Principal components analysis of Cerenkov photon distributions from extensive air showers applied to GeV gamma–proton discrimination

High-energy interactions of cosmic rays with Earth atmosphere and the resulting extensive air showers have been simulated by means of the CORSIKA Monte Carlo code. Statistical fluctuations of the two-dimensional secondary Cerenkov photon density distributions at ground level produced by GeV cosmic γ rays and protons are studied by means of principal components analysis (PCA). This provides a decreasing sequence of covariance matrix eigenvalues that may be fitted to an analytical expression allowing to distinguish among different primary cosmic rays. A very efficient discrimination method of gamma ray induced showers from proton simulated extensive air showers is proposed as a result of this analysis. A cutting parameter is calculated, and the efficiency of the cutting procedure for γ–proton separation is evaluated under experimental conditions.