2005 Volume 1 Pages 177-180
A series of numerical experiments were conducted to investigate the sensitivity of the WRF Model to the advection scheme. The differences in the simulated results between third- and fifth-order schemes are not large, however the vertical velocity field of the former is somewhat smoother. On the other hand, fourth- and sixth-order schemes easily produce grid-scale computational noise, although the time step is set to be smaller than the recommended one. Artificial second-order diffusion damps the noise, however it negates the advantage of the high-order advection scheme. Explicit fourth-order diffusion damps the noise, however it smoothes the field more than the inherent, implicit diffusion in the third-order scheme. Contrary to this, explicit sixth-order diffusion damps noise, but maintains detailed structure and high energy spectral density of the vertical velocity. We recommend the fifth-order upwind scheme for use in the WRF Model.