Experiments in Atmospheric Predictability: Part I. Initialization

Open Access
Publisher Website
Google Scholar
Abstract
Predictability experiments are carried out with a divergent barotropic model that describes the evolution of quasi-geostrophic planetary waves and high-frequency gravity-inertia waves. Error growth, relative to a model-determined control state, is initiated by an initialization procedure that is not compatible with the model equations. An analysis of error growth due to improper representation of the physics incorporated in prediction models is also carried out with the present model. The error growth rate and the range of predictability determined from these experiments, based on a simple triad solution of the nonlinear forecast equation, compare very well with the results from experiments carried out with multi-level numerical models. The mechanism of predictability decay by nonlinear energy exchange is shown to differ from the corresponding mechanism discussed by Lorenz and Leith, which is based on a model of two-dimensional turbulence.

This publication has 0 references indexed in Scilit: