Responses of climatic resource use efficiency and wheat yield to green manure planting on the Loess Plateau

Green manure is increasingly used to replace fallow because of its ecosystem benefits, yet its effects on subsequent crops remain uncertain in water-limited regions with high climatic variability. Accordingly, we conducted an eight-year field experiment on the semi-arid Loess Plateau to compare three cropping systems: winter wheat (Triticum aestivum L.) with summer bare fallow, soybean (Glycine max (L.) Merr.), or sorghum-sudangrass (Sorghum bicolor × S. sudanense). We quantified climatic resource use efficiency based on biomass and wheat yield responses across dry, normal, and wet years, and assessed associated changes in soil properties and grain quality. Green manure biomass inputs averaged 4035 kg ha⁻¹ for soybean and 6366 kg ha⁻¹ for sorghum-sudangrass, increasing total system biomass. Relative to the wheat–fallow system, precipitation, accumulated temperature, and radiation utilization efficiencies increased by 17.2%–18.5% in wheat–soybean and by 30.2%–35.9% in wheat–sorghum-sudangrass. However, in dry years, soybean and sorghum-sudangrass reduced wheat yield by 37.1% and 34.0%, respectively; in normal years, wheat yield tended to decline, but the difference was not significant, whereas the effect on wheat yield was overall neutral in wet years. Hierarchical partitioning showed stronger climatic control of interannual yield variability in the wheat–green manure system than in the wheat–fallow system. In addition, soybean increased soil available nitrogen and grain protein content, while sorghum-sudangrass led to greater increases in soil carbon and grain starch content; correspondingly, the soil quality index increased by 27.7% and 33.2%, respectively. Overall, replacing fallow with green manure increased interannual uncertainty in wheat yield responses, but improved soil properties support it as a viable option for sustainable rainfed management on the Loess Plateau.