Thursday, March 29, 2007 ... Français/Deutsch/Español/Česky/Japanese/Related posts from blogosphere

Borehole climate reconstructions & hockey stick revolution in 1998

Borehole climate reconstructions are based on the assumption that if you drill a very deep hole and measure the temperature profile i.e. the dependence of temperature on the z-coordinate - the depth - you will be able to reconstruct how the surface temperature looked like in the past. That's because you can simulate the propagation of heat by partial differential equations and it's very slow. When you're finished, you assume that the reconstructed surface temperature is correlated with the air temperature.

But let me get to the point which is the following:

Boreholes 1997

In August 1997, Huang, Pollack, and Shen published a paper in Geophysical Research Letters. Click the picture to see the full paper. Using the borehole paradigm, they reconstructed the temperature in the last 20,000 years.

The curves a,b,c distinguish different - increasing - weights given to the data, allowing different degrees of variation. The c-curve in particular leads to a very warm holocene climate optimum (much warmer than today) - a warm and pleasant period around 8,000 years ago that lasted for about 3 millenia. The picture shows a cool period 2000 years ago followed by the medieval warm period 800 years ago (warmer than the present) and the little ice ago 300 years ago or so: quite a lot of natural variability that exceeds the recent variations.

Their work was using heat flow and 6000 sites. The graph also makes it clear that as you go to the past, the borehole method clearly loses resolution. But that doesn't imply that we have much better methods to reconstruct the climate thousands of years ago.

MBH 1998 ...

... and MBH 1999

During the same time, in May 1997, Mann, Bradley, and Hughes (MBH) submitted their multiproxy paper with the hockey stick graph (MBH98). They were pre-determined to revolutionize and overrun the field of paleoclimatology which they indeed did, although for 5 years only.

Their paper - now known to be seriously flawed (click the hockey stick graph for an article by an esteemed physicist) - was going to argue that there were virtually no natural variations of the climate in the past 600 years (and in 1999, millenium) or so and only in the industrial era of the 20th century, the temperatures suddenly started to skyrocket. The medieval warm period was eradicated. For example, on this graph, the red curve is IPCC 1990, the blue curve is MBH99, and the black curve is Moberg 2006. Quite a difference!

Their paper was going to become a new standard that defines who is a "real" expert. As soon as it was published in February 1998, it indeed did become.

Boreholes 1998

Quite suddenly, the same borehole authors - Pollack, Huang, Shen (note a subtle permutation here) - published a new, two-page-long paper in Nature: it appeared in October 1998. The paper contained a rather different graph than the graph from 1997 (see below). Can you spot the difference?

It also argued that the authors had found an "independent confirmation of the unusual character of 20th century climate that has emerged from recent multiproxy studies [MBH98]" (see the abstract). The new paper was using temperatures and 358 sites only instead of the 6000 sites used in 1997 (94 percent of sites eliminated) and it has erased 19,500 years out of 20,000 years (97.5 percent of the time interval eliminated) from the paper written in 1997 in order not to contradict Mann et al.

That's what they call "independence". Moreover, if someone wanted to extend the record as far as possible while avoiding any hints of a warmer period in the past such as the medieval warm period, he would have made the same cut: 500 years ago. What a coincidence. Suddenly, everyone knew that boreholes should be trusted exactly for the last 5 centuries: Beltrami 2002, Mann et al. 2003 (he appears in these global warming stories as frequently as the polar bear!), and others.

But let's get back to 1998 when the second borehole paper appeared. At that time, the second hockey stick paper, MBH99 that has extended MBH98 from 600 to 1000 years, was already waiting for publication. MBH99 was published in GRL 26/6, around March 1999.

Comparing boreholes 1997 and 1998: today

It turns out that no one is able to identify any feature of the 1997 borehole paper that would make it less trustworthy than the 1998 paper - except that it doesn't quite agree with a certain fashionable lore in the contemporary climate science. The authors also think that their 1997 paper was fine. But it just happens that almost no one refers to the 1997 paper today: the 1998 paper is preferred.

Why do you think that Pollack et al. published their very different & heavily truncated paper in 1998 and what do you think is happening right now in that particular field of science - namely paleoclimatology? We report, you decide.

Via William Connolley, see Disclaimer for his fellow alarmists: William Connolley is not responsible for any interpretations found in this article and should be treated as innocent in the heresy accusations and should only be punished for an unintended help to heretics. :-)

Add to Digg this Add to reddit

snail feedback (1) :

reader Peter said...

In the 1997 Huang, Pollack, and Shen, the authors first exclude all data from above 100m below the ground surface, hence removing almost all of the 20th century climate signal from the data. This is why the magnitudes of the recent (100 years) warming in the two (1997,1998) papers are so different.