In order to study the impact of variables vertical distribution on hydrostatic adjustment process, the equations sets describing the hydrostatic adjustment process are discretized on these five grids, the non-staggered N grid on which all variables will be placed on the layer; the Charney Phillips staggered grid (CP grid) on which the vertical velocity and temperature are put on the layer and the horizontal velocity, pressure and density variables are placed on the semi-layer; Lorenz staggered grids (L grid) on which the horizontal velocity, pressure and temperature are put on the layer and the vertical velocity and density are placed on the semi-layer; new Charney Phillips grid on which the density is put on the layer (CP_N grid), new Lorenz grid on which the density is placed on the layer (L_N grid). And the vertical grid spacings are set to 1 km, 0.5 km, 0.2 km and 0.01 km, respectively. The relative errors of frequency and vertical group velocity generated on these five grids were studied. The results show that (1) L_N grid is equivalent to the CP grid. (2) No matter how many vertical grid spacing is, the errors from CP grid and L grid are the smallest, followed by N grid, and the CP_N grid shows the maximum errors. (3) The errors generated on these grids are reduced with the decrease of the vertical grid spacing. Large errors are produced at the vertical short wave and horizontal long wave for the CP, L and N grids. Meanwhile, for the CP_N grid, the error is insensitive to the change of the horizontal wavelength; however the shorter the vertical wavelength is, the greater the errors become. (4) These grids are only sensitive to the changes of vertical wavelength but insensitive to the changes of the horizontal wavelength when the vertical grid spacing is 0.01 km. (5) The CP, L_N and L grids are suitable to describe both of the hydrostatic adjustment process and baroclinic geostrophic adjustment process due to their minimal errors. So they should be preferable in the numerical prediction model of non-hydrostatic full-compressible deep atmosphere.