Stochastic time series modelling of <i>TESS</i> photometry of classical T Tauri stars

ABSTRACT A standard approach to extracting information from classical T Tauri star (CTTS) light curves is to search for periodicities. This paper, by contrast, focuses on time domain fitting of standard linear stochastic time series models to such data. The necessary statistical time series analysis tools are covered in some detail. The theory is then applied to TESS 10-minute cadence photometry of 13 CTTSs. In most instances, completely satisfactory models could be fitted to the observations; in the remainder of cases models may still capture the gist of the variability. Models are quite parsimonious, requiring only one to six parameters. Statistical properties of the model residuals are also studied. The distributions of the residuals are generally highly non-Gaussian, with large positive kurtosis.