Speaker
Description
Calibrating the energy scale of surface detector arrays using fluorescence detector data is a primary source of systematic uncertainty at ultra-high-energy cosmic ray observatories. Standard procedures provide point estimates of the calibration function with no associated uncertainty. We present a hierarchical Bayesian model that introduces the true event energy as a latent variable, allowing measurement noise from both detectors to be accounted for simultaneously while inferring a full posterior over the calibration function. The posterior uncertainty grows naturally at high energies where hybrid coverage is sparse or absent, allowing calibration uncertainty to be propagated into any downstream analysis. The model is validated on synthetic data, demonstrating unbiased recovery of the calibration function and correct posterior coverage.
