从三维大气运动的基本控制方程出发,基下MPI消息传递接EI和区域分解的思想,建立了模拟三维可压缩大气中重力波传播过程的并行数值模式.在对垂直方向进行区域分解的基础上,针对跳点网格的特点,以垂直速度主格点线为子区域的下边界,状态变量主格点线为子区域的上边界进行区域分解.利用MPI消息传递接口来传递计算各子区域中变量时所要用到的相邻子区域上相应变量,从而顺利地完成对整个计算域的并行计算.根据线性重力波理论,通过模拟小振幅重力波的传播过程对所建模式进行了验证.结果表明,模式可以很好地模拟小振幅重力波在三维空间中的传播过程,模拟的重力波振幅随着传播高度的增加以指数形式增长,对能量传播路径以及有效扰动位能和扰动动能的模拟结果均与理论预测的结果吻合很好.另外,随着进程个数的增加,完成相同的计算量所用的计算时间也显著减少.这表明本文建立的并行数值模式不但能够很好的模拟重力波的传播过程,而且能够有效地节约计算时间.
Based on the control equations which describe the dynamics of three dimensional compressible atmosphere and the MPI (Message Passing Interface) and the idea of domain decomposition, a parallel numerical model is developed in this paper. The model is used to simulate the gravity waves propagation in three-dimensional compressible atmosphere. Firstly, on the basis of domain decomposition in the vertical direction and the characteristics of the staggered grid system, the main grid lines for vertical velocity are regarded as the lower boundary of the child domain, and the main grid lines for the state variables are regarded as the upper boundary of the child domain. Then,the data communication from one child domain to another is completed by MPI. As a result, the parallel computation on the whole computational domain can be completed successfully. Finally, according to the linear theory for small amplitude gravity waves, we simulated the small amplitude gravity waves propagation and the results indicated that, the propagation of small amplitude gravity wave in three-dimensional atmosphere can be simulated well by our model. The amplitude of gravity wave is increased in the form of exponential with its propagation height increased. The simulated results about the energy path, the available perturbative potential energy and the perturbative kinetic energy are well consistent with the gravity wave's linear theory. Additionally, with the number of processes increasing, the computational time decreases dramatically. All of these indicated that, the parallel numerical model developed in this paper not only can simulate the propagation of gravity waves very well, but also the model can save the computational coast effectively.