🔭 astrophysics

Exploring the impact of AGN feedback model variations on the Lyman- $\alpha$ Forest Flux Power Spectrum

Using the CAMELS suite to vary AGN feedback parameters in the Simba simulation, this study demonstrates that while increasing AGN feedback strength generally suppresses the Lyman- $\alpha$ forest flux power spectrum, the specific impact depends critically on how parameters like radiative efficiency and jet thresholds influence the population of massive black holes and the interplay between jet heating and feedback suppression.

Original authors: Megan Pirecki, Megan Taylor Tillman, Blakesley Burkhart, Stephanie Tonnesen, Simeon Bird

Published 2026-05-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Megan Pirecki, Megan Taylor Tillman, Blakesley Burkhart, Stephanie Tonnesen, Simeon Bird

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe as a giant, invisible ocean of gas stretching between galaxies. This "ocean" is mostly empty, but it's not perfectly smooth; it has ripples, clumps, and waves. Astronomers study this gas by looking at the light from distant quasars (super-bright galactic cores) as it passes through the gas. As the light travels, the gas absorbs specific colors, creating a "forest" of dark lines in the spectrum. This is called the Lyman-alpha forest.

This paper is like a massive cooking experiment. The researchers wanted to see how changing the "recipe" for how supermassive black holes (SMBHs) at the centers of galaxies interact with this gas ocean affects the final taste of the Lyman-alpha forest.

Here is the breakdown of their experiment in simple terms:

The Setup: The Cosmic Kitchen

The team used a supercomputer simulation called CAMELS (Cosmology and Astrophysics with MachinE Learning Simulations). Think of this as a giant, digital sandbox where they can build a universe. They used a specific version of this sandbox called Simba, which has a built-in "chef" that decides how black holes behave.

In this digital universe, black holes don't just sit there; they eat gas and shoot out powerful jets of energy (like a cosmic firehose) or radiate heat. These actions are called AGN feedback. The researchers wanted to know: If we tweak the settings on this cosmic firehose, how does the pattern of the gas forest change?

They tested five specific "knobs" or settings on the black hole model:

Momentum Flux: How hard the black hole pushes the gas.
Jet Speed: How fast the firehose shoots out.
Radiative Efficiency: How much energy the black hole releases as light/heat while eating.
Jet Velocity Threshold: How fast a jet must be moving before it gets super-heated.
Minimum Black Hole Mass: How big a black hole must be before it's allowed to turn on its firehose.

The Results: What Happened When They Turned the Knobs?

1. The Speed of the Firehose Matters Most
The most dramatic changes happened when they adjusted the speed of the jets.

The Analogy: Imagine a garden hose. If you turn the water pressure up (faster jets), the water sprays much further, soaking a larger area of the garden.
The Result: When the jets were faster, they heated up more of the gas ocean and pushed the gas further away. This "smoothed out" the forest, making the dark lines in the light spectrum less distinct (lowering the power). When they slowed the jets down, the gas stayed clumpier, and the forest looked "rougher" (higher power).

2. The Size of the Black Hole is a Gatekeeper
They found that only the biggest black holes really matter for this specific effect.

The Analogy: Think of the firehose as a heavy-duty industrial machine. Small black holes are like handheld water pistols; they just can't reach far enough to change the whole garden. Only the massive black holes (the industrial machines) have the power to reach out and heat the distant gas.
The Result: If they raised the "minimum size" required to turn on the firehose, the effect on the forest was huge because it stopped the medium-sized black holes from contributing. If they lowered the size limit, it didn't change much because the tiny black holes weren't powerful enough to do the job anyway.

3. The "Goldilocks" Zone of Heating
They discovered a tricky balance with heating the jets.

The Analogy: Imagine you are trying to clear snow from a driveway. If you use a little heat, it melts the snow. But if you use too much heat, you might accidentally melt the driveway itself or trigger a mechanism that turns off your heater.
The Result: Heating the jets helps remove gas (neutral hydrogen) from the forest. However, if you heat them too much, it actually stops the black holes from growing as fast. If the black holes don't grow, they can't shoot as many jets later on. So, too much heating ironically reduces the overall impact on the forest.

4. The "Push" vs. The "Heat"
They found that simply pushing the gas harder (momentum) had a limit.

The Analogy: Imagine pushing a swing. Pushing it a little harder makes it go higher. But if you are already pushing it as hard as the standard setting, pushing even harder doesn't make it go much higher because the swing is already at its limit.
The Result: Reducing the push made the forest clumpier (more power). But increasing the push beyond the standard setting didn't change the forest much more. The standard setting was already doing the maximum job.

The Big Takeaway

The paper concludes that to understand the universe's gas ocean, we can't just treat black holes as generic heaters. The speed of their jets and the size of the black holes are the most critical factors.

Furthermore, the way these black holes heat the gas is unique. It's not just a general warming of the universe (which could be explained by background radiation); it's a specific, localized "scorching" that happens right near the black holes and travels outward. This creates a unique fingerprint in the Lyman-alpha forest that standard background models can't replicate.

In short: The universe's gas structure is heavily influenced by the "firehoses" of the biggest, fastest black holes. If we want to understand the universe's history, we need to get the recipe for these firehoses exactly right.

