Network centrality as a potential method for prioritizing ports for aquatic invasive species surveillance and response in the Laurentian Great Lakes.

Ballast water is a leading vector for the introduction of aquatic invasive species worldwide and, once a novel species is established, regional ballast water exchange between ports can accelerate secondary spread. The importance of shipping induced invasions in the Laurentian Great Lakes has resulted in policies that require more stringent ballast water treatment standards for transoceanic shipping than is required of ships operating regionally within the Great Lakes. As a result, ballast water discharges within the Great Lakes are not well regulated, primarily because of the challenge of treating the high volumes of water carried by vessels that are confined to the waters of the Great Lakes. We used a discrete-time Markov chain model on a network with annual time-steps to simulate ballast water management scenarios at high-priority ports in the Great Lakes shipping network for two potential invaders, golden mussel (Limnoperna fortunei) and monkey goby (Neogobius fluviatilis). We chose high-priority ports by using graph-theoretic network analysis techniques to calculate six network centrality metrics for 151 ports in the network. Ports scoring high in network centrality scores have more ties with other ports or are positioned within the network such that they potentially have greater influence over the secondary spread of aquatic invasive species than other ports. We simulated secondary spread scenarios where hypothetical ballast water treatment was implemented at the top twenty ranked ports in each network centrality metric, as well as the top twenty busiest ports by ship arrivals. The results of each scenario were compared to a scenario where no management action was taken. Simulated secondary spread for both golden mussel and monkey goby resulted in significantly reduced infestation probabilities (p < 0.001) under all management scenarios when compared to unmanaged spread scenarios. Management at ports with inwardly directed ties to other ports reduced infestations by the greatest amount compared to other management scenarios; 65.4% for golden mussel and 74.6% for monkey goby. The indegree centrality of ports in the Great Lakes was found to be an important factor in governing secondary spread. Here we show that prioritized management, like high volume shore based treatment systems based on network centrality, is a potentially effective strategy for impeding the secondary spread of new or localized invasive species in the Great Lakes. [ABSTRACT FROM AUTHOR]