Accurately predicting the establishment and spread of invasive populations can greatly improve the efficacy of screening programs and allow for the optimal allocation of scarce control resources (Higgins and Richardson 1996, Higgins et al. 2000). Habitat modification by ecosystem engineers complicates predictions about spread, however. Climate matching and ecological niche models are important tools for evaluating the potential of species to establish and spread in a new range (Krticos et al. 2003, Peterson 2003). These models assume that environmental conditions set the constraints on species distributions, and that these vary independently from population dynamics. The hallmark of ecosystem engineers, however, is that their population dynamics change the abiotic environment. The engineering process can itself feed back on population dynamics or patterns of spread. One way this can happen is if engineering ameliorates conditions in unfavorable habitat patches. In a spatially implicit model Cuddington and Hastings (2004) showed that habitat-modifying invaders can have significantly faster population growth rates and ultimately higher population density in suboptimal habitats than invaders that do not modify their environment. Although no modeling studies have demonstrated this, it seems plausible that engineers could also slow their own expansion rates if they created unfavorable conditions. More generally, ecosystem engineering can alter habitat heterogeneity across a range of spatial and temporal scales, and this can greatly influence invasive spread (Hastings et al. 2007, Melbourne et al. 2007). Feedbacks such as this could lower the predictive ability of static climate matching or niche models parameterized with data from the native range. However, no studies have yet evaluated whether these models perform more poorly when predicting invasion patterns for engineers compared to non-engineers.
To be useful for specific management questions, predictive models that incorporate engineering feedbacks will need to be spatially explicit. This is because although engineering processes are often easily generalized, the consequences of engineering for recruitment or spread dynamics can be highly contingent on spatially varying traits such as environmental gradients (see Crain and Bertness 2006) or patterns of propagule supply. The S. alterniflora invasion in Willapa Bay provides a good example. In its native range, S. alterniflora stands reduce flow-related physical stress, facilitating the establishment of a suite of plant species in the cobble beach habitat immediately behind stands (Bruno and Kennedy 2000). In the invasion context of Willapa Bay, however, the majority of conspecific recruitment occurs on the open tideflat in front of established beds where there is little or no engineering influence. This recruitment pattern partly reflects the fact that seedling recruitment is strongly inhibited within and immediately adjacent to high-density stands by intense light competition in these areas (Lambrinos and Bando in press).
In low-density situations, where isolated individuals have established by long-distance dispersal, the reduced intraspecific competition could allow facilitation to operate. Here the leeward side of clones could provide a favorable recruitment environment similar to that seen in New England cobble beaches. Isolated Spartina individuals in Willapa Bay, however, are strongly pollen limited (Davis et al. 2004a, 2004b). This pollen limitation is overcome only as clones grow vegetatively and reach an adult density that also exerts a strong competitive inhibition on seedling survival. Moreover, although seeds can potentially disperse long distances on the tide, most seeds are retained within established beds (Figure 16.1). The same physical mechanisms that reduce hydrological flow and increase sedimentation within beds likely also increase seed deposition and retention. As a consequence, established Spartina beds may actually be slowing expansion rates in Willapa Bay by trapping potential propa-gules in an unfavorable recruitment environment.
Interestingly, in San Francisco Bay some hybrid genotypes exhibit high rates of self-fertilization. In this case, isolated individuals of these genotypes can produce abundant local seed shadows as well as localized habitat modification that could facilitate seedling establishment. Indeed, the lee sides of established clones appear to provide favorable microsites for recruitment in San Francisco Bay (Sloop and Ayres unpublished data). This suggests that species-specific life history traits can interact with the spatial patterning of engineering to drive spread patterns.
Was this article helpful?