Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG

arXiv:2601.18578 · astro-ph.GA, astro-ph.CO, astro-ph.HE · Submitted 2026-01-26 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: I'm Vera, and with me are Jocelyn and Subrahmanyan, guest researcher.

Jocelyn: Today's paper: "Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG".

Vera: The AIDA-TNG project utilizes "the AIDA-TNG cosmological simulation suite to predict the distributions of gas and neutral hydrogen (HI) in the CDM, Self-Interacting DM (SIDM), velocity-dependent SIDM (vSIDM),

Jocelyn: First, who's behind it and why it matters.

Title and authors: Vera: So, the core of this paper, "Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG," is that they found that for haloes with virial masses between ten nine and ten twelve solar masses, the gas structures don't show significant differences depending on which dark matter model you choose.

Jocelyn: I noticed they also point out a specific nuance: for more massive haloes, specifically those with a virial mass around ten thirteen solar masses, a flat central region starts to develop as the halo evolves within zero point zero two to zero point zero four of the virial radius, which they attribute to continuous baryonic feedback reducing the central gas density (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Subrahmanyan: That's interesting because it suggests that while dark matter models might be similar for smaller systems, more complex baryonic processes become important when you reach those larger halo masses (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Vera: And they also showed that in terms of neutral hydrogen content, the HI profiles show higher sensitivity to the dark matter models compared to the total gas profiles, which is a key finding for our observational targets (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Jocelyn: This means we should be focusing on tracing the neutral hydrogen content when we study galaxy haloes because that's where the variation between CDM and SIDM1, for instance, becomes more pronounced (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Subrahmanyan: The paper also notes that for haloes with masses of ten eleven and ten twelve solar masses, clear differences appear between SIDM1 and vSIDM, showing variations up to a factor of two compared to CDM and WDM3 (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Vera: So the summary is that the neutral hydrogen content provides clearer distinctions across dark matter models in more massive systems like those near ten thirteen solar masses (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Jocelyn: It really solidifies our observational focus on those higher mass systems where the differences between CDM and other models start to become statistically significant.

Subrahmanyan: This finding suggests that the effect of dark matter microphysics isn't just a small perturbation; it has clear signatures in the observable gas distribution, which is a strong indicator for future tests of particle physics (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

The paper's summary: Vera: Now looking ahead at what the authors suggest for future work, they are proposing several ways to make this research even more powerful using advanced machine learning techniques to move beyond just running the simulations.

Jocelyn: I noticed they suggest integrating derived parametric fitting functions for gas over-density and HI over-density with a Physics-Informed Neural Network architecture, which would allow the AI to act as a high-speed emulator that predicts these distributions across various models in milliseconds (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Subrahmanyan: That AI approach allows researchers to run many different scenarios for dark matter models quickly, which means we can explore the parameter space much faster than doing traditional, computationally expensive simulations (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Vera: Plus, they mentioned using a specialized attention mechanism in deep learning models specifically tuned to recognize those faint spectral features characteristic of WDM and peculiar velocities in Lyman-alpha spectra (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Jocelyn: That’s smart because it means the machine learns to look for those subtle imprints that might be buried in noisy spectroscopic data, which is a big hurdle for us when trying to constrain WDM models.

Subrahmanyan: And they also propose using generative models, like a GAN or Variational Autoencoder, to help separate the noise from the true physical signals related to baryonic feedback intensity when analyzing observational data (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Vera: That de-biasing of data is exactly what we need when dealing with complex astrophysical systems; if the AI can clean up the background noise caused by AGN activity, it could give us much cleaner statistics on dark matter effects (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Jocelyn: I’m really looking forward to seeing how those proposed observational strategy intelligence tools actually translate into practical plans for telescope scheduling and survey design, which is a huge practical step (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

The paper's improvements: Vera: So to wrap up on "Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG," we've seen how the neutral hydrogen content offers clearer distinctions in massive systems when considering SIDM1 compared to other models.

Jocelyn: It really highlights that for us observing the sky, we need to focus our observational efforts on tracing those HI profiles and looking at those higher mass haloes if we want to find these dark matter distinctions (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Subrahmanyan: I think this paper reinforces that dark matter microphysics has a measurable effect on baryonic structure in the most massive haloes when we look at neutral gas; it's a strong indicator for future tests of particle physics (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Vera: We’re definitely going to be keeping an eye on these simulation results as we push our observational capabilities further into the Lyman-alpha forest regime for this paper, focusing on those specific halo mass ranges (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Jocelyn: I’m excited to see how those proposed improvements, like the optimized observational strategy intelligence, might actually translate into better data collection plans for telescopes (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Subrahmanyan: The integration of these simulation insights with advanced AI techniques and targeted observational strategies is exactly what moves the field forward in constraining alternative dark matter models (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Vera: That’s our summary for today on "Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG."

Jocelyn: I'm really looking forward to seeing how those proposed improvements might actually translate into better data collection plans for telescopes.

Subrahmanyan: This work is a vital piece in linking fundamental particle physics to the large-scale structure of the universe through detailed hydrodynamical modeling (Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG).

