Indicators of healthy rhizosphere soils exhibit limited cross-system generalizability under diverse cropping patterns and management strategies

Abstract

Soil abiotic and biotic environment plays a crucial role in plant health. Developing prediction models based on universal soil indicators is of great significance for effective plant health management. In this study, we evaluated plant health indicators by comparing the rhizosphere soils of healthy and Fusarium wilt diseased watermelon plants, and investigated the consistency of these indicators across diverse cropping patterns and soil management strategies. Our results indicated that healthy soils exhibited significant differences in many physicochemical properties, biological community structures compared to the diseased soils in specific cropping pattern and soil management strategy. However, there are extremely limited common differential characteristics across systems. Notably, healthy soils had consistently higher available potassium contents. Enhanced complexity and stability in soil microbial networks were observed in healthy soils, characterized by improved topological parameters, robustness, and reduced vulnerability. The negative cohesion of bacterial-fungal network was significantly heightened in healthy soils compared to diseased ones. Furthermore, we developed a straightforward and highly effective prediction model for watermelon health with over 85.0 % accuracy, integrating potassium levels and the relative abundances of two Mortierella species. Overall, our study revealed the limited consistency of most soil healthy indictors across various cropping systems and highlighted the importance of developing universal plant health prediction models.