Title: Universality of noise-induced resilience restoration of ecological systems

Abstract: As ecosystems might lose their biodiversity and even change from the functional state to the extinct state due to environmental degradation, resilience restoration thus plays an increasingly important role in recovering the service and functions of the systems. The existence of alternative stable states and the switches between them in ecological systems containing a single dynamical variable has been substantially studied. Impeded by the difficulty of high dimensionality, nonlinearity, and stochastic effects, however, there is little understanding of the bistability and transitions for spatially-extended systems in noisy environment. We bridge the gap by successfully combining nucleation theory and the theory about noise-induced transition, with which the transition and resilience restoration can be explored in multi-variable systems. After applying it to four types of lattice-based ecosystems, two cluster modes that exhibit different transition patterns are distinguished by the system size and noise strength. Large systems under strong noise take less time to complete transition, exhibiting the multi-cluster transition mode. In contrast, small systems subjected to weak noise require more time, leading to the single-cluster mode. We show both analytically and numerically that the recovering time for the single-cluster mode follows an exponential distribution and that for the multi-cluster mode is almost fixed by spatial self-averaging. Our study advances the understanding of the role of noise in ecological restorations. More importantly, this approach is not restricted to ecosystems, but it can be generally used to study resilience and transitions in various dynamical systems with alternative stable states.