1997A&A...318..687S

In a series of previous papers we have considered the reconstruction of the surface mass density of a cluster of galaxies from images of lensed faint background galaxies. We showed that the reconstructed surface mass density is not uniquely determined, but that there exists a global invariance transformation that leaves the shape of the images of the lensed galaxies unchanged. Because of this, only lower limits on the total mass of a cluster can be derived if no further information besides image ellipticities are used. Throughout these papers we used the simplifying assumption that all sources are at the same redshift. In this paper we account for a redshift distribution of the faint galaxies, and in particular, some of these galaxies can lie in front of the cluster or can be cluster members. We show how the mass distribution of a cluster of galaxies can be obtained from images of these faint galaxies, if the redshift distribution of these galaxies is known. We demonstrate that for the reconstruction of non-critical clusters we need less information on the redshift distribution of the galaxies, i.e., we only need to know two or three moments of the distribution. We show that the mean mass density across the data field is still a free variable, i.e., there remains a global invariance transformation of the resulting mass density field. For non-critical clusters we can derive the transformation explicitly and it is similar to that derived previously for the case of a single redshift of the sources. We discuss several theoretical ideas to break the mass degeneracy; of those considered, we find that only the magnification effect on the number density of galaxy images can be used successfully in practice.