Using Gaussian processes to model light curves in the presence of flickering: the eclipsing cataclysmic variable ASASSN-14ag

TL;DR

This paper introduces a novel Gaussian process-based method for modeling flickering in cataclysmic variable star light curves, improving the accuracy of system parameter estimation.

Contribution

The study develops and applies a Gaussian process approach to eclipse modeling, effectively accounting for flickering noise in CV light curves, which enhances parameter determination accuracy.

Findings

Accurately determined system parameters of ASASSN-14ag

Demonstrated Gaussian processes effectively model flickering noise

Improved eclipse modeling methodology for CV stars

Abstract

The majority of cataclysmic variable (CV) stars contain a stochastic noise component in their light curves, commonly referred to as flickering. This can significantly affect the morphology of CV eclipses and increases the difficulty in obtaining accurate system parameters with reliable errors through eclipse modelling. Here we introduce a new approach to eclipse modelling, which models CV flickering with the help of Gaussian processes (GPs). A parameterised eclipse model - with an additional GP component - is simultaneously fit to 8 eclipses of the dwarf nova ASASSN-14ag and system parameters determined. We obtain a mass ratio $q$ = 0.149 $\pm$ 0.016 and inclination $i$ = 83.4 $^{+0.9}_{-0.6}$ $^{\circ}$. The white dwarf and donor masses were found to be $M_{w}$ = 0.63 $\pm$ 0.04 $M_{\odot}$ and $M_{d}$ = 0.093 $^{+0.015}_{-0.012}$ $M_{\odot}$, respectively. A white dwarf temperature $T_{w}$ = 14000 $^{+2200}_{-2000}$ K and distance $d$ = 146 $^{+24}_{-20}$ pc were determined through multicolour photometry. We find GPs to be an effective way of modelling flickering in CV light curves and plan to use this new eclipse modelling approach going forward.