Updated Air-Shower X max Moment Parametrizations for UHECR Composition with Latest Hadronic Interaction Models
summary
The gist
The mass composition of ultra-high-energy cosmic rays (UHECRs) is commonly inferred from the first two moments of the depth of shower maximum, Xmax…
In short
This work provides updated mathematical models for describing ultra-high-energy cosmic ray (UHECR) showers' depth of maximum ($X_{\rm max}$) for nuclei from proton to iron. Researchers used simulations from three advanced hadronic interaction models (Epos LHC-R, Sibyll 2.3e, QGSJet-III-01) to create parametrizations for the mean and variance of $X_{\rm max}$. The resulting Gumbel distribution fitting accurately reflects simulation results with small residuals, making these models useful for precise composition studies.
Key concepts
- Xmax
- This refers to the depth at which an air shower reaches its maximum particle density. It is a key observable used to infer the mass (atomic weight) of the primary cosmic ray particle that initiated the shower. Measuring $X_{\rm max}$ allows scientists to determine what type of nucleus (like a proton or iron) caused it.
- Hadronic Interaction Models
- These are sophisticated computer simulations that model how high-energy cosmic rays interact with and produce secondary particles in the Earth's atmosphere. The paper tested three state-of-the-art models: Epos LHC-R, Sibyll 2.3e, and QGSJet-III-01. These models are crucial because they generate the raw data (CONEX) used to predict $X_{\rm max}$ for different primary masses.
- Generalized Gumbel Function
- This is a specific statistical function used to model the full distribution of $X_{\rm max}$ values. It is a flexible mathematical tool that allows researchers to describe the shape of the data, capturing both the mean and variance simultaneously. Fitting this function to experimental data helps create an accurate representation of how $X_{\rm max}$ varies for a given energy and mass.
- Moment Parametrization
- This involves creating mathematical formulas that describe how key statistical measures—specifically the mean ($μX) and variance ($σ^2)—of the $X_{\rm max}$ distribution change as the primary cosmic ray energy and mass increase. These parametrizations act as a bridge between complex simulation data and simpler analytical tools needed for composition analysis.
Terminology used across episodes
This episode discusses
- Updated Air-Shower X max Moment Parametrizations for UHECR Composition with Latest Hadronic Interaction Models · Paper Radio
- Measurements of the Cosmic Ray Composition with Air Shower Experiments
- Interpretation of the Depths of Maximum of Extensive Air Showers Measured by the Pierre Auger Observatory
- Road map for the tuning of hadronic interaction models with accelerator-based and astroparticle data
- Measurement of the Depth of Maximum of Air-Shower Profiles with energies between 10 18.5 and 10 20 eV using the Surface Detector of the Pierre Auger Observatory and Deep Learning
- Depth of Maximum of Air-Shower Profiles at the Pierre Auger Observatory: Composition Implications
- Reinterpreting the development of extensive air showers initiated by nuclei and photons
- On the parametrization of the distributions of depth of shower maximum of ultra-high energy extensive air showers
- QGSJET-III model of high energy hadronic interactions: II. Particle production and extensive air shower characteristics
- Impact of the EPOS.LHC-R hadronic interaction model on the Centaurus A ultrahigh-energy cosmic-ray scenario
The paper
Updated Air-Shower X max Moment Parametrizations for UHECR Composition with Latest Hadronic Interaction Models · Read on arXiv
Carmelo Evolia, Igor Vaimana, Sergio Petreraa, Francesco Salamidab
Gran Sasso Science Institute (GSSI) · University of Aquila · INFN-Laboratori Nazionali del Gran Sasso (LNGS)
The mass composition of ultra-high-energy cosmic rays (UHECRs) is commonly inferred from the first two moments of the depth of shower maximum, X max, measured by fluorescence and hybrid detectors. Such analyses require fast and accurate mappings between the moments of X max and those of the logarithmic mass, A, based on realistic air-shower simulations. In this work we provide updated parametrizations of the X max moments and distributions for air showers initiated by nuclei from proton to iron, simulated with CONEX for three state-of-the-art hadronic interaction models: Epos LHC-R, Sibyll 2.3e, and QGSJet-III-01. We parametrize the mean depth X max and the variance σ 2(X max) as functions of energy and mass. For the variance we compare a second-order polynomial model with an exponential model. In addition, we model the full X max distributions with a three-parameter generalized Gumbel function. The Gumbel parameters are fitted using an unbinned likelihood and are validated by comparing the implied mean and variance with the raw CONEX samples and with the moment parametrizations. Across the full energy range considered, residuals between the parametrizations (or the Gumbel representation) and the simulations are at the level of a few g cm-2 for the mean and a few (g cm-2) squared for the variance, making these parametrizations suitable for precision UHECR composition studies and forward-folding analyses of X max distributions.
