Spatial Modeling as a Decision-making Tool for Invasive Species Management in the Great Lakes

Due to recent recognition that ballast water is playing an important role in the spread of invasive species within the Great Lakes, there has been increasing interest in implementing management strategies that include a secondary spread component for ballast discharge. Using ballast water data for ships visiting U.S. ports in the Great Lakes, I created a dynamic spatial model to simulate the spread of invasive species based on recent shipping patterns. My goal in producing this model was to provide information to natural resource managers, scientists, and policy-makers to help effectively regulate invasive species issues. In testing the model, I determined that including the number of discharging ship visits that a location receives from previously infested areas and the ability of an organism to survive in the ballast tank were important in more accurately identifying the past spread of the fish virus, viral hemorrhagic septicemia virus (VHSV), zebra mussel (Dreissena polymorpha), and Eurasian Ruffe (Gymnocephalus cernuus), than discharge location alone. I also included and tested a localized spread distance that simulated the dispersal of an invasive species upon being discharged at a location. I first applied the model to identify if ballast water played a role in the secondary spread of VHSV. Results indicated that ballast water movement has contributed to the spread of VHSV in the Great Lakes, albeit it is not the only vector of secondary spread. However, ballast water management would be an important part of any plan in preventing the future spread of VHSV in an ecosystem. Next, I applied the model to predict the future spread of Eurasian Ruffe, which already occurs in the Great Lakes, and two species that do not, golden mussel (Limnoperna fortune) and killer shrimp (Dikerogammerus villosus). The results of the prediction models are intended to be used to help direct early detection monitoring efforts. The Eurasian Ruffe results are currently being used by The Nature Conservancy in their eDNA monitoring efforts, and have led to the positive detection of ruffe eDNA in a location where ruffe has previously not been detected. Finally, I applied the model to identify potentially "safe" ballast water exchange (BWE) sites in Lake Michigan. The purpose of this exercise was to locate mid-lake sites where ships could exchange and flush their ballast tanks, so as to reduce the probability that species are able to survive and establish new populations in the Great Lakes. Potential BWE sites were identified by inputting the results of Lake Michigan circulation models into the ballast water model to determine which sites led to no or minimal spread throughout the Great Lakes. Results of model applications have led to specific predictions for species and management scenarios identified by invasive species managers that have previously not been made for ballast water management in the Great Lakes before.