New insights from old cosmic rays: A novel analysis of archival KASCADE data

TL;DR

This paper reanalyzes archival cosmic ray data from KASCADE using modern machine learning to improve mass composition understanding, anisotropy searches, and gamma-ray candidate identification in the PeV range.

Contribution

It introduces a novel machine learning-based analysis of archival KASCADE data, providing updated spectra and anisotropy results with enhanced techniques.

Findings

Spectra for primary nuclei groups are presented.

Anisotropy in event arrival directions is analyzed considering mass composition.

Search for PeV gamma-ray candidates yields new insights.

Abstract

Cosmic ray data collected by the KASCADE air shower experiment are competitive in terms of quality and statistics with those of modern observatories. We present a novel mass composition analysis based on archival data acquired from 1998 to 2013 provided by the KASCADE Cosmic ray Data Center (KCDC). The analysis is based on modern machine learning techniques trained on simulation data provided by KCDC. We present spectra for individual groups of primary nuclei, the results of a search for anisotropies in the event arrival directions taking mass composition into account, and search for gamma-ray candidates in the PeV energy domain.