Steve: the Mann 1209 is illusory in respect to the medieval warm period. It relies on bristlecones and/or upside down Tiljander. Whatever reasons exist for rejection of LM apply equally to other studies.

- Every sample has a different response to climate. Even two stalagmites ten centimeters apart in the same cave will have different climate signals.
- Mon-linear responses should be expected due to the non-linear hydrology of karst (soluble rock).

Especially interesting since one of the series in the paper, Søylegrotta, Norway, was used to construct figure 6.11 of AR4.

It has probably been covered here before, but here is the paper:

http://docs.google.com/viewer?a=v&q=cache:YQwd9hleR20J:www.gsf.fi/esf_holivar/baker.pdf+s%C3%B8ylegrotta+mo+i+rana&hl=no&gl=no&pid=bl&srcid=ADGEESjSo-71tXaLberZydyBdSrCPrcTSaWkRwyVyGBHMxgFbytRntl6P8jABf6XRt9ZoGvxc6ZCUc0RauLOPGzIbLtAIRk7JqhjS6hpS-tjWZxjiNcO2dZkVtbBhbJ-5AImeM029RAa&sig=AHIEtbT0UkR2nzBoy8qu6HDz4Ou4cGR6rw

I’ll have a look at your functions and try to decipher a bit what i’m seeing.

RegEM infilling for years 1986-1995… This is based on Schneider’s (Tapio?) algorithm?

Steve: yes. however, in the rather different application of proxy reconstructions, the inherent linear algebra can be dramatically simplified. I’ve had a post in the works on this for quite a while but haven’t posted the results. The analysis will also clarify some of the bewildering behavior of this technique.

I have a quick question about Mann 2008… [I am interested in] whether or not the values recorded as annual anomalies were smoothed prior.

If it is helpful, my discussion of the four (actually two) Tiljander data series can be found here. At that post, there is a link to the file “Tiljander03-Calculated_LS+DS.xls”, which contains the annual values from 0 (sic) through 1985 as archived by A.E.K. Ojala. The data as archived for Mann08 are also presented.

Note the RegEM-based infilling for the years 1986-1995, as Steve stated at 1:49 PM above. Also note that Ojala reported a very thick (~13 mm) varve for the year 1326. The Mann08 archive appears to have put the averages of 1325 and 1327 in the place of the original entries for 1326. This substitution is not described in the text or SI of Mann08.

Mann has very quirky and ad hoc splicing methods – that have been noted here but not fully parsed in CA posts. I’ve pretty much figured it out , but I don’t think that I’ve written it up yet.

As to “which is the appropriate one to use” – hard to say.

I’ll try to revisit this and write an overview.

“If you want a stick to beat a dog, any stick will do.”