MPA network design based on graph network theory and emergent properties of larval dispersal
Despite the recognised effectiveness of networks of Marine Protected Areas (MPAs) as a biodiversity conservation instrument, nowadays MPA network design frequently disregards the importance of connectivity patterns. In the case of sedentary marine populations, connectivity stems not only from the stochastic nature of the physical environment that affects early-life stages dispersal, but also from the spawning stock attributes that affect the reproductive output (e.g., passive eggs and larvae) and its survivorship. Early-life stages are virtually impossible to track in the ocean. Therefore, numerical ocean current simulations coupled to egg and larval Lagrangian transport models remain the most common approach for the assessment of marine larval connectivity. Inferred larval connectivity may be different depending on the type of connectivity considered; consequently, the prioritisation of sites for marine populations' conservation might also differ. Here, we introduce a framework for evaluating and designing MPA networks based on the identification of connectivity hotspots using graph theoretic analysis. We use as a case of study a network of open-access areas and MPAs, off Mallorca Island (Spain), and test its effectiveness for the protection of the painted comber Serranus scriba. Outputs from network analysis are used to: (1) identify critical areas for improving overall larval connectivity; (2) assess the impact of species' biological parameters in network connectivity; and (3) explore alternative MPA configurations to improve average network connectivity. Results demonstrate the potential of graph theory to identify non-trivial egg/larval dispersal patterns and emerging collective properties of the MPA network which are relevant for increasing protection efficiency.