When an ultra-high-energy cosmic ray (UHECR) collides with an atomic nucleus in Earth's upper atmosphere, it triggers a cascade of billions of secondary particles known as an Extensive Air Shower (EAS). Because we cannot deploy massive particle detectors directly into space to catch these rare, high-energy events before they break apart, astrophysics relies on ground-based arrays to observe these atmospheric showers.
To reconstruct the properties of the original cosmic ray—such as its energy and mass composition—from the secondary particles detected at the surface, researchers depend entirely on Monte Carlo simulations. However, as shown by Rasha Abbasi et al. in their seven-year analysis of Telescope Array data, these reconstructions are fundamentally limited by systematic uncertainties embedded within high-energy hadronic interaction models. This simulation-to-reality gap represents one of the most significant bottlenecks in modern high-energy particle astrophysics.
The Extrapolation Gap
The core challenge of modeling hadronic interactions in air showers is an energy scale discrepancy. The standard models used in simulations (such as QGSJET, SIBYLL, and EPOS) are calibrated using data from terrestrial particle colliders like the Large Hadron Collider (LHC). The LHC operates at a maximum center-of-mass energy ($\sqrt{s}$) of approximately 13 to 14 TeV. However, ultra-high-energy cosmic rays routinely strike the atmosphere at energies exceeding $10^{17}$ eV, which translates to center-of-mass energies far beyond the reach of any human-made machine.
As a result, simulation frameworks must extrapolate physical processes—such as forward particle production and cross-sections—into uncalibrated energy regimes. As detailed by Rasha Abbasi et al. in their mass composition study, these theoretical uncertainties introduce substantial systematic errors that frequently outweigh the statistical precision of modern experimental hardware.
The Muon Production Puzzle
The most prominent physical manifestation of this modeling gap is the "muon puzzle." When a cosmic ray initiates a shower, hadronic interactions produce mesons—primarily pions and kaons. Neutral pions ($\pi^0$) decay almost instantly into photons, driving an electromagnetic cascade. Conversely, charged pions ($\pi^\pm$) continue to interact hadronically until the shower thins out, at which point they decay into muons. Because muons do not interact via the electromagnetic or strong forces as intensely as other particles, they travel relatively unattenuated to the ground, acting as a clean tracer of the shower's hadronic history.
When researchers compare real-world ground measurements to simulated showers, they find a persistent deficit in the models' predictions of muon content:
- The Pierre Auger Observatory: In an analysis of inclined air showers, P. Abreu et al. demonstrated that observed muon densities exceed standard model predictions by 23% to more than 100% (The Pierre Auger Observatory I).
- The Telescope Array: Similarly, Rasha Abbasi et al. confirmed that real-world showers contain roughly 1.7 times more muons than predicted by standard hadronic models (Study of muons from ultrahigh energy cosmic ray air showers).
This systematic underestimation indicates that current theoretical frameworks fail to accurately capture particle production dynamics at extreme energy scales.
How Uncertainty Propagates Through Analysis
Because hadronic interaction models form the foundation of cosmic ray data analysis, their uncertainties propagate through multiple downstream observables, biasing our understanding of the universe.
Mass Composition and Flux Reconstructions
Determining whether a cosmic ray is a light proton or a heavy iron nucleus depends heavily on composition indicators like the depth of shower maximum ($X_{\text{max}}$) and muon density. Fahim Varsi et al. note that discrepancies between theoretical models and actual high-energy collision physics lead to systematic shifts in the reconstructed chemical composition of primary particles (Updated results on the cosmic ray energy spectrum and composition from the GRAPES-3 experiment). Furthermore, Rafael Alves Batista et al. point out that a model's ability to successfully fit an observed energy spectrum does not guarantee an accurate representation of chemical composition (Cosmic ray propagation with CRPropa 3).
Calorimetry and Jet Energy Scales
The problem is not unique to ground arrays. In space-based calorimetry experiments like DAMPE, which directly measure cosmic ray fluxes, the choice of hadronic interaction model represents the single largest source of systematic uncertainty, introducing an error of 12% to 15% at the tens of TeV scale (Francesco Alemanno et al. in Measurement of the Cosmic Ray Helium Energy Spectrum). This mirrors challenges in terrestrial collider experiments; for example, A. Bhatti et al. documented how discrepancies in modeling forward detector regions and spectator particle behavior limit the precision of jet energy scales at the CDF experiment (Determination of the jet energy scale).
Machine Learning and Background Subtraction
Modern astrophysics relies on machine learning classifiers to isolate rare signals—such as ultra-high-energy photons or electrons—from an overwhelming background of hadronic cosmic rays. If the underlying simulations used to train these classifiers are built on flawed proton models, the resulting classification boundaries become biased.
Rasha Abbasi et al. demonstrated that this dependency introduces systematic errors when setting upper limits on diffuse photon fluxes (Constraints on the diffuse photon flux) and searching for point sources (Search for point sources of ultra-high-energy photons). Similarly, ground-based Cherenkov telescopes like MAGIC face strict limitations in background subtraction and spectral slope precision due to these atmospheric cascade modeling constraints (D. Borla Tridon et al. in Measurement of the cosmic electron spectrum; P. Temnikov et al. in Protons Spectrum from MAGIC Telescopes data).
Constraining Models via Hybrid Observations
To resolve these uncertainties, researchers are designing observational strategies that isolate specific physical parameters to test and validate competing hadronic models.
One effective strategy involves measuring the attenuation length of the muon content in air showers across different zenith angles. By analyzing these distributions with the LHAASO-KM2A experiment, Z. Cao et al. showed that experimental data explicitly favors the EPOS-LHC hadronic interaction model over QGSJET-II-04 and SIBYLL 2.3d (Measurement of attenuation length of the muon content). While models may align well at specific intermediate energy scales like 100 TeV (Guang-Guang Xin et al. in Study on the muon lateral distribution), testing their predictive accuracy across wider energy ranges remains critical.
[UHECR Primary Particle]
│
▼ (Extrapolation Gap: Energies > LHC scale)
[Hadronic Cascade] ───► Mesons (Pions, Kaons)
│ │
│ (Electromagnetic) │ (Hadronic Skeleton)
├──► Neutral Pions └──► Charged Pions
│ │ │
│ ▼ ▼
└──► Photons/Electrons [Muons] ──► (Observed Surplus: 1.7x Model Predictions)
Another approach uses the air showers themselves as a high-energy laboratory to measure fundamental particle physics parameters directly. For example, Rasha Abbasi et al. bypassed historical model dependencies by analyzing the spatial "tail" of cosmic ray shower distributions to directly infer the inelastic proton-air cross section at a center-of-mass energy of 95 TeV (Measurement of the proton-air cross section).
By leveraging hybrid detection modes—such as combining surface particle arrays with atmospheric fluorescence or muon detectors—the community continues to generate empirical constraints that force the refinement of hadronic interaction models, gradually closing the gap between simulation and reality (Rasha Abbasi et al. in Study of Ultra-High Energy Cosmic Ray composition; D. Kurashige et al. in Sensitivity of the Tibet hybrid experiment).
Key Foundational Work
The Pierre Auger Observatory I: The Cosmic Ray Energy Spectrum and Related Measurements (2011) — contributed 4 concept elements to this direction.
Measurement of the cosmic electron spectrum with the MAGIC telescopes (2011) — contributed 2 concept elements to this direction.
Determination of the jet energy scale at the Collider Detector at Fermilab (2006) — contributed 1 concept elements to this direction.