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

imports. We intentionally set the global zero cereal import dependency constraint as a thought experiment to explore the impact of global trade restrictions on multi-sector outcomes. We are aware that this extreme experiment may not occur in the real world; however, our objective is not to perform a specific policy analysis but rather to analyze and understand the linkages embedded in the coupled systems using this experiment. 2 Materials and methods GCAM version 5.3, available at https://github.com/JGCRI/gcam-core/releases, is modified and employed in this study. GCAM is an open-source global-scale integrated model that cap- tures economic decisions and dynamic interactions between multiple systems. It is a market equilibrium model, in which the prices for all markets are solved simultaneously, such that supply equals demand. The model is well validated and documented with carefully designed model structure, open-source data inputs and model codes that support reproducibility. The model calibration periods span from 1975 to 2015. The final calibration year, 2015, is also called the base year, from which projections at 5-year time steps are made through 2100 to model long-term trends of different system outcomes. The model operates at 32 geopolitical regions for energy and economic systems overlayed with 235 water basins, resulting in 384 regions for water and land. Land use for competing purposes (e.g., crop land versus forest and land allocation among different crops) with price-induced intensification (e.g., irrigation) are

included in the model’s market equilibrium to balance against agricultural demand (e.g., crop demand) with endogenously solved prices and quantities. International trade for agricultural commodities (including crops) among the 32 geopolitical regions is modeled with regionally differentiated markets (i.e., different regions can have different exporting prices for the same commodity). and Armington-style preferences between domestic and imported commodities [24, 25]. Regional exports and imports are linked through one global trading pool, where regional trading flows into and out of the pool (i.e., export to and import from the pool) are traced for each agricultural commodity. Thus, the global importing price reflects the producer prices from each exporting region weighted by the exporting quantities and all regions import from the same global pool with the same global price. We do not model the bi-lateral import and export explicitly in this study. GCAM v5.3 models 11 aggregate crop commodities, including four cereal commodities (i.e., corn, rice, wheat, and other grains) and seven other crop commodities, which cover all Food and Agriculture Organization (FAO) reported crops (See S1 Table in S1 File for a list of GCAM agricultural commodities). The food demand model is derived based on neoclassical econom- ics, which emphasizes the importance of consumers’ utility when economic decisions are made (e.g., purchasing food). The responses of food demand in the model to changes of consumer

income and commodity price (i.e., income and price elasticities) evolve as income varies. In this model, demand is modeled for staple and non-staple foods. Staple food demand is supplied by caloric consumption of crops including cereals as well as roots and tubers, while consumption of all other agricultural commodities for food contributes to non-staple food demand. The food demand model is calibrated and cross-validated to empirical observations [26]. To introduce a perturbation to the human international trade system, we enforce a net zero cereal import requirement in GCAM, measured by the cereal import dependency ratio [CIDR; 27], which is the ratio of domestic net import over the sum of domestic production and the net import of cereals (Eq 1). The ratio can be deduced as one minus the ratio of domestic produc- tion over domestic consumption of cereals in total (Eq 2). Under global cereal import depen- dency constraints, we enforce the total of cereal production (in metric tons) in a region to be equal to or greater than the total of cereal consumption in the region, regardless of whether 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 3 / 18