Replies

There are 4 fundamental problems that make the models useless for predicting conditions:

Data. The historical data is simply not available. The data that is available, is not sufficiently accurate to make these fine, nuanced predictions. Much of the data has to be scaled and adjusted by factors to account for such things as changing land use patterns around the weather station. These scaling factors and adjustments are bigger than the "trends" discovered in the data. Think about that for a minute.
Unknown interactions. As good as the scientists are, and as much as is now known about how the climate works, there is still a lot we simply do not know. That means the models are built upon an admittedly incomplete understanding of how climate works. Look at it this way, if everything about climate science were known, then what are all those climate scientists "researching" anyways? ;-)
Iterations. The models work by predicting what conditions will be like just a little bit in the future (a time step) based on the state of the model at the "current" time. The smaller the timestep, the more accurate the predicted result. Moving any significant delta into the future can require millions of iterations. Inaccuracies build up. Rounding errors build up. Sensitivity to initial conditions builds up. Turns out these kinds of models are only valid for small ranges of time/iterations. After that their connection to reality is tenuous at best.
Mathematical size of the problem. Going back to #2 somewhat, we don't know all the interactions of elements of the climate, flora and fauna, geology, the sun, etc. There are literally millions of interacting elements. The mathematics becomes a combinatorial explosion. We literally do not have enough computer memory nor processing power on the entire planet to solve this kind of problem (even if we understood it and could express it) before the sun goes nova.

Well, we can address #2. We are increasing our understanding of how the climate works. Will we ever completely understand it? I doubt it. But I'll give the scientists the benefit of the doubt, let's say #2 is solvable.

Number #3 might even be workable with good programming techniques. Hey, I'm a software (and modeling) guy, I can see ways to address this, but they are going to make the models even bigger, even slower. Let's give me and my kind equal benefit of the doubt and say we're clever guys/gals, we can figure this out.

The problem, well, problems are #1 and #4. We just aren't ever going to get better data. I guess if we wait around a hundred years or so we'll have a hundred years of pretty good, pretty accurate data. So let's say #1 can be addressed with patience. Of course, the AGW types don't want that. These hucksters want your money and they want it NOW!

#4 is the real problem child. You just can't get around the mathematics. You can eliminate variables that do not appear to contribute, but this introduces subtle errors that become significant during those millions of iterations (see #3). There is no free lunch, no shortcut.

The upshot is, the models are completely worthless for what many are trying to use them for - a means to "prove" their pet theory and excuse for a money/power grab.

George E. P. Box:

"Remember that all models are wrong; the practical question is how wrong do they have to be to not be useful."

"Essentially, all models are wrong, but some are useful."

Thanks for your remarks.

I’m not sure your item #3 can be solved. The climate seems to be a chaotic system. When you have to do thousands of iterations, one little difference can take you in a completely wrong direction. Before I would trust a climate model, I would have to see proven results over a long period of time. Longer than any human being can live. I would certainly not base government policy on climate models that have been developed in the past thirty years or so.