Improving crop production for food security and improved livelihoods on the East India Plateau. I. Rainfall-related risks with rice and opportunities for improved cropping systems

Rainfed transplanted rice (Oryza sativa) is the staple crop of the East India Plateau (EIP), where it is low yielding and drought-prone despite high annual rainfall (>1200 mm). Although grown traditionally on lowlands associated with drainage lines, population pressure has forced rice onto terraced slopes (medium-uplands) that now comprise >80% of the rice area, and the only rice land for many families. Crop monitoring, soil water measurement and soil water-balance modelling in Pogro watershed (West Bengal) were used to explore rainfall-related risks associated with rice-fallow on medium-upland and to examine opportunities for using rainfall more effectively. The analysis was extended to three more EIP locations by using the model with long-term rainfall. Rice depends on sustained ponding for transplanting and good yields, but in Pogro, failure to meet this condition on medium-uplands led to delayed or failed transplanting and/or periodic or premature draining of fields in five years from 2005 to 2011. Modelled ponding duration was more variable than rainfall. Most farmers have adapted to variable ponding by growing medium-duration varieties on medium-uplands, rather than the longer types grown in lowlands. However, the average ponding duration of 65 days over all four locations was well short of the ~90 days required. Even shorter-duration varieties would provide only a partial solution as ponding was <50 days in ~25% of years, and transplanting impossible in 10% of years. Watershed development (WSD) is unlikely to deliver food security from transplanted rice, because in dry years there is little or no runoff to capture for irrigation and shallow groundwater becomes available too late; although modelling confirmed the potential for WSD to promote post-monsoon cropping in wetter years. Significantly, in every year at all locations there was enough soil water for non-flooded crops (rarely <140 continuous days of >30 mm available soil water). Duration of available soil water was the least variable measure of water security, confirming that perceptions of ‘drought’ arise from experience with transplanted rice that depends on ponding. We conclude that ‘aerobic’ (un-puddled) rice culture on medium-uplands should provide food security from this staple crop. Modelling identified opportunities to intensify and diversify cropping systems with manageable climate risk, even without WSD, by opportunistically using residual water after rice, shorter-duration varieties to maximise residual water, and minimal supplementary irrigation from shallow groundwater that apparently is recharged every year. The suggested developments have implications for river basin hydrology and WSD that require research and a reappraisal of policy.