A Tale of three trees

When we bought our house, there were three trees at the end of the driveway, planted when the house was built. Over the years the central tree died, and I planted a vegetable garden behind the trees. As a result of the reduced competition the tree on the right of the trio grew dramatically relative to the other.

Two trees in our drive, planted at the same time. (Note the difference in size of trunk)

I mention this series of events since it helps illustrate a very significant problem that appears to have arisen over the credibility of the data used to generate many of the curves predicting serious global warming over the next century. In order to determine temperature over the past thousand years climate scientists have relied on the quality of the rings laid down within the tree structure each year, with the assumption that, among other things, the local temperature controls the quality of the tree ring for that year. However, if the tree on the right in my driveway had been cored to look at its rings, then it would have shown a significant uptick in growth when the competing tree beside it died. It would now get more light and nutrition, and as a result may very well exhibit a growth rate that produced successive ring widths that went like this (units are 0.01 mm):

200; 350; 240; 210; 280; 350; 280; 390; 610; 740; 970

This would be in marked contrast to the poor tree sitting in its shadow, which remains shaded and does not increase the tree ring widths at all. Then consider that the first tree, because of all the extra sunlight and nutrition grows increasingly faster than the other tree, so that it now protrudes considerably above it and the surrounding canopy of trees (as you can see in the photo which I took in the last hour). It therefore in the more recent stages of its life (it is now around 40 years old) gets sunlight over a longer time period each day, and so its ring widths may go up even more markedly perhaps to:

710: 840; 900; 1090; 1220; 1600; 2020; 1150; 1020; 1920; 2120; 2500

These patterns of differential growth are, according to those who know forestry ( here and here) to be expected for individual trees in the forest, when local neighbors die, allowing both additional exposure to sunlight from the survivor, and also some additional nutrient from the corpse of the victim.

What makes this particularly germane at the moment is that such a tree record turns out to have played a very significant part in the development of the “hockey stick” curve that was made so much of in the 2004 IPCC report, and in subsequent reconstructions of temperatures over the past miilennium.

The numbers that I quoted above actually came from a tree that was found in the Yamal Peninsula (yes the home of much of Russia's future production of oil and gas) is known as YAD06 and its record was used by Keith Briffa in determining the rise in temperatures over the past century. If you look at the data for that tree, you can see (with the vertical scale normalized) that it shows a very strong increase in ring sizes since 1900.

Yamal tree ring growth (after Steve McIntyre)

Now one tree, obviously doesn’t create a viable record on its own, and in the original work Briffa and co-authors had used a dozen trees to generate the trend. And, when the data from them is examined, it does appear (although our tree is much more sensitive than the others) that there is some form of trend in the data.

Tree ring (and hence temperature) data for the 12 trees used by Briffa as a foundation to Mann’s hockey stick curve. (Source after Steve McIntyre)

However, what has not been clear until this week (there is a long story of efforts to block publication of the information which has been described by Bishop Hill and Ross McKitrick) is that there was a significantly larger body of data from which the above 12 tree chronologies were selected. When that larger body of data is examined it turns out that the trees that were not selected (of which there were a greater number) did not show this upward trend.

Tree ring (and hence temperature) measurements after being normalized, that were not included in the Briffa choice for his paper. (after Steve McIntyre)

It might be surmised after looking at this comparison that it would be more appropriate to use a more average value for the data from the region, rather than the smaller data set with the peculiarity of including our friendly example tree. But when one does that then hockey stick disappears. Briffa’s data was used extensively by other climate scientists in the past decade, and one might have thought that there would be some concern now that there is question on the validity of the underlying information used by the community.

Unfortunately the response to date at Real Climate seems to indicate, as has become increasingly obvious, that this is not about a search for truth, but rather part of the propaganda play in a political struggle. It is a curious statement that spends its time excoriating those that note the differences between the data selected and that admitted, rather than the meaning of the wider data set to the interpretation of how climate is changing. (I am however amused to see that we do see some evidence of a Medieval Warming Period now beginning to creep back into their posts).

Sadly I don't think that this is going to change the climate debate at all. Papers such as the Guardian are apparently blocking all attempts to blog about the topic on their Website (and a Google search of their site showed no hits) . and I imagine that the hope of the alarmists is that it will just go away. And in about a year they will come back with a "oh, yes we addressed that issue at the time" - citing their current post - and refuse to debate it further. It has worked for them before.