Stochastic variability and transitions to chaos in a hierarchical three-species population model

A variability of the dynamic behavior in stochastically forced multi-species population models is studied. We address how noise can generate complex oscillatory regimes with transitions between attractors and order-chaos transformations. For the parametric analysis of noise-induced transitions, we utilize a semi-analytical technique based on the stochastic sensitivity analysis of attractors and confidence domains method. This approach is used in the study of the fairly realistic three-species population model describing the interaction of prey, predator and top predator. We consider in detail the parametric zone where the system is monostable with excitable limit cycle, or bistable with coexisting limit cycle and chaotic attractor. These zones are separated by the crisis bifurcation point. Noise-induced transitions between regular and chaotic attractors in the bistability zone are analysed by the confidence ellipses method. In the monostability zone, a mechanism of the transition from regular periodic to multimodal chaotic oscillations is studied. (C) 2019 Elsevier Ltd. All rights reserved.