R-CCS Cafe 第205回 第3部

Recent development of supercomputers will enable us to perform global atmospheric simulations with horizontal grid-spacing of O(10~100 m) near the future. In such high spatial resolutions, the large-scale eddies in the planetary boundary layer can be explicitly resolved by Large-Eddy simulation (LES). This improves the representation of interaction between turbulence and low-clouds. However, in terms of numerical methods for fluid dynamics, the state-of-the-art atmospheric models often adopt relatively low-order schemes (e.g., second-order accuracy). The numerical errors are possible to dominate the sub-grid scale effect by the turbulent models. Although the problem has historically been concerned in computational fluid dynamics, it is still difficult to say that the problem is well understood in meteorological situations. Based on this background, Kawai and Tomita investigate how high we need the order of numerical accuracy for the LES of atmospheric boundary layer turbulence, focusing on the advection terms. Based on the turbulent theory, we derive two numerical criteria. According to them, as the resolution increases by the target resolution of O(10 m), we should adopt much higher-order schemes than second-order accuracy. In this presentation, I will show the overview of major results.