The Challenges of High Order Methods in Numerical Weather Prediction

Abstract

This paper reports on the communications made at the 2009 ICOSAHOM meeting minisymposium on the challenges of high order methods in numerical weather prediction, with contributions from mathematicians as well as atmospheric and ocean modelers. Motivation for an investment in high order method development for numerical weather prediction is given in terms of the potential payoff in light of the current challenges in the field. Among other issues, the implementation of physical parameterizations with high order methods stands out as a yet-unexplored and potentially difficult challenge to resolve. Adaptivity is also expected by some to significantly advance the state-of-the-art but no consensus seems to be reached that it will be feasible. Among the recommendations expressed at the workshop are the need for demonstrated efficiency comparisons between high order and low order methods for a desired level of accuracy in resolving waves.