Abstract
Achieving "zero hunger," the second sustainable development goal, is challenging due to climate change, weather and climate extremes, and unabated human population growth. Understanding likely changes in spatial crop suitability and yield for maize, rice, wheat, and soybean using the FAO-Ecocrop and eXtreme Gradient Boosting (XGBoost) models is imperative for sustainable agri-food systems. Our results reveal a northward shift in climate suitability as the optimal and suitable regions decrease and the unsuitable and marginal areas expand more profoundly in the far future (2061-2100) than in the near future (2021-2060) relative to the 1970-2000 baseline for all crops under all considered emission scenarios. Heterogeneous yield changes are observed in most parts of the globe, though with consistent regional outcomes. Across all emission scenarios, projections for all four staple food crops show spatially consistent declines in climate suitability and yield within major tropical producing zones, contrasted by yield gains in extratropical regions. Currently, rice can potentially be optimized in over 40% of the agricultural land, followed by wheat (30%), maize (25%), and soybeans under the integrated food systems. Future climate is anticipated to heterogeneously alter these land fractions by ~12%, ~8%, and ~13% in maize, rice, and wheat by the end of 21st century, consequently affecting the global food trade and food security. To reduce crop yield uncertainties, climate-smart agricultural interventions, improved agronomic practices, and the optimization of technological advancements are paramount in the global south. Improvement in climate models' parameterization schemes and integration of multi-crop modelling approaches is imperative in designing food policies that bolster global food security.