Long-Term Fertilisation Regimes, Crop Yield and Rhizosphere Microecology in a Wheat–Soybean Rotation System

Long-term fertilisation must sustain crop yield and rhizosphere microbial functions in wheat–soybean rotations, but interannual responses remain unclear. We examined 12 treatments in a long-term field experiment established in 1979 and continuing through 2025 in Heilongjiang, China; three within-plot soil composites per treatment represented the 2025 microbiome snapshot. Long-term wheat and soybean grain and straw yield, rhizosphere and non-rhizosphere metagenomes, bacterial diversity, functional gene profiles, taxonomic composition, yield–microbe correlations, and structural equation modelling were assessed. The displayed wheat grain yields differed among treatments, with the median approaching 2500 kg ha−1 under manure plus medium-rate mineral fertiliser. Soybean grain yield exceeded 2000 kg ha−1 under manure plus low- or medium-rate mineral fertiliser and was higher than under high-rate mineral fertiliser alone. Separate exploratory treatment-only PERMANOVAs returned R2 values of 0.7578 and 0.8227 for rhizosphere and non-rhizosphere functional profiles, respectively; the profiles were within-plot subsamples. Mesorhizobium, Lysobacter, Gemmatirosa, and Pseudolabrys were positively correlated with crop yield. Structural equation modelling linked available phosphorus to wheat yield and rhizosphere diversity and linked available potassium to soybean yield. These descriptive findings suggest crop-specific fertilisation patterns and possible associations of yield with soil nutrients and rhizosphere microbiota, indicating that productivity and microbial responses should inform long-term nutrient management.