Speaker
Description
We use the reconstructed properties of individual UHECR events to constrain the location of their unknown sources via approximate Bayesian computation. All important propagation effects, including deflections in both Galactic and extragalactic magnetic fields, are implemented via CRPropa 3. We define priors over key parameters, the source position, distance, particle energy at the source, and extragalactic magnetic field properties, inferring constraints where possible and otherwise marginalising over uncertainties. This approach allows us to jointly consider selection effects due to both the particle energy and arrival direction, leading to a more self-consistent inference. We present our updated results for the Amaterasu particle, including the updated UF23 Galactic magnetic field model, and explore the impact of mass composition and energy scale uncertainties on the interpretation. Our results identify several candidate sources beyond the Local Void. We also discuss ongoing efforts to accelerate our inference method, enabling applications to larger event samples and higher-dimensional parameter spaces.
