Scale-dependent patterns and mechanisms of grazing-induced biodiversity loss: evidence from a field manipulation experiment in semiarid steppe

CONTEXT: Although many studies have demonstrated that grazing may increase or decrease plant diversity of grasslands at small scales, few studies have examined the patterns and mechanisms of grazing effects on biodiversity across multiple scales. OBJECTIVE: Our study tested the scale dependence of grazing effects on plant diversity based a 7-year grazing manipulation experiment with seven levels of grazing intensity (0–9 sheep ha⁻¹) in a typical steppe of the Inner Mongolia grassland. METHODS: Species area relationships (SARs) were used to analyze the scale dependence of species loss. SAR decomposition approaches were followed to examine the contribution of four potential mechanisms to changes in the slope of SARs, including species aggregation, overall species richness, total number of individuals, and species abundance distribution. RESULTS: The proportional species loss increased with sampling area (4–1024 m²), which was evidenced by decreasing intercepts and slopes of SARs with grazing. Reduction in the slope of SARs was mainly caused by changes in overall species richness and species abundance distribution, with the relative minor effect of changes in number of individuals. CONCLUSIONS: The negative effect of grazing on overall species richness was mainly attributed to the loss of grazing-sensitive rare species from species pool. Compared with flat systems, plant diversity in slope systems was more sensitive to grazing at low intensities. However, the responses of plant diversity to grazing tended to converge between the flat and slope systems at high levels of grazing intensity. Our study has important implications for adaptive ecosystem management and biodiversity conservation in arid and semiarid grasslands.