Abstract
Due to salinity intrusion in the dry season, most of the rice land in the coastal zones of Bangladesh is grown with a single crop during the rainy season (aman), mainly with low-yielding local varieties. This study aimed to provide the basis for increasing the productivity of the coastal zones by developing technical packages that would allow double croppings (an aus-aman cropping system) in the rainy season. Under this system, the aus rice crop is sown at the onset of the rainy season, followed by aman rice. A field experiment (Experiment 1) was conducted in 2006-2008 in the silty clay soil of Batiaghata, Bangladesh, to test the hypothesis that high and stable productivity can be achieved by (i) dry seeding of the aus crop, (ii) using short-duration rice varieties and supplemental irrigation during crop establishment of the aus crop, and (iii) replacing local rice by a high-yielding variety in the aman season. The experimental layout was a split-plot design with three water regimes (rainfed, I; supplemental irrigation, I2; and full irrigation, I3) in the main plots and three seeding dates (D1 = early (5-10 April); D2 = mid (15 - 20 April); and D3 = late (25-30 April) in the subplots. In the aman season, two high yielding rice varieties (late-maturing and photoperiod-sensitive BR11 and early-maturing PSVT2) were transplanted 1 week after the aus crop harvest. The findings of Experiment 1 led to a supplemental experiment (Experiment 2) in 2009. The aim was to test the hypotheses that late transplanting (when rainfall has become steady) will secure good aus establishment, and that late planting of aman rice (due to late aus rice harvest) will maintain high yield with use of a suitable photoperiod-insensitive variety. The aus rice (OM1490) was established in a strip-plot design with establishment method (direct seeding [M] and transplanting [M]) in the main plots and two seeding dates (30 April [D] and 10 May [D2]) in the subplots. The photoperiod-sensitive variety BR46 and - insensitive variety BR49 were transplanted in the aman season 1 week after the harvest of the transplanted aus crop. Furthermore, the long-term probabilistic rice yield and water requirements of the aus crop were estimated by the crop model ORYZA2000. The effects of irrigation and seeding date treatments on aus crop yield varied across years because of variation in rainfall distribution. Aus crop yields under I, and I were comparable (4 t ha1) with but greater than those of I, (2 t ha1) when the aus crop was seeded during dry spells. But the yield of I2 was similar to that of I, when there was sufficient rainfall after seeding. I required only 100-200 mm of irrigation water, whereas I needed 660-1042 mm. The average incremental water productivity varied from 0 (in the year of adequate rainfall) to 1.3 kg grain ha1 mm'. Treatment D3 delayed the planting of aman rice, thus decreasing the growth duration of PVST2 and reducing its yield compared with D1 and D2. BRII did not change its growth duration and had yield similar to that (>4 t ha1) of PVST2 in D,. In Experiment 2, grain yield of aus rice under transplanting and direct seeding methods were similar for the first planting on 30 April. In the late seeding treatment, high rainfall submerged the transplanted rice, reducing its yield greatly, as the transplanted strips were kept in the low-lying part of the experimental plot. For both establishment dates, BR49 had higher yield than BR46. Long-term simulation results of ORYZA2000 confirmed that, to ensure 75% probability of good crop establishment and yield under rainfed condition, the aus crop should be seeded in late April. With earlier seeding dates, rice with supplementary irrigation (50-100 mm of water) performed better than rainfed rice, with the simulated yield of OM1490 approaching 5 t ha'. In favorable conditions, the aus-aman cropping system gave yields ranging from 8 t ha yr (OM1490 aus, ID2 followed by BR11) to 9.6 t ha1 yr1 (late seeding of OM1490 followed by BR49). The aus crop, however, was subjected to heavy pest infestation. This work showed that, with timely establishment of the aus season crop (seeding near the end of April), and the use of a modern, short- duration rice variety in the aus season and of a medium-duration variety in the aman season, the aus- aman system can yield as much as 8-9 t ha' yr'. The stability of the aus-aman cropping system can be enhanced by supplementary irrigation (where limited groundwater is available) to ensure good crop establishment of the aus crop. In water-scarce areas, seeding of the aus crop can be delayed until 10 May (to ensure adequate rainfall for establishment), followed by growing a photoperiod-insensitive rice variety in the aman season. Late seeding, however, may expose the aus crop to submergence during its crop establishment period.