Deep-learning based reconstruction of the shower maximum 𝑿max using the water-Cherenkov detectors of the Pierre Auger Observatory

THE PIERRE AUGER COLLABORATION (P. HANSEN PERTENECE A LA COLABORACION-ORDEN ALFABETICO)

JOURNAL OF INSTRUMENTATION

IOP PUBLISHING LTD

Lugar: Londres; Año: 2021

1748-0221

The atmospheric depth of the air shower maximum 𝑋max is an observable commonlyused for the determination of the nuclear mass composition of ultra-high energy cosmic rays. Directmeasurements of 𝑋max are performed using observations of the longitudinal shower developmentwith fluorescence telescopes. At the same time, several methods have been proposed for an indirectestimation of 𝑋max from the characteristics of the shower particles registered with surface detectorarrays. In this paper, we present a deep neural network (DNN) for the estimation of 𝑋max. Thereconstruction relies on the signals induced by shower particles in the ground basedwater-Cherenkovdetectors of the Pierre Auger Observatory. The network architecture features recurrent long shorttermmemory layers to process the temporal structure of signals and hexagonal convolutions toexploit the symmetry of the surface detector array. We evaluate the performance of the networkusing air showers simulated with three different hadronic interaction models. Thereafter, we accountfor long-termdetector effects and calibrate the reconstructed 𝑋max using fluorescence measurements.Finally, we show that the event-by-event resolution in the reconstruction of the shower maximumimproves with increasing shower energy and reaches less than 25 g/cm2 at energies above 21019eV