A Two-Dimensional Sky Background Model for LAMOST Based on Improved KICA

To accomplish the sky-subtraction task in a key reduction step upon the large sky area multi-object fiber spectroscopic telescope (LAMOST) data, the sky background is modeled by the LAMOST 2D pipeline using the one-dimensional (1D) extracted spectra of sky fibers. The spectrum of 'super sky' that represents the local background, ignoring the variation in this field, is obtained in each square degree by co-adding approximately 20 sky fiber spectra with a spline function, which could contribute to the difference in the emission line ratios in the spectra of celestial targets, especially at low galactic latitudes. In this paper, we model the sky by using the two-dimensional (2D) CCD images of sky fibers, and the local sky is dynamically simulated by considering the positions of the sky fibers relative to any object. To accelerate the subtraction process in 2D images, we exploit the improved KICA to separate the sky from object spectra. Given the dynamic 2D sky background and the improved KICA, this would be another option for sky-subtraction in CCD images.