Ecosystem multifunctionality shows unique stability responses to global change drivers

Abstract Ecological research increasingly investigates the dimensionality of ecosystem functioning and stability through the multivariate relationships among their constituent functions and stability components. Yet, it is unclear how this dimensionality jointly shapes multifunctionality and overall ecological stability, and how human-driven changes affect this higher order complexity. Here, we show that global change drivers generally destabilize ecosystems by unevenly affecting multiple stability components across different functioning levels. We conducted a 160-day multifactorial experiment in freshwater mesocosms to test how global change drivers (climate warming, nutrient addition and litter loss) interactively affect the dimensionality of ecosystem stability (DS) and the resulting overall ecosystem stability (OES), comparing the stability of individual ecosystem functions with that of effective multifunctionality. Although all drivers decreased or increased the stability of different ecosystem functions, OES decreased only in response to the interaction between warming and nutrients. In contrast, DS increased only under warming, reflecting weakened correlations among stability components. Whereas warming effects on DS were prevalent at single-function levels, DS for multifunctionality showed no changes across disturbances. Together, these results indicate that multifunctionality exhibits more consistent, directionally stable responses across environmental changes than its constituent functions, an insight with direct implications for predicting and managing multiple ecosystem processes.