Evaluation of CMIP6 model performance in simulating historical biogeochemistry across the southern South China Sea

Abstract

Abstract. This study evaluates the ability of Earth System Models (ESMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6) to simulate biogeochemical variables in the southern South China Sea (SCS). The analysis focuses on key biogeochemical variables: chlorophyll, phytoplankton, nitrate, and oxygen based on their availability in the selected models at annual and seasonal scales. The models' performance is assessed against Copernicus Marine Environment Monitoring Service (CMEMS) data using statistical metrics such as the Taylor diagram and Taylor skill score. The results show that the models generally capture the observed spatial patterns of surface biogeochemical variables. However, they exhibit varying degrees of overestimation or underestimation in their quantitative measures. Specifically, their mean bias error ranges from −0.02 to +2.5 mg m−3 for chlorophyll, −0.5 to +1 mmol m−3 for phytoplankton, −0.1 to +1.3 mmol m−3 for nitrate, and −2 to +2.5 mmol m−3 for oxygen. The performance of the models is also influenced by the season, with some models showing better performance during June, July, and August than December, January, and February. Overall, the top five best-performing models for biogeochemical variables are MIROC-ES2H, GFDL-ESM4, CanESM5-CanOE, MPI-ESM1-2-LR, and NorESM2-LM. The findings of this study have implications for researchers and end users of the datasets, providing guidance for model improvement and understanding the impacts of climate change on the southern SCS ecosystem.