Responses of spring wheat yield and growth period to different future climate change models in the yellow river irrigation area based on CMIP6 and WOFOST models

It is crucial to understand and forecast future climate change and its impact on regional crop yield. This study employed the optimized World Food Studies (WOFOST) model and Coupled Model Intercomparison Project Phase 6 (CMIP6) data to assess the effects of climate change in different emission scenarios [e.g., ssp126 [sustainable development with low greenhouse gas (GHG) emissions], ssp245 (balanced development with medium GHG emissions), and ssp585 (conventional development with high GHG emissions)] on the spring wheat yield and growth period in Northwest China. Lasso regression, Support Vector Machine (SVM), Recurrent Neural Networks (RNN), Random Forests (RF), and RNN showed superior optimization performance for radiation amount surface downwelling shortwave radiation flux (RSDS), maximum temperature (Tmax), minimum temperature (Tmin), wind speed (sfcWind), and precipitation (PR). These optimizations decreased the root mean square error (RMSE) of RSDS, Tmax, Tmin, sfcWind, and PR by 1.0285, 2.6396, 2.7839, 2.3763, and 0.3020, respectively, compared to the original data. Long-term annual PR will increase by 415 mm in the ssp126 scenario, whereas the increases were estimated to be 257 mm and 249 mm under the ssp245 and ssp585 scenarios, respectively. Regarding temperature, Tmax and Tmin are expected to rise by 0.97 °C and 1.08 °C, respectively, under the ssp126 scenario. These values are also projected to be 2.29 °C/1.70 °C and 1.81 °C/3.36 °C for Tmax/Tmin under the ssp245 and ssp585 scenarios, respectively. The optimized crop parameters can effectively improve the yield simulation accuracy of the WOFOST model and reduce RMSE. Yinchuan and Huinong are projected to experience an increase of 19.8 % and 15.8 % in their average yield under the ssp585 scenario in the short-term future, compared to the average yield during the baseline period. Qingtongxia is expected to reach a 20.2 % increase under the ssp126 model, while Zhongning will undergo an increase of 12.3 % under the ssp585 model, and Litong will observe an 11.2 % increase under the ssp245 model. Qingtongxia showed the highest long-term production growth, with increases of 18.0 % and 15.3 %, respectively, under the ssp126 and ssp245 scenarios. Higher temperatures, radiation levels, and PR are typical features of the future climate in the study area. The CMIP6 data simulations and the optimized WOFOST model predict that spring wheat yield will increase soon but will not increase as much in the long run.