Performance of AutoRegressive Planet Search methodology on simulated PLATO lightcurves

Abstract

AIM. This study examines the PLATO mission sensitivity to small transiting planets (radii between 0.5 and 2.0 R$_\oplus$) using Autoregressive Planet Search (ARPS) methodology based on simulated lightcurves. Methods. Several thousand simulations of PLATO-like lightcurves from magnetically active stars expected from the PLATO Long Observation Phase South 2 (LOPS2) field with injected transiting planets smaller than 2.0 R$_\oplus$ radius are examined. The lightcurves are analyzed with ARPS that uses autoregressive modeling for filtering of nonplanetary signals and the Transit Comb Filter periodogram for planetary transit detection. False alarms of nonexistent periodic signals are reduced by thresholds on two ARPS parameters, peak periodogram power, and transit duration. Results. When applied to PLATO's simulated first three months observations of bright stars (P1 and P2 samples), at least 50% of Earth-size planets are recovered for periods shorter than 40 d. In two years of LOPS2 observation, 25% are recovered for periods less than 240 d. Methodological implications on detrending, periodograms, False Alarms, vetting, and catalog evaluation are discussed. The utility of a positive predicted value heat map after detection and vetting is demonstrated. Conclusions: While the ARPS planet detection procedure is not entirely well-suited for PLATO analysis, this study provides methodological insight into the sensitivity and reliability of small transiting planet detection.