I wrote recently on the bizarre spaghetti graph from Guliya – where three inconsistent versions have been used in 2006 articles. I think that I may have a good first step at decoding this mess, as illustrated in the comparison below of the PNAS 2006, Climatic Change 2004 and Yang 2002 versions (used again in Yang 2006).
Figure 1. Scaled Guliya dO18 values for versions from Thompson et al 2006 (PNAS), Thompson et al 2004 (Climatic Change) and Yang et al 2002 (GRL).
Squint first at the Clim Chg 2004 version and compare it to the 2nd panel – it looks to me like that dates of the core have been made younger in the 2nd panel – for example, the blue values in the 18th and 19th century in the top panel all seem to be in the 19th century in the PNAS version. It’s a bit of a guess, but I think that the 1604 start date of the PNAS 2006 version corresponds to about 1400 in the Climatic Change version. On that basis, the Climatic Change version would not include any values from the MWP from 1000-1200.
The Yang version also looks to me like it’s dated younger than the Climatic Change 2004 version. The downspike in about 1250 in the Climatic Change version looks like it occurs about 1450 in the Yang 2002 version. The 20th century values in the Yang version are at “anomalous” levels, but are not anomalous in the more recent PNAS version. What accounts for the difference?
And of course, Thompson won’t show the sample data. Now Al Gore has used Thompson’s hockey stick, so I doubt that there’s anyway short of a court order under which Thompson will release the sample data to reconcile this mess.