Here is a script for collating Loehle proxies into an R-list in which each item is a table with two columns – one the AD year and the other the temperature (except for Holmgren where it is a native value as temperatures were not estimated in the underlying article.) This has not been coordinated with Craig Loehle (who’s undertaken to send me his values as used) so it will be instructive to see how close I’ve come to replicating his data as used.
I’m going to write a longer post on this, but I do wish to note that I’ve done quite a few calculations over the past couple of years showing how slight changes in proxy selection alter medieval-modern relationships, mostly focusing on the impact of bristlecones and foxtails, Yamal and the Arabian Sea G Bulloides. Typically I’ve observed that these variations cannot be distinguished from the canonical versions in the calibration period.
One such example is shown below for Briffa 2000, where using the Polar Urals Update instead of the Yamal series (substituted by Briffa instead of using the updated Polar Urals information) altered the medieval-modern relationship. (A virtually identical result applies to the medieval network of D’Arrigo et al 2006 which is nearly identical to Briffa 2000.) You can see that any verification statistics that apply for one variation will apply to the other and thus, inspection of verificaiton statistics will not enable anyone to determine which recon is “better”.
Figure 1. Briffa 2000-type reconstruction: black – canonical version; red- using Polar Urals update instead of Yamal.
I’ve done numerous similar variations – for Moberg, Esper, Jones et al 1998 and, of course, for MBH98-99, all with similar results. Slight variations yield different medieval-modern relationships, often with indistinguishable verification statistics (including similar problems.) Someone inquired about our apple-picking calculation presented to NAS. Here’s a plot of our apple picking version together with the Loehle recon in similar format. A number of the selections are in common: Sargasso, Conroy, Yang, Mangini; and the GRIP borehole and Cuffey recon have much commonality.
Here’s a quick plot of the series used in Loehle 2007. I’ll update when I get digital versions from Craig.