Conclusion: Vera: So we've walked through "Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG," and it’s clear that tracing neutral hydrogen profiles is a really powerful way to test these different dark matter theories.

Jocelyn: I totally agree, Vera; focusing on those HI profiles gives us concrete data points to compare against our observational targets in the Lyman-alpha forest.

Subrahmanyan: And from a theoretical standpoint, it confirms that dark matter microphysics definitely leaves a signature on how gas settles within massive haloes.

Vera: Exactly, and I’m really keen to see how these simulation results guide what we should actually be looking for in upcoming surveys.

Jocelyn: That's exciting because it gives us a specific target; instead of looking everywhere, we know exactly which mass scales and which tracers are most informative.

Subrahmanyan: It’s a significant contribution because this work sets clear benchmarks for what future observational experiments need to achieve to actually test these different dark matter scenarios.

Jocelyn: I’m excited to see how those proposed improvements, like the optimized observational strategy intelligence, might actually translate into better data collection plans for telescopes.

Vera: That’s our summary for today on "Gas distributions inside and around haloes in the alternative dark matter simulations AIDA-TNG." Thanks to everyone for joining us.

Jocelyn: I'm really looking forward to seeing how those proposed improvements might actually translate into better data collection plans for telescopes.

Subrahmanyan: This work is a vital piece in linking fundamental particle physics to the large-scale structure of the universe through detailed hydrodynamical modeling.

Chi Zhang, Enrico Garaldi, Giulia Despali, Matteo Viel, Lauro Moscardini, Mark Vogelsberger

Key Laboratory of Dark Matter and Space Astronomy, Purple Mountain Observatory, Chinese Academy of Sciences · School of Astronomy and Space Science, University of Science and Technology of China · SISSA - International School for Advanced Studies · Kavli IPMU (WPI), UTIAS, The University of Tokyo · Center for Data-Driven Discovery, Kavli IPMU (WPI), UTIAS, The University of Tokyo · IFPU - Institute for Fundamental Physics of the Universe · INAF - Osservatorio Astronomico di Trieste · Dipartimento di Fisica e Astronomia "Augusto Righi", Alma Mater Studiorum Università di Bologna · INAF-Osservatorio di Astrofisica e Scienza dello Spazio di Bologna · INFN-Sezione di Bologna · INFN - Sezione di Trieste · ICSC - Centro Nazionale di Ricerca in High Performance Computing, Big Data e Quantum Computing · Department of Physics, Kavli Institute for Astrophysics and Space Research, Massachusetts Institute of Technology

astro-ph.GA, astro-ph.CO, astro-ph.HE

Submitted: 2026-01-26

Updated: 2026-09-11

Comments: 21 pages, 11 figures. Published in Physical Review D. Find more information on AIDA-TNG here:https://gdespali.github.io/AIDA/

Journal ref: Phys. Rev. D 114, 063020 (2026)

DOI: 10.1103/gssw-j3sv

Project page: https://gdespali.github.io/AIDA

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 57/100

The gist: The AIDA-TNG project utilizes "the AIDA-TNG cosmological simulation suite to predict the distributions of gas and neutral hydrogen (HI) in the CDM, Self-Interacting DM (SIDM), velocity-dependent SIDM

Key concepts

AIDA-TNG: The AIDA-TNG project uses the AIDA-TNG cosmological simulation suite to predict the distributions of gas and neutral hydrogen (HI) across different dark matter models, including CDM, SIDM, and vSIDM.
Neutral Hydrogen Content (HI): The paper found that HI profiles show higher sensitivity to dark matter models compared to total gas profiles. This makes tracing neutral hydrogen content a key focus when studying galaxy haloes for observational targets.
Baryonic Feedback: For more massive haloes, continuous baryonic feedback is responsible for developing a flat central region in the halo as it evolves within zero point zero two to zero point zero four of the virial radius, reducing central gas density.
Physics-Informed Neural Network (PINN): The authors propose using PINNs to create high-speed emulators that predict gas and HI over-densities across various dark matter models in milliseconds. This allows researchers to explore parameter space much faster than traditional simulations.

Terminology

Dark Matter — 58× in this paper
AIDA-TNG — 48× in this paper · explained in 1 episode
AGN — 37× in this paper · explained in 1 episode
Core — 24× in this paper · explained in 2 episodes
AGN Feedback — 22× in this paper · explained in 1 episode
alpha — 7× in this paper · explained in 1 episode
Baryonic Feedback — 3× in this paper · explained in 3 episodes
Gravitational Lensing — 3× in this paper · explained in 2 episodes
Benchmark — 2× in this paper · explained in 2 episodes
Metallicity — 2× in this paper · explained in 3 episodes
Abundance — 1× in this paper · explained in 1 episode
Attention Mechanism — 1× in this paper · explained in 2 episodes
Clustering — 1× in this paper · explained in 2 episodes
Degeneracy — 1× in this paper · explained in 3 episodes
Generative Model — 1× in this paper · explained in 2 episodes
Generative Models — 1× in this paper · explained in 2 episodes
Warm Dark Matter (WDM) — 1× in this paper · explained in 3 episodes

Summary

The AIDA-TNG project utilizes the AIDA-TNG cosmological simulation suite to predict the distributions of gas and neutral hydrogen (HI) in the CDM, Self-Interacting DM (SIDM), velocity-dependent SIDM (vSIDM), and Warm DM (WDM) models. The study finds that the DM models investigated have very limited impact on the median gas and HI profile of haloes, though specific nuances exist depending on mass scales and model types.

