Comparing methods for measuring the rate of spread of invading populations

Authors

  • Marius Gilbert,

  • Andrew Liebhold


M. Gilbert (mgilbert@ulb.ac.be), Biological Control and Spatial Ecology, Univ. Libre de Bruxelles CP160/12, Av FD Roosevelt 50, BE-1050 Brussels, Belgium, and Fonds National de la Recherche Scientifique, rue d'Egmont 5, BE-1000 Brussels, Belgium. – A. Liebhold, Northern Research Station, USDA Forest Service, Canfield St. 180, Morgantown, WV 26505, USA.

Abstract

Measuring rates of spread during biological invasions is important for predicting where and when invading organisms will spread in the future as well as for quantifying the influence of environmental conditions on invasion speed. While several methods have been proposed in the literature to measure spread rates, a comprehensive comparison of their accuracy when applied to empirical data would be problematic because true rates of spread are never known. This study compares the performances of several spread rate measurement methods using a set of simulated invasions with known theoretical spread rates over a hypothetical region where a set of sampling points are distributed. We vary the density and distribution (aggregative, random, and regular) of the sampling points as well as the shape of the invaded area and then compare how different spread rate measurement methods accommodate these varying conditions. We find that the method of regressing distance to the point of origin of the invasion as a function of time of first detection provides the most reliable method over adverse conditions (low sampling density, aggregated distribution of sampling points, irregular invaded area). The boundary displacement method appears to be a useful complementary method when sampling density is sufficiently high, as it provides an instantaneous measure of spread rate, and does not require long time series of data.

Ancillary