Implementation of an end-to-end model of the Gulf of Lions ecosystem (NW Mediterranean Sea). I. Parameterization, calibration and evaluation

Abstract An end-to-end model named OSMOSE-GoL has been built for the Gulf of Lions, the main French Mediterranean fishing area. This spatialized dynamic model links the coupled hydrodynamic and biogeochemical model Eco3M-S/ SYMPHONIE(LTL – low trophic levelmodel) to OSMOSE (HTL – high trophic levelmodel). It includes 15 compartments of living organisms, five from the LTL model ( i.e . nanophytoplankton, microphytoplankton, nanozooplankton, microzooplankton and mesozooplankton) and ten from the HTL model ( northern krill, southern shortfin squid, European pilchard, European anchovy, European sprat, Atlantic horse mackerel, Atlantic mackerel, blue whiting, European hakeand Atlantic bluefin tuna). With the exception of northern krilland European sprat, all HTL species are commercially exploited and undergo fisheries mortality pressure. The modeled species represent more than 70% of annual catches in this area. This paper presents the parameterization, calibration and evaluation of this model with satellite data for phytoplankton and with biomass, landings, diet and trophic leveldata for HTL groups. For most species, the diets in output of OSMOSE-GoL are similar to field and literature data in terms of dominant prey groups and species. However, some differences were observed. Various reasons may explain the mismatch between the modeled diet and field data. Benthic prey sometimes observed in the stomach content of the HTL predators were not modeled in OSMOSE-GoL. Field studies were carried out at specific periods and locations, while our data concern the period 2001–2004 and the entire modeled domain. Inter- and intra-annual variations in spatial distribution and density of prey may also explain these differences. The model estimates trophic levelvalues similar to those cited in the literature for all the HTL compartments. These values are also close to the trophic levelsestimated by a previous Ecopath model for the same area and period. Even though some improvements are still possible, this model may already be of use to explore fishery or Marine Protected Areasscenarios for socio- ecosystem managementissues.