Abstract
By combining the ideas of the iterative time integration scheme and the time filtering, a time integration scheme for a primitive equations model is formulated. This scheme turns out to be identical with the generalized two time level iterative time integration scheme. However, it is shown that a weight parameter used at the corrector step of time integration is permissible to exceed unity. It is also shown that a weight parameter larger than unity has the character to damp high-frequency noises more efficiently than that less than unity.
Along the Kurihara and Tripoli's idea, terms in the primitive equations are divided into two parts, that is, terms which contribute to the relatively slow temporal variation and those which yield the high-frequency noises. By adapting a weight parameter less than unity to the former terms and one larger than unity to the latter ones, it is shown that the amplitude of low-frequency meteorological waves are preserved fairly well while the high-frequency noises are damped effectively.
Numerical examples corresponding to the differences of a weight parameter are presented and the results are compared with each other. The results of application of this scheme to the actual numerical prediction model are also presented.
