Quasar variability in the Palomar Transient Factory Survey

We quantify the quasar variability in the Palomar Transient Factory survey. This is the data survey of unprecedented size with over 2 million observations of 30000 quasars brighter than r=19 over 6 years. Using custom recalibration we greatly improve on the photometric quality of the survey reducing the unaccounted errors to less than 1% at r=19. We split our sample in the redshift, mass and luminosity bins and using the ensemble structure function analysis we find anti-correlation of variability with luminosity, no correlation with redshift and only very weak correlation with mass. We also notice the dependence of the slope of the structure function with mass and luminosity indicating that random walk is not equally good description of QSO variability at all masses and luminosities. We extend the analysis using the power spectral density (PSD) analysis with CARMA models. We observe breaks in the PSD but after extensive simulation effort we conclude that these are artifact arising from the insufficient length of the light curves and non-uniform cadence. We observe the dependence of PSD slope with mass/luminosity further corroborating our finding in the structure function analysis that type of the process that drives variability is dependent on the physical parameters.