Résumé
High mortality of farmed oysters during summer leads to major economic losses. Vibrio aestuarianus is one of the pathogens associated with such recently observed severe mortality outbreaks in Pacific oysters, in France as well as in other parts of the globe. Our objective is to better understand, using a multiscale modelling framework, how this bacterium spreads in a spatially explicit metapopulation of oysters to evaluate its impact and, on a longer term, control measures to limit losses for farmers. We design a discrete time stochastic compartmental model, with farm level subpopulations connected through water neighboring relationships and anthropogenic movements of oysters between different farming sites. The model accounts for the diversity in inter-farm infection dynamics by coupling epidemiological models at intra-farm scale through a hydrodynamic particle transport model and a set of rules representing anthropogenic movements between sites based on farming practice. Farming sites are identified by land registers. The intra-farm host-pathogen dynamics is represented using three epidemiological compartments for oysters together with an infectious particle density compartment. Subpopulations exchange the pathogen particles shed by infected oysters at two scales: within a neighborhood through a local diffusion process, and on long distances using a hydrodynamic particle transport model. A mechanistic contact network is designed to simulate the anthropogenic movements of oysters between farming sites, as part of farm management practices, based on the information obtained from interviews. The model is applied on the oyster farming layout of the Baie des Veys area in Normandy, France, using observed oceano-environmental inputs for the hydrodynamic transport model, aiming to explore pathogen dynamics in a real case study. Scenarios of the initiation of pathogen outbreak are simulated to assess the respective roles of local and long distances dynamics. The next step will be to validate the model using data collected during a mortality event in summer 2014. We propose a generic modelling framework to represent pathogen spread in a metapopulation of oysters, coupling interactions at multiple scales. The specific part of the model concern the farming system practices in the area considered, which have to be adapted before using the model for different locations or types of farm management