To nitpick: this is only mostly correct. The predominant detection method is actually having arrays of particle detectors on the ground that register through go

To nitpick: this is only mostly correct. The predominant detection method is actually having arrays of particle detectors on the ground that register through going particles. Because you have an array of them, you can use detection timings of a number of triggered detectors to compute arrival detection/etc. These more "indirect" detectors get a 10x data collection advantage over fluorescence telescopes because they run 24x7 whereas the telescopes run in clear, moonless nights. (The moon in the field of view can actually burn out the camera.) Source: subject of my thesis was trying to calibrate the energy measurement of the fluorescence telescopes by using models of the air shower profile and relative light yield from fluorescence vs Cherenkov effects to infer a scalar fluorescence light yield parameter.