DOI: 10.1016/j.astropartphys.2026.103239
Transcript
Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.
Vera: We're looking at the paper: "Updated Air-Shower X max Moment Parametrizations for UHECR Composition with Latest Hadronic Interaction Models".
Jocelyn: Here's the gist of the paper. The mass composition of ultra-high-energy cosmic rays (UHECRs) is commonly inferred from the first two moments of the depth of shower maximum, Xmax, measured by fluorescence and hybrid detectors This work provides updated parametrizations of the Xmax moments and distributions for air showers initiated by nuclei from proton to iron, simulated with CONEX for three state-of-the-art hadronic interaction models: Epos-LHC, Sibyll,2.3e, and QGSJet-III-01 The authors parametrize the mean depth of maximum as a function of energy and mass using a generalization of models following the Heitler model They use a parametrization with four free parameters (a0, a1, b0, b1) to describe the mean depth of maximum as µX(ϵ, Y) = a′(ϵ) + b′(ϵ) Y For the variance of the distributions, two alternative parametrizations are considered: a second-order polynomial model based on ϵ and a quadratic expansion in Y The exponential parametrization is based on an exponential dependence on Y, which guarantees σ2X > 0 for all ϵ and Y Gumbel Distribution Fitting The full Xmax distributions are modeled with a three-parameter generalized Gumbel function, G(X; µ, σ, λ) = 1/σλΓ(λ)exp−λz + e−z, z≡X − µσ, z≡X - µσ The Gumbel parameters are fitted using an unbinned likelihood and validated by comparing the implied mean and variance with the raw CONEX samples and with the moment parametrizations The resulting representation reproduces simulated mean and variance with residuals at the level of a few g cm−2 and a few (g cm−2)2, respectively across the full energy range considered Model Comparison and Validation The study compares Xmax observables’ parametrization between latest and previous versions of each hadronic interaction model, quantifying the impact of the update on observables EPOS and QGSJet show more pronounced updates, with EPOS predicting systematically deeper ⟨Xmax⟩ by up to O(10)–O(30) g cm−2 (depending on energy) Consistency Checks and Invertibility The generalized Gumbel parameterization provides a faithful and internally consistent description of the Xmax distributions, yielding first and second moments that are fully consistent with those from the underlying CONEX simulations within the accuracy required for combined spectrum and composition analyses The polynomial model for variance leads to a simple analytical inversion when mapping σ2Xmax to σ2ln A in composition analyses, while the exponential model can be inverted numerically following a procedure described in this work These parametrizations provide a practical interface between air-shower simulations and experimental composition analyses, enabling efficient forward-folding studies and systematic comparisons across hadronic interaction models All data products associated with this note are made publicly available through a dedicated Zenodo repository The released material includes the fitting parameters of derived parameterizations, in plaintext format for ease of use The paper concludes that the generalized Gumbel parameterization provides a robust analytical representation of the Xmax distributions and is sufficiently accurate for use in combined spectrum and composition analyses The size and structure of the residuals are 12 Over the full range of energies and masses considered, the Gumbel-based moments track the CONEX results at the level of a few g cm−2 for the mean and a few percent for the variance, with no significant systematic trend as a function of E or A The comparison is summarized in Fig.s 8 and 9, which shows the residuals between the CONEX moments and those obtained from the Gumbel representation, together with the residuals from the direct moment parameterizations Table B.5 presents best-fit parameters for µX and σ2(Xmax) for EPOS LHC-R, Sibyll 2.3e, and QGSJet III-01 Table B.4 reports average residuals between the CONEX simulations and the analytical parametrizations for σ(Xmax) Table B.3 shows average residuals between the CONEX simulations and the analytical parametrizations for µX Figure 7 displays empirical cumulative distribution functions (CDFs) of Xmax from CONEX simulations compared to the CDFs predicted by the generalized Gumbel parameterization Figure 9 shows residuals between the CONEX moments and those obtained from the Gumbel representation The best-fit coefficients for the three hadronic interaction models are reported in Table B.5 The best-fit parameters for the µX parametrization (Eq.
Vera: 4) and for the σ2(Xmax) parametrizations corresponding to the second-order polynomial model (pol2, Eq.
Jocelyn: 5) and the exponential model (exp, Eq.
