Abstract: 【Objective】 By applying climate projections based on 30 Atmosphere-Ocean General Circulation Models (AOGCMs) under representative concentration pathway (RCP) scenarios in the Coupled Model Inter-comparison Project Phase 5 (CMIP5), the effects of climate change on wheat yield in China were assessed in terms of ensemble method. 【Method】 The impact assessment of climate change on crops is typically based on daily data. However, significant uncertainties exist among the AOGCM outputs, particularly in daily data. In this paper, a pseudo global warming (PGW) method was assumed to be a linear coupling of contemporary weather fields and the difference component of climate perturbation signals by AOGCMs. CERES-Wheat model was employed to stimulate the wheat yield in the future and a probabilistic approach is used to address the uncertainties. 【Result】 Warming is expected in all representative stations during the wheat growth period. Temperature increase under the RCP8.5 scenario is higher than that under the RCP2.6 scenario. The temperature in the representative stations of winter wheat is projected to increase by 2.7-2.9℃, and increase by 3.0-3.3℃ in the representative stations of spring wheat at the end of the 21st century. The precipitation rate is projected to significantly increase in the future. Compared with the baseline, the observation data collected from 1996 to 2005 show that the climate-change-induced wheat yield reduced in all representative stations under irrigation conditions. The reduction probabilities increased with climate change. The irrigated yield reduction in the representative stations of spring wheat was greater than that in the representative stations of winter wheat. By the end of the 21st century, the yield in the representative stations of winter wheat is projected to be decreased by 2% under the RCP 2.6 scenario. The yield reduction will be decreased by approximately 6% under the RCP 4.5 scenario and decreased by 9% under the RCP 8.5 scenario with a probability of 85%. In the representative stations of spring wheat, yield will be decreased by 5% under the RCP 2.6 scenario, by more than 8% under the RCP 4.5 scenario, and by more than 15% under the RCP 8.5 scenario with a probability of 90%. In comparison with the baseline, the rain-fed yield in the representative stations of winter wheat will be increased significantly. By the end of the 21st century, the yield in winter wheat is projected to be increased by more than 21% under the RCP 2.6 scenario, more than 22% under the RCP 4.5 scenario, and more than 25% under the RCP 8.5 scenario with a probability of 90%.【Conclusion】The ensemble of daily data were obtained through the PGW method, which efficiently reserve the contemporary weather information, particularly that of extreme weather events. Effects of climate change on wheat yield under RCP2.6, RCP4.5, and RCP8.5 scenarios were assessed through the ensemble method. The results indicate that, with the increasing greenhouse gas emissions, the climate-change-induced yield-reduction probabilities of irrigated wheat in China gradually increased. Rain-fed wheat yield will be increased in the future, with large uncertainties.

Key words: CMIP5 , wheat yield , ensemble simulation , probabilistic assessment , climate change