Evaluating the time-lagged extinction risk of a migratory species
Surya, Gautam S.
- Species extinction is currently reducing the world's biodiversity at an alarming rate. The species-area relationship predicts a decline in the number of species in a landscape as the area of available habitat is reduced. This time-lagged extinction predicted for the future is known as the extinction debt of the landscape. The incidence function model (IFM) is a single-species approach to the multi... read morespecies extinction debt model. However, how one should apply it particularly to migrant species is unclear. To date, applications of the IFM have treated migrants as residents, in that they are assumed to return to their natal sites and then disperse. We here apply the IFM to the bobolink (Dolichonyx oryzivorus), a grassland obligate breeding bird that migrates every year to South America. Our modeled study system is Massachusetts grasslands, which exist as a highly fragmented landscape. Under the IFM, the probability of a patch of habitat being occupied is a function of the quality of the habitat, and of the connectivity of the patch within the landscape. Connectivity, in turn, is a function of the dispersal capacity of the species, as well as the spatial arrangement of habitat patches within the landscape. We used the bobolink to explore a variety of scenarios for determining extinction debt for migrants. We investigated the sensitivity of the IFM to five measures of habitat quality and four dispersal scenarios. Presence-absence data were obtained from the Massachusetts Audubon Society (MAS) Grassland Bird Database for 59 sites, for the 2005 survey year. One of the dispersal scenarios mimicked bobolink dispersal, which was estimated from published mark-recapture studies, and corrected to account for unmeasured long-distance dispersal (LDD) events. Other dispersal distances evaluated included: no dispersal, unlimited dispersal, and a published dispersal distance. Quality measures evaluated included area of the grassland patch, available habitat within 1, 3, and 5 km from each patch, and the distance to the nearest patch. We found that varying estimates of dispersal capacity, while keeping the quality measure constant, had a large effect on predicted bobolink persistence as measured by the percentage of habitat patches occupied. Our measures of quality, in contrast, had a small effect on estimated persistence for the estimated dispersal distance obtained from mark-recapture studies without accounting for LDD events. We conclude that obtaining accurate data regarding dispersal, along with data regarding the manner in which the species of interest evaluates its environment, is critical to predicting the persistence of a population on a landscape. In addition, we believe that these results constitute evidence that the IFM should be applied in a species-specific manner, and the results from the IFM for one species cannot be generalized to include all species in a landscape.read less