通用地球系统模式(CESM)是美国国家大气研究中心最新推出的地球系统耦合模式, 对解决气候(地球)系统建模中所涉及的新挑战和新问题具有很大的帮助。首先介绍CESM模式的结构框架以及最新版本的重要更新; 然后结合具体的应用实例和使用经验, 重点讨论如何在高性能计算机上对模式进行移植和合理的CPU配置, 并比较不同配置之间的优劣性, 从而确定模式最佳负载平衡和最优效率, 对模式新用户的使用具有极大的帮助; 最后对模式进行一系列的稳定性测试和验证, 结果表明模式具有较好的稳定性, 可以进行数值模拟和科学研究。同时对地球系统耦合模式的发展进行了总结, 并对模式发展中存在的问题提出了一些建议。

The Community Earth System Model (CESM) is a fully-coupled global climate model, and is maintained by the National Center for Atmospheric Research (NCAR). Composed of several separate models simultaneously simulating the earth's atmosphere, ocean, land surface, sea-ice, land-ice, river transport and wave, and one central coupler component, the CESM allows researchers to conduct fundamental research into the earth's past, present and future climate states. CESM1 contains totally new infrastructure capabilities, the implementation of a coupling architecture, and model parameterization development. These permit new flexibility and extensibility to address the challenges involved in earth system modeling with ultra high resolution simulations on High Performance Computing (HPC) platforms using tens-of-thousands of cores. Firstly, the infrastructure of the model is introduced, and also the notable improvements. The CESM1 coupling architecture provides “plug and play” capability of data and active components and includes a user-friendly scripting system and informative timing utilities. Then, the processor (PE) layout is customized for the load balancing on high-performance computers to optimize the throughput or efficiency of a CESM experiment. At the end of the paper, the port validation and model verification are made for the ocean model—the Parallel Ocean Program version 2 (POP2) which has properly ported to the machine—Polaris at Ocean University of China. The POP2 model output is subsequently verified to be a successful port, and CESM1 POP2 ocean-model solutions are the same as solutions generated on a trusted machine—bluefire at NCAR. Together, it enables a user to create a wide variety of "out-of-the-box" experiments for different model configurations and resolutions and also to determine the optimal load balance for those experiments to ensure maximal throughput and efficiency. The results and experiments will provide useful experience and method to the new CESM users to make simulations and load balancing of the model.