Review of Chapter 2, The Hockey Stick Illusion by A. W. Montford

In an attempt to see both sides of the debate in the question of human caused climate change, I am starting to read this book: The Hockey Stick Illusion: Climategate and the Corruption of Science (Independent Minds) by Andrew Montford


 As I read it, I will give a chapter by chapter review comprising quotes and commentary. So, if this is an area of interest, keep a watch.

All material quoted from the book is in italics. My comments are in plain text in brackets.

One problem that occurs in calibration is that any relationship we might have been able to calculate could have arisen purely by chance. In other words, just because tree ring width multiplied by ten happened to be equal to temperature in the twentieth century, that doesn’t mean that it was always that way. For example, twentieth century ring widths might have been lower than normal, say because of insect infestation in that particular set of trees, or maybe even in trees in general. Another possibility is that the tree isn’t actually responding to temperature at all, but to something else. If any of these issues were really affecting tree growth, it might be that the normal relationship is actually an extra 0.15 mm of growth from a rise in temperature of 1°C.

[This is nothing new and climate researchers have already spent considerable effort addressing the problem. Montford is just grasping at straws here. As just one example of the ongoing research in this area, see Dendroclimatology: Progress and Prospects edited by M. K. Hughes, Thomas W. Swetnam, and Henry F. Diaz (2011)]

If you have a large number of proxies, there are two main ways in which you can go about the calibration.26 The first of these has been described as ‘the Schweingruber method’, or composite-plus-scale (CPS), and involves taking proxies that are expected to be temperature sensitive, calibrating them against local temperatures and essentially taking an average. The other way, the ‘Fritts method’, or climate field reconstruction, involves taking lots of proxy series, which are sometimes not even responding to their local temperatures, and seeing if some sort of correlation can be found with temperature measurements somewhere in the wider vicinity. What emerges from this latter method is essentially a weighted average of the full proxy set, with the temperature sensitive proxies having a much higher weight than the non-temperature-sensitive ones. It’s this Fritts method that is the relevant one for our story. The Fritts method involves a certain leap of faith to trust that trees that are not responding to their own local temperature can nevertheless detect a signal in a wider temperature index. You have to believe in the existence of something called ‘teleconnections’, whereby temperatures in a possibly distant part of the world affect the climate in the locale of the tree in such a way as to affect its growth, and in a consistent manner. If this sounds implausible to you, then you are not alone. However, the reality of the mechanism is accepted by the paleoclimate community and for the purposes of our story that’s what you need to know.

[Firstly, The work of Fritts (Tree Rings and Climate, 1976) has stood the test of time and his methods have been tested and retested by literally hundreds of scientists. Seems rather arrogant for an armchair climate “expert” to sweep such work aside so cavalierly. His innuendo concerning the silliness of “teleconnections” is beyond the pale. Teleconnections are not some hocus pocus. There is more than a half century of empirical evidence for them, and there are some good hypotheses about the underlying atmospheric processes. As an example, here is a quote from The Global Climate System: Patterns, Processes, and Teleconnections by Howard A. Bridgman and John E. Oliver:

“Teleconnection is a term used to describe the tendency for atmospheric circula­tion patterns to be related, either directly or indirectly, over large and spatially non-contiguous areas. The AMS Glossary of Weather and Climate (Geer 1996) defined it as a linkage between weather changes occurring in widely separated regions of the globe. Both definitions emphasize a relationship of distant pro­cesses. However, the word "teleconnection" was not used in a climate context until it appeared in the mid 1930s (Angstrom 1935), and even until the 1980s was not a commonly used term in the climatic literature.

“As stressed throughout this book, teleconnections are often associated with atmospheric oscillations. Any phenomenon that tends to vary above or below a mean value in some sort of periodic way is properly designated as an oscillation. If the oscillation has a recognizable periodicity, then it may be called a cycle, but few atmospheric oscillations are considered true cycles. This is illustrated by the early problems in predicting the best-publicized oscillation, the Southern Oscillation and El Niño (Chapter 2). Were this totally predictable then many of its far-reaching impacts could be forecast.”]

The PCs are often described as being like the shadow cast by a three-dimensional object. Imagine you are holding an object, say a comb, up to the sunlight, and it is casting a shadow on the table in front of you. There are lots of ways you could hold the comb, each of which would cast a different shadow onto the table, but the one which tells you the most about the object is when you expose the face of the comb to the light. When you do this, the sun passes between the teeth and you can see all the individual points. You can tell from the shadow that what is being held up is a comb. This shadow is analogous to the first PC. Now rotate the comb through a right angle, so that you are pointing the long edge of the comb to the sun. If you do this, the shadow cast is just a long thin line. You can see from the shadow that you are holding a long thin object, but it could be just about anything. This would be the second PC. It tells us something about the object, but not as much as the first PC. You can rotate through a right angle again and let the sunlight fall on the short edge of the comb. Here the shadow is almost meaningless. You can tell that something is being held up, but it’s impossible to draw any meaningful conclusions from it. This then, is the third PC.

[Montford spends considerable time in this chapter explaining PCA (Principal Components Analysis) in lay terms. This is a worthy undertaking, but some of his description is a bit of an oversimplification. The “shadow” metaphor is a crude way to explain what PCA results mean. At least he makes some mention of variance, considering that it is variance that is being partitioned, i.e. what proportion of the variance can be explained by each component. It also is a bit annoying for him to abbreviate PCA as just PC, when PCA is the standard abbreviation used in most publications.]

And that’s it: that’s all you need to know. Throughout the months and years of bitter argument over Mann’s Hockey Stick, this simple step was the only part of the PC analysis that was in dispute. For the purposes of the story it is not really necessary to understand anything else about how the subsequent calculations work. You can think of PC analysis as a big black box which takes the centred data and churns out as many patterns as are felt necessary. It is, however, useful to understand just a little of the detail of what happens to the centred data, as it will help explain just why centring is so important.

[This comes after Montford describes what centering is, a data transformation method used in the PCA analysis by Mann. What he forgets to point out, is that although there was a lot of attention paid to this detail of Mann’s analysis, the results are not even dependent on that approach. More than a dozen independent analyses since Mann’s first hockey stick, using a variety of different approaches and data sets, have produced results largely the same. This seems to be just another example of how climate change contrarians pick some minor detail that they then consider a fatal defect to the interpretation of the results.]

P. S.

[I found a second case of the incorrect use of the word "data" in this chapter (and more in Chapter 3 which I am now reading), which leads me to believe that Montford really is not a scientist. I know, I know, some will accuse me of being nitpicky, but careful analysis and correct writing go hand in hand in science, and not learning that "data" is a plural word, and then repeatedly using it as a singular when discussing scientific research shows a degree of sloppiness and lack of scientific professionalism. Here is the quote, at any rate:]

"It is important to realise, however, that this result is only achieved if the data is first centred. Because of this, centring is considered an integral part of PC analysis."