Andrew J. Tanentzap, William G. Lee, Katharina A. C. Schulz

• A unified understanding of the relationship between disturbance and biodiversity is needed to predict biotic responses to global change. Recent advances have identified the need to deconstruct traditional models of disturbance into intensity and frequency to reconcile empirical studies that appear to generate contradictory associations between species diversity and disturbance. We integrate results from theoretical simulation modelling, field-based surveys of 5176 vegetation plots from 48 transects across 6 sites, and experimental pot-based manipulations of flooding to identify how disturbance drives species diversity within ephemeral wetlands in South Island, New Zealand. We find empirical, hump-shaped and positive relationships between species diversity and both disturbance intensity and frequency, mirroring patterns from a simulation model in which species differed in their demographic responses to disturbance. More generally, our simulations show that the relationships between diversity and disturbance shift from positive toA unified understanding of the relationship between disturbance and biodiversity is needed to predict biotic responses to global change. Recent advances have identified the need to deconstruct traditional models of disturbance into intensity and frequency to reconcile empirical studies that appear to generate contradictory associations between species diversity and disturbance. We integrate results from theoretical simulation modelling, field-based surveys of 5176 vegetation plots from 48 transects across 6 sites, and experimental pot-based manipulations of flooding to identify how disturbance drives species diversity within ephemeral wetlands in South Island, New Zealand. We find empirical, hump-shaped and positive relationships between species diversity and both disturbance intensity and frequency, mirroring patterns from a simulation model in which species differed in their demographic responses to disturbance. More generally, our simulations show that the relationships between diversity and disturbance shift from positive to hump-shaped to negative as species that are favored at low disturbance because of their resistance strategies, defined by low mortality and recruitment, decline within communities relative to resilient species. Resilient species with higher mortality and recruitment rates are instead favored as disturbance intensity and frequency intensify. Our theoretical findings suggest that sites must also have a third group of unique species with intermediate resilience and resistance. Analyses of community composition along our disturbance gradients support this prediction, emphasizing that shifts in community-level resistance and resilience drive empirical associations between diversity and disturbance. Overall, terrestrial plants may be unable to resist intense and frequent flooding, even with specialized traits. Only fast-growing species with high regeneration from seed may respond once flooding subsides and dominate community composition in these situations, especially on nutrient-rich soils. However, different strategies can co-occur at intermediate disturbance, ultimately increasing species richness. As disturbances become more pervasive globally, our results suggest that differences in the niches of species, rather than demographic stochasticity, drive biodiversity patterns. These niche-based processes may especially prevail, without accompanying losses in species richness, where sites are initially dominated by resistant taxa or life history strategies that balance resistance and resilience.…