Home | Poster Submission | Posters | Archives | Contacts |
New ways to discretize the atmosphere for numerical weather prediction Rainer Bleck, Jian-Wen Bao, Stan Benjamin, Jin Lee, Sandy MacDonald, Jacques Middlecoff, Ning Wang Global weather prediction requires solving a set of fluid dynamics equations on a sphere. Doing this in spherical coordinates leads tonumerical problems at the two poles. An increasingly popular alternativeis a grid layout based on approximating the sphere by an icosahedron andjudiciously subdividing its 20 triangles. This poster presents firstresults from a high-resolution (30 km mesh size) global atmospheric model developed at ESRL featuring an icosahedral grid in the horizontal and aquasi-Lagrangian, or "flow-following", grid in the vertical. The special vertical grid is meant to reduce vertical as well as horizontal numerical dispersion errors that affect traditional "fixed" grids during simulation of 3-dimensional transport and mixing processes. |