Effects of long-term substitution of chemical fertilizer with Chinese milk vetch on soil phosphorus availability and leaching risk in the double rice systems of Eastern China

The cultivation of Chinese milk vetch (MV) as green manure is an effective practice for improving paddy soil fertility and rice yields in China. However, few studies have investigated the effects of MV on phosphorus (P) availability and the risk of P loss. An 11-year field experiment was conducted to analyze the effects of returning MV to paddy soil on soil P forms, late rice yield and soil fertility, as well as the related mechanisms. The experiment included seven treatments: (1) F0M0 (no-fertilizer and no-MV); (2) F1M0 (chemical fertilizer alone), (3) F2M0 (70% of chemical fertilizer), and F2MV1-4 (70% of chemical fertilizer and 7500 kg ha-1 MV, 15000 kg ha-1 MV, 22500 kg ha-1 MV, 30000 kg ha-1 MV). The Olsen-P of F2M1-4 were significantly higher (38.3-54.5%) than that of F2M0 and the CaCl2-P levels of F2M1-4 were significantly lower (47.7-61.6%) than that of F1M0. The Olsen-P was mainly influenced by soil total nitrogen (TN), total phosphorus (TP), pH, free Fe/Al (FeDCB/ AlDCB), organic bound Fe (Feo), noncrystalline amorphous Al (Alox) and exchange Mg2+, while the CaCl2-P was mainly influenced by TP, pH Alox, FeDCB, AlDCB, and exchange Mg2+/Ca2+. Partial least squares path model analysis indicated that chemical fertilizer played more important role on the Olsen-P and the CaCl2-P. MV application had the negative effect (the total effect value-0.48) on the CaCl2-P and decreased the risk of P loss yet. Two-segment regression analyses indicated the critical Olsen-P levels were 19.39 mg kg-1 for late rice yield and 31.94 mg kg-1 for soluble P, respectively. In conclusion, the continuous application of MV not only improved P availability to enhance late rice yield, but also reduced the risk of P loss in paddy soil in double rice systems.