Reply #76: Yes they are

[Nederland]

Hindcasting proves absolutely nothing. Claiming that hindcasting proves anything is one of the very few things that I would say is an out and out lie in the IPCC AR4 report. Hindcasting is curve fitting, pure and simple. You run your model, compare your results to the historical record, make some alterations to better fit the data, and then repeat until something looks close. It proves absolutely nothing about whether or not your model is correct. The simple truth of the climate models is that you have 15-25 different non-linear inputs coming together to produce one output: temperature. It is a mathematical fact that in this situation there are literally trillions of possible ways to have those 15-25 inputs interact to get the right answer. Simply because you managed to come up with one set that produces a match to the historical record does NOT mean it is the correct one.

What makes this situation even worse is that for more than half of those 15-25 inputs that climate models use we do not have all the data for the entire historical record. Cloud cover data for example only exists since 1979, and it is a far more potent factor than CO2. So what data do the models use for all of these inputs that they don't have a full record for? They guess at the values. Doesn't that give you a nice warm and fuzzy feeling?