Sibyll$^\bigstar$: ad-hoc modifications for an improved description of muon data in extensive air showers

TL;DR

This paper explores ad-hoc modifications to the Sibyll~2.3d model to better match observed muon data in extensive air showers, aiming to address discrepancies in muon content predictions.

Contribution

It introduces specific final state modifications in Sibyll~2.3d to improve the simulation of muon production in air showers, providing a potential pathway to resolve existing data-model mismatches.

Findings

Modified models show improved muon content agreement with data

Increased muon production reduces proton-iron shower differences

Model adjustments may aid machine learning analysis of air shower data

Abstract

Current simulations of air showers produced by ultra-high energy cosmic rays (UHECRs) do not satisfactorily describe recent experimental data, particularly when looking at the muonic shower component relative to the electromagnetic one. Discrepancies can be seen in both average values and on an individual shower-by-shower basis. It is thought that the muonic part of the air showers isn't accurately represented in simulations, despite various attempts to boost the number of muons within standard hadronic interaction physics. In this study, we investigate whether modifying the final state of events created with Sibyll~2.3d in air shower simulations can achieve a more consistent description of the muon content observed in experimental data. We create several scenarios where we separately increase the production of baryons, $\rho^0$, and strange particles to examine their impact on realistic air shower simulations. Our results suggest that these ad-hoc modifications can improve the simulations, providing a closer match to the observed muon content in air showers. One side-effect of the increased muon production in the considered model versions is a smaller difference in the predicted total muon numbers for proton and iron showers. However, more research is needed to find out whether any of these adjustments offers a realistic solution to the mismatches seen in data, and to identify the precise physical process causing these changes in the model. We hope that these modified model versions will also help to develop improved machine-learning analyses of air shower data and to estimate sys.{} uncertainties related to shortcomings of hadronic interaction models.