Characteristic Timescales in the Gamma-ray Power Spectra of Blazars

Abstract

I study the γ-ray variability features of 13 blazars by estimating their power spectral densities (PSDs) with continuous-time autoregressive moving average (CARMA) models. The analyzed blazars were observed with the Fermi/Large Area Telescope during the first ∼ 6 years of the Fermi sky survey, and were selected due to their lightcurves being the most complete. Several methods of estimating the PSDs of lightcurves were compared, and CARMA was found to give the most accurate approximation. CARMA also has the advantage of offering likelihoods for the models, and can account for irregular sampling and measurement errors. Break frequencies in the PSDs are estimated from the CARMA model, which correspond to characteristic timescales τ0 separating regions of the PSD with differing slopes. I find that in each of these regions, the PSD resembles a power law 1/f, with differing α values. Using CARMA, characteristic γ-ray timescales of variability are detected in all 13 sources, and are in partial agreement with earlier measurements, and values of α are calculated. Evidence of quasi-periodic oscillations (QPOs) is also found in several blazars. Interpretations of my results are discussed under theoretical models of blazar variability.