This claim is in the same category of Al-Gore's response that Artic Ice would vanish by 2020.
And failing in a single prediction does not invalidate magically all the other models. Ecological problems are complex and chaotic problems by nature. Much more complex that building an app. Being wrong is totally normal, and even necessary to understand and improve the model.
So this would be like discarding an entire program of millions of lines because, gasp, one bug was found. Bugs are expected and even welcomed. They are tests that allow us to improve much faster
If the problem is what was said by <"politician"> from <"random color"> that you dislike, you are attacking the problem in all the wrong ways.