GA-NIFS GTO Survey: Galaxy Assembly & AGN Feedback

Updated 13 December 2025

The paper presents the GA-NIFS GTO Survey, a JWST/NIRSpec IFS initiative that spatially resolves galaxies and AGN to study assembly, feedback, and chemical enrichment over 2<z<11.
The survey employs high spectral and spatial resolution modes with advanced data reduction to map ISM diagnostics, stellar populations, and kinematics at sub-kiloparsec scales.
Key results include evidence of AGN-driven outflows, dual/triple AGN occurrences, and detailed metallicity stratifications that challenge existing galaxy evolution models.

The GA-NIFS GTO Survey (“Galaxy Assembly with NIRSpec Integral Field Spectroscopy”) comprises a series of JWST Guaranteed Time Observations and complementary ground-based campaigns dedicated to spatially resolving the physical, kinematic, and chemical properties of galaxies and AGN across cosmic time, from the local Universe up to $z > 10$ . Integrating high spectral and spatial resolution near-infrared integral field spectroscopy over a statistically robust, multi-phase sample, GA-NIFS addresses key questions in galaxy assembly, AGN feedback, chemical enrichment, and merger-driven evolution by mapping ISM diagnostics, stellar populations, and gas flows at sub-kiloparsec scales.

1. Survey Design, Selection, and Scientific Motivations

GA-NIFS was conceived to exploit the unique combination of JWST/NIRSpec IFU sensitivity (reaching line fluxes $\lesssim10^{-21}$ erg s $^{-1}$ cm $^{-2}$ ) and spatial resolution ($0.05''$–$0.2''$ native spaxels, $\sim0.26$ –$1$ kpc at $z = 3$ –7), targeting a heterogeneous set of star-forming galaxies, AGN hosts, bright quasars, and group environments over $2Jones et al., 2024, Marconcini et al., 2024, Perna et al., 2023, Riffel et al., 2022). The primary selection draws from well-characterized extragalactic fields (COSMOS, GOODS, CANDELS, JADES) and incorporates prior photometry/spectroscopy (e.g., ALMA [C II]/[O III] detections, HST/JWST imaging, and X-ray/IR AGN identification). Both highly magnified ( $\mu\sim10$ –30, e.g. MACS1149-JD1 at $z=9.11$ ) and unlensed sources are targeted to maximize spatial fidelity and probe faint, compact systems.

The survey addresses several intertwined science objectives:

Mapping ISM thermodynamic conditions (electron temperature $T_e$ , density $n_e$ ), dust obscuration [E(B–V)], and gas-phase metallicities over resolved regions.
Charting the assembly, disruption, and feedback processes tied to mergers, starbursts, and AGN-driven outflows, including dual/triple AGN incidence at high redshift (Perna et al., 2023).
Constraining the chemical enrichment cycle with “direct” and strong-line methods (including abundance stratification and rare-element diagnostics out to $z\sim10$ ), thereby anchoring the mass–metallicity and abundance scaling relations.
Enabling kinematic decomposition—rotation, turbulence, binarity, and outflows—over galaxies, groups, and proto-cluster environments (Lamperti et al., 2024, Pino et al., 2023).
Systematically relating black hole and host galaxy properties in the epoch of reionization (Marshall et al., 2023).

2. Observational Strategies and Instrumentation

GA-NIFS employs JWST/NIRSpec IFS in at least two modes per field:

