🔭 astrophysics

A stochastic forward model for the intergalactic dispersion-measure distribution of Fast Radio Bursts

This paper introduces \turbofrb, a semi-analytic stochastic forward model that accurately reconstructs the intergalactic dispersion-measure distribution of Fast Radio Bursts by explicitly resolving contributions from diffuse IGM, halos, and filaments, thereby enabling precise host-excess quantification and cosmological parameter recovery.

Original authors: Jéferson A. S. Fortunato, Valerio Marra, Wiliam S. Hipólito-Ricaldi, Amanda Weltman

Published 2026-08-19

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Jéferson A. S. Fortunato, Valerio Marra, Wiliam S. Hipólito-Ricaldi, Amanda Weltman

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

Fast Radio Bursts are brief, intense flashes of radio waves that arrive from deep space, traveling across billions of light-years before reaching our telescopes. As these signals journey through the universe, they pass through a vast, invisible ocean of ionized gas that fills the space between galaxies. This gas slows down the radio waves slightly, and the amount of slowing depends on how much gas the signal has passed through. Astronomers measure this effect as a "dispersion measure," a number that tells them the total density of free electrons along the path. In a perfectly smooth universe, this number would rise steadily with distance, offering a clean way to map the expansion of the cosmos. However, the universe is not smooth; it is a complex web of empty voids, dense clusters of galaxies, and long, thin bridges of gas connecting them. Because of this clumpiness, two signals traveling the same distance can encounter very different amounts of gas, making the dispersion measure a noisy and unpredictable guide. Understanding exactly how this gas is distributed is crucial for using these cosmic flashes to measure the universe's expansion rate and the total amount of normal matter it contains.

A team of researchers has developed a new tool called turboFRB to solve this problem of cosmic noise. Instead of trying to force the messy reality of the universe into a simple mathematical formula, they built a computer model that simulates the journey of a radio signal step by step. They treat the path of the signal as a journey through three distinct types of environments: the vast, thin, and relatively uniform gas that fills most of space; the dense clouds of gas surrounding individual galaxies; and the long, filamentary bridges of gas that connect galaxy clusters. The model generates thousands of possible journeys, randomly deciding how many galaxies and filaments a signal might cross, and how much gas it would pick up in each encounter. By running these simulations, the researchers created a detailed map of what the dispersion measure should look like for a signal coming from any given distance.

The team tested their model against a massive, high-resolution computer simulation of the universe that tracks the movement of gas and dark matter in extreme detail. They found that their new tool could reproduce the results of this complex simulation with remarkable accuracy. The model matched the average amount of gas encountered by signals at different distances to within a few percent. More importantly, it correctly captured the shape of the distribution, including the rare but important events where a signal happens to pass through a particularly dense region, resulting in a much higher-than-average dispersion measure. Previous attempts to describe this distribution often relied on simplified curves that failed to account for these rare, high-density encounters or broke down at certain distances. The new model, by explicitly separating the contributions of the smooth gas, the galaxy halos, and the filaments, showed that the smooth gas forms the main body of the distribution, while the dense halos and filaments are responsible for the long, heavy tail of extreme values.

One of the key discoveries from this work is how the structure of the universe changes the signal as it travels. The model revealed that as a signal travels further, it is more likely to cross through the dense halos of galaxies and the connecting filaments. At lower distances, the signal mostly travels through the smooth gas, but at greater distances, the number of galaxy encounters increases significantly. The researchers found that the average number of galaxy halos a signal crosses grows from less than one at relatively close distances to nearly three at the farthest distances they studied. This increase in encounters is what causes the distribution of dispersion measures to broaden and develop that heavy tail of high values. The model also introduced a way to account for the fact that some lines of sight pass through denser regions of the cosmic web than others, effectively linking the number of galaxy and filament encounters to a single, hidden variable representing the local density of the universe along that specific path.

The researchers then applied their model to real Fast Radio Bursts that have been located in the sky and whose distances are known. For some of these bursts, the observed amount of gas was much higher than what the model predicted for their distance, suggesting that the signal passed through an unusually dense environment or that the host galaxy itself contributed a large amount of gas. For others, the observation fit perfectly within the range of expected values. The model provided a way to quantify exactly how unusual a particular burst was, calculating the probability that its high dispersion measure was just a random fluke of the cosmic web versus a sign of a dense local environment. In one case, a burst at a very high distance was found to be consistent with the average expectation, showing that even at extreme distances, the model can distinguish between typical and anomalous signals.

Beyond analyzing individual bursts, the model offers a powerful new way to measure the expansion rate of the universe. Because the amount of gas a signal encounters depends on how the universe has expanded over time, the distribution of dispersion measures contains information about the Hubble constant, a value that describes the current rate of cosmic expansion. The researchers tested this by creating a fake set of radio bursts using their model and then trying to recover the expansion rate they had used to create them. The model successfully recovered the correct value, demonstrating that the approach is internally consistent and ready to be applied to real data. This work does not claim to have solved the mystery of the universe's expansion, but it provides a much more reliable and physically transparent tool for doing so. By breaking down the journey of a radio signal into its physical components, the researchers have given astronomers a clearer lens through which to view the invisible matter that fills the cosmos.

