Refined crop modelling reveals intensifying impacts of droughts and heatwaves on global wheat productivity.

Extreme weather events, including heatwaves and droughts, are becoming more frequent and intense, yet their impact on crop productivity remain highly uncertain. Here we jointly improve the structure and parameterization of a crop model to enhance its ability to reproduce climate impacts on global wheat productivity at 10-km resolution. Compared with a multimodel ensemble, the improved parameterization increases model skill by 30%, 20% and 7% in simulating global wheat yield losses from droughts, heatwaves and their compound events, respectively, during 1981-2015. Over the same period, the improved modelling reveals global average wheat yield losses of 9.7%, 5.8% and 13.3% caused by droughts, heatwaves and their compounds, respectively. Further analysis underscores that irrigation represents an effective adaptation option to offset these adverse impacts; however, such countervailing effects has markedly weakened over time with disaster intensifying. Our framework enhances crop models' ability to assess the impacts of extreme weather on global crop productivity, supporting climate-risk assessment and the development of effective adaptation strategies.