Inferring landscape effects on dispersal from genetic distances: how far can we go?

Authors

  • J. JAQUIÉRY,

    1. Department of Ecology and Evolution, University of Lausanne, CH-1015 Lausanne, Switzerland
    2. Institut National de Recherche Agronomique, UMR 1099 Biology of Organisms and Populations applied to Plant Protection, Domaine de la Motte, 35653 Le Rheu Cedex, France
    Search for more papers by this author
  • T. BROQUET,

    1. Department of Ecology and Evolution, University of Lausanne, CH-1015 Lausanne, Switzerland
    2. Team Diversity and Connectivity in Coastal Marine Landscapes, Roscoff Biological Station, UMR 7144 CNRS, Pierre and Marie Curie University, 29682 Roscoff, France
    Search for more papers by this author
  • A. H. HIRZEL,

    1. Department of Ecology and Evolution, University of Lausanne, CH-1015 Lausanne, Switzerland
    Search for more papers by this author
  • J. YEARSLEY,

    1. Department of Ecology and Evolution, University of Lausanne, CH-1015 Lausanne, Switzerland
    2. School of Biology & Environmental Science, University College Dublin, Belfield Dublin 4, Ireland
    Search for more papers by this author
  • N. PERRIN

    1. Department of Ecology and Evolution, University of Lausanne, CH-1015 Lausanne, Switzerland
    Search for more papers by this author

Julie Jaquiéry, Fax : +33 2 23 48 51 50;
E-mail: Julie.Jaquiery@gmail.com

Abstract

Functional connectivity affects demography and gene dynamics in fragmented populations. Besides species-specific dispersal ability, the connectivity between local populations is affected by the landscape elements encountered during dispersal. Documenting these effects is thus a central issue for the conservation and management of fragmented populations. In this study, we compare the power and accuracy of three methods (partial correlations, regressions and Approximate Bayesian Computations) that use genetic distances to infer the effect of landscape upon dispersal. We use stochastic individual-based simulations of fragmented populations surrounded by landscape elements that differ in their permeability to dispersal. The power and accuracy of all three methods are good when there is a strong contrast between the permeability of different landscape elements. The power and accuracy can be further improved by restricting analyses to adjacent pairs of populations. Landscape elements that strongly impede dispersal are the easiest to identify. However, power and accuracy decrease drastically when landscape complexity increases and the contrast between the permeability of landscape elements decreases. We provide guidelines for future studies and underline the needs to evaluate or develop approaches that are more powerful.

Ancillary