As I mentioned yesterday, Malcolm Hughes and/or the University of Arizona Laboratory for Tree-Ring Research had gone to the trouble of blocking my IP address from accessing their website. Thanks to the help of a couple of CA readers, I was able to circumvent the block. This is the sort of petty behavior that gives (or should give) the Team a bad name in a couple of senses – it both makes them look bad and it’s pointless because it’s readily circumvented. Given the relative ease of circumvention, you’d think that they wouldn’t bother, but they don’t seem to care. One CA reader wrote and asked me if he could locate what I was looking for. Originally I was just going to check their listserv, which I did, and there was nothing new there. But since I’d been blocked, I went to a little extra trouble to check the site and it was worth it.
As I’ve mentioned previously, the Sheep Mountain bristlecones were re-sampled in 2002 but nothing has been published about it. Based on my experience with mining promotions – delayed results nearly always are bad as promoters find way of ensuring that good results are released on time or early, I’ve predicted that the Sheep Mountain update would not show elevated growth rates. (Yeah, there will be a “good” reason why any results took more than 5 years to be published, but “good” results would have been accelerated so that they were available for AR4. My bet offer was discussed by James Annan here.
An LTRR PhD student, Linah Ababneh, has been working on bristlecones (she got their “Bristlecone Prize” in 2002). I’ve emailed her in the past without any response and, from time to time, I google or otherwise check for information on her work. In yesterday’s browse through the LTRR website, I checked the page for theses and, sure enough, Linah Ababneh’s thesis was now online. It’s a must-read for CA readers.
Her starting point is Graybill and Idso 1993 and their reported difference between chronologies for whole-bark and strip-bark trees. She sampled 100 bristlecones – 50 at Sheep Mountain and 50 at nearby Patriarch Grove, with 25 whole-bark and 25 strip-bark – for each site. She has an overhead photo showing the location of each site. In addition to making a composite chronology, she made 4 chronologies for the bark x site combinations. She said that the chronologies extended from 1071-2003 and would be archived at ITRDB “after December 2006”; they aren’t.
Before I discuss how she reported her findings, let me simply show two illustrations of her chronologies. The first figure shows all 4 chronologies from 1600 to 2003 (with an 8 year smooth). BTW her thesis has 3 chapters which are drafted in journal style – two of the chapters are aimed at a dendro audience and a third is aimed at an archaeological audience (as she associates information from tree rings with settlements in the White Mountains.) In the two dendro chapters, she only illustrates her chronology for the 1600 on period. Seems odd, given the interest in millennial chronologies. Anyway here is the composite. Look at the right hand side of the chronology for the precipitous decline in Sheep Mountain ring widths in the past 20 years – which affected both strip-bark and whole-bark chronologies.
Abaneh Figure 3.Sheep and Patriarch strip and whole bark tree-ring width chronologies compared with Briffa et al. 1992 summer reconstruction based on tree-ring density. All chronologies are filtered after converting to z-scores to match the scale. Tree-ring data and Briffa et al. (1992) data are filtered using an 8-year low-pass filter (Fritts 1976). Confidence Intervals are based on the Briffa et al. (1992) series.
In her chapter linking to archaeology (scheduled for submission to Quaternary International), she has the following illustration of her long composite chronology (including strip bark), which she compares to Lamarche 1974 (but not Graybill’s chronology used in the MBH PC1: note- I’ll plot this up later or you can see Graybill’s Sheep Mountain chronology in the MM articles.) As you see, for whatever reason, her chronology differs materially from the Lamarche chronology (and even more from Graybill.)
Ababneh chapter 3 Fig.4. LaMarche (1974) overlaid by Aggregated chronology from Strip-bark and whole-bark of Patriarch Grove and Sheep Mountain develop in this study for visual comparisons. Both chronologies are tree-ring width indices.
Here is the same chronology plotted a little differently. IF you look closely, you can see that this is the same as the red series in Figure 4.
