Climate change is a hot topic, with everyone wondering how the climate will change in the future. Since climate change is a slow process, it is hard to study based on short-term observations. Using climate models has become the most common method.

A climate model is a set of software running on a super computer that can solve the physical formulations that describe variations in atmosphere, ocean, etc. via numerical algorithms, thus helping us to study climate changes.

In recent decades, climate models have developed from simple models to complex systems, and their performances keep on improving.

The climate system model FGOALS-f3-L, developed at the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics (IAP) of the Chinese Academy of Sciences, is one of the latest climate models in China. It is currently participating in phase six of the Coupled Model Intercomparison Project (CMIP6).

Evaluations of the experimental outputs show that the model is able to reasonably simulate climatic elements in historical experiments, including the long-term trends and climatological patterns.

Besides, improvements are apparent compared with the previous version. For the simulated ocean and sea ice, the most obvious biases are cold biases and overestimation of sea ice in the Arctic.

According to Prof. YU Yongqiang, the corresponding author of this recently published study in Advances in Atmospheric Sciences, "the obvious improvements of this latest model in simulating the climate system are meaningful for us to study climate dynamics and enhance our ability to project climate change, and of course make new contributions to the development of autonomous climate models in China."

Simulated sea surface temperature and its biases by FGOALS-f3-L (Image by GUO Yuyang)