Rice-wheat (RW) is the dominant cropping system in north-western India and is of immense importance for national food security. However, the sustainability of the RW system is threatened by water shortage and nutrient mining. Permanent bed RW systems with crop residue retention have been proposed as a means of reducing irrigation water use, improving soil properties and reducing the cost of crop establishment. A field experiment was conducted over 4 years in Punjab, India, to compare conventional and permanent bed RW cropping systems, with and without retention of crop residues, in terms of crop performance and nitrogen use efficiency (NUE). Two methods of rice establishment (transplanting and dry seeding) were included on both beds and flats with four N application rates (0, 80, 120, 160 kg N/ha). Rice grain yield increased significantly as N rate increased up to 160 kg N/ha irrespective of method of rice establishment. Puddled transplanted rice (PTR) was always superior to all other establishment methods in terms of biomass, yield and NUE. At 120 kg N/ha, yield of transplanted rice on permanent beds (TRB) was 29% lower than yield of PTR, while yield of direct-seeded rice on permanent beds (DSRB) was even lower (44% lower than yield of TRB). Wheat straw mulch further reduced yield of DSRB by 26% on average, but there was no effect of mulching on yield of TRB. Dry-seeded rice on flats and beds was prone to severe iron deficiency and root nematode infestation. Yield of DSRB relative to yield of PTR declined as the beds aged but there was no trend in relative yield of TRB. Recovery of fertiliser 15N in the straw plus grain was 30% in PTR compared with 14% for TRB and 17% for DSRB. The majority (65-83%) of the crop N uptake was derived from the soil in all treatments despite the application of urea at 120 kg N/ha. Total N losses from the urea N applied to rice ranged from 52% to 60% in TRB and DSRB compared with 38% in PTR. Wheat yield increased with N rate up to 120 kg N/ha, with further significant response to 160 kg N/ha in 2 of the 4 years. Wheat grain yield on permanent beds after TRB and DSRB was 75-96% of that of conventionally tilled wheat (CTW), with no trend in relative yield over time as the beds aged. Grain yield of wheat was similar in CTW and direct-drilled ('zero-till') wheat (DDW) on the flat. The 15N recovery in the wheat plants in all flat and bed treatments was similar. Straw mulch had no effect on yield or NUE of wheat. Recoveries of applied N in the wheat plants (27-38%) and soil (45-59%) were much higher than in rice. Total fertiliser N losses were much lower in wheat (mean 14-21%) compared with rice (mean 38-60%). After eight crops, soil organic C, total N and available K were significantly higher with straw mulch compared with no mulch. Permanent beds for RW seem to have limited potential under the soil and climatic conditions of Punjab, India, with current technology, even with full residue retention for both crops. Further research on permanent raised beds should focus on selection of rice and wheat cultivars that are better suited to beds; soil health issues such as nematodes and iron deficiency; weed control; N, water and residue management; and machinery development and practices.
