Nitrogen reduction effects in double rice by planting and returning Chinese milk vetch to the field in Northern Hunan Province

  Objectives  To evaluate the effects of reducing the amount of nitrogen fertilizer on rice yield and soil properties under the mode of planting and returning Chinese milk vetch (Astragalus sinicus L., CMV) to the field in the following rice season. The study aims to provide a theoretical basis for environmentally friendly and efficient double-cropping rice production in Northern Hunan Province.

  Methods  The long-term experiment, early rice–late rice–CMV rotation, started in 2009. There were 7 treatments, including no fertilization (CK), conventional fertilization (N100), fertilization combined with N reduction at 0, 20%, 40%, and 60% (MvN100, MvN80, MvN60, MvN40) of conventional N rate and CMV alone (Mv). Rice yield, N, P, and K absorption, soil physicochemical properties and six hydrolase activities related to C, N, and P cycle (α-glucosidase, β-cellobiosidase, β-glucosidase, β-xylosidase, L-leucine aminopeptidase, alkaline phosphomnoesterase) were determined.

  Results  Compared with the N100, all the Mvs increased yield and saved fertilizer, with the highest yield increase recorded in MvN100 (10.2% in the whole year). In the early rice season, planting and returning CMV (P<0.05) improved the N absorption and utilization efficiency of rice to ensure the stability of soil total N. Compared with N100, CMV combined with 0–60% less N fertilizer increased N uptake by 16.6%–43.4% and N use efficiency by 8.4%–37.8%. The hydrolases activities related to C, N, and P cycling were the highest under Mv treatments in the whole year. Compared with N100, MvN80 and MvN60 (P < 0.05) increased α-glucosidase, β-cellobiosidase, β-glucosidase, and β-xylosidase activity. The results of correlation analysis showed that available soil P and K (P < 0.05) correlate with rice yield. Aggregation enhancement tree analysis (ABT) further confirmed that the contribution rate of available soil P to early and late rice yield was 30.4% and 50.2%, respectively.

  Conclusions  The 10-years field experiment proved that planting and turnover of CMV could replace 60% N fertilizer in early rice and 40% in late rice without impacting rice yields, due to increased N, P, and K accumulation in rice and soil enzyme activities. The experimental soil’s high available P and K content greatly contribute to the replacement ratio mentioned above.