Ecological coherence of Marine Protected Areas: New tools applied to the Baltic Sea network

Spatial connectivity is an essential process to consider in the design and assessment of Marine Protected Areas (MPAs). To help maintain and restore marine populations and communities MPAs should form ecologically coherent networks. How to estimate and implement connectivity in MPA design remains a challenge. Here a new theoretical framework is presented based on biophysical modelling of organism dispersal, combined with a suite of tools to assess different aspects of connectivity that can be integrated in MPA design. As a demonstration, these tools are applied to an MPA network in the Baltic Sea (HELCOM MPA). The tools are based on the connectivity matrix, which summarizes dispersal probabilities, averaged over many years, between all considered areas in the geographic target area. The biophysical model used to estimate connectivity included important biological traits that affect dispersal patterns where different trait combinations and habitat preferences will produce specific connectivity matrices representing different species. Modelled connectivity matrices were used to assess local retention within individual MPAs, which offers indications about the adequacy of size when MPAs are considered in isolation. The connectivity matrix also provides information about source areas to individual MPAs, e.g. sources of larvae or pressures such as contaminants. How well several MPAs act as a network was assessed within a framework of eigenvalue perturbation theory (EPT). With EPT, the optimal MPA network with respect to connectivity can be identified. In addition, EPT can suggest optimal extensions of existing MPA networks to enhance connectivity. Finally, dispersal barriers can be identified based on the connectivity matrix, which may suggest boundaries for management units. The assessment of connectivity for the HELCOM MPA are discussed in terms of possible improvements, but the tools presented here could be applied to any region.