Problem Statement
Active Galactic Nuclei (AGN) feedback is a critical component in cosmological simulations, yet its specific impact on the intergalactic medium (IGM) at low redshifts remains difficult to disentangle from other heating sources, such as the metagalactic ultraviolet background (UVB). While previous studies have utilized the Lyman- $\alpha$ (Ly $\alpha$ ) forest to constrain cosmological parameters and dark matter models, the unique signature of AGN feedback on the 1D transmitted flux power spectrum ( $P_{1D}$ ) requires further exploration. Specifically, it is unclear how variations in specific subgrid parameters of AGN feedback models—particularly those governing jet feedback versus radiative feedback—alter the thermal state and ionization fraction of the IGM. Current simulations often fail to reproduce observed $b$ -value distributions at $z < 1$ , and the degeneracy between AGN feedback effects and UVB scaling complicates the interpretation of observational data.

Methodology
This study utilizes the Cosmology and Astrophysics with Machine Learning Simulations (CAMELS) suite, specifically the CAMELS-Simba simulations, to isolate the effects of five distinct AGN feedback parameters. The simulations employ the AREPO code with a baseline cosmology of $\Omega_b = 0.049$ , $h = 0.6711$ , and $n_s = 0.9624$ . The authors analyze variations in the following parameters:

$A_{\text{AGN1}}$ : Momentum flux normalization for both radiative and jet modes.
$A_{\text{AGN2}}$ : Maximum speed normalization for AGN jets.
$\text{BHRadiativeEff}$ : Radiative efficiency of the black hole accretion disk.
$\text{BHJetTvirVel}$ : Velocity threshold above which ejected jets are heated to the host halo's virial temperature.
$\text{BHJetMassThr}$ : Minimum supermassive black hole (SMBH) mass required to activate jet feedback.

Synthetic spectra were generated using the Fake Spectra Flux Extractor (FSFE) code, placing 5,000 random sightlines per simulation box to minimize sample variance. The 1D flux power spectrum ( $P_{1D}$ ) was calculated for redshifts $z = 0.1, 0.4, 0.7, 1.0, 1.25,$ and $2.0$. To address potential degeneracies with the UVB, the authors applied a post-processing correction to match the mean transmitted flux of the fiducial simulation, effectively isolating the impact of feedback on the power spectrum shape rather than just normalization. The authors acknowledge that the simulation box size (25 Mpc/ $h$ ) limits convergence on small scales and precludes direct comparison with observational data, focusing instead on relative trends between parameter variations.

Key Results
The analysis reveals that while all parameters affect the $P_{1D}$ , the magnitude and nature of the impact vary significantly:

Jet Speed ( $A_{\text{AGN2}}$ ): This parameter exhibits the most dramatic impact. Increasing jet speed suppresses the $P_{1D}$ (by up to 50% at $z=0.1$ ) by heating the IGM and removing neutral hydrogen. Conversely, decreasing jet speed enhances the $P_{1D}$ significantly. The effect is most pronounced at high wavenumbers ( $k$ ), corresponding to small physical scales.
Jet Mass Threshold ( $\text{BHJetMassThr}$ ): Increasing the minimum SMBH mass required for jet feedback leads to a substantial increase in the $P_{1D}$ (up to 100% at $z=0.1$ ), indicating that the most massive SMBHs ( $M_{\text{BH}} > 10^7 M_\odot$ ) are the primary drivers of jet feedback effects on the Ly $\alpha$ forest. Decreasing the threshold has a negligible effect, as lower-mass SMBHs produce less energetic jets.
Momentum Flux ( $A_{\text{AGN1}}$ ): Decreasing the momentum flux significantly enhances the $P_{1D}$ at low redshifts. However, increasing the flux beyond the fiducial value yields diminishing returns, suggesting the fiducial model already maximizes the feedback effect on the forest.
Radiative Efficiency ( $\text{BHRadiativeEff}$ ) and Jet Velocity Threshold ( $\text{BHJetTvirVel}$ ): Variations in these parameters produce smaller effects, generally crossing the 10% threshold only at $z \leq 0.4$ . Interestingly, both increasing and decreasing these values relative to the fiducial model result in an enhanced $P_{1D}$ . The authors attribute the increase from higher radiative efficiency to a self-suppressive effect on SMBH growth, which reduces the overall number of active AGNs and thus the total feedback energy injected.
Jet Heating: Heating jets to the virial temperature at injection aids in removing neutral hydrogen but can inhibit further jet feedback by suppressing SMBH growth, leading to a complex, non-monotonic relationship between heating efficiency and the final $P_{1D}$ .

Significance and Conclusions
The paper concludes that AGN feedback, particularly via the jet mode, significantly suppresses the Ly $\alpha$ forest $P_{1D}$ at low redshifts, but only if the feedback does not inadvertently suppress the population of massive, jet-producing SMBHs. The study demonstrates that the impact of AGN jets cannot be fully captured by adjusting the assumed UVB, as jets uniquely affect the high- $k$ end of the power spectrum through thermal broadening and localized heating, whereas the UVB primarily acts as a normalization factor.

The authors emphasize that future studies of AGN feedback must carefully explore the unique aspects of specific subgrid models and their interactions. They suggest that to fully understand the astrophysical impacts of SMBH feedback, models must be calibrated not only to galaxy stellar mass functions but also to larger-scale statistics like the Ly $\alpha$ forest. The results imply that current simulations may require higher resolution and larger box sizes to fully resolve the convergence of these effects, and that future observational constraints (e.g., from the Habitable Worlds Observatory) will be necessary to distinguish between different feedback parameter variations.

The Setup: The Cosmic Kitchen

The Results: What Happened When They Turned the Knobs?

The Big Takeaway

More like this