ecological modelling 203 ( 2 0 0 7 ) 521–526 available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/ecolmodel Short communication Influence of dispersal, stochasticity, and an Allee effect on the persistence of weed biocontrol introductions Ian D. Jonsen a, * , Robert S. Bourchier b , Jens Roland a a Department of Biological Sciences, University of Alberta, Edmonton, AB, Canada T6G 2E9 b Agriculture & Agri-Food Canada, Research Centre, Lethbridge, AB, Canada T1J 4B1 article info Article history: Received 14 November 2005 Received in revised form 30 November 2006 Accepted 11 December 2006 Published on line 12 February 2007 Keywords: Emigration Immigration Invasion Landscape Simulation Weed biocontrol abstract An important problem encountered in biocontrol is the failure of introduced populations to establish and persist. Recent biocontrol studies focus on the roles of environmental stochas- ticity and Allee effects in determining introduction persistence but few studies consider the role of dispersal. We use a spatially explicit simulation model that incorporates dispersal and spatio-temporally random population growth to show that elevated emigration rates can exacerbate the negative influences of environmental stochasticity and Allee effects on introduction persistence. However, successful immigration can compensate partly for the otherwise reduced persistence that occurs when environmental stochasticity is high. These results illustrate that dispersal can have antagonistic effects on the persistence of biocon- trol introductions and failure to consider the dispersal ability and typical emigration rates of biocontrol agents may yield misleading predictions regarding successful establishment. © 2006 Elsevier B.V. All rights reserved. 1. Introduction The practice of biological weed control often is criticized for reliance on a trial-and-error approach and little use of eco- logical theory (McEvoy and Coombs, 2000). This approach has arisen in part because much of the potentially relevant theory is developed from insect biocontrol systems and applications for weed biocontrol are not always apparent or even possible. The field of weed biocontrol, therefore, needs more theoretical development in order for the practice to become more rigor- ous, especially with regard to designing effective and efficient strategies for introduction of biocontrol agents (Memmott et al., 1998; Grevstad, 1999a; Shea and Possingham, 2000). * Corresponding author at:Department of Biology, Dalhousie University, 1355 Oxford St., Halifax, Nova Scotia, Canada B3H 4J1. Tel.: +1 902 494 3910; fax: +1 902 494 3736. Previous studies (Grevstad, 1999b; Shea and Possingham, 2000) address the issue of optimal biocontrol agent intro- duction strategies under different regimes of demographic and environmental stochasticity and/or Allee effects. We are aware of few studies, however, that consider explicitly how weed biocontrol agent dispersal may affect population persistence (see Hopper and Rousch, 1993, for an insect biocontrol example). Dispersal may play two antagonistic roles during the early phase of population increase following an introduction. First, successful movement to unoccupied locations can establish new populations and spread the risk of overall extinction or failure of the release (den Boer, 1968). Second, emigration can act as a drain on local population 0304-3800/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ecolmodel.2006.12.009