The Munich Near-Infrared Cluster Survey - IV. Biases in the completeness of near-infrared imaging data

We present the results of completeness simulations for the detection of point sources as well as redshifted elliptical and spiral galaxies in the K-band images of the Munich Near-Infrared Cluster Survey (MUNICS). The main focus of this work is to quantify the selection effects introduced by threshold-based object detection algorithms used in deep imaging surveys. Therefore, we simulate objects obeying the well-known scaling relations between effective radius and central surface brightness, for both de Vaucouleurs and exponential profiles. The results of these simulations, while presented for the MUNICS project, are applicable in a much wider context to deep optical and near-infrared selected samples. We investigate the detection probability as well as the reliability for recovering the true total magnitude with Kron-like (adaptive) aperture photometry. The results are compared with the predictions of the visibility theory of Disney and Phillipps in terms of the detection rate and the lost-light fraction. Additionally, the effects attributable to seeing are explored. The results show a bias against detecting high-redshifted massive elliptical galaxies in comparison to disc galaxies with exponential profiles, and that the measurements of the total magnitudes for intrinsically bright elliptical galaxies are systematically too faint. Disc galaxies, in contrast, show no significant offset in the magnitude measurement of luminous objects. Finally, we present an analytic formula to predict the completeness of point sources using only basic image parameters.