---
title: 'AESOPICA Simulations: Early AGN in Dwarf Galaxies'
url: https://www.emergentmind.com/topics/aesopica-simulations
type: topic
---

# AESOPICA Simulations: Early AGN in Dwarf Galaxies

Searching arXiv for AESOPICA and closely related simulation papers to ground the article.
Search query: AESOPICA simulations Koudmani FABLE low-mass AGN high redshift
Searching arXiv…
AESOPICA simulations are a new suite of large-volume cosmological simulations built upon the FABLE galaxy formation model and used to place two unusual low-mass galaxies at \(z\sim 7.66\) into a broader theoretical context of early AGN-host dwarfs [2507.14936]. In the current literature, AESOPICA appears not as the primary engine of observational inference, but as an interpretive framework for asking whether the measured outflows in these systems are plausibly associated with efficient black-hole growth and AGN feedback. The comparison is focused narrowly on outflow observables, especially \(v_{\rm out}/v_{\rm esc}\) and \(\eta=\dot M_{\rm out}/{\rm SFR}\), rather than on metallicity gradients, AGN classification, or a full posterior matching of simulated analogues [2507.14936].

## 1. Definition, lineage, and scientific role

AESOPICA is introduced as “a new suite of large-volume cosmological simulations (Koudmani et al., in prep.) built upon the Fable galaxy formation model” with targeted changes designed to model the growth of infant supermassive black holes at early times [2507.14936]. The underlying framework is therefore FABLE, while AESOPICA specifically extends the black-hole sector of the physics toward early-Universe, low-mass systems. The published description is explicit that the suite explores black-hole growth and AGN feedback in low-mass galaxies and is meant to capture early SMBH seeding and accretion physics, but it does not spell out a full hydrodynamic subgrid inventory in detail [2507.14936].

In the observational study that currently provides the public description, AESOPICA functions as an interpretive bridge rather than as the core analysis machinery. The JWST/NIRSpec/IFU data establish the existence of low metallicities, flat metallicity profiles, and kinematical components decoupled from galactic rotation; AESOPICA is then used to test whether the measured outflow properties resemble those of simulated AGN-host galaxies at the same epoch [2507.14936]. This role is methodologically important: AESOPICA is deployed as a qualitative-to-semiquantitative context for physical interpretation, not as a strict inference engine.

## 2. Black-hole growth prescriptions and simulation modes

The suite explores three key modifications relative to more standard fiducial prescriptions: efficient accretion in the low-mass regime following Koudmani et al. (2022), super-Eddington accretion, and a broad range of seed masses, \(10^2\)–\(10^5\,M_\odot\), with seed evolution followed from \(z\sim 20\) [2507.14936]. The purpose of these changes is to test whether dwarf and other low-mass galaxies at very high redshift can host sufficiently efficient BH growth to generate strong AGN-driven outflows.

| Mode | Prescription | Interpretation in the comparison |
|---|---|---|
| **Fiducial** | fiducial accretion following FABLE | standard or less efficient AGN growth |
| **BoostAcc** | boosted accretion, allowing efficient AGN activity in low-mass galaxies following Koudmani et al. (2022) | efficient BH feeding in dwarfs |
| **SuperEdd+BoostAcc** | boosted accretion plus super-Eddington accretion, up to 10 times the Eddington limit | efficient dwarf accretion with super-Eddington episodes |

These modes define the main theoretical contrast used in the paper. The distinction is not between different galaxy-formation backbones, but between different assumptions about how readily BHs in low-mass galaxies can accrete and drive feedback. This suggests that AESOPICA is best understood as a targeted extension of a FABLE-based galaxy-formation framework toward early, low-mass AGN phenomenology rather than as a wholly separate simulation ecosystem [2507.14936].

## 3. Observational systems used to contextualize AESOPICA

The comparison is anchored in two galaxies, SMACS J0723 NIRSpec-ID6355 and ID10612, both at \(z\approx 7.66\), with stellar masses from Curti et al. (2023) of \(\log(M_*/M_\odot)=8.72\pm0.04\) and \(8.08\pm0.04\), and with \(12+\log({\rm O/H})<8\) [2507.14936]. The paper reports direct-\(T_e\) integrated metallicities of \(7.90^{+0.30}_{-0.21}\) for ID6355 and \(7.68^{+0.25}_{-0.17}\) for ID10612, with strong-line metallicities of \(7.81\pm0.11\) and \(7.41\pm0.15\), respectively. Their nominal strong-line metallicity gradients, \(-0.008\pm0.010\ {\rm dex\,kpc^{-1}}\) and \(-0.085\pm0.011\ {\rm dex\,kpc^{-1}}\), are argued to be consistent with flat gradients [2507.14936].

