Presentation #106.45 in the session “AGN (Poster)”.
X-ray reverberation mapping is a powerful technique used to map the innermost accretion flows of Active Galactic Nuclei. Emission from the central corona irradiates the inner accretion disk, leading to delayed correlated variability. By measuring the time delays resulting from light-travel time between the coronal and reprocessed variability, we gain spatial information of the disk. A Fourier frequency-resolved approach has been especially fruitful for isolating the time lags due to variability on different timescales and, thus, distinct physical processes. However, the lowest frequency one can probe with this approach is limited by the length of one’s longest continuous observation — a significant limit for data collected with instruments in low-Earth orbit.
We present new simultaneous XMM-Newton and NuSTAR observations of variable Seyfert I Ark 564, including NuSTAR’s longest single exposure to-date, which we combine with archival observations for 1.4 Ms of data in total. In order to apply Fourier techniques, we model observed variability using Gaussian processes (GPs) from which we generate continuous realizations of our NuSTAR lightcurves, including data in the low-Earth orbit gaps. We best train each GP with the first employment of multi-task learning for application in AGN timing by training a single description of short- and long-timescale variability using all of our observations. We present constraints for various properties of the disk by fitting both the time lags and spectra with reverberation model RELTRANS, from which we find consistency between fitting the spectra and time lags separately and obtain a mass estimate by fitting the time lags.