Ensemble Dynamical Classification of X-ray Binaries

Accreting compact object systems such as X-ray Binaries (XRBs) exhibit variability in their luminosity on many timescales ranging from milliseconds to hundreds of days. The fundamental structure of the accretion flow and how it evolves to result in the observed variability is not well known. The expectation is that there are simultaneous, and interconnected, random and chaotic or periodic components in the light curves of accretion disk systems, operating on different timescales. The strength of each component is dependent on the dominant sources of instability in the accretion flow and the magnitude of stochastic fluctuations. Metrics derived from recurrence analysis specifically probe the strength and nature of nonlinearity, or lack thereof, in time series and consequently can provide information about the dominating physical mechanisms at play. We perform a pilot study of canonical XRBs monitored by RXTE and MAXI over two decades and compare their measured dynamical variability via recurrence analysis to an ensemble of simulated classical dynamical systems including first and second order continuous autoregressive processes, multi- and quasi-periodicity, noise, and chaos. We find XRBs across spectral classes align with distinct dynamical systems, which suggests that the spectral state of the accretion flow correlates with the dynamical classification of the light curve.