Drivers of ɑ-diversity
Overall, our models explained a relatively large amount of variance in species richness (ɑ-diversity: mean conditional Rc2 = 0.54, ranging between 0.31 - 0.76). A considerable amount of this variance was explained by site, as indicated by the difference in model fit between the conditional and marginal R2 (ɑ-diversity variation explained by site: mean Rc2–Rm2 = 0.34, range 0.14-0.56). Site was a particularly strong predictor of ɑ-diversity of springtails, protists and fungi, where it explained at least twice as much variation as all other variables combined (Table 2). Explanatory soil variables differed strongly between groups of soil organisms (Table 2). Species richness of all microbial groups (bacteria, fungi and protists) was most strongly and positively related to soil pH and to a lesser extent to litter biomass, soil bulk density and moisture content. Species richness of nematodes was also related to pH, albeit less strongly than for soil microbes. Nematode species richness was also explained by soil carbon content, and interaction effects between pH*grazing removal and litter biomass*grazing removal. Conversely, species richness of springtails and mites was unrelated to soil pH. Species richness of springtails was negatively related to bulk density and the interaction effect between grazing removal and cover of grasses and herbs, soil bulk density and mineral nitrogen concentration, while mite richness was explained by the interaction effect between soil moisture*grazing removal.
Effects of abandonment on β-diversity .
Cessation of grazing caused a significant decrease in β-diversity for protists and fungi (both approximately 5%), whereas β-diversity of plants and bacteria was unaffected (Figure 3B). In contrast, cessation of grazing caused an increase in β-diversity for mites (5%), springtails (15%) and nematodes (15%).