Moderate to high resolution (G395M/H, F290LP, $R\sim1000$ –2700): capturing rest-frame UV/optical nebular lines ([O II] $\lambda\lambda$ 3727,29, [Ne III] $\lambda$ 3869, Balmer series, [O III] $\lambda$ 5007, [N II] $\lambda$ 6584, [S II] $\lambda\lambda$ 6716,31, [S III] $\lambda$ 9531). Typical FOVs span $3.1'' \times 3.2''$ ( $\sim20$ –$30$ kpc at $z > 3$ ) with spatial sampling as fine as $0.05''$ (Jones et al., 2024, Marconcini et al., 2024).
Low-resolution PRISM ( $R\sim100$ ): continuous $\lambda=0.6$ – $5.3\,\mu$ m coverage for broadband SED fitting, Ly $\alpha$ /UV continuum, and weak line detection (Marconcini et al., 2024).
Dithering (8-point “medium pattern”) to optimize spatial registration and mitigate bad pixels.
Exposure times per field range from 4–20 ks, selected to reach S/N $\gtrsim10$ on key lines in diffuse extra-nuclear regions.
Complementary ground-based spectroscopy (e.g., VLT/MUSE, VIMOS) is incorporated for rest-UV and Ly $\alpha$ line coverage at $z\lesssim6$ (Perna et al., 2024, Ji et al., 2024).

Data are reduced with the latest STScI pipeline releases (v1.8.2–v1.15.0 and CRDS 1105–1241) augmented by custom routines for cosmic-ray (“snowball”) masking, $1/f$ noise correction, open-shutter leakage elimination, and optimal astrometric alignment ( $\lesssim0.1''$ ) to auxiliary JWST/NIRCam and ground-based imaging. Final cubes are constructed via drizzle resampling to $0.05''$ spaxels.

3. ISM Diagnostics and Analysis Methodology

GA-NIFS leverages a comprehensive suite of nebular line diagnostics, full-spectrum synthesis, and multi-Gaussian spectral fitting:

Key line ratios for physical conditions:
- $T_e$ from [O III] $\lambda$ 4363/ $\lambda$ 5007,
- $n_e$ from [O II] $\lambda$ 3727/ $\lambda$ 3729 or [S II] $\lambda$ 6716/ $\lambda$ 6731,
- Reddening from Balmer decrement [e.g., H $\alpha$ /H $\beta$ ].
- Ionization parameter from [O III]/O II, [Ne III]/O II.
Strong-line metallicity calibrations utilizing O32, R23, R2, R3, and Ne3O2 [Curti et al. 2020].
Spectral fits adopt single or two-component Gaussians per emission line (narrow: FWHM $\leq$ 350 km s $^{-1}$ ; broad: 350–1000 km s $^{-1}$ ), with fixed atomic flux ratios (e.g., [O III] $\lambda$ 4959/ $\lambda$ 5007 = 0.301).
Continuum models are single power-laws or SED-derived templates, with dust extinction set via Calzetti et al. and fixed E(B–V) in cases where the fit is otherwise degenerate.
Spaxel-by-spaxel fitting enables 2D mapping of resolved SFR (from H $\beta$ /H $\alpha$ ), metallicity, velocity ( $v_{50}$ ), line width ( $w_{80}$ ), and higher-order kinematic moments (Jones et al., 2024, Marconcini et al., 2024, Pino et al., 2023).

Physical parameter maps are directly compared to ancillary ALMA data (e.g. [C II]158 $\mu$ m, CO) for multi-phase ISM constraints and morphological context.

4. Key Results: Exemplars and Physical Insights

Multiple high-redshift galaxies observed by GA-NIFS illustrate the survey's core capabilities:

