OzDES Reverberation Mapping Program: Lag recovery reliability for 6-year CIV analysis

TL;DR

This paper develops statistical methods and simulations to assess the reliability of lag recovery in the OzDES reverberation mapping program, focusing on CIV emission line analysis over six years.

Contribution

It introduces a set of customizable simulations and criteria to evaluate the accuracy and reliability of lag measurements in reverberation mapping data.

Findings

Recovered R-L relation slope within 1σ for all quality levels

Gold standard yields the lowest dispersion in R-L slope

Simulations enable assessment of lag recovery reliability

Abstract

We present the statistical methods that have been developed to analyse the OzDES reverberation mapping sample. To perform this statistical analysis we have created a suite of customisable simulations that mimic the characteristics of each source in the OzDES sample. These characteristics include: the variability in the photometric and spectroscopic lightcurves, the measurement uncertainties, and the observational cadence. By simulating the sources in the OzDES sample that contain the CIV emission line, we developed a set of criteria that rank the reliability of a recovered time lag depending on the agreement between different recovery methods, the magnitude of the uncertainties, and the rate at which false positives were found in the simulations. These criteria were applied to simulated light curves and these results used to estimate the quality of the resulting Radius-Luminosity relation.We grade the results using three quality levels (gold, silver and bronze). The input slope of the R-L relation was recovered within $1\sigma$ for each of the three quality samples, with the gold standard having the lowest dispersion with a recovered a R-L relation slope of $0.454\pm 0.016$ with an input slope of 0.47. Future work will apply these methods to the entire OzDES sample of 771 AGN.