Characterizing glitches in high-energy pulsars. A Bayesian approach to pulsar timing

Glitches are seen in more than 50 young gamma-ray pulsars detected by the Fermi-LAT, but traditional timing techniques fail to accurately characterize gamma-ray glitch parameters. In this thesis we discuss the case of the variable Fermi-LAT pulsar PSR J2021+4026, an isolated gamma-ray pulsar that shows repeated changes in its gamma-ray flux and spin-down rate. We report on a multi-wavelength spectral and timing analysis. The results suggest that the phenomenon must be related to a global change in the geometry of the magnetic field. We propose a semi-quantitative model that assumes curvature radiation in a quasi-force-free dissipative magnetosphere. We explore different configurations of a multipolar magnetic field in vacuum, and we find a combination of parameters that is qualitatively consistent with the observations. Motivated by this example, we propose a new analysis approach to pulsar timing that aims to characterize glitches in Fermi-LAT pulsars by means of Bayesian inference. Our procedure starts with unbinned and weighted Fermi-LAT photons and runs a nested sampling algorithm to jointly infer rotational and profile model parameters. We have implemented GLIMPSE, a modular Python package dedicated to pulsar monitoring and glitch characterization. We describe the main components of GLIMPSE and its implementation principles. We test the efficiency of our algorithm and discuss its applications in pulsar astrophysics and multi-messenger astronomy.