Community performance curves predict community stability despite interaction effects

Predicting whether aggregate community properties remain stable as environments fluctuate is a central challenge in ecology. We asked how well species' fundamental niches predict this aspect of community stability. We developed community performance (CP) curves by aggregating species-specific performance across environmental conditions and tested whether their variability explains temporal variability in communities. In Lotka-Volterra and mechanistic consumer-resource simulations, CP variability explained, on average, 89% of variation in total abundance variability, and explanatory power remained above 75% across gradients of interaction strength and coexistence conditions. In a purpose-designed microcosm experiment, CP variability explained 67% of variation in total biomass variability. CP curves thus retain substantial predictive information even in strongly interacting systems, providing a tractable basis for forecasting community stability when species interactions are unknown.