Fishing inside or outside? A case studies analysis of potential spillover effect from marine protected areas, using food web models

International audienceMarine protected areas (MPAs) are implemented worldwide as an efficient tool to preserve biodiversity and protect ecosystems. We used food web models (Ecopath and EcoTroph) to assess the ability of MPAs to reduce fishing impacts on targeted resources and to provide biomass exports for adjacent fisheries. Three coastal MPAs: Bonifacio and Port-Cros (Mediterranean Sea), and Bamboung (Senegalese coast), were used as case studies. Pre-existing related Ecopath models were homogenized and ecosystem characteristics were compared based on network indices and trophic spectra analyses. Using the EcoTroph model, we simulated different fishing mortality scenarios and assessed fishing impacts on the three ecosystems. Lastly, the potential biomass that could be exported from each MPA was estimated. Despite structural and functional trophic differences, the three MPAs showed similar patterns of resistance to simulated fishing mortalities, with the Bonifacio case study exhibiting the highest potential catches and a slightly inferior resistance to fishing. We also show that the potential exports from our small size MPAs are limited and thus may only benefit local fishing activities. Based on simulations, their potential exports were estimated to be at the same order of magnitude as the amount of catch that could have been obtained inside the reserve. In Port Cros, the ban of fishing inside MPA could actually allow for improved catch yields outside the MPA due to biomass exports. This was not the case for the Bonifacio site, as its potential exports were too low to offset catch losses. This insight suggests the need for MPA networks and/or sufficiently large MPAs to effectively protect juveniles and adults and provide important exports. Finally, we discuss the effects of MPAs on fisheries that were not considered in food web models, and conclude by suggesting possible improvements in the analysis of MPA efficienc