There's one from Japan, one from England, Canada, Germany and 3 from the United States, GFDL and 2 from NCAR. You see that this is a scenario A2 and the responses for global warming all over the map. Somewhere from modest change of a 1.5 degree to relatively characteristic change a 5.5 degree. It really doesn't make any different right now. I disagree with one thing Jerry said yesterday. One of the reasons that people, politicians, others are not going to do anything is that they do not know what to do. If it's at the top curve, you got to do a lot, if it's at the bottom curve, you might not have to do much at all. We need to know what the real answer is, or, at least, have better approximation of what the real answer is. And, that's seems to me the goal number. Try to understand where the uncertainty is coming from in the modeling and in the forcing data. That's point one.
Now, what I saw over the last couple of days is very heartening to me and I'm really pleased with how your projects have started. I'm really impressed with the diagnostic studies and climate variability. I hope that you have on your minds the possible implementation of prediction schemes. That requires modeling, hopefully the same kind of models that you're going to use, for your global change studies. The confidence in a model to predict one thing is in its ability to predict something else. It's not the universal truth but it certainly has some reality that model that does good El Nino, for example, is more credible than the model that clearly doesn't. A model that has a good prediction of monsoons, or simulation of monsoons, has more credibility than one that clearly does not. So I really appreciate the work village developing here on every aspect of climate variability and making sure that your model can do a descent job of that. Furthermore, I think that societal impact in the short term of that kind of modeling capability might be more beneficial than what you are get for global change which as we see from that curve I had up there. It's really a second half of 21st century kind of problem or the 22nd century. So, solving some these prediction problems will be greatly beneficial to society and I encourage you to continue with that.
I agree with Jerry's comments about rewards. One of the things you need is not only model development but model diagnostics, thorough program of model testing and model experiments. They are not experiments that you can think of just based on the observational studies that we've already seen. Those experiments could be done by various collaborations among this group of talented young people that you are trying to assemble where everybody gets the chance to be first author on a paper and trying to make their scientific theme where contribute to the overall goals.
I Iike the general ideas that Sugi talked about with variety of experiments, model's different complexity. That makes a lot of sense. One of the disappointment with that kind of approach though.