Vera: 6), obtained from CONEX simulations for the earlier version of the three hadronic interaction models discussed in the main text are reported in Table B.2 The best-fit coefficients for the generalized Gumbel parameterization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 The best-fit coefficients for the generalized Gumbel parametrization of the Xmax distributions are reported in Table B.5 <ref:2602.
Paper summary: Vera: So, this paper is about updating how we measure the mass composition of ultra-high-energy cosmic rays using shower maximum depth, X max. It's looking at the first two moments—the mean and variance—to map those onto the logarithmic mass.
Jocelyn: Exactly. The main point here is that we need fast and accurate mappings between those moments of X max and the logarithmic mass, A, based on realistic air-shower simulations. This work provides updated parametrizations of these X max moments and distributions for nuclei from proton to iron, simulated with CONEX for three specific hadronic interaction models: Epos LHC-R, Sibyll 2 point 3e, and QGSJet-III-one <ref:2602.18118#pg1,nuclei from proton to iron, simulated with CONEX for three>.
Subrahmanyan: It matters because the way we infer composition relies heavily on these mappings, and since the underlying hadronic interaction models have uncertainties, you need something robust to handle those differences between them two <ref:2602.18118#pg1>. This paper tackles that uncertainty head-on by providing these updated parametrizations for X max and sigma two(X max) as functions of energy and mass <ref:2602.18118#pg1>.
Vera: It claims they use a few different ways to parametrize the mean depth, X max, depending on energy and mass, using generalizations of models like Heitler. They use a parametrization with four free parameters, a zero a one b zero and b one to describe mu X(epsilon, Y) = a'(epsilon) + b'(epsilon) Y <ref:2602.18118#pg3>.
Jocelyn: And for the variance of those distributions, they look at two alternatives. They compare a second-order polynomial model based on epsilon with an exponential model. The exponential parametrization is interesting because it guarantees that the variance, sigma squared X, stays positive for all energy and mass values <ref:2602.18118#pg1>.
Subrahmanyan: That comparison between the polynomial and exponential models for the variance is key because it gives you two different mathematical ways to describe how much the shower depth spreads out as you change energy or primary mass. I wonder which one fits better across the whole range of what we're seeing.
Paper summary: Vera: Beyond just describing X max and sigma two(X max), they model the full X max distributions using a three-parameter generalized Gumbel function, G(X; mu, sigma, lambda) = one/sigma lambda (lambda) (-lambda z) + e-z, where z is defined in terms of mu and sigma <ref:2602.18118#pg1>. They fit the parameters of this Gumbel function using an unbinned likelihood.
Jocelyn: The validation step is pretty solid; they compare the implied mean and variance from this Gumbel representation against the raw CONEX samples and also against those moment parametrizations they built earlier. They report that across the full energy range considered, the residuals between these parametrizations or representations and the simulations are at a level of a few g cm-two for the mean and a few (g cm-two) squared for the variance.
Subrahmanyan: That residual level is quite good, it makes these parametrizations suitable for precision UHECR composition studies and forward-folding analyses of X max distributions. It means we can use this stuff to get much tighter constraints on the composition than before.
Vera: They also compared how different versions of the hadronic interaction models changed when they updated them. EPOS and QGSJet show more pronounced updates, with EPOS predicting systematically deeper X max by up to O(ten) to O(thirty) g cm-two depending on the energy three.
Jocelyn: That difference in the mean depth between models is significant, and it shows how much the choice of interaction model impacts your final composition inference. It also points to where we still have sizable and difficult-to-quantify systematic uncertainties in air-shower physics beyond terrestrial accelerators two <ref:2602.18118#pg1>.
Subrahmanyan: So, for someone who only listens to this show, what does this mean? It means that if you are trying to figure out the cosmic ray composition right now, you need these updated tools to account for the differences between the models we use. The paper shows how these new parametrizations provide a practical interface between those simulations and the experimental analysis.
Paper summary: Vera: So, thinking about what this paper actually delivers in terms of utility, it’s providing a way to do efficient forward-folding studies and systematic comparisons across all three hadronic interaction models using these new tools. The authors also noted that the polynomial model for variance leads to a simple analytical inversion when mapping sigma squared X max to sigma squared A in composition analyses, while the exponential model requires a numerical inversion procedure <ref:2602.18118#pg1>.
Jocelyn: It’s about making the process more tractable for experimentalists, giving them concrete tools based on these simulations that they can actually use with their data. And when you look at the Gumbel representation, it provides a faithful and internally consistent description of the X max distributions, yielding first and second moments fully consistent with those from the underlying CONEX simulations within the accuracy required for combined spectrum and composition analyses <ref:2602.18118#pg3>.
