Soil microbial community structure explains the resistance of respiration to a dry–rewet cycle, but not soil functioning under static conditions

Summary Soil ecosystem functions underpin many ecosystem services. Understanding the drivers of these functions and their stability in response to disturbance is important given land-use intensification and the predicted increase in disturbance frequency with climate change. Changes in land use and soil abiotic conditions are likely to impact on soil ecosystem functions and their stability, but the effects of changes in soil microbial community structure (i.e. traits, diversity, composition and biomass) are poorly understood. We measured soil abiotic properties, microbial community structure, ecosystem functions (carbon and nutrient cycling) and the stability of functions in response to a dry–rewet cycle, across a land-use gradient consisting of natural forest, planted forest, high- and low-producing grassland, and vineyards. Results from additive modelling, where explanatory variables were added in a hierarchical manner and retained if significant (land use followed by abiotic variables followed by biotic variables), showed that measures of soil microbial community structure were the only variables able to explain variation in several measures of stability. In contrast, microbial community structure was rarely able to explain further variation in functions measured under static conditions once land use and abiotic factors were accounted for. Consistent with current theory, correlation analysis indicated that fungal : bacterial ratios were negatively related to measures of carbon and nutrient cycling, whereas the Gram-positive : Gram-negative ratio was positively related to measures of nutrient cycling and stability. In contrast to current theory, the fungal : bacterial ratio was negatively related to the resistance of soil water-soluble N contents to drying. Our results show that measures of microbial community structure can be good indicators of a wide range of soil functions and their stability in response to a dry–rewet cycle. Further, they suggest that soil microbial community structure may have a far greater influence on process rates under disturbed, compared to static, conditions. The delivery of ecosystem services in the future may therefore be strongly dependent on the characteristics of the soil microbial community.