Multigrid Methods for Meteorological Models

数值模式是大气科学研究的重要工具。大气运动所涉及的广泛的时间和空间尺度，就如何在数值模型中处理这些不同的尺度而言，提出了一个真正的挑战。无论是在研究中还是在操作环境中，成功的建模工作的一个基本要素是计算效率高的数值方法的可用性。多重网格方法有望克服这些数值问题。在以前的国家科学基金会的支持下，富尔顿教授研究了多网格方法在各种大气模型中的应用。他现在计划将他之前的成果扩展到其他领域，并着手进行一些新的应用。具体来说，他将继续研究(1)平衡模型的多网格求解器，(2)飓风和锋面模型的自适应网格细化，以及(3)有限区域和全球光谱模型的多网格方法。

Numerical models are key tools for research in atmospheric science. The wide range of temporal and spatial scales that are involved in atmospheric motions presents a real challenge in terms of how these different scales are handled in a numerical model. An essential element of a successful modeling effort, whether in a research or an operational context, is the availability of computationally efficient numerical methods. Multigrid methods hold promise for overcoming of these numerical problems. Under previous NSF support, Professor Fulton investigated applications of multigrid methodology to a variety of atmospheric models. He now plans to extend his previous results to other areas and embark on some new applications. Specifically, he will pursue studies of (1) multigrid solvers for balanced models, (2) adaptive mesh refinement in models of hurricanes and fronts, and (3) multigrid methodology for spectral models, both limited area and global.