Regarding gas distributions, for haloes with M vir in [10 9, 10 12] M... the four DM models do not exhibit significant differences. However, in more massive haloes (M vir 10 13 M), a flat central region develops within 0.02 to 0.04 as the halo evolves, a feature attributed to the action of a continuous baryonic feedback that reduces the central gas density. In terms of dark matter specifics, SIDM1, instead, shows a prominent core radius already at z = 3, accompanied by a much slower growth, so that by z = 0 its gas core is only slightly larger than for the other models.

The neutral hydrogen (HI) content shows higher sensitivity to dark matter models than total gas. The HI profiles show more variety across dark matter models compared to the gas profiles. For haloes with masses of 10 11 M and 10 12 M, We find clear differences for SIDM1 and vSIDM, which show differences up to a factor of 2... compared to CDM and WDM3. In the most massive haloes (M vir about 10 14 M), SIDM1... preserves more HI than other DM models in the central part of the halo, which is consistent with SIDM1 inducing a dark matter cored profile that provides a physical pathway connecting DM physics to BH accretion and feedback.

The role of AGN feedback is central to the observed variations, as the halo-to-halo variation in the HI profiles is explained by AGN feedback, and that the specific characteristics of DM model is largely subdominant. In massive haloes (M vir 10 13 M), AGN injects large amounts of energy into the surrounding gas, resulting in a strong suppression of central HI density across all DM models.

In terms of observational prospects via the Lyman- alpha forest, the study investigates the galaxy-Lyman- alpha cross-correlation function (GaL alpha CC) for different halo masses, redshift and observation strategies. While standard random sightlines are dominated by cosmic variance, an optimised sightline placing allows for better discrimination. The researchers conclude that at z = 0 vSIDM can be distinguished from CDM in haloes with 10 12 M vir 10 13 M, while SIDM1 can be distinguished from CDM in haloes with M vir 10 13 M. To achieve a statistically-robust detection, the study estimates that it is necessary to sample about 160 haloes with about 20 sightlines each, a task that can be achieved with current and future facilities like WEAVE, 4MOST, PFS, ELT and WST.

Improvements for AI systems

1. Physics-Informed Surrogate Emulators for Cosmological Hydrodynamics

Improvement: Integrate the derived parametric fitting functions (Eq. 1 for gas over-density and Eq. 2 for HI over-density) and the quantified correlation between cumulative AGN feedback energy (E tot,AGN) and HI profile suppression into a Physics-Informed Neural Network (PINN) architecture.
Capability: The AI can act as a high-speed emulator that predicts complex, non-linear gas and neutral hydrogen (HI) distributions across various dark matter models (CDM, SIDM, vSIDM, WDM), redshifts, and halo masses in milliseconds. This bypasses the need for computationally prohibitive hydrodynamical simulations while maintaining physical accuracy regarding baryonic feedback effects.

2. Targeted Feature Extraction for Spectroscopic Survey Pipelines

Improvement: Implement a specialized attention mechanism in deep learning models trained on Lyman- alpha and Lyman- beta synthetic spectra, specifically weighted to recognize the tail-like structures characteristic of WDM and the specific absorption shifts caused by peculiar velocities.
Capability: The improved system can automate the identification of subtle dark matter imprints within massive spectroscopic datasets (e.g., from WEAVE, 4MOST, or DESI). It can distinguish between DM models by focusing on high-sensitivity spectral features that are typically lost in generic signal-to-noise processing.

3. Degeneracy-Aware Generative Models for Parameter Estimation

Improvement: Develop a generative model (such as a Variational Autoencoder or GAN) trained on the AIDA-TNG suite to explicitly model the joint probability distribution of baryonic feedback intensity (E tot,AGN) and dark matter microphysics.
Capability: The AI can perform de-biasing of observational data by disentangling the stochastic variations caused by AGN feedback (which dominates halo-to-halo HI variation) from the systematic signatures of alternative dark matter models. This allows for more precise cosmological parameter estimation even when baryonic physics acts as a significant noise source.

4. Optimized Observational Strategy Intelligence (OSI)

Improvement: Incorporate the halo-centered sightline sampling logic into an AI-driven optimization engine used for telescope scheduling and survey design.
Capability: The system can calculate the optimal spatial distribution and impact parameters for background sightlines around detected galaxies to maximize the Galaxy-Lyman- alpha cross-correlation (GaL alpha CC) signal. This enables future facilities like the Wide Field Spectroscopic Telescope (WST) to achieve statistically robust dark matter detection with minimal required sampling (e.g., targeting about 160 haloes with about 20 sightlines each).