Modeling the Secondary Spread of Viral Hemorrhagic Septicemia Virus (VHSV) Via Commercial Shipping in the Laurentian Great Lakes

We used dynamic spatial models to assess if commercial ship ballast water discharge is a vector of secondary spread of viral hemorrhagic septicemia virus (VHSV) in the Great Lakes. We contrasted the results of a null model with a random shipping model to determine if ballast water discharge locations were related to VHSV occurrences. Additionally, a weighted shipping model was constructed to determine if discharge locations were more likely to be infested if they received more visits from ships discharging ballast water from VHSV infested locations. The weighted model was also modified to test which of three different locations was the most likely initial introduction location in the Great Lakes. All models included a local spread component, ranging from 10 to 30-km per year to represent the spread of the disease via natural vectors. To quantify how our models performed compared to the reported pattern of VHSV observations, we used sensitivity, specificity, and a weighted kappa statistic. Our results suggest that ballast water discharge has played a role in the spread of VHSV, and that Montreal, as opposed to the initially observed location of Lake St. Clair, was the initial infestation location.