B14-65666 at $z=7.15$ : Resolved into two major nuclei and multiple clumps, shows a mass $\log_{10}(M_*/M_{\odot})=9.8\pm0.2$ and a total SFR $308\pm39$ $M_{\odot}$ yr $^{-1}$ derived from H $\beta$ luminosity. Metallicities ( $Z_g \simeq 0.24$ – $0.47~Z_\odot$ ) place it on the high- $z$ mass–metallicity relation, and broad [O III] components (FWHM $\gtrsim$ 650–760 km s $^{-1}$ ) localized to one nucleus reveal active tidal disturbance or outflow signatures. ALMA velocity fields ([C II], [O III]) mirror those from NIRSpec (Jones et al., 2024).
MACS1149-JD1 at $z=9.11$ : Lensed system with two stellar/gas clumps, electron densities $n_e\sim1\times10^3$ cm $^{-3}$ , $T_e\sim1.6\times10^4$ K, direct-method metallicities of $-1.2$ to $-0.7$ dex solar, and a 3 $\sigma$ anti-correlation between $Z$ and SFR density, indicative of gas inflow-driven dilution fueling starbursts.
GS133 AGN at $z=3.47$ : Spatially resolves a $\sim$ 3 kpc, $v\sim1000$ km s $^{-1}$ , twin-lobed, galaxy-scale ionized outflow with a mass outflow rate $\dot{M}_\text{out}\sim 200$ $M_\odot$ yr $^{-1}$ , mass-loading $\eta=1$ –10, and coupling efficiency $\epsilon=0.1$ –2%, matching energy-conserving wind model predictions for AGN feedback (Perna et al., 2024).
Dual/Triple AGN fraction: Among 16 NIRSpec-observed high- $z$ systems, 3–5 show evidence of dual/triple AGN (f $_\text{dual}=0.19$ – $0.31\pm0.13$ ), higher than the $\sim10\%$ fraction predicted by simulations, underlining frequent merger-driven SMBH fueling (Perna et al., 2023).
Chemical stratification: E.g., GS_3073 at $z=5.55$ exhibits extreme N/O enhancement, but only in dense nuclear regions ( $n_e\sim10^5$ cm $^{-3}$ ), with more “normal” abundances traced by optical lines in extended, lower-density gas—evidence for rapid, local enrichment by massive stars or AGN disk winds (Ji et al., 2024).

5. Comparative Studies at Low Redshift: Gemini NIFS AGNIFS Survey

A parallel effort with Gemini/NIFS targets 18–29 nearby (z < 0.022) Swift-BAT-selected Seyfert galaxies using NIR IFS at 0.12–0.20 $''$ ( $\sim$ 10–80 pc) (Riffel et al., 2022, Riffel et al., 2017, Riffel et al., 2017). This enables:

Mapping stellar population gradients and kinematics, revealing a nuclear dominance by intermediate-age (0.5–2 Gyr) stars with suppressed very recent star formation within $\lesssim50$ pc (“nuclear rejuvenation”).
Quantifying circumnuclear gas distribution: ionized-gas ( $\Sigma_\mathrm{H\,II}$ ) and warm molecular ( $\Sigma_\mathrm{H_2}$ ) mass profiles, showing steeper concentration of ionized gas toward the nucleus, while H $_2$ is more extended.
Deriving inflow and outflow rates via dynamical modeling, correlating stellar/gas kinematic features with AGN luminosity (elevated velocity residuals in more powerful AGN).
Establishing that massive gas reservoirs ( $M_{\mathrm{gas}}\sim10^{7}$ – $10^{9}~M_\odot$ ) are present on 100–500 pc scales, sufficient to support both sustained AGN fueling and local star formation.

6. Broader Implications and Legacy Value

GA-NIFS has established a new empirical framework for resolved studies of ISM physics, star formation, and feedback in both normal and extreme galactic environments across 10 billion years of cosmic history:

It has demonstrated that massive, metal-enriched galaxies, mergers, and AGN-driven feedback are all present and dynamically relevant already at $z>7$ (Jones et al., 2024, Marconcini et al., 2024).
The detection of strong outflows, frequent AGN pairs, and spatially complex metallicity fields challenge simplified chemical evolution models and inform next-generation hydrodynamical simulations.
The survey's systematic approach to ISM mapping across mass and redshift ranges provides a statistical baseline to trace the coevolution of star formation, black hole growth, and galaxy assembly.

Its high-precision physical parameter maps and uniquely deep emission-line inventories will serve as calibration and testing grounds for galaxy-formation models and future legacy surveys.