That depends on the theory. If the theory says “X causes Y all the time”, then one exception where X does not cause Y would invalidate that theory. But if the theory is probabilistic in nature, it cannot be defeated by one exception. Here is an example:
Theory: If you flip a coin 10 times, it will tend to land on heads 5 times more often than any other number…
Here is one exception: I flipped a coin 10 times, and it landed on tails all 10 times.
That one exception does not disprove the theory. But if you flipped the coin 10,000 times and it landed on heads all 10,000 times, that would put the theory in serious doubt, and I would begin to suspect that the coin was rigged with both sides tails.
Climate change predictions are also probabilistic. They have bands of uncertainty, expressed in statistical terms. Without a strong grounding in the science of statistics, it is hard to say if a theory is validated by a particular graph of data.
What’s that in answer to?
Only if you don’t understand what it means for a probabilistic model to be invalidated.
If it is so obvious, please explain it to me in scientific or statistical terms.