Location: Northern China.

Major taxa: Plant and soil microbes of grassland ecosystems.

Time period: 2018 and 2019.

Methods: We conducted a transect survey across grasslands to measure plant diversity, plant traits, and soil microbial diversity. High throughput sequencing was used to assess soil microbial diversity for bacterial 16S ribosomal RNA (16S) and fungal internal transcribed spacer (ITS) regions on an Illumina MiSeq. The random forest algorithm was used to determine the important spatial and environmental variables in predicting plant and microbial diversity, and structural equation modelling was used to examine the direct and indirect effects of climatic and edaphic variables on plant and microbial diversity.

Results: Plant diversity was positively correlated with the diversity of soil fungi, particularly for predicted arbuscular mycorrhizal fungi (AMF) and saprotrophic fungi, and they were positively related to soil nutrients and texture. However, the correlation between plant and bacterial diversity varied by phyla and functional guilds, resulting in decoupling between plant and soil bacterial diversity. Community weighted mean leaf C:N ratio indirectly decreased soil fungal diversity through a negative relationship with soil total nitrogen. Soil bacterial and fungal diversity increased with increasing functional richness of specific leaf area and stem density, respectively.

Main conclusions: These findings have contributed to unravelling the direct and indirect linkages between plant and soil fungal diversity, highlighting particularly strong linkages between plant diversity and predicted AMF and saprotrophic fungi diversity. However, we failed to detect an overall linkage between plant and soil bacterial diversity. Still, our findings suggest that integrating soil fungi into the framework of plant diversity conservation is conducive to biodiversity restoration in degraded grassland ecosystems.