Optimal selection of marine protected areas based on connectivity and habitat quality

Networks of nature reserves and protected areas are important instruments to protect biodiversity, including harvested populations. Selection of marine protected networks (MPA) will depend on both the connectivity of concerned species and the habitat quality of individual sites. We explore the relative effect of connectivity and habitat quality on solutions for optimal networks of MPA using eigenvalue perturbation theory and a metapopulation model. Based on analyses of both synthetic networks and realistic connectivities for a sessile invertebrate with planktonic larvae in the Baltic Sea, we show that connectivity is expected to be more efficient than habitat quality as a selection criterion for MPA networks with realistic probabilities of local recruitment. In a second series of analyses we explore the effect of temporal variability of connectivity on the selection of optimal MPA networks. We show that optimal solutions of MPA networks converged when based on 8–10 years of connectivity information, corresponding to the time scale of the North-Atlantic oscillation. In conclusion, this study indicates that connectivity may be more important than habitat quality as selection criterion for MPAs when targeting species with long-distance dispersal that is typical for many marine invertebrates and fish. Our study also shows that connectivity patterns may be relatively consistent in time which suggests that the recent progress in biophysical modelling can offer a framework for optimal selection of MPA networks based on connectivities, which should improve guidelines for the design of functional MPA networks.

► We demonstrate the importance of connectivity for prioritizing marine protected areas.
► Connectivity is the dominant factor under a wide range of situations.
► Only at high levels of self-recruitment is, e.g. habitat quality a more important factor.
► Temporal variations in connectivity have limited effect on the protection network.