Proton Flux Measurement from Neutron Monitor Data Using Neural Networks

TL;DR

This paper introduces a novel deep learning approach using neural networks to accurately estimate cosmic ray proton flux from neutron monitor data, providing continuous measurements including during AMS data gaps and enabling hourly flux analysis for solar activity studies.

Contribution

The study presents a new method combining CNN-based data pre-processing with deep learning to derive proton flux from neutron monitor data, including hourly flux calculations for the first time.

Findings

Proton flux measurements show strong agreement with AMS data from 2011 to 2024.

The method provides reliable proton flux estimates during AMS operational gaps.

Hourly proton fluxes are successfully calculated for short-term solar activity analysis.

Abstract

Accurate measurements of cosmic ray proton flux are essential for studying the modulation processes of cosmic rays during the solar activity cycle. A proton flux measurement method, based on ground-based neutron monitor (NM) data and deep learning techniques, is presented. After the necessary pre-processing of ground-based NM data using a convolutional neural network (CNN) model, we simulate the relationship between NM observations and proton flux measured by the Alpha Magnetic Spectrometer (AMS). The daily proton flux data, ranging from 1 GV to 100 GV, are obtained for the period from 2011 to 2024, showing strong agreement with the observed values. In addition, daily proton flux measurements are provided for periods when AMS data were unavailable due to operational reasons. For the first time, hourly proton fluxes as a function of rigidity are calculated dedicated to the short-time solar activity studies.