Ensemble estimates of future extreme event impacts on global food production

Abstract:

Future changes in weather variability can affect yield fluctuations and global food system stability more than long-term mean climate change impacts. Extreme events such as heat waves and droughts are expected to become both more severe and more frequent, but associated implications for agricultural productivity remain insufficiently understood. Here we use new CMIP6 climate model projections from 1850 to 2100 to drive an ensemble of state-of-the-art mechanistic crop models at global scale to evaluate historical and future impacts of heat waves and droughts on the productivity of various staple crops. Climate projections are bias-adjusted to preserve mean and variability, based on a new quantile-mapping approach developed within the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP). Extreme event selection is based on a consistent rule-based metric, largely matching EM-dat disaster records. We evaluate multiple RCP trajectories, not only against historical conditions, but also against preindustrial control runs by the same climate model spanning the entire time period until 2100, which allows for climate change attribution. Results robustly show that the frequency of heat waves and agricultural droughts significantly increases under unabated climate change. Composite analyses indicate that yield losses associated with these events substantially increases as well, irrespective of the uncertainty linked to CO₂ fertilization. Hotspots of adversely changing heat wave and drought impacts include major breadbasket regions in the U.S., Europe, and South Asia.