Productivity, water use efficiency and hydrology of wheat on beds and flats in Punjab, India

Summary Rice-wheat rotations are major contributors to both rice and wheat production in South Asia. However their sustainability is at risk due to associated problems including groundwater depletion, waterlogging and salinity, deteriorating soil physical conditions, declining soil organic matter content, groundwater and air pollution. Over the past decade it has been shown that wheat can be grown successfully on beds in rice-wheat systems, with many advantages including substantial irrigation water savings and higher yields. However, the nature of the water savings has not been quantified, and guidelines for irrigation management for wheat on beds, based on understanding of the system, have not been developed. This report presents the results of field experiments to evaluate irrigation scheduling options for wheat on beds in comparison with the recommended practice for conventionally tilled wheat. Replicated, small plot experiments were established on three soil types (sandy loam, loam, reclaimed sodic loam) in Punjab, India in November 2002. Irrigation treatments were commenced following an initial common irrigation 2-4 weeks after sowing. Irrigation of beds was to be scheduled when cumulative pan evaporation minus rain (CPE-R) reached 40, 60 or 80 mm (B40, B60, B80) and irrigation of flats when CPE-R reached 80 mm (recommended practice) or 120 mm (F80, F120). Irrigation application rates were 40 mm on beds and 80 mm on flats. Mid-season conditions were unusually overcast and cold, and the crops virtually stopped growing during the month prior to anthesis, resulting in lower than expected yields. Grain yields ranged from 3.5 to 3.6 t/ha on the reclaimed sodic loam, and 4.1 to 4.9 t/ha on the other two soils. Lodging prior to heading and waterlogging on the reclaimed sodic loam probably contributed to the lower yields. Rainfall was much higher than average and reasonably well distributed, and there were no differences in yield between irrigation treatments and layouts, with one exception. On the sandy loam yields on the flats were generally significantly higher than on the beds, by around 0.5 t/ha. This appeared to be due to failure of the crop on beds to produce sufficient tillers to compensate for the lower sowing rate associated with the wider row spacing between beds. The number of tillers per plant on the flats and beds on the sandy loam was similar, whereas at the other two sites the number of tillers per plant was 15-30% higher on the beds. The soil matric potential data suggest that the crop on the sandy loam suffered from water deficit stress during the period when tillers were forming as a result of delaying the first irrigation until 26 days after sowing, together with the more rapid drying of the topsoil on the beds. The recommended practice of delaying the first irrigation until 3-4 weeks after sowing may need to be revised to 2-3 weeks after sowing for wheat on beds on sandy loam soils. There was a substantial amount of residual soil water in all treatments at all sites at the time of harvest, more so on the two loam soils, largely due to the heavy rain (90-125 mm) prior to anthesis and over irrigation, particularly on the loam soil. The pattern of crop water extraction was quite different between the bed and flat treatments. The degree of drying at each site was relatively uniform to depth in the flat treatments. In contrast, drying was much greater in the surface layers of the beds, more so on the loam soils, where there was less drying at depth. The crops extracted water to at least 180 cm in all treatments on the sandy loam, and in the flat treatments on the two loam soils. However, extraction was confined to the upper 100 and 180 cm