Maximum-entropy weak lens reconstruction: Improved methods and application to data

P. J. Marshall, M. P. Hobson, S. F. Gull, S. L. Bridle

    Research output: Contribution to journalArticlepeer-review

    Abstract

    We develop the maximum-entropy weak shear mass reconstruction method presented in earlier papers by taking each background galaxy image shape as an independent estimator of the reduced shear field and incorporating an intrinsic smoothness into the reconstruction. The characteristic length-scale of this smoothing is determined by Bayesian methods. Within this algorithm the uncertainties owing to both the intrinsic distribution of galaxy shapes and galaxy shape estimation are carried through to the final mass reconstruction, and the mass within arbitrarily shaped apertures can be calculated with corresponding uncertainties. We apply this method to two clusters taken from n-body simulations using mock observations corresponding to Keck LRIS and mosaicked Hubble Space Telescope (HST) WFPC2 fields. We demonstrate that the Bayesian choice of smoothing length is sensible and that masses within apertures (including one on a filamentary structure) are reliable, provided the field of view is not too small. We apply the method to data taken on the cluster MS 1054-03 using the Keck LRIS and HST, finding results in agreement with previous work; we also present reconstructions with optimal smoothing lengths, and mass estimates that do not rely on any assumptions of circular symmetry. The code used in this work (LENSENT2) is available from the web.
    Original languageEnglish
    Pages (from-to)1037-1048
    Number of pages11
    JournalMonthly Notices of the Royal Astronomical Society
    Volume335
    Issue number4
    DOIs
    Publication statusPublished - 1 Oct 2002

    Keywords

    • Cosmology: theory
    • Dark matter
    • Galaxies: clusters: general
    • Gravitational lensing
    • Methods: data analysis

    Fingerprint

    Dive into the research topics of 'Maximum-entropy weak lens reconstruction: Improved methods and application to data'. Together they form a unique fingerprint.

    Cite this