Ababneh Fig. 5. Cold and warm periods as inferred from tree ring widths chronology (Ababneh, 2006, This study) fluctuations above and below the mean after normalizing, whole-bark and strip-bark chronologies are grouped together from two sites Patriarch Grove and Sheep Mountain.
For comparison, I’ve plotted up the archived Lamarche (ca046) and Graybill (ca534) chronologies for Sheep Mountain as shown below. The Lamarche chronology ends in 1970; the Graybill chronology goes to 1987, but Mann only uses it to 1980. There’s something inconsistent between the Lamarche version illustrated by Abaneh and the archived version: they look very similar, but the archived version is about 20% higher (it touches 1.8 while the Ababneh illustrated version only goes to 1.6). The archived Graybill version goes as high as 2.4 – why wouldn’t Ababneh have illustrated and discussed it. I’ve marked the 1980 value of the Graybill Sheep Mountain series, as this is the last year used in Mann’s PC1. I’ve plotted the values from 1980-1987 in grey and one sees an incipient divergence problem. But the more interesting issue is that Ababneh’s chronology isn’t just calling into question the Divergence Problem; it’s calling into question the reality of the Lamarchel growth pulse – and even more dramatically, the Graybill growth pulse (which is never discussed.)
Her dissertation committee included Malcolm Hughes and Jeffrey Dean, but not a whisper about the “Divergence Problem”. She reported that she had confirmed a significant difference between strip-bark and whole-bark chronologies, but was unable to explain the difference.
I noted that, at Almagre, measurement data is missing for some Almagre trees where tags have been located, but measurements have not been archived. Ababneh made the following interesting comment about sampling differences:
As shown in Figure 13 (5 samples ), whole-bark trees from Patriarch Grove seem to have a mixed form of growth, i.e. three trees have an increase in tree-ring width, while two do not show such an increase. In comparison, Figure 8 (18 samples) illustrates that none of the whole-bark trees have an increase in tree-ring widths over the period shown. This shows how inadequate sampling might lead to different conclusions.
In chapter 2, Ababneh assessed climatic relationships between these chronologies, finding little relationship. She states:
Because the results of this study do not provide enough evidence to support the hypothesis that whole-bark tree growth is controlled by temperature, a temperature reconstruction cannot be established based on the chronologies and analyses developed in this study….
Until these limitations [other factors limiting growth] are taken into consideration when modeling tree growth, and until further measurements are obtained that involve a longer instrumental climate record from the same elevation of the research sites, the positive or negative effects of temperature cannot be substantiated.
The following paragraph is amusing. After re-iterating the evidence against a temperature signal in these chronologies one more time, she cites IPCC 2011, MBH99 and Mann et al 2000 as “a new and growing body of literature … that upper forest border trees respond …to the observed global temperature increase”:
The results reported in this paper are partially in accordance with other studies that support the hypothesis of a moisture related signal in bristlecone pine. Wright and Mooney (1965) concluded that bristlecone pine responds to a precipitation gradient rather than to cool alpine temperatures. Similar results were reported from a three year ecophysiological study of bristlecone pine by Fritts (1969). Such findings are not unexpected since the White Mountains are xeric with annual precipitation not exceeding 41.5 cm. Graybill and Idso (1993) and Graybill and Funkhouser (1999) compared tree-ring widths from the western United States, including one of the sites investigated in this study (Patriarch Grove), and found similar results: a low correlation with temperature that prevents use for temperature reconstruction, a negative correlation with the previous year’s temperature, and a highly significant tree growth response to spring precipitation. Bunn et al. (2003) and Tang et al. (1999) implicated soil moisture sensitivity especially in the strip-bark trees. However, a new and growing body of literature supports the other part of the results provided within; that is, upper forest border trees respond differently to climate variability including the observed global temperature increase (IPCC 2001; Mann et al. 1999; Mann et al. 2000), and a growing season temperature signal can be detected on decadal time scales (Hughes and Funkhouser 2003).