China’s Rice Yield Sensitivity to Extreme Cold is Underestimated

Climate change is projected to increase the frequency, intensity, and spatial extent of extreme climate events. Among these, extreme cold impacts on crop yield are often overlooked from historical and future analyses. To address this issue, a unique national dataset detailing 2,490 field-identified extreme cold days at 212 sites was assembled to quantify stage-specific crop responses to extreme cold. Results show that extreme cold  affected 27% of China’s rice seasons during 1999-2012, resulting in an average yield reduction of 12.1±3.2%. This is mainly attributed to extreme cold during the transplanting-stem elongation and the heading-flowering stages, which reduces the total grain number per panicle and yield. In contrast, current global gridded crop models underestimate the cold sensitivity by 60% and a board range of model sensitivity. The constrained estimates show with >95% probability that rice yield would be reduced by extreme cold in stage of transplanting-stem elongation (−3.8±1.2% day⁻¹), heading-flowering (−3.6±1.0% day⁻¹), and milking grain-mature grain (−1.6±0.9% day⁻¹). Uncertainties associated with modelled sensitivities were reduced by 36-44%. The national rice yield losses decrease by 9.1 ± 2.4% under the scenario of SSP1-2.6 by the end of this century, approximately twice as large as the unadjusted model estimates. This research highlights the underappreciated role of extreme cold in reducing crop yield under climate change.