxy population:

• Galaxy sizes decrease sharply with redshift: effective radii scaling as r_eff ∝ (1+z)^–0.635 (rest-optical, resolved to sub-kpc at z > 7) (Carreira et al., 22 Jan 2026).
• Bulge and disk evolution are decoupled, with bulges growing only marginally (r_eff,bulge ∝ (1+z)^–0.23), and disks growing rapidly (r_eff,disk ∝ (1+z)^–1.09).
• Clumpy, multi-knot, and extended morphologies are already common out to z ≳ 13–14, supporting rapid assembly and hierarchical growth (Hainline et al., 22 Jan 2026).
• Star-forming main sequence and non-parametric SFHs show widespread rising, bursty, and quenched evolutionary tracks at high redshift.

Notably, JADES identifies a striking candidate galaxy overdensity at z ≈ 10.5 with an enhancement of a factor four in number density and evidence for accelerated growth, close interactions, and associated local reionization bubbles (tentatively mapped via spatially varying Lyα transmission), providing the first robust laboratory for environmental effects and structure formation in the first 500 Myr (Wu et al., 22 Jan 2026).

5. Extragalactic Populations and Cosmic Dawn

JADES catalogs at z > 8, built from both SED-based template fitting and dropout techniques, establish the largest uniform high-redshift samples:

• Over 2000 photometric galaxy candidates with za > 8, including 19 at za > 14; robust selection criteria yield low contamination (Hainline et al., 22 Jan 2026).
• UV luminosity distributions span M_UV ≈ –22 to –16, with blue rest-UV slopes steepening with redshift (β–M_UV relationship steepens at z > 11).
• >25% of sources at z > 8 are morphologically extended (beyond the PSF) or multi-knot, indicating early assembly of galactic structure and star formation in-situ or by mergers.
• Faint-end completeness and cross-validation with NIRSpec/FRESCO provides direct calibration of the UV luminosity and mass functions to the lowest accessible limits (Hainline et al., 2023).

JADES also delivers the first systematic mapping of dusty star-forming galaxies (DSFGs) and AGN at high-z, quantifying stellar/dust mass, SFRs, and bolometric properties, and resolving the emergence of dusty, massive, and obscured populations previously inaccessible in optical/NIR surveys (Mitra et al., 10 Jun 2025).

6. Time-domain Science and Black Hole Demographics

The JADES Transient Survey utilizes repeat ultra-deep NIRCam imaging to identify and classify high-z supernovae (SNe), with a discovery rate of 1–2 arcmin⁻² yr⁻¹ at m_AB = 30 and redshifts up to z > 5, including SNe types Ia and core-collapse at z ≃ 2.9–3.6. This demonstrates JWST's capability as a time-domain discovery engine and highlights the need for temporal filters against high-z dropout contamination by SNe (DeCoursey et al., 2024).

Stacked NIRSpec spectra reveal the existence of a pervasive population of low-mass (M_BH ∼ 10⁶ M⊙) black holes at 3 < z < 7, accreting at low Eddington ratios (L_bol/L_Edd ≃ 0.02–0.1). Many of these black holes align with the local M_BH–M_* relation, contrasting prior results that favored overmassive early black holes. The observed black-hole mass function at 3 < z < 5 rises steeply to low masses and supports the necessity of short super-Eddington growth epochs in early BH assembly (Geris et al., 27 Jun 2025).

7. Statistical Anomalies and Cosmological Implications

An analysis of spiral galaxy rotation in the JADES GOODS-S field finds a 50% excess of galaxies rotating opposite to the Milky Way (relative sense), statistically significant at p ≃ 7 × 10⁻⁴ (3.4σ) (Shamir, 26 Feb 2025). Spin-asymmetry is maximal near the Southern Galactic pole. Inter-field comparisons (HST UDF, DESI, JWST early-release fields) and empirical confirmation of method symmetry corroborate this result.

Interpretations include primordial large-scale anisotropy (e.g., cosmic vorticity, rotating Universe), local observational/selection effects (e.g., Doppler-brightening bias), or unknown spin–emission mechanisms. These may have relevance to the origin of the H₀ tension and the formation of early massive spirals at z > 12.

8. Legacy and Future Prospects

JADES public data releases (DR1–DR5) provide the community with fully reduced mosaics, multi-band photometry, morphological and spectroscopic catalogs, and value-added stellar population grids for statistical and detailed applications. The survey’s depth, completeness, and wavelength leverage ensure its foundational status for calibrating galaxy evolution, cosmological parameters, black-hole–galaxy co-evolution, and forecasting future missions.

Planned extensions include targeted NIRSpec kinematic follow-up (direct spin–velocity correlation studies), Northern polar field mapping for spin-bias reversal, and deep, wide-area Euclid/Roman imaging over the polar caps. JADES datasets are expected to enable further advances in the characterization of faint, diffuse, and rare extragalactic populations, environmental structure at early epochs, and rigorous investigation of cosmological anomalies and tensions.

(Eisenstein et al., 2023, Rieke et al., 2023, D'Eugenio et al., 2024, Scholtz et al., 1 Oct 2025, Johnson et al., 22 Jan 2026, Alberts et al., 22 Jan 2026, Robertson et al., 22 Jan 2026, Carreira et al., 22 Jan 2026, Wu et al., 22 Jan 2026, Hainline et al., 22 Jan 2026, Wu et al., 22 Jan 2026, Duan et al., 20 May 2026, Mitra et al., 10 Jun 2025, Geris et al., 27 Jun 2025, Shamir, 26 Feb 2025, DeCoursey et al., 2024)