Since the rise of oyster farming in the 17th century, French aquaculture has been recurrently impacted by emerging pathogens. After successive cultivation failures of native and introduced species, the Pacific oyster (Magallana gigas) has become the dominant species since the 1970s. However, the emergence in 2008 of a highly virulent variant of the Ostreid herpesvirus type 1 ( OsHV- 1 µVar) has led to recurrent mass mortality events in spats, with rates reaching 100% in some production sites. This still ongoing crisis threatens the long-term viability of the industry and underscores the urgent need to better understand the viral transmission dynamics to inform disease management strategies.

To this end, we developed a compartmental epidemiological model tailored to the transmission dynamics of OsHV- 1 µVar in M. gigas. Building on the classical SEIR framework, our SWEIRD model introduces a viral water compartment to capture environmental transmission and explicitly accounts for phenotypic host tolerance by distinguishing tolerant and non-tolerant individuals through separate epidemiological pathways. Model parameters were estimated from targeted experimental trials measuring incubation periods, infectious durations, and viral shedding rates in both tolerant and non-tolerant oysters, as well as virus decay in seawater. These parameters were implemented using stochastic processes to reflect inter- and intra-individual variability.

Simulations revealed distinct epidemic dynamics between tolerant and non-tolerant pathways and highlighted key temporal patterns of mortality and environmental contamination. Moreover, an Approximate Bayesian Computation (ABC) approach, based on a validation dataset, was used to refine parameter distributions and improve the model's fit to observed dynamics. In addition, a sensitivity analysis identified the most influential parameters driving infection peaks and transmission speed. Our integrative approach offers a robust framework to explore mollusc pathogens transmission dynamics and provides a basis for future spatial modeling and in situ data interpretation during upcoming field sampling.