Alternative stable states: Evidence from natural ecosystems and implications for ecological restoration

Abstract Human activities and climate change increasingly push ecosystems towards abrupt degradation, or even collapse, posing urgent challenges for ecological restoration. Many degraded ecosystems exhibit nonlinear responses to disturbance, in which recovery does not simply retrace the trajectory of degradation. Understanding such dynamics is crucial for improving restoration outcomes under accelerating environmental change. We synthesize theoretical and empirical studies of alternative stable states across diverse ecosystems, review methods for their identification, examine the feedback processes that maintain them and evaluate how key concepts related to alternative stable states, especially critical thresholds, hysteresis and positive feedbacks, have been applied in ecological restoration. Across diverse ecosystems, positive feedbacks between biotic components and between biotic and abiotic components play a central role in trapping degraded states. Consequently, restoration efforts targeting only external drivers often fail due to hysteresis. Empirical evidence shows that restoration succeeds when management disrupts feedback that traps degradation or reinforces those that promote recovery. These findings highlight that effective restoration in ecosystems governed by alternative stable states requires moving beyond linear recovery expectations. Applying the alternative stable states theory, by identifying critical thresholds and targeting key feedbacks, can improve restoration outcomes. Practical implication: Restoration in ecosystems governed by alternative stable states requires recognizing nonlinear recovery and the potential existence of critical thresholds. Even without precise threshold estimates, managers should consider that recovery may be constrained by positive feedbacks rather than external drivers alone. Restoration strategies should therefore target key feedback processes through small‐scale interventions that weaken degradation‐reinforcing feedbacks and strengthen recovery‐promoting feedbacks. Such local interventions may trigger autocatalytic recovery processes, allowing restoration to progressively expand across larger areas.