I validated a model by comparing its output from a set of inputs to the actual results. If they are different, the model is wrong, or at least there is a factor that is not accounted properly.
And that's where climate modeling has gone wrong: with one exception, every other model failed to predict the actual results in the past decade.
In my work, that was enough reason to toss the entire model and start over. But, rather than doing that, the climate modelers still insist the past decade was simply an aberration. Or, they claim the data needs "adjusting".
I'll also note that someone experimented with the original model that generated the "hockey stick" graph, kicking off the entire debate. He found that no matter what data set he used, he still got the "hockey stick". He even tried random data, and still got the hockey stick.
Very true, and if you ask most people who are not hardcore IPCC types but familiar with weather data, they will agree that modelling has been a spectacular bust. The real trends since about 1990 amount to this — a very slight increase with a spike in the 1997-98 El Nino event to about 2006, then a flat-lined variable period with no discernible slope at all.
This is about what you would predict (from 1990) if you assumed that human contribution to the complex result was a very small one (0.3 C deg) and that declining solar activity that was already somewhat predicted (more often since 2000) would compensate for that.
Almost every model I have seen is basically overcooked by a factor of 5 to 10 and has the flaw of accelerating warming when many natural factors would lead to the conclusion that even if the warming was real, it would slow down over time.
But people make the mistake of thinking that when they see a computer simulation on the TV news, they are seeing some sort of very sophisticated prediction. It amounts to little more than a technically savvy person programming a computer to show accelerating warming, it’s not really a prediction it’s an illustration of a prediction. And the illustration has already begun to fail so what use could it be?
Being an “anti-science” conservative (despite my inexplicable past awards in the physical sciences), I operate by a simple axiom:
The model is never the thing itself. A model is either more or less accurate - especially in comparison with a competing model. If a model were equatable with reality, then it would, in effect, be an ostensive definition of a reflexive syllogism: “I am that I am.”
I do not expect a model (such as Relativity) to have a perfect one-to-one correspondence to reality; I expect it to have a high degree of agreement, and to be reliably useful as a tool.
The so-called climate models in vogue with the communists are - to use a technical term - crap, tools fit only for global redistribution (which, hey, come to think of it,...).