A New Approach to the Internal Calibration of Reverberation Mapping Spectra

TL;DR

This paper introduces a Bayesian-based calibration method for optical AGN reverberation mapping spectra, significantly improving calibration precision and providing a Python package for community use.

Contribution

It proposes a novel calibration approach that enforces constant spectral resolution, models resolution changes flexibly, and assesses uncertainties with Bayesian methods.

Findings

Achieved calibration precision of ~1 millimagnitude, five times better than traditional methods.

Demonstrated the method on MCG+08-11-011 data, improving calibration accuracy.

Provided an open-source Python package, mapspec, for community use.

Abstract

We present a new procedure for the internal (night-to-night) calibration of time series spectra, with specific applications to optical AGN reverberation mapping data. The traditional calibration technique assumes that the narrow [OIII]$\lambda$5007 emission line profile is constant in time; given a reference [OIII]$\lambda$5007 line profile, nightly spectra are aligned by fitting for a wavelength shift, a flux rescaling factor, and a change in the spectroscopic resolution. We propose the following modifications to this procedure: 1) we stipulate a constant spectral resolution for the final calibrated spectra, 2) we employ a more flexible model for changes in the spectral resolution, and 3) we use a Bayesian modeling framework to assess uncertainties in the calibration. In a test case using data for MCG+08-11-011, these modifications result in a calibration precision of $\sim\! 1$ millimagnitude, which is approximately a factor of five improvement over the traditional technique. At this level, other systematic issues (e.g., the nightly sensitivity functions and FeII contamination) limit the final precision of the observed light curves. We implement this procedure as a {\tt python} package ({\tt mapspec}), which we make available to the community.