Spatial models of foraging in clonal plant species

Abstract

Ramets of some clonal plant species alter their internode lengths or their frequency of lateral branching in response to their immediate microenvironment. Such “plant foraging” responses are thought to allow clones to concentrate in favorable portions of their environment. Despite widespread interest among ecologists in plant foraging, few realistic models have been developed to examine conditions under which plant foraging responses are likely to provide clones with ecological benefit. In this paper, we develop spatially explicit, stochastic simulation models to examine consequences of both empirical and hypothetical plant foraging responses. We construct a hierarchical series of models in which we incorporate effects of resource heterogeneity on spacer lengths, angles of growth, and lateral branch production. We also vary the number, size, and arrangement of patches, and the presence or absence of ramet mortality. Simulations based on hypothetical data demonstrated the potential importance of shortening spacer lengths in favorable habitat. In these simulations, ramet crowding increased significantly, implying a potential cost to plant foraging responses whose magnitude is large enough to cause ramets to concentrate in favorable patches. Models calibrated with empirical data suggest that when clonal plants were able to concentrate in favorable habitat, this was usually caused by increased daughter ramet production in the favorable habitat. Variation in clonal growth angles had little impact on the ability of ramets or clones to locate favorable patches, but did increase the ability of clones to remain in favorable patches once found. Alterations in the number and size of patches strongly influenced the effectiveness of the foraging response. The spatial arrangement of patches also was important: clumped distributions of patches decreased the success with which plants located favorable patches, especially at the genet level and when the number of patches was low. Finally, when ramet mortality varied with patch quality, there was an increase in the percentage of ramets located in favorable patches; differential ramet mortality also lessened the impact of other effects, such as the decreased success of clones when patches are clumped. Overall, our models indicate that the effectiveness of plant foraging responses is variable and is likely to depend on a suite of environmental conditions.