Bayesian model averaging of the RegCM temperature projections: a Canadian case study. Issue 2 (26th October 2021)

• Record Type: Journal Article
• Authors: Song, Tangnyu; Huang, Guohe; Wang, Guoqing; Li, Yongping; Wang, Xiuquan; Lu, Chen; Shen, Zhenyao
• Is Part Of: Journal of water and climate change. Volume 13:Issue 2(2022)
• Page Start: 771
• Page End: 785
• Publication Date: 2021-10-26
• Subjects: BMA -- Canada -- ensemble projections -- RegCM -- temperature; Water -- Periodicals; Hydrology -- Periodicals; Climatic changes -- Periodicals; Climatic changes; Hydrology; Water; Electronic journals; Periodicals; 333.9116
• DOI: 10.2166/wcc.2021.393
• Languages: English
• ISSNs: 2040-2244

Abstract

Abstract: The choices of physical schemes coupled in the Regional Climate Model version 4 (RegCM4), the input general circulation model (GCM) results, and the emission scenarios may cause considerable uncertainties in future temperature projections. Therefore, the ensemble approach, which can be used to reflect these uncertainties, is highly desired. In this study, the probabilistic projections for future temperature are generated at 88 Canadian climate stations based on the developed RegCM4 ensemble and obtained Bayesian model averaging (BMA) weights. The BMA weights indicate that the RegCM4 coupled with the holtslag PBL scheme driven by the HadGEM can provide relatively reliable temperature projections at most climate stations. It is also suggested that the BMA approach is effective in simulating temperature over middle and eastern Canada through taking advantage of each ensemble member. However, the effectiveness of the BMA method is limited when all the models in the ensemble cannot simulate the temperature robustly. The projected results demonstrate that the temperature will increase continuously in the future, while the temperature increase under RCP8.5 will be significantly larger than that under RCP4.5. HIGHLIGHTS: The RegCM simulations coupled with different PBL schemes driven by multiple GCMs have been conducted. The probabilistic projections for future temperature are generated at climate stations over Canada through the Bayesian model averaging method. The temperature will increase continuously in the future, and the temperature increase under RCP8.5 scenario would be significantly larger than that under RCP4.5 scenario.