Multiple Stability Mechanisms Act Independently or in Concert to Maintain the Temporal Stability of Natural...

The maintenance mechanisms underlying community temporal stability represent a pivotal concern in ecology. However, empirical evidence on how multiple mechanisms independently or synergistically stabilize natural communities, and how their importance responds to external factors and evolves over time, remains limited. Leveraging a 12-year (2007-2018) manipulative experiment involving clipping and fertilization in an alpine meadow, we assessed the relative contributions of four mechanisms, namely, species asynchrony (compensatory dynamics among species), the portfolio effect (statistical averaging of species' fluctuations), the selection effect (dominance of stable species), and interspecific interactions, across treatments and temporal scales. Stability was quantified as the reciprocal of the coefficient of variation in community coverage. Asynchrony was a ubiquitous foundation of stability across all treatments and time periods. The portfolio effect was a critical positive driver in the initial phase but was suppressed by fertilization over time. In contrast, interspecific interactions and the selection effect emerged as central determinants of long-term stability in later stages. Fertilization amplified the portfolio effect and fostered weak interactions while reducing the fluctuation disparity between dominant and non-dominant species. Clipping enhanced stability mechanisms by preserving species richness and asynchrony. Structural equation modelling revealed that treatments indirectly influenced stability by "reprogramming" the causal pathways among these mechanisms. Our study demonstrates that community stability is upheld by multiple coordinated mechanisms, whose relative importance is contingent on treatment and time scale. Grassland management should therefore move beyond a singular focus on species richness and adopt strategies that promote the synergistic functioning of multiple stability mechanisms.