Directional Coincidence Assessment in Astronomy with Bayesian Multilevel Models

• Activity Number: 377
• Type: Invited
• Date/Time: Tuesday, July 31, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Thomas J. Loredo*+ and David F. Chernoff and David Ruppert and Kunlaya Soiaporn and Ira Wasserman
• Companies: Cornell University and Cornell University and Cornell University and Cornell University and Cornell University
• Keywords: Bayesian statistics ; directional data ; astronomy ; multilevel models ; hierarchical bayes

Abstract

Astronomers observe the sky in many regions of the electromagnetic spectrum, and in forms of non-electromagnetic radiation, including energetic particles (cosmic rays, neutrinos) and gravitational radiation. It has become a truism of multi-wavelength and multi-messenger astronomy that leaps in understanding often follow the discovery of counterparts to a source in multiple radiation regimes. But when there is significant uncertainty in the measured directions to sources on the crowded celestial sphere, distinguishing genuine physical associations from mere directional coincidences becomes challenging. We describe a Bayesian approach for directional coincidence assessment based on multilevel modeling, where upper levels in a model describe properties of potentially associated source populations and radiation propagation, and lower levels describe measurement errors and survey selection effects. We apply the approach to the search for sources of ultra-high energy cosmic rays (UHECRs), atomic nuclei accelerated to near the speed of light by unknown astrophysical processes believed to involve supermassive black holes in the nuclei of active galaxies.