Technical Summary: A Stochastic Forward Model for the Intergalactic Dispersion-Measure Distribution of Fast Radio Bursts

Problem Statement
Fast Radio Bursts (FRBs) serve as probes of ionised baryons via their observed dispersion measures (DM). While the mean relation between DM and redshift (the Macquart relation) has been used to infer cosmological parameters, it is insufficient for individual FRB sightlines. At a fixed redshift, different lines of sight sample distinct environments—diffuse intergalactic medium (IGM), halo outskirts, and cosmic filaments—resulting in a broad, skewed, and non-Gaussian conditional distribution, $P(\text{DM}_{\text{IGM}} | z)$ . Existing phenomenological descriptions often rely on closed-form parametric families (e.g., skewed distributions with redshift-dependent widths) that fail to capture the physical mechanisms generating the distribution's shape, particularly the high-DM tail. Furthermore, recent ray-tracing studies of hydrodynamical simulations (e.g., Konietzka et al. 2025) have revealed that previous analytic fits suffer from artifacts and biases, necessitating a modeling framework that resolves the physical channels (diffuse IGM, halos, filaments) explicitly rather than assuming a PDF shape a priori.

Methodology: The turboFRB Stochastic Engine
The authors present turboFRB, a semi-analytic stochastic forward model that constructs the line-of-sight DM contribution as the sum of three physically motivated channels:

Diffuse Channel ( $\text{DM}_{\text{smooth}}$ ): Represents the smooth, low-density IGM. It is sampled from a mean-preserving lognormal distribution to account for multiplicative fluctuations in electron density along the path, ensuring the ensemble mean remains anchored to the analytic Macquart relation.
Halo Channel ( $\text{DM}_{\text{halo}}$ ): Models encounters with dark matter halos as a Poisson process. Halos are drawn from a mass function ( $10^{12} - 10^{14.8} M_\odot$ ) and integrated through Navarro–Frenk–White (NFW) gas profiles. The encounter rate depends on redshift and is modulated by a latent environmental variable.
Filament Channel ( $\text{DM}_{\text{fil}}$ ): Models encounters with cosmic filaments as Poisson-sampled finite cylinders with isothermal $\beta$ -model gas profiles.

Coupling and Calibration:
The halo and filament encounter rates are coupled by a latent line-of-sight environmental variable, $E$ , drawn from a Gaussian distribution. This variable modulates the Poisson rates of both structural channels via a mean-preserving exponential factor, ensuring that overdense lines of sight intersect more structures while preserving the ensemble mean.

The model is calibrated against the corrected, continuous ray-traced IllustrisTNG catalogue (Konietzka et al. 2025). The calibration strategy isolates the distribution shape by working in the normalized variable space $\Delta \equiv \text{DM}_{\text{IGM}} / \langle \text{DM}_{\text{IGM}}(z) \rangle$ . Only four effective parameters are tuned via a grid search:

$\sigma_0$ : Amplitude of the diffuse log-scatter.
$\gamma_s$ : Redshift exponent of the diffuse scatter.
$\sigma_E$ : Amplitude of the environmental modulation.
$n_{\text{hbg}}$ : Effective comoving background halo number density.

All other inputs (mass pools, gas fractions, profile shapes, geometric factors) are fixed based on literature values, engine defaults, or budget closure constraints.

Key Contributions and Results

Physical Decomposition: Unlike closed-form parametric descriptions, turboFRB explicitly resolves the contributions of the diffuse IGM, halos, and filaments. The model demonstrates that the diffuse channel sets the body of the distribution, while halos and filaments populate the shoulder and high-DM tail.
Benchmark Agreement: The calibrated model matches the corrected TNG benchmark to the percent level in mean DM (ratios of 0.982–0.990 across $z=0.5–2.5$ ). The Jensen–Shannon divergence in $\Delta$ -space is at most $5 \times 10^{-3}$ , indicating a high-fidelity reproduction of the distribution shape.
Scatter Behavior: The directly measured normalized scatter, $\sigma_\Delta(z)$ , crosses the TNG benchmark near $z \sim 1.5–1.8$ and tracks it within $\lesssim 15\%$ for $z \gtrsim 1$ . This contrasts with some analytic parameterizations that lie systematically above the benchmark. The model reveals that the effective halo encounter rate increases significantly with redshift (from $\langle N_h \rangle \approx 0.47$ at $z=0.2$ to $\approx 2.9$ at $z=2.5$ ), while the scatter per encounter increases moderately.
Event-Level Diagnostics: Applied to representative localised FRBs, the forward likelihood quantifies the "tail probability" ( $P_{\text{tail}}$ ) of an event's observed DM. This allows for the independent identification of "host-excess" events (where the observed DM is unlikely given the cosmological distribution and priors) without relying on specific astrophysical signatures. For example, FRB 20190520B is identified as requiring a host excess, while FRB 20240304B ( $z=2.148$ ) is consistent with a typical sightline.
Cosmological Consistency: A closed-loop test on a synthetic catalogue of 100 FRBs demonstrates that the forward likelihood can recover the injected Hubble constant ( $H_0$ ) within $1\sigma$ , verifying the internal consistency of the inference machinery.

Significance and Claims
The paper claims that turboFRB provides a natural tool for Bayesian inference of cosmological parameters and for diagnosing individual FRB sightlines within a physically transparent framework. By resolving the physical mechanisms (diffuse gas, halo encounters, filament crossings) rather than assuming a PDF shape, the model avoids the limitations of closed-form approximations, particularly at low redshifts where scaling laws diverge.

The authors emphasize that the model is constrained by the physics of the cosmic web; with only four effective nuisance parameters calibrated against a hydrodynamical benchmark, the model cannot arbitrarily fit any distribution but is instead tied to the encounter physics of halos and filaments. The work establishes the forward machinery necessary for future cross-correlation analyses (e.g., FRB DMs with galaxy positions or weak-lensing maps) and for joint cosmological parameter inference, though the authors note that a full application to real data with marginalisation over host-galaxy models and baryon fractions is reserved for future work. The model is implemented as an open-source package (turboFRB).

More like this