Moberg #3: Some First Comments

I’ve collated public versions of Moberg data to the extent possible and have posted up some comments from a first inspection of the data, including the apparent avoiding of ice core dO18 series, prominent in Hockey Team studies. I’ll post up a script collating the data into R if anyone’s interested. Moberg doesn’t directly use any dO18 ice core series (either from Thompson or anyone else).

Thompson’s dO18 series are a staple of Hockey Team multiproxy studies. The tropical ice core dO18 series have struck me as highly problematic: the lowest dO18 is in the summer; the highest in the winter. This is the opposite of what one would expect from dO18-temperature relationships. The reason is that summer dO18 comes from monsoon rainout. If you have a change in annual dO18, you can’t tell whether it’s from a change in summer-winter proportions or a change in annual temperature. Thompson simply arm-waves through this problem. Moberg seems to avoid Thompson series. A couple sneak in through the back door – Moberg uses Yang’s temperature reconstruction (#10), where Thompson data is included, but the effect of Thompson data is much attenuated in Moberg relative to the Hockey Team.

Supplementary Figure 1 shows the 11 time series used for low-frequency. 9 of the 11 were calibrated for temperature; the 2 that were not calibrated for temperature are the ones with 20th century blades: Arabian Sea % bulloides and Agassiz % melt. If these are non-linear, then that would affect the results somewhat. I’ve just gone through the process of collating digital versions of the data into R – I’m missing series #7, 8 which are pers. comms. I’ve posted up a digital version of #2 GRIP, which was used in Jones and Mann [2004]. Màƒⵢerg et al. only use tree ring series for high-frequency. No data citations (as URLs) is given for the tree ring series, but 3 bristlecone pine series are identifiable at WDCP – Indian Garden (nv515) ending in 1980; Methuselah Walk (ca535) ending in 1979 and White Mountain Master (ca506) ending in 1962. I think that the White Mountains Master series is an old combination of Campito Mt and Sheep Mt – it’s not clear why they used this old series. Moberg (SI#1) adopts the position that these series are "primarily interpreted as reflecting annual precipitation variability" and that they have a negative correlation (-0.57) with warm-season temperature. The versions in Fig. S2 are upside down relative to the archived chronologies and presumably are the images of the chronologies from the regression against temperature.

Walk data has been updated and I’ve recently (about 7 days ago) obtained updated information through Climatic Change, where Hughes and Funkhouser [2003] was published. No digital citation is given for the Tornetrask, Yamal, Taimyr and Yakutia series. Briffa has posted his own RCS versions for Tornetrask, Yamal and Taimyr at his website from the authors cited in Moberg and the Briffa versions look similar to Figure S2 #4-6. Briffa has a Yakutia version from 1600 on, but not the longer version cited here. In summer 2004, Hughes archived some Yakutia ring width information (possibly in response to my complaints that this data was unarchived) but has not archived a chronology.

I’ll post some more on this over the next few days.


Get every new post delivered to your Inbox.

Join 3,416 other followers

%d bloggers like this: