Microbial network and its relationship with plants reflect the restoration of grasslands during a five-year period.

Introduction: The ecological restoration of degraded grassland has been widely studied in terms of plants, soil, and microbes. The microbial network and its relationship with plants are essential indicators of the grassland restoration process. Across restoration statuses, how the microbial network affects plant-microbe interactions and reflects grassland restoration remains unclear. Methods: We examined changes in soil microbial networks, plant biomass, and soil properties across early, moderate, and late restoration, as well as stable community status, over a 5-year restoration period. We calculated the grassland restoration index (GRI) to evaluate the restoration process and constructed microbial co-occurrence networks to assess changes in microbial communities over 5 years. Results: Our results showed the dominant species shifting from Kochia sieversiana to Chloris virgata in early restoration status, and the bacterial network complexity fluctuated over the 5 years, indicating a restoration in early restoration status. In moderate and late restoration statuses, the complexity of bacterial networks gradually increased, suggesting the increased recovery potential in the grassland. However, the GRI showed that the moderate and late restoration statuses did not progress to the next successional status over the 5 years. Moreover, the random forest regression model revealed that high soil pH may limit the restoration process by regulating soil microbial networks. Conclusion: Our study illuminates the signaling roles of soil microbial networks in the restoration process and suggests that moderate and late restoration statuses may take longer to recover to a stable community status. We emphasized that soil pH may be a key constraint in the different restoration statuses, warranting further experimental testing of pH regulation as a restoration intervention.