Role of predator-prey reversal in Rock-Paper-Scissors models

Abstract: In this letter we consider a single parameter generalization of the standard three species Rock-Paper-Scissors (RPS) model allowing for predator-prey reversal. This model, which shall be referred to as $\kappa$RPS model, incorporates bidirectional predator-prey interactions between all the species in addition to the unidirectional predator-prey interactions of the standard RPS model. We study the dynamics of a May-Leonard formulation of the $\kappa$RPS model using lattice based spatial stochastic simulations with random initial conditions. We find that if the simulation lattices are sufficiently large for the coexistence of all three species to be maintained, the model asymptotically leads to the formation of spiral patterns whose evolution is qualitatively similar to that of the standard RPS model, albeit with larger characteristic length and time scales. We show that there are in general two distinct scaling regimes: one transient curvature dominated regime in which the characteristic length of the population network grows with time and another where it becomes a constant. We also estimate the dependence of the asymptotic value of the characteristic length of the population network on the likelihood of predator-prey reversal and show that if the simulation lattices are not sufficiently large then predator-prey reversal can have a significant negative impact on coexistence. Finally, we interpret these results by considering the much simpler dynamics of circular domains.