Negative year-to-year agricultural yield extremes projected to occur more frequently under global warming

Studies on projected agricultural yields focus on end-of-century scenarios. Simulations from the Global Gridded Crop Model Intercomparison (GGCMI) Project phase 3b show conflicting results for global and regional changes of different crops by the end of the century. Here, we interrogate the same simulations, focusing on year-to-year variations of agricultural yields in the important staple crops maize, rice, soybean and wheat.

An ensemble of GGCMI models shows a larger agreement on the variations of crop yields than for the long-term trend. Year-to-year variations of projected crop yields become more pronounced over time, especially so for negative crop yield anomalies. As a result, the frequency of negative global crop yield extremes increases with global warming. We show that these negative yield extremes may occur for individual or multiple crops at the same time, and may originate from individual or multiple regions. North America dominates global maize and soybean yield extremes (57% and 44% of all significant global extremes, respectively), and South East Asia and South Asia are important for rice extremes (24% and 22%, respectively), while regional results are inconclusive for wheat. Multi-crop extremes occur most commonly for the combination of maize and soybean, and are dominated by the North America region. Based on these findings, we show that depending on the region and crop, persistent spring or summer drought, cold or heat can be associated with years of global and regional negative agricultural yield extremes.

Our results show how specific climatic boundary conditions can lead to year-to-year extremes in important staple crops, highlighting the potential to anticipate such events in the future.