TESTS OF THE EFFECT OF GRID RESOLUTION IN A GLOBAL PREDICTION MODEL

M. SANKAR RAO Institute for Space Studies, Goddard Space Flight Center, NASA, New York, N.Y.

Search for other papers by M. SANKAR RAO in
Current site
Google Scholar
PubMed
Close
and
LUDWIG UMSCHEID JR. Computer Applications, Inc., New York, N.Y.

Search for other papers by LUDWIG UMSCHEID JR. in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Kurihara and Holloway have proposed an integration scheme that offers advantages in the problems of geophysical fluid dynamics by rigorously conserving mass and energy. We have attempted to investigate the accuracy of the Kurihara and Holloway method by numerical experiments, applying it to a problem for which an approximate analytic solution is available.

For the case in which the planetary wave number is 4, we find that with the equal-area grid and with a latitude grid spacing of 4.5°, the planetary wave is destroyed by truncation errors within 5 days. In order to achieve a solution with acceptable accuracy, in which the planetary wave character is retained for a minimum of 10 days, the grid spacing near the Pole has to be decreased by a factor of 9.

Abstract

Kurihara and Holloway have proposed an integration scheme that offers advantages in the problems of geophysical fluid dynamics by rigorously conserving mass and energy. We have attempted to investigate the accuracy of the Kurihara and Holloway method by numerical experiments, applying it to a problem for which an approximate analytic solution is available.

For the case in which the planetary wave number is 4, we find that with the equal-area grid and with a latitude grid spacing of 4.5°, the planetary wave is destroyed by truncation errors within 5 days. In order to achieve a solution with acceptable accuracy, in which the planetary wave character is retained for a minimum of 10 days, the grid spacing near the Pole has to be decreased by a factor of 9.

Save