Orateur
Description
Large time-domain surveys are turning AGN variability into a population-scale probe of both supermassive black-hole accretion and cosmology. UV/optical variability depends on rest-frame timescale, wavelength, and luminosity: at fixed wavelength and time lag, more luminous sources vary less. This luminosity anti-correlation motivates Type 1 AGN variability as a potential distance indicator beyond the redshift range densely populated by Type Ia supernovae, offering an independent route to test the expansion history at high redshift.
I will present a hierarchical Bayesian framework that combines Gaussian-process modelling of AGN light curves with joint inference of variability-luminosity population parameters and cosmology. The method propagates source-level light-curve uncertainties into a population model relating variability to luminosity, rest-frame wavelength, intrinsic scatter, selection, and the assumed distance-redshift relation. Survey-like mock catalogues are used to validate the inference pipeline and quantify the impact of finite baselines, magnitude limits, chromatic variability, and calibration degeneracies.
I will discuss applications to Gaia G-band and ZTF g/r-band light curves for ∼200,000 SDSS AGN, highlighting current constraints, key degeneracies, and the role of multi-band data in separating luminosity, wavelength, and cosmological effects. I will close with prospects for AGN variability cosmology with Gaia DR4, DESI-selected AGN, and Rubin/LSST-era time-domain surveys.