Planting a deep-rooted green manure (GM) (more than 1.0 m depth) greatly improves soil fertility and reduces the loss of nutrients. However, few studies have examined the response of soil nitrogen (N) distribution in the soil profile and subsoil N recovery to the long-term planting and incorporation of deep-rooted GM. Based on a 12-year (2009–2021) experiment of spring maize-winter GMs rotation in the North China Plain (NCP), this study investigated the effects of different GMs that were planted over the winter, including ryegrass (RrG, Lolium L.) (>1.0 m), Orychophragmus violaceus (OrV, Orychophragmus violaceus L.) (>0.8 m), and hairy vetch (VvR, Vicia villosa Roth.) (>1.0 m), on the spring maize yield, N distribution in the deep soil profile, N use efficiencies, functional gene abundances involving soil nitrification–denitrification processes and N₂O production. Compared with the winter fallow, the maize yield significantly increased by 11.6% after 10 years of green manuring, and water storage in 0–200 cm soil profile significantly increased by 5.0–17.1% at maize seedling stage. The total N content in the soil layer at 0–90 cm increased by 15.8–19.7%, while the nitrate content in the deep soil layer (80–120 cm) decreased by 17.8–39.6%. Planting GM significantly increased the N recovery rate (10.4–32.7%) and fertilizer N partial productivity (4.6–13.3%). Additionally, the topsoil N functional genes (ammonia-oxidizing archaea amoA, ammonia-oxidizing bacterial amoA, nirS, nirK) significantly decreased without increasing N₂O production potential. These results indicated that long-term planting of the deep-rooted GM effectively reduce the accumulation of nitrates in the deep soil and improve the crop yield and N use efficiencies, demonstrating a great value in green manuring to improve the fertility of the soil, increase the crop yield, and reduce the risk of N loss in NCP.

The winter planting of green manure (GM) is widely used in South China to reduce chemical nitrogen (N) fertilizer use, improve soil fertility, and maintain rice yields, but its effect on N runoff loss in paddy fields remains unclear. This study combines multi-site field experiments with a process model (WHCNS-Rice) to assess how GM with reduced N fertilizer impacts N runoff loss and its forms in the Yangtze River’s middle and lower reaches, considering different rainfall years. The network field experiments included four treatments: conventional fertilization (FR), conventional fertilization plus straw return (FRS), GM with a 40% N reduction (MR), and GM-straw combined return with a 40% N reduction (MRS). Monitoring the results showed that compared to the winter fallow treatment, the GM treatments reduced the peak and average total N (TN) concentrations by 11.1–57.9% (average 26.9%) and 17.1–27.3% (average 22.3%), respectively. The TN runoff loss under the GM treatment decreased by 3.50–10.61 kg N ha⁻¹ (22.5–42.1%). GM primarily reduced the runoff loss of dissolved inorganic N (DIN), with reductions at different sites ranging from 0.22 to 9.66 kg N ha⁻¹ (8.4–43.4%), indicating GM effectively decreases N runoff by reducing DIN. Model simulations of ponding water depth, runoff, TN concentration in surface water, and TN loss in paddy fields produced the consistency indices and simulation efficiencies of 0.738–0.985, 0.737–0.986, 0.912–0.986, and 0.674–0.972, respectively, indicating that the model can be used to evaluate water consumption and N runoff loss in the GM-paddy system. The simulations showed that GM with a 40% N fertilizer significantly reduced N runoff loss under all rainfall conditions, with the greatest reductions in wet years. Under wet, normal, and dry conditions, the GM treatments significantly reduced average TN loss by 0.37–5.53 kg N ha⁻¹ (12.77–29.17%), 0.21–5.32 kg N ha⁻¹ (9.95–24.51%), and 0.02–3.2 kg N ha⁻¹ (1.78–23.19%), respectively, compared to the winter fallow treatment. These results indicate that the combination of GM and a 40% reduction in N fertilizer can significantly reduce N runoff loss from paddy fields, demonstrating good effectiveness under various rainfall conditions, making it a green production model worth promoting.