The noise of cluster mass reconstructions from a source redshift distribution
Institüt für Astrophysik und Extraterrestrische Forschung, Auf dem Hügel 71, 53121 Bonn, Germany
Corresponding author: M. Lombardi, email@example.com
Accepted: 13 November 2001
The parameter-free reconstruction of the surface-mass density of clusters of galaxies is one of the principal applications of weak gravitational lensing. From the observable ellipticities of images of background galaxies, the tidal gravitational field (shear) of the mass distribution is estimated, and the corresponding surface mass density is constructed. The noise of the resulting mass map is investigated here, generalizing previous work which included mainly the noise due to the intrinsic galaxy ellipticities. Whereas this dominates the noise budget if the lens is very weak, other sources of noise become important, or even dominant, for the medium-strong lensing regime close to the center of clusters. In particular, shot noise due to a Poisson distribution of galaxy images, and increased shot noise owing to the correlation of galaxies in angular position and redshift, can yield significantly larger levels of noise than that from the intrinsic ellipticities only. We estimate the contributions from these various effects for two widely used smoothing operations, showing that one of them effectively removes the Poisson and the correlation noises related to angular positions of galaxies. Noise sources due to the spread in redshift of galaxies are still present in the optimized estimator and are shown to be relevant in many cases. We show how (even approximate) redshift information can be profitably used to reduce the noise in the mass map. The dependence of the various noise terms on the relevant parameters (lens redshift, strength, smoothing length, redshift distribution of background galaxies) are explicitly calculated and simple estimates are provided.
Key words: cosmology: dark matter / gravitational lensing / large-scale structure of universe / galaxies: clusters: general / methods: statistical
© ESO, 2002