The kinematic decomposition is based on broad and narrow components of the [O III] \(\lambda 5007\) line. The adopted observational outflow definition is
\[
v_{\rm out}=|v_{\rm broad}-v_{\rm narrow}|+2\sigma_{\rm broad,deconv},
\]
with \(\sigma_{\rm broad,deconv}\) corrected for instrumental broadening [2507.14936]. For ID6355, \(\sigma_{\rm broad,deconv}=210\pm20\ {\rm km\,s^{-1}}\) and \(\Delta v_{\rm outflow}=92\pm23\ {\rm km\,s^{-1}}\), giving \(v_{\rm out}=512^{+83}_{-84}\ {\rm km\,s^{-1}}\). For ID10612, \(\sigma_{\rm broad,deconv}=120\pm10\ {\rm km\,s^{-1}}\) and \(\Delta v_{\rm outflow}=30\pm10\ {\rm km\,s^{-1}}\), giving \(v_{\rm out}=255^{+18}_{-27}\ {\rm km\,s^{-1}}\) [2507.14936].

Escape speeds are estimated from dynamical masses inferred from ionized-gas dispersion and structural parameters. The reported values are \(v_{\rm esc}=480\pm190\ {\rm km\,s^{-1}}\) for ID6355 and \(210\pm50\ {\rm km\,s^{-1}}\) for ID10612, yielding
\[
\frac{v_{\rm out}}{v_{\rm esc}}=1.06\pm0.45
\]
and
\[
\frac{v_{\rm out}}{v_{\rm esc}}=1.23\pm0.31,
\]
respectively [2507.14936]. The mass outflow rate is estimated with
\[
\dot M_{\rm out}=3\,M_{\rm out}\,v_{\rm out}/r_{\rm out},
\]
and the normalized outflow strength is
\[
\eta \equiv \frac{\dot M_{\rm out}}{{\rm SFR}}.
\]
The inferred values are \(\dot M_{\rm out}=40^{+57}_{-18}\,M_\odot\,{\rm yr}^{-1}\) and \(21^{+24}_{-8}\,M_\odot\,{\rm yr}^{-1}\), corresponding to \(\eta=0.5^{+0.9}_{-0.3}\) and \(1.2^{+1.5}_{-0.5}\) [2507.14936].

These observational measurements are the point of contact with AESOPICA. The metallicity structure and AGN diagnostics motivate the interpretation of the galaxies, but they are not the axes on which the simulation comparison is performed.

## 4. Observation–simulation matching and analogous measurement strategy

The simulated comparison sample is selected narrowly. The authors state that they select simulated AGN host galaxies at \(z\sim 7.6\), matching the redshift of the observed sources, and that the simulated galaxies “roughly match the dynamical and stellar mass of our targets” [2507.14936]. No metallicity cut, SFR cut, halo-mass cut, or explicit BH-luminosity threshold is reported. The matching is therefore deliberately broad and does not constitute a strict nearest-neighbor analogue search.

Only two direct observation–simulation metrics are used in the main text: \(v_{\rm out}/v_{\rm esc}\) and \(\eta=\dot M_{\rm out}/{\rm SFR}\) [2507.14936]. A central technical issue is that native simulation outputs are not automatically comparable to broad-line-based outflow measurements. The paper therefore emphasizes that mass-weighted outflow velocities directly measured in simulations are biased low relative to outflow velocities inferred from mock observations of the same simulations. To mitigate this, AESOPICA does not use the simple mass-weighted mean velocity as the analogue of the observed broad-line outflow speed; it adopts the 95th percentile of the mass-weighted gas-velocity distribution as “a more reliable tracer of the broad-line based outflow velocity,” following Martin-Alvarez et al. (2025) as described in the paper [2507.14936].

The simulated outflow-rate estimate is also made to mimic the observational estimator. The same form,
\[
\dot M_{\rm out}=3\,M_{\rm out}\,v_{\rm out}/r_{\rm out},
\]
is used, integrating over all outflowing mass between the observational resolution limit, taken conservatively as \(\sim 40\ {\rm km\,s^{-1}}\), and the outflow velocity based on the 95th percentile [2507.14936]. This is not a full mock-spectral decomposition, but an observationally motivated approximation. In Appendix D, the simulations additionally show \(v_{\rm out}/v_{\rm vir}\) as a function of stellar mass, following Koudmani et al. (2022) in using virial velocity as a proxy for escape speed. That proxy is not identical to the observational \(v_{\rm out}/v_{\rm esc}\) definition based on \(M_{\rm dyn}\) and \(R_e\), and the distinction is explicitly noted in the paper [2507.14936].

The comparison protocol is therefore notable less for precision matching than for measurement harmonization. It attempts to reduce, rather than eliminate, the theory–data mismatch.

