Print PageThe goal of conservation planning is to design networks of reserves that will ensure the long-term persistence of biodiversity in landscapes undergoing habitat conversion. Common quantitative approaches to conservation planning use complementarity-based algorithms to represent as many species as possible in a reserve network. These approaches, however, often select widely dispersed sets of reserves that do not effectively support long-term biodiversity persistence. Dynamic population models provide an alternative framework that can directly optimize reserve networks for biodiversity persistence, but these models traditionally require extensive empirical research on each species in a landscape.
We introduce a new, general approach to reserve network design based on the concept of an “allometric community", a community in which a species' life history parameters are determined solely by its body size. We use empirical relationships from the literature to estimate population density, dispersal distance, and longevity for generic mammal species of various body sizes. We combine these data with two additional parameters controlling the probability of local population extinction to create stochastic, demographic population models for each species. Finally, we use a stepwise heuristic to select a set of reserves that maximizes the mean time to extinction for each species or the entire community.
Results/Conclusions
We find that for species with very small and very large body sizes, the optimal reserve network consists of the largest individual sites available for protection, regardless of their locations in the landscape. For intermediate body sizes, however, optimal reserve networks are significantly more clustered, suggesting that connectivity most benefits those species with moderately high risks of local extinction and moderately large dispersal distances. We also find that, across all body sizes, optimal reserve network designs become more clustered as the probability of local extinction within each site in the network increases. Finally, we find that “core-satellite” network designs perform well for a variety of species and across many levels of local extinction risk. This design could be recommended for landscapes in which very little is known about the resident species or levels of environmental stochasticity.
Beyond identifying these general principles, our approach also provides a theoretically-grounded, quantitative method for prioritizing conservation investment to prevent biodiversity loss in the face of habitat loss and climate change. We demonstrate this application of our approach by suggesting specific designs for a reserve network in Sonoma County, California.
