RESULT FRONTIER Newsletter No.11 July.2000

On the current activities in Integrated Modeling Research Program

Hirofumi Sakuma
Integrated Modeling Research Program Acting Director

For better or worse, our program in FRSGC is certainly a new attempt in promoting collaboration with diverse fields relating to numerical modeling activities. We may say that an organizational effort in developing or improving a sophisticated climate model such as a state-of-the-art general circulation model is a sort of big enterprise. And we decided to embark on that a year ago from scratch. In order for the program to be successful, we believe that a clear incentive to concrete research goals in climate studies is the most important factor. In this respect, we first focused on the development of "human network" among traditional climate researchers in other programs and new comers from other fields and on sharing the common research goals through informal study meetings. Luckily enough, not few people willingly joined the "network" and we have progressed steadily though slowly.

Recently we established a research collaboration with Center for Climate System Research at University of Tokyo and National Institute for Environmental Studies to accelerate the development of a high-resolution coupled model. At the same time, we are closely working with the Earth Simulator Development Center to pursue the common goals of high performance computation on our developing general circulation models. Our current major activities can be classified into three different kinds of researches summarized in what follows.

[1] Coupled model development
In this activity, we have three research groups : Group 1, 2 and 3. The mission of Group1 is to improve the latest version of CCSR/NIES-AGCM for higher-resolution simulations, which includes, for instance, the improvement of cloud-radiation process, sub-grid-scale parameterizations and associated parameter tuning.Group 2 is working on the preparation of high-resolution ocean simulations using MOM3 in which the free surface of the oceans is explicitly treated. This function of the ocean model will be advantageous not only to the evaluation of a certain tracer field in the oceans and data assimilation studies but to high performance computations in the sense that it circumvents the use of the discrete Poisson equation that needs a time-consuming iterative operations. Since the performance of a sophisticated coupled model is quite sensitive to the details of the model physics and sub-grid-scale parameterizations, Group 3 started from a medium-resolution model on which we already have ripe experience.

[2] Next generation model
The problems we are now tackling on are :
1) The development of advection schemes preserving the Lagrangian property of fluid motion. Among others, we are pursuing the possibility of applying CIP advection schemes to our OGCM. We expect that tracer fields in the oceans are improved considerably with such a shape-preserving advection scheme.
2) One of the unavoidable problems in high-resolution simulations using longitude- latitude coordinates is a fact that the longitudinal grid interval becomes smaller and smaller when we approach the poles. As a new dynamical core free from such shortcomings, we are developing a system of quasi-equidistant grid configurations that uniformly covers the entire globe. (Fig. A and B)
3) Cloud resolving models. There are not many to mention at present but that we are now forming a research group challenging difficult problems for the development of such models.

[3] Oceanic data assimilation
One of the main goals of this group is to provide relatively reliable oceanic data set that can be used for a variety of climate researches. Now we are developing the adjoint code of MOM3, which is the core part of our data assimilation procedures. In the adjoint method, assimilation is done in a dynamically consistent fashion so that it is an elaborate way of suppressing unfavorable computational noises arising from the errors introduced otherwise.

A-1. Conformal cubic grid configuration B-1. Icosahedral grid configuration
A-2. Wave propagation experiment on the above grid system. B-2. Same as A-2 except for grid system.

Prev. Cover Next