Big data analysis for evaluating bioinvasion risk

Prioritizing management of high-risk routes and ports by vessel type to improve marine biosecurity efforts

The shipping industry constitutes the main vector of marine bioinvasions. Over 90,000 vessels world-wide create a highly complex shipping network that requires appropriate management tools. Here we characterized a novel vessel category, Ultra Large Container Vessels (ULCV), in terms of potential contribution to the dispersal of Non-Indigenous Species (NIS) in comparison to smaller vessels traveling similar routes. Such approach is essential for providing precise information-based risk analysis necessary to enforce biosecurity regulations and reduce the adverse global effects of marine NIS. We used Automatic Identification System (AIS) based websites to extract shipping data that will enable us to test for differences in two vessel behaviors linked to NIS dispersal: port visit durations and voyage sailing times. We then examined the geographic spread of ULCVs and small vessels, quantifying the accumulation of new port visits, countries, and ecoregions for each vessel category. Finally, Higher Order Network (HON) analysis revealed emergent patterns within shipping traffic, species flow, and invasion risk networks of these two categories. Compared to the smaller vessels, ULCVs spent significantly longer time in 20% of the ports and were more geographically constrained, with fewer port visits, countries, and regions. HON analysis revealed that the ULCV shipping species flow and invasion risk networks were more similar to each other than to those of the smaller vessels. However, HON port importance shifts were discernible for both vessel categories, with major shipping hubs not necessarily being major invasion hubs. Overall, compared to smaller vessels, ULCVs behave differently in ways that potentially increase biofouling risk, albeit in a smaller set of ports. Future studies using HON analysis of other dispersal vectors appears critical for prioritizing management of high-risk routes and ports.

Application of multi-region input-output analysis to examine biosecurity risks associated with the global shipping network

The vast majority of globally traded cargo is transported via maritime shipping. Whilst in port for loading and unloading, these ships can pick up local marine organisms with internal ballast water or as external biofouling assemblages and subsequently move these to destination far beyond their natural ranges. Over the past decades, this mechanism has led to the establishment of hundreds of non-indigenous species (NIS) around global coastlines. Marine NIS cause significant environmental, economic, cultural and human health impacts. Taking effective steps to preventing their dispersal and establishment is an enduring challenge for governments and conservation agencies around the world. Here we use international commodity trade data and a Nobel-Prize-winning economic analysis technique to develop a novel approach for assessing global marine NIS transfer risks. We show that by tracing the origins and destinations of seaborne trade connections, and the nature of the traded commodities, we can predict the strength of shipping vectors and associated marine biosecurity risks. We demonstrate the utility of our approach via a case-study, where we trace the spread of a hypothetical marine NIS from Japan and show the congruence of our model results with documented invasion histories from that region. Our study demonstrates that biosecurity risk can be assessed using established economic modelling frameworks on the basis of monetary transaction data alone, and without the need for detailed itineraries of the many thousand vessels making up the global commercial fleet. Novel, cost-effective tools are needed to mitigate biosecurity risks associated with maritime trade, and to meet conservation goals while enabling economic prosperity. The modelling framework presented here can be expanded to incorporate future risk factors, life-history traits of particular NIS of concern, and even adapted to simulate the dispersal of terrestrial pests or disease agents.