An individual-based numerical model of medusa swimming behavior

Abstract

Scyphomedusae are ubiquitous in marine and estuarine systems, where they frequently play an important role in trophodynamics. Many scyphomedusae are cruising predators, and feeding rates depend, in part, on swimming behavior. Yet, no model of medusa swimming exists. An individual-based correlated random walk (CRW) model of medusa swimming behavior in three dimensions was developed. The model was validated using a previously published dataset of the swimming of 19 Chrysaora quinquecirrha (Desor, 1848) medusae that were observed in the presence or absence of zooplankton prey in laboratory mesocosms in August–October 1998 (Matanoski et al. in Mar Biol 139:191–200, 2001). In the presence of prey, medusae swam at a constant moderate rate in looping trajectories. In the absence of prey, medusae alternated periods of slow and fast swimming in more linear trajectories. In the model, looping trajectories were reproduced only when changes in movement by a medusa were oriented to its current position and orientation; more linear trajectories were reproduced by movement oriented to a fixed framework. This suggests that medusae change from swimming behavior oriented to local stimuli (e.g., contact with prey) to long-range stimuli (e.g., gravity) depending on the availability of prey. The model reproduced cyclical changes in swimming speeds by medusae in the absence of prey by simulating switching in the behavior controlling the strength of swimming bell pulsations using a probabilistic function. Model results also demonstrated that medusae tend to swim toward the surface, avoid contact with the bottom, increase time spent in prey patches if they alter swimming patterns in the presence of prey, and exhibit significant periodicities in swimming patterns that are the result of deterministic behavior. The model will permit the simulation of the complex behavior of medusae.