Improving Air Shower Simulations by Tuning Pythia 8/Angantyr with Accelerator Data

Abstract: We present a combined analysis of the Pythia 8 event generator using accelerator data and evaluate its impact on air shower observables. Reliable simulations with event generators are essential for particle physics analyses, achievable through advanced tuning to experimental data. Pythia 8 has emerged as a promising high-energy interaction model for cosmic ray air shower simulations, offering well-documented parameter settings and a user-friendly interface to enable automatic tuning efforts. Using data from collider and fixed-target experiments, we first derive tunes for each domain separately, before tuning both domains simultaneously. To achieve this, we define a core set of observables and quantify their dependence on selected parameters. The tuning efforts are based on gradient descent and Bayesian methods, the latter providing a full uncertainty propagation of the parameters to the observables. Results for the impact of a combined analysis for the Pythia 8/Angantyr event generator on air shower observables, such as particle densities at ground level and energy deposit profiles, are presented.