Decision-making for rabi cropping in the High Barind Tract of Bangladesh: a farmer perspective

A farmer perspective on improving agricultural productivity in the High Barind Tract was developed through the analysis of farmer decision-making at both the plot and whole-farm levels. A decision-tree was constructed to model the decision to sow dryland rabi crops. The tree incorporated 11 separate decisions that included a combination of climatic, physical, and socioeconomic variables. A validation exercise showed that the tree correctly predicted 78% of the outcomes. Household survey data were used to analyze socioeconomic factors that influenced rabi cropping. Harvest dates were the same for both long-duration Swarna and short-duration MVs (modern high-yielding varieties), suggesting that farmers minimized transaction costs when hiring harvest labor. The harvest labor contract delayed rabi sowing by one week because of the need for straw to dry. Resource-poor farms were less likely to grow MV T. aman rice because they had more highland with poorer soils and because they needed to maximize yields on sharecropped plots. Rabi cropping was also most intensive on resource-poor farms because of the need to make the best use of limited land. Farmers recognized the potential to increase the area under rabi crops if rice was harvested earlier. However, time of harvesting was determined by Swarna, which occupied 80% of the area planted to T. aman rice. We conclude that rice-based interventions will have limited impact on agricultural productivity in the rabi season until breeders have developed a variety with a shorter field duration that will be as widely adopted as Swarna.