Guiding large-scale management of invasive species using network metrics

Abstract

Invasive species harm biodiversity and ecosystem services, with global economic costs of invasions exceeding $40 billion annually. Widespread invasions are a particular challenge because they involve large spatial scales with many interacting components. In these contexts, typical optimization-based approaches to management may fail due to computational or data constraints. Here we evaluate an alternative solution that leverages network science, representing the invasion as occurring across a network of connected sites and using network metrics to prioritize sites for intervention. Such heuristic network-guided methods require less data and are less computationally intensive than optimization methods, yet network-guided approaches have not been bench-marked against optimal solutions for real-world invasive species management problems. We provide the first comparison of the performance of network-guided management relative to optimal solutions for invasive species, examining the placement of watercraft inspection stations for preventing spread of invasive zebra mussels through recreational boat movement within 58 Minnesota counties in the United States. To additionally test the promise of network-based approaches in limited data contexts, we evaluate their performance when using only partial data on network structure and invaded status. Metricbased approaches can achieve a median of 100% of optimal performance with full data. Even with partial data, 80% of optimal performance is achievable. Finally, we show that performance of metric-guided management improves for counties with denser and larger networks, suggesting this approach is viable for large-scale invasions. Together, our results suggest network metrics are a promising approach to guiding management actions for large-scale invasions.