Subrahmanyan: And if you look at Figure nine which shows the residuals between the CONEX moments and those obtained from the Gumbel representation, they track those CONEX results at the level of a few g cm-two for the mean and a few percent for the variance, with no significant systematic trend as a function of energy or mass. That consistency is what makes these tools reliable for structure formation studies.
Vera: It’s about establishing this robust analytical representation of the X max distributions using the Gumbel function, which they say is sufficiently accurate for use in combined spectrum and composition analyses <ref:2602.18118#pg3>. This level of accuracy is what allows us to move forward with using these measurements to constrain the UHECR sources.
Jocelyn: It’s about giving experimentalists a way to efficiently compare results across different models, which is crucial because those models don't agree perfectly on every single detail. The data products are made public through a dedicated Zenodo repository, so the fitting parameters are available in plaintext format for easy use.
Subrahmanyan: So, the big picture is that this work gives us a more reliable way to connect what we see at Earth—the X max measurements—to what’s happening out there in the cosmos by providing these calibrated tools against which we can test different physics scenarios. It sets up a clear path for future composition studies using these updated moment parametrizations.
Conclusion: Vera: So, this paper is giving us updated ways to measure cosmic ray mass composition using X max moments, specifically focusing on how the mean and variance of that depth change with energy and nucleus type.
Jocelyn: It's updating those mappings based on three different hadronic interaction models—EPOS LHC-R, Sibyll two point 3e, and QGSJet-III-one—which is important because those models don't all agree perfectly on the physics of the air shower itself <ref:2602.18118#pg2>.
Subrahmanyan: The authors are providing these new parametrizations for the mean depth X max and the variance sigma two(X max) using different mathematical tools, like a polynomial model or an exponential model for that variance.
Vera: They use a generalized Gumbel function to fit the whole distribution of X max, and they say these fitted results match their simulation data pretty well, with residuals being only a few g cm-two for the mean.
Jocelyn: That means they have a solid way to translate what we see on the detector—the raw shower depths—into a more reliable estimate of which nucleus is hitting us.
Subrahmanyan: For us in cosmology, this helps because if we can do that conversion accurately across different models, it gives us tighter constraints on the composition and how those sources are distributed out there.
Vera: It’s about making the link between air-shower simulations and real data much more practical for someone doing a forward-folding analysis of X max distributions.
Jocelyn: The authors also pointed out that EPOS, for instance, predicts a systematically deeper mean depth than some of the others by up to ten or thirty g cm-two depending on the energy.
Subrahmanyan: That systematic shift is significant because it shows where the physics in those models still has large uncertainties that need addressing.
Vera: The conclusion is that these parametrizations offer a robust analytical representation for the X max distributions and are accurate enough for combined spectrum and composition analyses.
Jocelyn: It’s essentially giving experimentalists a very useful bridge between the simulation side and the observation side of UHECR studies.
Subrahmanyan: So, what this means is that we can start using these tools to push our constraints on the sources of ultra-high-energy cosmic rays more tightly.
Vera: If you want to see exactly how they did all this work, you can find the fitting parameters and the full details in a dedicated Zenodo repository.
More episodes
- 2607.08835-Hybrid Simulations of Proton Acceleration at Oblique High- beta Shocks
- 2606.29552-Improving the Precision of Line-by-Line Radial Velocities: A Data-Driven Iterative Algorithm for Spectral Line Selection
- 2609.05312-Evidence for LP 890-9d via Transit Timing Variations
- 2609.10319-Extending the Little Red Dot population at intermediate redshift with VIPERS
- 2611.14233-Probing for anisotropies at different angular scales in the arrival directions of the cosmic rays detected at the Pierre Auger Observatory
- 2610.21722-Hydrostatic Equilibrium Estimate of the Fuzzy Dark Matter Mass from Dwarf Spheroidals: Method and Systematics
- 2610.40550-An Empirically Calibrated [C II] Luminosity-Halo Mass Relation for Line Intensity Mapping
- 2607.27328-PANTERA. I. Hunting for the Most Metal-Rich Stars in the Solar Neighborhood with High-Resolution Spectroscopy
- 2607.09850-Low-hard to high-soft spectral state transitions in the faintest early-X-ray-detected optical tidal disruption event TDE 2025aarm
- 2611.14028-Constraints on Lorentz Invariance Violation from PeV Gamma Rays of Cygnus X-3 Observed by LHAASO