Environmental Pseudointeraction: The Effects of Ignoring the Scale of Environmental Heterogeneity in Competition Studies

Spatially explicit models of competition such as neighborhood and area-of-influence models are useful approaches for gaining insight into interactions between individual organisms. The environments in which these models are parameterized may be spatially heterogeneous. We demonstrate that, even in the absence of interaction between individuals, nonzero estimates of competitive or facilitative coefficients can result from environmental heterogeneity when densities are not experimentally imposed. We show this using a model of individual position and size in a simple environment composed of two patch types. Suppose two noninteracting species are intermixed in the study area and species 1 becomes established more readily in the patch type in which species 2 grows larger and has higher fecundity. Analyses that ignore environmental heterogeneity may indicate that species 1 facilitates species 2 because larger individuals of species 2 will tend to be associated with greater numbers of species 1. Similarly, environmental pseudocompetition can result if species 2 is more frequent among small individuals of species 1. The strength of the environmental pseudointeraction depends on (a) the magnitude of the size and fecundity differences between patch types, (b) the magnitude of the establishment differences between patch types, and (c) the size, relative to the environmental grain, of the neighborhood or quadrat used for sampling the number of competitor individuals.