Soil inoculation for enhancing agricultural grassland conversion

Restoring species-rich grasslands on agricultural land requires recovering both plant and soil communities simultaneously, yet soil biotic communities recover slowly and can constrain the rate of vegetation recovery through plant-soil feedbacks. Whole-soil inoculation offers a means to actively accelerate this process by steering soil community composition toward that of a species-rich donor grassland. The transfer of soil from a donor grassland to a degraded recipient site, known as whole-soil inoculation, offers a means to actively steer soil community composition, but its effectiveness under agricultural conditions remains poorly understood. We combined a controlled greenhouse experiment and a two-year field experiment to investigate whether whole-soil inoculation induces detectable shifts in microbial community composition and functional group structure during early stages of grassland conversion. In the greenhouse, inoculum origin and seed mixture both significantly affected microbial community composition, with containersreceiving donor soil developing communities most similar to the original donor, confirming that living biota drove community shifts. In the field, however, no detectable inoculation signal emerged in microbial community composition, regardless of seed mixture. Instead, sampling period explained the majority of variation in soil microbial community composition, reflecting gradual microbial changes during the conversion from arable to grassland. Both prokaryotic and fungal communities shifted directionally towards the donor community over time in all plots, independently of inoculation, suggesting that seed mixture also had a significant part in driving the belowground changes. Inoculation also did not lead to detectable shifts in microbial functional group composition in either the greenhouse or field experiment, although functional guilds of the microbial community changed in abundance over time in the field. We discuss methodological constraints, including thin inoculum layers, insufficient prior disturbance, and short duration of the soil monitoring that may have limited detection of inoculation effects. Our results suggest that under realistic agricultural conditions, whole-soil inoculation in combination with sowing is insufficient to redirect microbial community assembly two years after inoculation. Hence longer monitoring periods, larger inoculum quantities, and integration with topsoil management are needed to further improve the potential of inoculation-based grassland restoration.