That's not far off liberal scientific methodology, to be fair, but it seems to put the cart before the horse. You might want to look up 'falsifiability', 'confidence factors', and, if you have the time and inclination, Karl Popper's Objective Knowledge.
This won't give you everything but it should go some way to explaining the scientific method in more detail.
The process is roughly as follows:
- Make a hypothesis that is capable of being disapproved;
- Test the hypothesis;
- Refine the hypothesis based on the findings;
- Test the hypothesis;
- And so on.
The more times the hypothesises is not disproved, the more likely it is too be correct, the more confident the prediction. According to this theory, it's impossible to prove anything; we can only be confident that knowledge is objectively true if we have tried and failed to disprove it. This is a bit of a blunt summary.
If you don't trust this method, I wouldn't ever get on a plane or take any medication.
The key point being that a prediction won't become the consensus until it has a fairly high confidence factor (i.e. lots of people have tried and failed to disprove the prediction). Climate change is one of those things. Every time someone conducts another experiment, the new data strengthens the view that global warming cannot be disapproved.
Just to put all my cards on the table, I think Popper is wrong. But he sets the scene for a lot of liberal conceptions of science. It's his ideas that underpin many of the kinds of predictions that you're talking about, I think. (When I say liberal, I'm referring to the main ideology of capitalism, not to the 'left' brand of US politics.)
That is, climate change about as 'true' as things can get, and so it is predicted. But even 'prediction' in this sense, makes it seem as if we're taking about something in the future (I couldn't help but challenge the Popperian model just a little bit, I'm afraid). But climate change is already here. It's the present. The prediction only concerns how bad it's going to get.
