Managing tillage, sowing rate and nitrogen top-dressing level to sustain rice yield in a low-input, direct-sown, rice-vetch cropping system

Direct sowing of rice into drained soil is a relatively new farming practic e, replacing transplanting, being developed for use in External Low Input S ustainable Agriculture in Korea. Field experiments were conducted to invest igate the importance of tillage, sowing rate, and top-dressing of fertilise r nitrogen (N) to rice growth and grain yield in a direct-sown, rice-vetch cropping system. From 1993 to 1997, Chinese milk vetch (Astragalus sinicus L. Vetch) was gro wn during the winter season (October-May) at Uiryong, Korea. Until 1995 tra nsplanted rice (Oryza sativa L.) was grown during the summer season, but in 1996 and 1997 the direct-sowing method was used. In 1996, the treatments a pplied to rice were either tillage or no-tillage of the paddy before the ri ce crop was sown, combined with 2 rates (0 or 2 g/m(2)) of fertiliser N top -dressed onto the paddy just after the panicle initiation stage. In 1997, t he treatments were either tillage or no-tillage of the paddy before the ric e crop, combined with 2 rice sowing rates (400 or 800 seeds/m(2)), and 3 ra tes (0, 2 or 4 g/m(2)) of fertiliser N top-dressed just after the panicle i nitiation stage. Rice yielded 243-435 g/m(2) in 1996 and 493-678 g/m(2) in 1997, the lower y ield in 1996 was attributed to fewer established seedlings and productive p anicles. There was no effect of tillage treatment. Across years, yield was correlated with shoot N content, and grain N-use efficiency exceeded 50 g g rain/g shoot N content. Rice residues contained about 5 g N/m(2) and vetch residues released 6-8 g N/m(2) during the 4 weeks after submergence, theref ore the release of N from residues was potentially sufficient to supply the rice crop demand. In both years, however, the yield of unfertilised rice w as N limited so increasing the shoot N content appeared to be the most impo rtant issue to be resolved in sustaining yield in External Low Input Sustai nable Agriculture. A low rate of top-dressed fertiliser N (2-4 g N/m(2)) in creased yield by increasing the number of panicles and spikelets, and to a lesser extent, 1 000-grain weight. Growing rice using no-tillage practices did not limit the yield of rice, except in 1996 when no fertiliser N was ap plied and the N uptake by rice was low. A high sowing rate combined with to p-dressed fertiliser N in 1997 reduced yield through crop lodging and an in crease in the number of unproductive tillers. This problem may be overcome by the use of a variety less susceptible to lodging.