Controlled experiments of habitat fragmentation: a simple computer simulation and a test using small mammals.

Habitat fragmentation involves a reduction in the effective area available to a population and the imposition of hard patch edges. Studies seeking to measure effects of habitat fragmentation have compared populations in fragments of different size to estimate and area effect but few have examined the effect of converting open populations to closed ones (an effect of edges). To do so requires a shift in spatial scope-from comparison of individual fragments to that of fragmented versus unfragmented landscapes. Here we note that large-scale, "controlled" studies of habitat fragmentation have rarely been performed and are needed. In making our case we develop a simple computer simulation model based on how individual animals with home ranges are affected by the imposition of habitat edges, and use it to predict population-level responses to habitat fragmentation. We then compare predictions of the model with results from a field experiment on Peromyscus and Microtus. Our model treats the case where home ranges/territories fall entirely within or partially overlap with that of sample areas in continuous landscapes, but are restricted to areas within habitat fragments in impacted landscapes. Results of the simulations demonstrate that the imposition of hard edges can produce different population abundances for similar-sized areas in continuous and fragmented landscapes. This edge effect is disproportionately greater in small than large fragments and for species with larger than smaller home ranges. These predictions were generally supported by our field experiment. We argue that large-scale studies of habitat fragmentation are sorely needed, and that control-experiment contrasts of fragmented and unfragmented microlandscapes provide a logical starting point.