Eliminating Noise at the Box-fitting Spectrum

Abstract

Abstract. Non gaussian sources of erros need to be taken into consideration when searching for planetary transits. Such phenomena are mostly caused by the impact of high energetic particles on the detector (Pinheiro da Silva et al. 2008). The detection efficiency of transits, therefor, depend significantly on the data quality and the algorithms utilized to deal with these errors sources. In this work we show that a modified detrend algorithm CDA (CoRoT Detrend Algorithm; Mislis et al. 2010) using a robust statistics and an empirical fit, instead of a polynomial one, can eliminate more efficiently gaps in the data and other long-term trends from the light-curve. Using this algorithm enables us to obtain a reconstructed light-curve with better signal-to-noise ratio that allows to improve the detection of exoplanet transits, although long term signals are destroyed. The results show that these modifications lead to an improved BLS (Box-fitting Least Squares; Kovács, Zucker & Mazeh 2002) algorithm spectrum. At the end we have compared our planetary search results with CoRoT (Convection, Rotation and planetary Transits) satellite chromatic light-curves available in the literature.