[alloysteel]

Application of statistical methods to climate modeling may be a useful tool for limiting the parameters to deal with, and when no important aspect is overlooked, may be expected to yield reproducible results.

But to get the results that fit the observations, there is something introduced called a "finagle factor", that when applied, voila!, the "right answer" magically appears. This is a universal practice that runs throughout ALL scientific endeavor, when there are more unknowns than knowns, the human tendency to build myth as an explanation.

In integral calculus, the process is the reverse of differential calculus, but to cover variations in outcome, there always has to be room for the "jigger factor", expressed as k, from the German word konstant.

[Publius6961]

This is a universal practice that runs throughout ALL scientific endeavor, when there are more unknowns than knowns, the human tendency to build myth as an explanation.

In all fairness, that is not true of real science. A real scientist goes where the data takes him, no matter how silly the result. A real scientist knows when the results are silly and seeks the reason why and iterates, and tries again, or seeks alternative approaches to a solution.

Junk science, on the other hand, science with an agenda, is right up there with political chicanery, and worse than useless.

[Frumious Bandersnatch]

Yeah, but to be valid, any climate model should be able to accurately forecast, not only the future, but the past as well. Now, this model seems to be a bit different, but still, it should be able to be applied against the known data of the last 20 or 30 years and give us reasonable results if the model is accurate.

Unfortunately, I know of no climate model which has been able to accurately forecast the past, much less the future.