Current food trade helps mitigate future climate change impacts in lower-income nations

Materials and methods Conceptual framework Every country (y) consumes calories (C) from crop (i), which can be written as: Cy;i ¼ Py;i Ey;i þ Iy;i ð1Þ where Py,i are the calories produced within the country, and E and I are the sum of all exports and imports of crop y to/from other countries. We consider production net of exports (Py,i−Ey,i) as production for domestic use P*(y,i) Cy;i ¼ P∗ y;i þ Iy;i ð2Þ The impact of climate change on a nation’s consumption (ΔC) is: DC ¼ DP∗ þ DI ð3Þ Our analysis compares the three terms of Eq 3 in relative terms, i.e., we compare ΔC/C (con- sumption impact), ΔP*/P*(domestic production impact), and ΔI/I (import impact), (see S1 Text for the mathematical derivation relating these 3 terms and S2 Table for the summary of variables). Climate change impacts on crop productivity We used estimates of climate impacts on global crop productivity from the latest process- based crop model simulations conducted as part of the Agricultural Model Intercomparison and Improvement Project’s (AgMIP) Global Gridded Crop Model Intercomparison (GGCMI) [24]. We used crop productivity data from 12 global crop models, forced by future climate sim- ulations from 5 global climate models from the Coupled Model Intercomparison Project phase 6, using the Shared Socioeconomic Pathways (SSP) 585 climate scenario, bias-adjusted and downscaled by Inter-Sectoral Impact Model Intercomparison Project [27]. GGCMI provides 60 crop-climate model simulations per crop, running from pre-industrial times to 2100. For additional details see Ja¨germeyr et al 2021 [24].

Global agri-food trade data We used bilateral trade data for wheat, maize, and rice for the year 2010 from Schwarzmueller and Kastner 2022 [23], who designed an approach to link consumption impacts to origins of production. They used bilateral trade matrices and crop production data (from FAOSTAT) and traced crop supply in a country to where the crops were grown, correcting for re-exports of crop commodities. For instance, if maize is exported from France to the Netherlands where it is partly processed into maize germ oil, which is further exported to Austria, the corrected data will link calorie supply in Austria back to crop production in France. This data is widely used in estimating the environmental impacts of food consumption and telecoupling research [28–32]. We converted the bi-lateral trade data from tonnes to calories using conversation fac- tors from GENuS [33]. Domestic production impact We estimated the domestic production impacts on calorie supply from climate change for each country from all 60 gridded crop-climate model simulations for the three crops. For each simulation we first estimated 10-year mean crop yields (at pixel level) around the year when global mean temperatures exceed 1˚C, 2˚C, and 3˚C above pre-industrial levels in the PLOS ONE Current food trade helps mitigate future climate change impacts in lower-income nations PLOS ONE | https://doi.org/10.1371/journal.pone.0314722 January 3, 2025 3 / 16