蛙跳(1eapfrog)时间差分格式采用Asselin—Robert时间滤波方案去除计算解能够降低原始方程组的时间差分格式的计算精度,采用二阶Adams-Bashforth格式构造的欧拉前差方案可弥补蛙跳格式的不足,即:在不存在计算解的条件下去除滤波的影响,更大程度的保持方程组的计算准确性。本文基于NCARCAM3.0(Community Atmosphere Model3.0)完善的软件平台,将原模式的三时间层蛙跳时间差分方案修改为两时间层二阶Adams-Bashforth时间差分格式,对与重力波有关项使用中央差隐式处理,以此构建半隐式大气环流谱模式。利用动力检验的方法探讨模式对垂直分辨率的敏感性,从而寻找模式在较小计算代价下提高计算效果的可能。通过斜压波实验发现,提高垂直分辨率使模式具有更强的斜压波模拟能力,其模拟效果甚至已经与提高水平分辨率的效果相当,可以作为一种弥补模式运算效率不足的可行方案。
Asselin-Robert time filter used in leapfrog time differencing scheme for deleting computing solution can degrade the accuracy of the calculations. The second-order Adams-Bashforth method is not subject to time splitting and has the same accuracy as leapfrog scheme. A new Eular-forward semi-implicit atmospheric general circulation model with the spectral method is developed from NCAR CAM3.0 ( Community Atmosphere Model 3.0). The second-order Adams-Bashforth method is used as the alternative method of leapfrog scheme and Crank-Nicholson scheme for implicit. By an idealized dynamical baroclinic wave test, we analysis the sensitivity of vertical resolutions for the new dynamical core, moreover, by the test, we try to find out a feasible way to get better simulation results with lower simulation cost. It shows that the new dynamic core with higher vertical resolution has better simulation ability for baroclinic wave development. The results of higher vertical resolution are even as good as the results of higher horizontal resolutions. So we consider that it can be a feasible way to make up the lack of simulation efficiency by increasing vertical resolution of numerical weather prediction model.