New approach to primary mass composition analysis with simultaneous use of ground and fluorescence detectors data

We study the possibility to reconstruct primary mass composition with the use of combinations of basic shower characteristics, measured in hybrid experiments, such as depth of shower maximum from fluorescence side and signal in water Cherenkov tanks or in plastic scintillators from the ground side. To optimize discrimination performance of shower observables combinations we apply Fisher's discriminant analysis and give statistical estimates of separation of the obtained distributions on Fisher variables for proton and iron primaries. At the final stage we apply Multiparametric Topological Analysis to these distributions to extract composition from prepared mixtures with known fractions of showers from different primary particles. It is shown, that due to high sensitivity of water tanks to muons, combination of signal in them with $\xmax$ looks especially promising for mass composition analysis, provided the energy is determined from longitudinal shower profile.