Future rice systems must produce more grain while minimizing the negative environmental impacts. A key question is how to orient agricultural research & development (R&D) programs at national to global scales to maximize the return on investment. Here we assess yield gap and resource-use efficiency (including water, pesticides, nitrogen, labor, energy, and associated global warming potential) across 32 rice cropping systems covering half of global rice harvested area. We show that achieving high yields and high resource-use efficiencies are not conflicting goals. Most cropping systems have room for increasing yield, resource-use efficiency, or both. In aggregate, current total rice production could be increased by 32%, and excess nitrogen almost eliminated, by focusing on a relatively small number of cropping systems with either large yield gaps or poor resource-use efficiencies. This study provides essential strategic insight on yield gap and resource-use efficiency for prioritizing national and global agricultural R&D investments to ensure adequate rice supply while minimizing negative environmental impact in coming decades.

Crop models are ideally suited to quantify existing climatic risks. However, they require historic climate data as input. While daily temperature and rainfall data are often available, the lack of observed solar radiation (Rs) data severely limits site-specific crop modelling. The objective of this study was to estimate Rs based on air temperature solar radiation models and to quantify the propagation of errors in simulated radiation on several APSIM/ORYZA crop model seasonal outputs, yield, biomass, leaf area (LAI) and total accumulated solar radiation (SRA) during the crop cycle. The accuracy of the 5 models for estimated daily solar radiation was similar, and it was not substantially different among sites. For water limited environments (no irrigation), crop model outputs yield, biomass and LAI was not sensitive for the uncertainties in radiation models studied here.

The upland rice (UR) cropped area in Brazil has decreased in the last decade. Importantly, a portion of this decrease can be attributed to the current UR breeding programme strategy, according to which direct grain yield selection is targeted primarily to the most favourable areas. New strategies for more-efficient crop breeding under non-optimal conditions are needed for Brazil’s UR regions. Such strategies should include a classification of spatio-temporal yield variations in environmental groups, as well as a determination of prevalent drought types and their characteristics (duration, intensity, phenological timing, and physiological effects) within those environmental groups. This study used a process-based crop model to support the Brazilian UR breeding programme in their efforts to adopt a new strategy that accounts for the varying range of environments where UR is currently cultivated. Crop simulations based on a commonly grown cultivar (BRS Primavera) and statistical analyses of simulated yield suggested that the target population of environments can be divided into three groups of environments: a highly favorable environment (HFE, 19% of area), a favorable environment (FE, 44%), and least favourable environment (LFE, 37%). Stress-free conditions dominated the HFE group (69% likelihood) and reproductive stress dominated the LFE group (68% likelihood), whereas reproductive and terminal drought stress were found to be almost equally likely to occur in the FE group. For the best and worst environments, we propose specific adaptation focused on the representative stress, while for the FE, wide adaptation to drought is suggested. ‘Weighted selection’ is also a possible strategy for the FE and LFE environment groups.