ENERGY-SPECTRA AND ELEMENTAL COMPOSITION DETERMINATION ON MOUNTAIN ALTITUDES AND SEA-LEVEL

The accurate measurement of the elemental composition and energy spect ra of primary cosmic rays in the energy range 10(14) - 10(17) eV is vi tally important. The information carried by Extensive Air showers (EAS ) in this energy range is limited by significant fluctuations of the s hower development, insufficient experimental sampling and uncertaintie s of the detector response. Additional uncertainties arise from differ ent simulation procedures and statistical methods used for inferring r esults from the measurements. We propose to combine an advanced EAS si mulation program like CORSIKA with the ANI multivariate statistical an alysis package for an event-by-event analysis and for a standardizatio n of the inference from EAS data. As an example, the possibilities of modern EAS installations at mountain altitudes and sea level are explo red. The accuracies of the elemental composition determination based b n measured components of EAS are presented. The Neural Networks classi fication and Bayesian Decision Making approaches are used and compared . The new nonparametric methods of primary energy estimations are disc ussed.