Minnesota aquatic invasive species (AIS) managers are tasked with preventing the spread of many AIS moving through a highly complex and interconnected waterways system, often with limited resources. A data-driven approach to identify and prioritize waterbodies of high risk of invasion is needed to help inform effective and efficient control programs throughout the state and, in time, the region.
Early meetings between stakeholders and the team at the University’s Minnesota Aquatic Invasive Species Research Center (MAISRC) identified a large number of needs as the state works to control AIS. One of these needs seemed to tie all of the issues together: Minnesota needs a decision tool that would help managers on the ground determine the best ways to control AIS with limited people and funds. In short, they needed to not only know where AIS are but also where they are most likely to go next.
Minnesota’s water system includes over 10,000 lakes with thousands of rivers and streams interconnecting them. Hundreds of these are already infested by AIS and thousands more are at risk of infestation. The state is also home to over 800,000 registered boats. Armed with over 1.6 million data points of reported boater movements, the Minnesota Department of Natural Resources infested waters list, and complex networks of natural water connectivity, an interdisciplinary team set out to determine which Minnesota lakes are most at risk and thus should be prioritized at the local and state level. They ultimately created two models: 1.) Introduction Risk for Surveillance and 2.) Prioritizing Watercraft Inspections. The former uses a Bayesian modeling approach to predict the likelihood a lake will be infested with zebra mussels or starry stonewort. The latter uses a GAMS-CPLEX optimization approach to rank lakes within a county that should be prioritized for watercraft inspection efforts and provides a figure that displays the optimal balance of inspection resources.
In November 2020, after five years of development and thousands of simulations to test accuracy, these models moved into an online dashboard: aisexplorer.umn.edu. The AIS Explorer allows local and state managers to optimize their limited resources by accessing science-based recommendations about where to focus surveillance efforts and which lakes inspectors should be placed at for the best effect. Ranked inspection lists can be generated and exported for zebra mussels, starry stonewort, Eurasian watermilfoil, spiny water flea and any combination thereof.
The models are able to forecast eight years into the future and are updated weekly with new data to account for new infestations and changing risk dynamics. Feedback from training workshops with county managers has been positive with many planning to incorporate AIS Explorer into their work stream. While currently focused on county and statewide AIS work, the team is working with additional stakeholders to incorporate data that will not only improve the current models’ accuracy but could help with region-wide AIS management decisions. Significantly, while built for managers, stakeholders across the board can learn and gain insights from it.
Minnesota is home to hundreds of AIS-infested waters and thousands more that need protection. Millions of dollars are spent annually on AIS activities but with such a broad waterway system a decision-making tool was needed to help prioritize time, people and funds. This project showcases how quantitative risk modeling, implemented in specialist software, can be directly linked with operational decision-making in biosecurity, and connect field work conducted by regional authorities with cutting edge science in a timely manner. The AIS Explorer supports stakeholder engagement and smart use of data for rapid detection and response to biosecurity threats in a practical and cost-effective way.