The optimization of model ensemble composition and size can enhance the robustness of crop yield projections

Climate change significantly impacts agriculture, yet the influence of ensemble configurations (model composition and size) on crop yield projections and uncertainty remains unclear. This study investigates these influences using Global Gridded Crop Models (GGCMs) and Global Climate Models (GCMs) under future climate change. Cluster analysis identified distinct groups of ensemble members based on projected outcomes, revealing unique patterns in crop yield projections and uncertainty levels, particularly for wheat and soybean. Results suggest that approximately six GGCMs and 10 GCMs suffice to capture modeling uncertainty, while a cluster-based selection of 3-4 GGCMs effectively represents the full ensemble. The contribution of individual GGCMs to overall uncertainty varies regionally and by crop type, highlighting the importance of model selection for local applications. This research emphasizes optimizing ensemble configurations to identify primary uncertainty sources in crop yield projections.

Linchao Li, Bin Wang, Puyu Feng, Jonas Jägermeyr, Sennethold Asseng, Christoph Müller, Ian Macadam, De Li Liu, Cathy Waters, Yajie Zhang, Qinsi He, Yu Shi, Shang Chen, Xiaowei Guo, Yi Li, Jianqiang He, Hao Feng, Guijun Yang, Hanqin Tian, Qiang Yu