Fig. 5: Relationships between community stability and its components in the biodiversity-manipulated experimental dataset.

These components include: a population stability (p < 0.001 for BigBio and BioCON), b asynchrony (p < 0.001 for both studies), c the compensatory effect (p < 0.001 for BigBio and p = 0.006 for BioCON), and d the statistical-averaging effect (p < 0.001 for BigBio and p = 0.019 for BioCON) within two studies (indicated by different colors). All values are log10 transformed. Regression and significance results formatted as in Fig. 2.