Abstract: The expected negative effects of global warming on crop production can be mitigated by changing the planting dates, which is both low-cost and easy to implement. The parameter set of the localized spring wheat model in the Qingtongxia irrigation region (QTXIR) was calibrated using a two-year field experiment. The changes of spring wheat yield and irrigation water requirement (IWR) under two future climate scenarios (SSP2-45 and SSP5-85) were analyzed based on 7 bias-corrected global circulation models (GCMs) and the well-tested SWAP model. Simultaneously, the yield and IWR of spring wheat in QTXIR in response to changing the sowing date were also quantitatively analyzed. The results indicate that the spring wheat growth in the QTXIR can be effectively simulated by the parameter-adjusted SWAP model. Under the SSP2-45 scenario, the multi-model average maximum temperature in the QTXIR in the next two periods (2021−2050 and 2051−2080) will increase by 1.6℃ and 2.6℃ respectively relative to the historical baseline period (1991−2020), while under the SSP5-85 scenario, the maximum temperature will increase by 1.8℃ and 3.6℃ respectively in the next two stages. The length of spring wheat growth duration will continue to shorten with the increasing temperatures in the future under current sowing date and variety, with a maximum shortening of 14.2 days occurring between 2051 and 2080 under the SSP5-85 scenario. The multi-model average yield for spring wheat will decrease by 9.6% and 12.9% for the next two periods under the SSP2-45 scenario and by 12.1% and 17.2% for the corresponding periods under the SSP5-85 scenario. At the same time, the changes in irrigation water requirements for different scenarios and periods are relatively small, not exceeding 3.5%. Sowing in advance can effectively reduce the yield loss of spring wheat in the QTXIR, but it cannot fully offset the negative effects of climate change. Under the SSP2-45 scenario, in the next two periods, sowing 23 days and 33 days in advance respectively can control the yield loss to 1.5% and 5.3 %. Under the SSP5-85 scenario, when sowing 30 days and 42 days ahead, the yield losses can be controlled to 2.9% and 5.4% respectively. Under the SSP5-85 scenario, the yield loss can be controlled to 2.9% and 5.4% when sowing 30 and 42 days ahead, respectively. Under the optimized sowing date, the harvest date for spring wheat in the QTXIR will be forward by an average of 5 days. At the same time, IWR for spring wheat will increase significantly in all future scenarios and periods, with rates ranging from 4.0% to 8.0%.

Key words: CMIP6, Climate change, Spring wheat, Adaption, Sowing date adjustment