LAUSR

Experiments in fields and micro-plots were conducted to investigate the optimal cropping system and nitrogen (N) fertilizer application rate and timing. The treatments consisted of Chinese milk vetch–rice (CMV–R) rotation with five N fertilizer application rates (0, 120, 180, 240, 300 kg N ha−1) during the rice-growing season, and fallow–rice (F–R) and wheat–rice (W–R) rotations with only one N application rate (240 kg N ha−1) each. Rice yield increased with increasing N fertilizer application rate under CMV–R rotation, and achieved highest yield under CMV–R180. There is a decreasing trend when N application rate exceeded 180 kg N ha−1. Rice yield was always higher under CMV–R240 compared to W–R240 and F–R240. During the 2012 rice season, the fertilizer N-use efficiency, residual N fertilizer in soil and N fertilizer recovery efficiency of CMV–R180 reached largest under CMV–R rotation with different N treatments. Furthermore, the fertilizer N-use and recovery efficiencies of CMV–R240 and F–R240 were far higher than those of W–R240. In 2013, fertilizer N-use efficiency was the highest (> 50%) at the heading stage, which was nearly twice as much as the efficiencies during the basal and tillering stages. The N fertilizer loss rate during the basal stage was significantly higher than that at the tillering and heading stages, which was up to 60%. CMV–R rotation with 180 kg N ha−1 achieved the highest rice yield of 9454 kg ha−1 and high fertilizer N-use efficiency (40.6%) under a relatively lower N application rate. Therefore, Chinese milk vetch–rice cropping system could be a promising approach for decreasing fertilizer inputs to prevent N pollution problems and increasing rice yield, especially for the intensive rice-based cropping systems in southern China.