On the Use of Hybrid Vertical Coordinates in Numerical Weather Prediction Models

Abstract

A comparison is made of several ways to combine isentropic with sigma coordinates in a numerical weather prediction model in order to benefit from the advantage of either coordinate. The resulting hybrid model versions appear to be well-behaved numerically. A noise index introduced to provide a measure of external gravitational imbalances shows that hybrid models are somewhat ?quieter? than isentropic coordinate models. Results from a related experiment suggest that this reduction in gravitational noise is not simply due to a fundamental noise difference between isentropic and sigma coordinate models but is likely to be a result of the hybridization.