Using Markov chains to quantitatively assess movement patterns of invasive fishes impacted by a carbon dioxide barrier in outdoor ponds

Natural resource managers use barriers to deter the movement of aquatic invasive species. Research and development of new invasive species barriers is often evaluated in pond and field scales using high-resolution telemetry data. Telemetry data sets can be a rich source of data about fish movement and behavior but can be difficult to analyze due to the size of these data sets as well as their irregular sampling intervals. Due to the challenges, most barrier studies only use summary endpoints, such as barrier passage counts or average (e.g., mean or median) fish distance from the barrier, to describe the data. To examine more fine-scale fish movement patterns, we developed a first-order Markov chain. We then used this model to help understand the impacts of a barrier on fish behavior. For our study system, we used data from a previous study examining how bighead and silver carp (two invasive fish species in the United States) responded to a carbon dioxide (CO2) barrier in a pond. Considerations for management Based upon our findings, management considerations include: Markov chains can be used to describe and quantify movement patterns of aquatic organisms responding to barriers. These models provide insight not possible with previously used summary approaches. The CO(2)barrier changed both the locations of fish and movement patterns of fish. The changes in movement were not captured using conventional endpoints. Specifically, fish had less random movement patterns during periods of CO(2)exposure. This provides evidence that bighead and silver carp actively avoided areas with CO(2)and also shows that our approach can help assess random versus nonrandom movements