Newsletter No.7 July1999@
Dr.Akinari Kaneko 
I joined the Integrated Modeling Research Program in February 1999. I am now at grips with the development and application of the oceanic generalcirculationmodel NJROGCMS in coordinates. NJROGCMS would be treated on the earth simulator which will be completed in 2002. For this purpose, it is divided into parallel lines. The simulator has the capability of longterm integrals with high speed on numerical calculation of oceanic prediction over a wide range with high resolutions. The model has not only to make high speed calculation possible,but to give correct numerical solutions physically. At the present, I make in various numerical experiments for this physical verification.
Originally I majored in physics,but as to my doctorate, my specialization was in the field of structural engineering,which covered analysis of simultaneous diffusion of chemical materials and heat, with consideration of chemical phenomena. For the last four years before I joined here, I had worked with the development of the oceanic model in coordinates,and was especially interested in/studied horizontal diffusion term in coordinates and numerical diapycnal mixing by horizontal pressure gradient terms.
In the future, I would like to study issues of numerical modeling relating to the atmosphereocean coupling model. Yet, I may change my interest into analysing physical phenomena, during the course of learning atmospheric physics and oceanophysics.
Mr.Kinji Baba 
I joined the Integrated Modeling Research Program last April. I am in charge of data assimilation. This technique is also developed in other Program research, and it is an indispensable technique to understand nature through models as an interpreter. Moreover, the data assimilation technique is also expected to train models.
In the field of mathematics, Hadamard clarified a wellposed conception at the turning point of the 20th century. That is, there is not always a correct answer to each problem. He showed that conditions should be set as if there would be an answer, and physical phenomena could be treated in such a way. Generally data assimilation belongs to the socalled inverseproblems, and becomes illposed problems. For example in the 19th century, it was thought to be optimistically possible to determine the shape of the earth from the lunar orbit. Up to the present, I have been engaged in the inverse problems in the engineering field with monitoring and nondestructive testing. Now I think the key point is how we should approximate to wellposed problems. 100 years after Hadamard (even though it is not much related), I am happy to take part in the research of inverse problems like data assimilation. From now on I intend to study high resolution and nonlinearity issues, keeping in mind the prediction system and reanalysis data.



