Of particular note is this graph.
The biggest takeaways are:
1). There are 180 different climate models listed here. All of them are getting warmer. And the global temperate has consistently fallen within the bounds of these models.
2). In the 1990's, global temperature tended toward the upper extremes of these models. In the 2010's it's been more toward the lower extremes. While the climate models failed to predict this, the duke study calculated a 70% likelihood of seeing these anomalies given a middle of the road warming scenario. But it is highly unlikely that we would have seen these same anomalies if either of the extreme ends of the models were correct.
3). And most importantly... The study showed that climate models by in large GET IT RIGHT.. but they tend to underestimate short term (decade long) climate variability.
In short, this study is a near airtight validation of human induced global warming with the most likely outcome being about a 1 deg c temperature rise by 2050.
(also.. while I applaud your enthusiasm, there's a reason scientists don't read about science in places like the Guardian. Lets just say that Journalists and Physicists don't take very many classes together. And from personal experience, having had various things covered by the press.... it's always been painful to read their descriptions of what we've done. 50% of the coverage tends to be a complete fabrication, the other 50% poorly understood and misstated.