There has been a certain amount of excitement on upholder blogs this week - the cause being the news that the National Center for Science Education (NCSE) has decided to involve itself in the climate change wars in the USA. NCSE has been at the forefront of efforts to keep intelligent design out of the classroom so this latest move represents a broadening of its focus.
Their website incorporates a shiny new climate change section, including a Climate Change 101 page. From there comes this paragraph on detection and attribution:
To ensure the accuracy of the models at projecting future climate trends, the models are often run backwards in time to “retrodict” past climate changes, and then compared with paleoclimate observations. The models through this process have become remarkably accurate and give the climate research community confidence that the future projections are robust.
Remarkably accurate? Is that right? And do we think they have portrayed the uncertainties in a reasonable fashion? Or in any fashion at all?