Bulletin of Mathematical Biology

, Volume 64, Issue 5, pp 913–958

Extinction dynamics in mainland-island metapopulations: An N-patch stochastic model

  • David Alonso
  • Alan McKane

DOI: 10.1006/bulm.2002.0307

Cite this article as:
Alonso, D. & McKane, A. Bull. Math. Biol. (2002) 64: 913. doi:10.1006/bulm.2002.0307


A generalization of the well-known Levins’ model of metapopulations is studied. The generalization consists of (i) the introduction of immigration from a mainland, and (ii) assuming the dynamics is stochastic, rather than deterministic. A master equation, for the probability that n of the patches are occupied, is derived and the stationary probability Ps(n), together with the mean and higher moments in the stationary state, determined. The time-dependence of the probability distribution is also studied: through a Gaussian approximation for general n when the boundary at n = 0 has little effect, and by calculating P(0, t), the probability that no patches are occupied at time t, by using a linearization procedure. These analytic calculations are supplemented by carrying out numerical solutions of the master equation and simulations of the stochastic process. The various approaches are in very good agreement with each other. This allows us to use the forms for Ps0) and P(0, t) in the linearization approximation as a basis for calculating the mean time for a metapopulation to become extinct. We give an analytical expression for the mean time to extinction derived within a mean field approach. We devise a simple method to apply our mean field approach even to complex patch networks in realistic model metapopulations. After studying two spatially extended versions of this nonspatial metapopulation model—a lattice metapopulation model and a spatially realistic model—we conclude that our analytical formula for the mean extinction time is generally applicable to those metapopulations which are really endangered, where extinction dynamics dominates over local colonization processes. The time evolution and, in particular, the scope of our analytical results, are studied by comparing these different models with the analytical approach for various values of the parameters: the rates of immigration from the mainland, the rates of colonization and extinction, and the number of patches making up the metapopulation.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Society for Mathematical Biology 2002

Authors and Affiliations

  • David Alonso
    • 1
    • 2
  • Alan McKane
    • 3
  1. 1.ICREA Complex Systems LaboratoryUniversitat Pompeu Fabra (GRIB)BarcelonaSpain
  2. 2.Department of Ecology, Facultat de BiologiaUniversitat de BarcelonaBarcelonaSpain
  3. 3.Department of Theoretical PhysicsUniversity of ManchesterManchesterUK

Personalised recommendations