Agriculture, bioenergy, and water implications of constrained cereal trade and climate change impacts

trade happens or not (i.e., CIDR�0). Note that in our model, crop consumption includes food, non-food, bioenergy, feed, and others (e.g., wastes). As we are not looking at shorter- term shocks that might cause regions to add or subtract to their stocks, we do not model crop storage change over time. This zero or negative CIDR means a region can satisfy its needs from domestic production, which, when enforced in all regions, is expected to lead to decreased participation in global markets (i.e., reduced trade). In contrast, a positive CIDR means a region needs to import cereals to satisfy its domestic cereal consumption. Because every region’s CIDR must be �0, no region can have a positive CIDR, so every region’s CIDR equals zero under this constraint. We enforce this constraint by introducing a cereal production credit created by domestic cereal crop producers. Each unit of domestic production of cereal crops creates one unit of the credit. This credit must be paid for each unit of cereal crop consumed in a region, regardless of whether it is domestically produced or imported. The payment of the credit is returned to the domestic cereal crop producers as a subsidy, which encourages domestic production to meet domestic consumption. cereal import dependency ratio ¼ cereal imports cereal exports cereal production þ cereal imports cereal exports ð1Þ Given market clearance conditions, where consumption equals to production plus net imports, Eq 1 can be transformed to the following: cereal import dependency ratio ¼ net cereal imports cereal consumption

¼ cereal consumption cereal production cereal consumption ¼ 1 cereal production cereal consumption ð2Þ The cereal import constraint in GCAM allows us to analyze a range of outcomes relative to the reference (i.e., without this constraint). As GCAM captures economic decisions for various activities across different sectors and regions, we are able to analyze the spatial pattern of model outcomes (e.g., crop land allocation, crop production, global trade, and consumption) under the influence of global zero cereal import dependency and the associated economic changes for producers and consumers in each region. We also explore the implications for irri- gation water consumption and crop-based ethanol production (including corn ethanol and sugarcane ethanol) in regions where these are consumed. Climate change impacts on crop yields are also considered in combination with global zero cereal import dependency. We use the global gridded climate outputs from the Hadley Centre Global Environmental Model [HadGEM; 28, 29] under RCP8.5 and processed them through Persephone [30] to estimate the corresponding crop yield changes under the climate impact for all GCAM crops. Persephone emulates crop yield changes based on yield response sensitivities seen in the Agricultural Model Intercomparison and Improvement Project [AgMIP; 31]. Coordinated Climate-Crop Modeling Project (C3MP) dataset. The cli- mate change impacts on crop yield are incorporated into GCAM as a multiplier to the base- line yield. The corresponding harvest area for each crop is endogenously decided in GCAM,

which uses a nested logit architecture in the land-use sector. For more details, please refer to the model documentation. This one climate scenario under RCP8.5 was selected to understand the underlying mechanism of spatially varying climate change impacts infiltrating across the coupled PLOS ONE Agriculture, bioenergy, and water implications of constrained cereal trade and climate change impacts PLOS ONE | https://doi.org/10.1371/journal.pone.0291577 September 15, 2023 4 / 18