Erlangen 2026 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 40: Cosmic Rays II

T 40.2: Talk

Tuesday, March 17, 2026, 16:30–16:45, KS 00.006

Transferability of 3D-Wave Reconstruction with ML Models^* — •Sven Querchfeld for the Pierre Auger collaboration — Bergische Universität Wuppertal, Gaußstr. 20, 42119 Wuppertal, Germany

Within the ErUM-Wave project, a Machine Learning (ML) model is developed to reconstruct three-dimensional wave fields. The goal is to predict the propagation of seismic waves on the basis of only a few measurements. To test the transferability of this method to other fields, it will be applied to the propagation of radio waves in the atmosphere. These waves are produced by cosmic-ray-induced extensive air showers and are measured by the Pierre Auger Observatory. First events have already been recorded by its radio detector, which finished deployment on the full array in November 2024. Training and validation are based on CORSIKA 7 simulations using the CoREAS extension. To obtain more realistic traces, a noise library was created using in-field measurements, utilising the small radio footprint of vertical showers to supplement the simulated events. Due to the CPU-intensive simulations, a preliminary test of the developed ML models for reconstructing 3D wave propagation is presented, using a limited set of showers.

^*Supported by BMBF Verbundforschung Astroteilchenphysik (Vorhaben 05A23PX1) and ErUM-Wave

Keywords: radio detection; CoREAS; CORSIKA; ErUM-Wave