日本財団 図書館


Dr. Jerry Mahlman (GFDL, Princeton University):

I'd like to start with a bit of regret. I really wish I'd have a time to give the discussion of GFDL's flexible modeling system, which we started about 5 years ago where you are now. I'd like to invite one of your people to come to our place. And perhaps, examine that the some detail, because we have a lot of things that you haven't even realized that you are going to have yet.

Several observations: I noticed out of the 18 frontier papers presented 5 were on climate change or global change. Clearly there has to be cohering strategy for coupling the longer time scale processes. Local process studies are important but are not sufficient to answer those questions. I'd like also point out that, in this couple problem really important point is to develop a proper sense of the level of complexity and a scaling sense to be at the next generation model. If we have order one physics in a model and we have now include level 4 physics partially there will be wasted complexity. You have to think very carefully about how you put the right physics to the next generation. I think also my impression is that you need a real strategy for incorporating the physical effects of the smaller scales. And it is clear to me, you need the original non-hydrostatic sub-modeling piece of your global model to be able to check out physics and perhaps make a local impact analysis.

I'd like to also make some institutional comments. As a former director of GFDL, I spent a lot of time working with whole bunch of scientists trying to figure out how to put all the scientists together. And this is much easier said than done. I call it the climate model dilemma. The dilemma is, this is the thing we know a lot scientific creativity that's absolutely crucial but the bureaucracy always considered this is a form of anarchy because it's not cooperative with the establishment. And so the other limit is what I call programmatic control which is fail socialism. So what we have to do is to figure out how to find the optimum between these. And what we've done in the GFDL is to develop flexible atmosphere ocean land surface core models with a lot of options. We don't know how to put these models together. We have to test a number of different things. And also we've rejected the concept "modernistic model" this is the "do everything model," where simulating everything on planet earth.

We also have worked very hard to open up process to investigators throughout the world so that they can contribute to the model when used to the diagnostic packages and the model data can be shared with wide variety of research collaborators. Also, the research collaborators from outside the organization have to be, for participations, getting for credit for their cooperative efforts. And this requires interaction. So this is of course easier said than done, because this need to have a situation where scientist can work together be rewarded for working together. And we recognize not only for the creativity but also are rewarded for the team efforts. This is my version what kind of organization has to work in order to deal with the problem of this level of complexity. This is Jerry's personal list (not shown).

 

 

 

前ページ   目次へ   次ページ

 






日本財団図書館は、日本財団が運営しています。

  • 日本財団 THE NIPPON FOUNDATION