Time is money: Spatially explicit system analysis for rice-wheat cropping systems of Eastern Indo-Gangetic Plains, India

Late sowing of wheat is a major reason for poor yields in eastern India (Bihar and Eastern Uttar Pradesh) due to a shorter growing period, and risk of terminal heat stress. Despite big losses and widespread awareness, late sowing of wheat is widely prevalent in the region. Why? Most wheat is sown after harvesting monsoon rice from the same plot. Later rice harvests interfere with timely wheat sowing. The aim of the study is to analyse the interdependence between the rice and wheat to optimize crop establishment dates at the cropping systems level for increasing yields, revenues and mitigating risks. We collected production practices data from 5021 plots sequentially cultivated with rice and wheat in the State of Bihar and the eastern districts of Uttar Pradesh. We use this unique data to implement a multivariate Bayesian geo-additive model and risk-return optimization framework to determine rice and wheat sowing dates that give the highest risk-adjusted economic gains to farmers. Early transplanting of rice and early sowing of wheat have spatially differentiated yield, revenue, and risk (minimal variance) benefits. We find that early transplanting of rice (between June 20 and July 20) and wheat (between November 1 and November 15) leads to a revenue gain of 5000–10,000 Rs ha −1 (∼62.5–125 US$ ha −1 ) at minimal risk and no revenue trade-offs (non-negative correlation). Conversely, late transplanting of rice has negative effects on correlation of rice and wheat yields therefore leading to a yield and revenue tradeoff. Evidence of spatially differentiated dependence between rice and wheat yield systems implies that analysing these crops separately may be suboptimal. Spatial intelligence on cropping system inter-dependence can help farmers select the appropriate crop management practices (e.g., variety duration, irrigation, fertilizer application) and adjust their sowing dates based on local conditions and constraints, thereby optimizing yields and incomes in the rice-wheat system. It can also help policy makers in implementing spatially differentiated entry points for increasing yields and farm incomes at minimum risks. • Most farmers know the optimal and recommended sowing dates for rice and wheat. • They deviate from these due to monsoon onset delay, groundwater, and access to irrigation. • There is spatially differentiated interdependence in yields between rice and wheat. • A multi-crop risk-return analysis shows that early sowing of rice and wheat gives highest returns at minimal risk.