CRITICAL THRESHOLDS IN SPECIES RESPONSES TO LANDSCAPE STRUCTURE

Critical thresholds are transition ranges across which small changes in spatial pattern produce abrupt shifts in ecological responses. Habitat fragmentation provides a familiar example of a critical threshold. As the landscape becomes dissected into smaller parcels of habitat, landscape connectivity-the functional linkage among habitat patches-may suddenly become disrupted, which may have important consequences for the distribution and persistence of populations. Landscape connectivity depends not only on the abundance and spatial patterning of habitat, but also on the habitat specificity and dispersal abilities of species. Habitat specialists with limited dispersal capabilities presumably have a much lower threshold to habitat fragmentation than highly vagile species, which may perceive the landscape as functionally connected across a greater range of fragmentation severity. To determine where threshold effects in species' responses to landscape structure are likely to occur, we developed a simulation model modified from percolation theory. Our simulations predicted the distributional patterns of populations in different landscape mosaics? which we tested empirically using two grasshopper species (Orthoptera: Acrididae) that occur in the shortgrass prairie of north-central Colorado. Increasing degree of habitat specialization and dispersal range of a species enhanced the level of aggregation-the degree of clumping exhibited by the population-in our simulations. The landscape threshold at which populations became aggregated was affected by dispersal range for habitat generalists, but not for habitat specialists. Habitat specialists exhibited aggregated populations when preferred habitat occupied <40% of the landscape. Habitat generalists with good dispersal abilities occurred as aggregated populations when <35% of the landscape contained suitable habitat; habitat generalists with limited dispersal only formed patchy distributions when the preferred habitat was a minor (20%) proportion of the habitat. In field tests, a large species of grasshopper (Xanthippus corallipes Haldeman) exhibited reduced rates of travel in two microhabitats that together comprised 35% of a grassland mosaic; a smaller species (Psoloessa delicatula Scudder) had the highest residence time in a rare microhabitat that occurred in only 8% of the landscape. On the basis of our simulation results, we predicted that the large species would be patchily distributed because the abundance of its associated habitat is at the critical threshold. The small species should be unable to aggregate, given the rarity of habitat for which it has an affinity, and thus should be randomly distributed across the landscape. The distribution of these two species in this grassland mosaic matched the predictions from our simulations. By providing quantitative predictions of threshold effects, this modelling approach may prove useful in the formulation of conservation strategies and assessment of land-use changes on species' distributional patterns and persistence.