Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 58: Cosmic Rays III

T 58.7: Vortrag

Mittwoch, 2. April 2025, 17:45–18:00, VG 3.102

Impact of adding simulations of ultra-heavy cosmic rays on neural network-based estimators using surface detector data of the Pierre Auger Observatory — •Steffen Hahn for the Pierre-Auger collaboration — Karlsruhe Institute of Technology, Eggenstein-Leopoldshafen, Germany

To understand the physics of ultra-high-energy cosmic rays (UHECRs), an accurate estimate of the masses of UHECR is crucial. Direct detection of UHECRs is not feasible and requires the study of air showers induced by the interaction of UHECRs with the atmosphere. The surface detector stations of the Pierre Auger Observatory (Auger) measure the front of such cascades, called the shower footprint. Analyzing the spatio- temporal information of these shower footprints is highly non-trivial. Neural networks (NNs) offer a convenient way to exploit the correlations in the footprints and improve the reconstruction of high-level shower observables. However, simulations of UHECRs face limitations due to incomplete understanding of the high-energy hadronic interactions. The most prominent discrepancy is the muon puzzle -- a systematic deficit of muons in simulations which complicates the application of simulation-trained NNs to Auger data. Typically, training data sets for Auger consist of a mix of proton, iron, and intermediate-mass nuclei. Since the number of muons produced in an air shower correlates with the mass of the UHECR, varying the mass composition in the training dataset could impact the transition to measurements. In this contribution, we show how the inclusion of heavier UHECRs affects NN-based estimators in simulations and measurements.

Keywords: Pierre Auger Observatory; ultra-high-energy cosmic rays; ultra-heavy cosmic rays; machine learning