Viruses, the most abundant microorganisms in the ocean, play important roles in marine ecosystems, mainly by killing their hosts and contributing to nutrient recycling. However, in models simulating ecosystems in real marine environments, the virus-mediated mortality (VMM) rates of their hosts are implicitly represented by constant parameters, thus ignoring the dynamics caused by interactions between viruses and hosts. Here, we construct a model explicitly representing marine viruses and the VMM rates of major hosts, heterotrophic bacteria, and apply it to two sites in the oligotrophic North Pacific and the more productive Arabian Sea. The impacts of the viral processes were assessed by comparing model results with the viral processes enabled and disabled. For reliable assessments, a data assimilation method was used to objectively optimize the model parameters in each run. The model generated spatiotemporally variable VMM rates, generally decreasing in the subsurface but increasing at the surface. Although the dynamics introduced by viruses could be partly stabilized by the ecosystems, they still caused substantial changes to the bacterial abundance, primary production and carbon export, with the changes greater at the more productive site. Our modeling experiments reveal the importance of explicitly simulating dynamic viral processes in marine ecological models.
Keywords: data assimilation; heterotrophic bacteria; marine ecological model; marine virus.