If you actually look at what those predictions have been, you'll find that they're pretty consistently on the mark. https://www.carbonbrief.org/analysis-how-well-have-climate-m...
back
1 comments
There is no quantification of the predictive power of any of these models presented in the linked-to article.
In fact it appears highly misleading, since they directly contrast (a) data from 1970-2017, with (b) a model written in 1973 predicting out to 2016, and (c) a model written in 2013 that can be compared to the data up until 2017.
So for this "pretty consistently on the mark" model, it only encompasses predictions for the years 2014-2017, based on observations from 2013. Many ARMA-type regression models do that well, since on average each year is mostly similar to the previous one.
In short, this doesn't support the credibility of this body of work.
You aren't giving any indication in what you've written that you've read and understood the article.