## 5. Physical interpretation of the AESOPICA comparison

The principal simulation result is restrained but clear. In the relevant low-mass regime, the high outflow velocities observed in the targets are “quite rare” in the Fiducial model, which lacks efficient AGN activity, whereas they are “much more common” in the BoostAcc and SuperEdd+BoostAcc runs [2507.14936]. This is the strongest simulation-based claim attached to AESOPICA in the current literature.

The outflow-ratio comparison shows that the observed values of \(v_{\rm out}/v_{\rm esc}\) are more commonly reproduced in BoostAcc and SuperEdd+BoostAcc than in Fiducial. The mass-loading factors \(\eta\) can also be reproduced by some simulated systems of similar mass, although the bulk of the simulated galaxies tend to have larger \(\eta\) [2507.14936]. In Appendix D, the broader simulated population indicates that, at fixed stellar mass, higher outflow velocities and larger mass-loading factors are associated with overmassive black holes, roughly \(M_{\rm BH}\sim 10^6\,M_\odot\) in the stellar-mass range relevant to the observed galaxies [2507.14936].

The resulting physical picture is that efficient BH accretion in low-mass galaxies provides a more natural explanation for the observed outflow properties than a more standard fiducial accretion prescription. Because the observed galaxies have \(v_{\rm out}/v_{\rm esc}\sim 1\), and because AESOPICA produces comparable values in AGN-host dwarfs, the simulations support the idea that AGN feedback in low-mass \(z\gtrsim 7\) galaxies can generate outflows capable of reaching or marginally exceeding the escape speed [2507.14936]. A plausible implication is that such outflows may eject gas from the halo, enrich the circumgalactic medium, and suppress future star formation over long timescales.

The paper is more circumspect on quenching than on outflow plausibility. It states that low observed \(\eta\) does not imply negligible long-term impact because AGN feedback can be cumulative over many episodes. AESOPICA is therefore used in support of the statement that AGN-driven outflows may contribute to quenching, not that the two observed systems are undergoing instantaneous quenching at the time of observation [2507.14936].

## 6. Scope, limitations, and common misconceptions

Several limitations are emphasized and define the correct scope of AESOPICA in this context. First, the comparison is explicitly “not apple-to-apple”: the paper states that a true apple-to-apple comparison would require full mock observations of the AESOPICA outputs, and that this is beyond the scope of the exploratory work [2507.14936]. AESOPICA is therefore used for qualitative or semi-quantitative interpretation, not for precision inference.

Second, the observational outflow estimates carry large systematics. The outflow mass depends on assumed density, metallicity, and geometry. The paper adopts \(n_{\rm out}=380\ {\rm cm^{-3}}\) and assumes the outflow metallicity equals the galaxy’s direct-\(T_e\) metallicity, while also noting that real outflows may be more metal-enriched than the narrow-line gas. If \(Z_{\rm out}\) were closer to solar, \(\dot M_{\rm out}\) would drop by a factor of \(\sim 7\)–10, and \(\eta\) is stated to have systematic uncertainties up to \(\sim 1\) dex [2507.14936]. Third, the simulation-side outflow definitions are themselves uncertain, especially because broad-line observables do not map uniquely onto native gas-kinematic quantities.

A common misconception would be to treat AESOPICA as a classifier of AGN or as a simulator that quantitatively explains the flat metallicity gradients. The paper does neither. AGN diagnostics are observational, including [O III] \(\lambda 4363\) auroral-line diagrams from Mazzolari et al. (2024) and previous evidence such as high-ionization \(\mathrm{C\,IV}\) emission in ID6355, while metallicity gradients are measured from annular strong-line and direct-\(T_e\) analyses [2507.14936]. AESOPICA enters only after AGN presence is already favored observationally, and only to ask whether the outflow properties are plausible for AGN-host low-mass galaxies at \(z\sim 7.6\).

A second misconception would be to read the comparison as proof that AGN driving is required. The paper does not make that claim. Its strongest conclusion is that AGN feedback is favored or consistent, especially for ID6355, because that galaxy is the stronger observational AGN case, its \(v_{\rm out}\approx 512\ {\rm km\,s^{-1}}\) is rare in the Fiducial model, and such velocities are much more common in BoostAcc and SuperEdd+BoostAcc [2507.14936]. For ID10612, where both the AGN identification and the outflow interpretation are less secure, AESOPICA shows plausibility rather than preference.

In this sense, AESOPICA presently occupies a specific epistemic niche. It is a FABLE-based cosmological simulation suite designed to follow infant SMBH growth in early low-mass galaxies, and its current published use is to show that the outflow properties of two \(z\sim 7.66\) galaxies fall within the broad distribution of simulated AGN-host dwarfs when efficient accretion is allowed [2507.14936]. The suite therefore provides theoretical context for early dwarf-galaxy feedback, while leaving metallicity structure, AGN identification, and precision theory–data matching to future work.

Source: https://www.emergentmind.com/topics/aesopica-simulations