At Climate Audit, Something odd has been discovered about the provenance of the work associate with the Marcott et al paper. It seems that the sharp uptick wasn’t in the thesis paper Marcott defended for his PhD, but is in the paper submitted to Science.
Steve McIntyre writes:
Reader ^ drew our attention to Marcott’s thesis (see chapter 4 here. Marcott’s thesis has a series of diagrams in an identical style as the Science article. The proxy datasets are identical.
However, as Jean S alertly observed, the diagrams in the thesis lack the closing uptick of the Science. Other aspects of the modern period also differ dramatically.
Here is Figure 1C of the Science article.
Now here is the corresponding diagram from the (Marcott) thesis:
The differences will be evident to readers. In addition to the difference in closing uptick, important reconstruction versions were at negative values in the closing portion of the thesis graphic, while they were at positive values in the closing portion of the Science graphic.
I wonder what accounts for the difference.
Read the full report at Climate Audit
This story just got a lot more interesting. I wonder if we don’t have a situation like with Yamal, and sample YAD06 which when included, skewed the whole set. Perhaps there was some screening in the thesis and that didn’t include part of the proxy datasets, or later for the Science paper maybe there was some Gergis sytyle screening that made hockey sticks pop out. It might also be some strange artifact of processing, perhaps some Mannian style math was introduced. Who knows at this point? All we know is that one paper is not like the other, and one produces a hockey stick and the other does not.
Some additional detective work is sorely needed to determine why this discrepancy exists and if anyone in the peer review process asked any similar questions.