AUTHOR=Pan Xiaoyu , Liu Chengyan , Wang Zhaomin , Xu Zhongfeng , Liang Xi , Li Xiang 

TITLE=Evaluating the performance of CMIP6 models in the Southern Temperate Zone with a multivariable integrated evaluation method

JOURNAL=Frontiers in Marine Science

VOLUME=Volume 12 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2025.1651187

DOI=10.3389/fmars.2025.1651187

ISSN=2296-7745

ABSTRACT=The Southern Temperate Zone (STZ, 30°S–55°S) plays a crucial role in global energy, water, and carbon cycles. While the Earth System Models (ESMs) of phase 6 of the Coupled Model Intercomparison Project (CMIP6) provide essential data for climate research focused on the Southern Hemisphere, significant inter-model discrepancies still necessitate a comprehensive evaluation, especially in the STZ. This study employs a multivariable integrated evaluation (MVIE) method to assess 17 CMIP6 ESMs in simulating the near-surface atmospheric fields and the oceanic temperature and salinity fields over the STZ, enabling holistic assessment of multiple variables. The multi-model ensemble (MME) mean of the near-surface atmospheric fields exhibits systematic biases, including overestimated westerly winds, northerly winds, and specific humidity. For the oceanic fields, pervasive warm biases in the potential temperature have been found in the deep ocean, whereas fresh biases in the salinity have been identified in the deep layer. According to the results of the MVIE, ten models show relatively good performance in simulating climatological annual means. Based on integrated statistical indices, eight models (ACCESS-ESM1-5, CanESM5, CanESM5-CanOE, CNRM-ESM2-1, GFDL-ESM4, MRI-ESM2-0, NorESM2-LM, NorESM2-MM) rank ahead among 17 CMIP6 ESMs. Evaluation of the seasonal climatology indicates that ESMs generally exhibit better performance during the austral summer than in winter. GFDL-ESM4 performs best in summer and autumn, whereas MPI-ESM1-2-HR and NorESM2-MM excel in winter, and MPI-ESM1-2-HR leads in spring. The study reveals persistent challenges in CMIP6 ESMs for simulating deep-ocean processes in the STZ.