Active galactic nucleus time-variability analysis and its caveats

TL;DR

This paper critically examines statistical methods for analyzing AGN variability, revealing overlooked caveats and demonstrating how sampling biases and algorithm misuse can distort timescale measurements.

Contribution

It provides a practical review of statistical issues in AGN light-curve analysis, highlighting new caveats and testing their impact through simulations.

Findings

Biased sampling affects power spectral density measurements.

Incorrect use of PDF-matching algorithms can lead to errors.

Sampling bias towards flares influences variability analysis.

Abstract

In this study, we demonstrate some of the caveats in common statistical methods used for analysing astronomical variability timescales. We consider these issues specifically in the context of active galactic nuclei (AGNs) and use a more practical approach compared to mathematics literature, where the number of formulae may sometimes be overwhelming. We conducted a thorough literature review both on the statistical properties of light-curve data, specifically in the context of sampling effects, as well as on the methods used to analyse them. We simulated a wide range of data to test some of the known issues in AGN variability analysis as well as to investigate previously unknown or undocumented caveats. We discovered problems with some commonly used methods and confirmed how challenging it is to identify timescales from observed data. We find that interpolation of a light curve with biased sampling, specifically with bias towards flaring events, affects its measured power spectral density in a different manner than those of simulated light curves. We also find that an algorithm aiming to match the probability density function of a light curve has often been used incorrectly. These new issues appear to have been mostly overlooked and not necessarily addressed before, especially in astronomy literature.