This is such a bollocks argument. The reductio ad absurdum of your argument is that you cannot model anything and strict experimental evidence is the only standard for science. Really? Is that what you really believe?
(Though I have to say, it's a truism that all models are wrong, it's just that some models are less wrong than others)
In the first place, we do have the temperature records anyway, as well as all the other biological and geological evidence that climate change is occurring. But sticking with the 'flawed' models, what alternative can you offer?
You can back-cast the models, without having to "fit" them, just using the data and the calculations, and in fact they fit very well with observations. That works even in the 1800s. So obviously something is going right.
Risby et al. (2014) checked model accuracy just this year, and found them to have excellent predictive power.
Keep coming up with zombie ideas, we've heard them all before, they've all been thoroughly investigated and refuted in the scientific literature.