Prioritizing agronomic practices and uncertainty assessment under climate change for winter wheat in the loess plateau, China

Abstract

CONTEXTEnhancing the climate resilience of local food systems with adaptation options related to cultivar, irrigation, sowing, and fertilization presents significant opportunities for ensuring food security under climate change. The climate-crop modeling method is one of the main ways to customize climate adaptation strategies. However, there is a current lack of prioritization and uncertainty assessment regarding the potential of various adaptations feasible to local farmers. OBJECTIVEThis study aimed to investigate the prioritizing agronomic practices (shifting the thermal time of cultivars, irrigation, topdressing schedules, and sowing date) and uncertainty assessment under climate change for winter wheat in the Loess Plateau, China. METHODSHence, this study integrated eight crop models (CMs), six global climate models (GCMs), and four different types of adaptation options under two Shared Socioeconomic Pathway (SSP) 245 and SSP585 emission scenarios. We assessed the potential for adaptation during the periods of 2031–2060 and 2071–2100 at three representative sites (Changwu, Linfen, and Yangling) in the Loess Plateau, China. The ANOVA analysis was used to quantify the uncertainties in wheat yield projections caused by adaptation measures (ADP), CM, GCM, and climate change scenarios (Scen). RESULTS AND CONCLUSIONSWe found that optimizing irrigation and topdressing timing positively impacted winter wheat yields more than adjusting planting dates or prolonging the reproductive stage across all three sites. By implementing irrigation during the booting or flowering stage, the ensemble of climate-crop models projected yield increases ranging from 7.1% to 8.5% at Changwu, 18.2% to 20.2% at Linfen, and 13.5% to 17.3% at Yangling. Crop models dominated the projection uncertainty, with values over 50% at all three sites. However, adaptation strategies would dominate the uncertainty in yield projection when the number of crop models used was less than five. Furthermore, the uncertainty in yield projection due to individual crop models varied depending on the study sites and adaptation options. SIGNIFICANCEOur findings could provide valuable guidance to modelers in selecting appropriate climate-crop models to develop effective adaptation options for addressing climate change challenges. Additionally, our findings will provide guidance to producers in the Loess Plateau to optimize food production under climate change.