Climate of the Past has published a new study looking at models' ability to accurately capture temperatures during the last interglacial period from 131,000 to 114,000 years ago. In response, the Hockey Schtick has a post that not only misinterprets the study, but also employs some painfully bad logic.
Ecologist Richard Telford, blogs about the study in a post titled: "The perils of just reading the abstract (and of not understanding it)." While the study is technical, Telford highlights its main takeaway: "models are performing well at the global scale." The bottom line is that because the last interglacial period is too far back to radio carbon date, you have to line up the hottest temperatures from different regions in order to match them together. Because of this rudimentary but necessary warm bias, models tend to overestimate temperatures by 0.4 ± 0.3°C. The new study supports scientists' estimates of warming bias by finding that models overestimated by 0.67°C. Comparing the two figures, you can see models are actually pretty accurate.
This, of course, is the opposite conclusion reached in The Hockey Schitck post, which claims models can't fully simulate temperatures. The post goes on to say the models "failure" to simulate past warming must mean present-day warming could be entirely the result of natural processes. Of course we all know this is a bit like saying that because people die naturally, there's no such thing as murder.