A method for enhanced gamma-proton discrimination with imaging atmospheric Cherenkov telescopes

Abstract

Purpose: Very-high-energy gamma-ray astronomy is an important window to study the extreme astrophysical processes in the universe, and how to effectively distinguish between gamma photons and background signals (mainly protons) is a key technical challenge to realize very-high-energy gamma-ray detection. Spaceborne calorimeters record secondary particle spatial distributions, achieving a background rejection power of 10000-100000 near 1 TeV. Ground-based imaging Cherenkov telescopes measure Cherenkov photon angular distributions, yet they only reach a rejection factor of about 10 in this energy range, leaving substantial room for optimization. Methods: We use CORSIKA and sim_telarray to simulate H.E.S.S. detector responses. Photon emission positions are reconstructed from Cherenkov shower images. Inspired by analysis techniques for space-borne calorimeter, we propose new observables derived from the transverse distribution of Cherenkov photon emission positions to separate gamma-ray signals from proton backgrounds. We employ a likelihood-ratio method to combine single-telescope variables across multiple telescopes and compare its performance with conventional Hillas-based analysis. Results: The results show that observables built from the transverse distribution of Cherenkov photon emission positions are less sensitive to Cherenkov image size and outperform the Hillas width parameter for single-telescope gamma-proton discrimination. Furthermore, the likelihood ratio combining width yields substantially better performance than the conventional mean reduced scaled width: the improvement reaches roughly one order of magnitude near 1 TeV, and this enhancement remains above 50% for energies above 10 TeV.