Soil microbial diversity, stability, and function are enhanced by cover cropping: A machine learning-based pooled analysis of Mississippi agroecosystems

Cover cropping has emerged as a pivotal strategy to enhance soil health and microbiome functionality across diverse agroecosystems. However, the extent to which cover crops reshape microbial diversity, composition, functional capacity, and ecological stability remains insufficiently understood, particularly within heterogenous environments like Mississippi. In this study, we conducted a comprehensive pooled analysis of 473 soil samples collected from multiple independent cover crop trials (2020-2024) across diverse cropping systems and edaphic contexts in Mississippi. Amplicon sequencing (16S rRNA and ITS2) coupled with machine learning, co-occurrence network modeling, and functional prediction tools were employed to evaluate microbial community responses to cover cropping. Results revealed that cover crops significantly elevated bacterial and fungal α-diversity, altered community composition, and enriched taxa associated with nitrogen fixation, organic matter turnover, and pathogen suppression. Notably, Proteobacteria, Acidobacteriota, and Chloroflexi were more abundant under cover cropping, whereas Firmicutes and Actinobacteriota dominated control plots. Fungal communities under cover cropping exhibited higher relative abundance of Rozellomycota and Chytridiomycota, indicating enhanced saprotrophic and decomposition potential. Functional predictions showed increased cellular and environmental processing functions in cover-cropped soils, alongside a marked reduction in KEGG pathways linked to human diseases and pathotroph dominance. Co-occurrence network analysis indicated increased connectivity, modularity, and robustness under cover cropping, suggested enhanced microbial interaction strength and ecological stability. Sloan's neutral model fitting and increased migration rates further revealed that stochastic processes played a greater role in microbial assembly under cover cropping. Random Forest model identified Bradyrhizobium, Bryobacter, and Solirubrobacter as top bacterial biomarkers enriched in cover-cropped soils, while Talaromyces, Purpureocillium, and Clonostachys were the most predictive fungal genera, known for their biocontrol and nutrient cycling roles. These findings underscore the ecological potential of cover crops to enhance soil microbial networks and sustainability within Mississippi production systems, while insights may inform similar humid, subtropical agroecosystems elsewhere.