Overview - The "Hockey Stick" and Historic Climate Change

Global climate is variable. Over extended periods regional and global average temperatures show wide variation driven by many factors, only some of which can be tied to human activity. This is particularly telling in past centuries prior to the industrial age when the human race had little or no ability to impact climate significantly. To access the reality of anthropogenic, or human-caused, global warming it is necessary to separate the impact of our activities from the natural variations in climate that have been occurring for millions of years.

Sorting out long-range historical climate, or paleoclimate, is no trivial matter. Instrument-based temperature measurements do not exist prior to the 19th century and what little exists before the 20th is less than reliable. As such, very long-term temperature trends have to be inferred indirectly through their effects on proxies—data for which historical records do exist (e.g. coral reef growth, glacier advance and retreat, tree-rings, ice cores, etc.). Studies like this have been done, and within the range of uncertainty all show that the global temperature trends of the last century have been unprecedented in the historical record for at least 600-1000 years if not longer. Furthermore, similar trends in pre-history all correlate with natural events that have no present-day counterparts. This is strong evidence that humans are impacting climate in dangerous ways. Graphical representations of this trend show temperatures fluctuating within a bounded range for tens of thousands of years, followed by an exponential rise in temperature in the last 100 years that greatly exceeds anything preceding it. Jerry Mahlmann of the NOAA Geophysical Fluid Dynamics Laboratory noted that the curve looks very much like a hockey stick, and the name has stuck ever since.

Other than the recently debunked controversy over troposphere temperatures, no other temperature record been subjected to more scrutiny and politicization than this one. The seminal hockey stick work was published in 1998 in the journal Nature by Mann et al. (MBH98). Mann’s team studied tree-ring, ice core, and ice melt proxy data supplemented with historical instrument records as far back as they could be obtained and compared the resulting Northern Hemisphere surface temperature trends with solar, volcanic, and greenhouse gas trends for the same period. They found that greenhouse gases (particularly those of human origin) have been the single most important driver of global temperature during the industrial age, and that the warming observed during the last 100 years is unprecedented at least as far back as 1400 AD.

To understand the controversy over these results we need to understand how Mann’s team evaluated their datasets. Unless a dataset gathered over time is completely random (few are) that variations in the data will “ride on the back” of an underlying trend that tells the real story, blurring the vision of it so to speak. Various mathematical techniques can be used to “sharpen the focus” and separate this underlying trend from the noise obscuring it. Mann’s team applied one such method to their proxy datasets called Principal Component Analysis, or PCA. PCA analysis reduces a number of unique mathematical building blocks called “components” and then removes the blocks that comprise the noise, leaving only the “foundation” blocks behind. The more components used in an analysis, the more complete and reliable the resulting reconstruction (foundation of component building blocks) is. This is not unlike sanding a rough block. Each successive increase in components goes over the block with increasingly fine “grit” sandpaper. Eventually, a level of smoothness is obtained for which increasingly fine sandpaper yields diminishing returns in smoothness—the perfect underlying shape of the block has been reached and we know that we’ve included fine enough sandpaper.

Nearly all criticism of the hockey stick has come from industry funded Astroturf organizations and Far-Right forums (the few that do not do not dispute the existence of the hockey stick itself). Most is little more than poisonous ad-hominem and politicization. However, at least two noteworthy attacks have come from this community, both of which claim that MBH98 did not “sand the block” enough to show that 20th century temperature variations are not unprecedented during the last millennium. The first appears in two papers published in 2003 by Willie Soon and Sallie Baliunas (both of whom are paid consultants for Astroturf fronts), one inClimate Research and another in Energy and Environment (the latter with co-authors). Soon and Baliunas and co-authors (SB) analyzed Northern Hemisphere temperature proxy records for the last 1000 years looking for unusually warm periods, or anomalies and concluded that 20th century warming was not unprecedented during this period. Mann’s team, they claimed, had either not done their PCA analysis correctly, or had omitted important data.

As it turned out, both papers were riddled with errors including flawed analysis methods and improper use of data. Most significantly SB had, The publication of the first paper generated much controversy in the scientific community leading to an investigation of the peer-review standards at Climate Research. Soon thereafter the senior editor and three other editors resigned in protest over the manner in which the SB paper had been reviewed. The journal’s publisher (Otto Kinne) eventually conceded that "[the conclusions drawn] cannot be concluded convincingly from the evidence provided in the paper."

Interestingly enough, this was not the first time that Soon and Baliunas had been embroiled in controversy over substandard work. Both were co-authors of the unpublished 1998 paper that accompanied the OISM Global Warming Petition Project. In addition to being filled with errors, incomplete references, and out-of-date calculations, that paper was involved in one of the most significant plagiarism scandals of the 90’s. The original mailing of the petition contained a print version of it that had been deliberately typeset in a publication format almost identical to that of the Proceedings of the National Academy of Sciences—a highly respected journal that had not published it—in an apparent attempt to leverage off of that journal’s credibility.

To make matters worse, one of the petition’s authors was a former president of the NAS, but made little effort to clarify his current role (or lack thereof) with the Academy. Shortly after its release the NAS released a strongly worded press release condemning the petition project, the paper, and the unprofessional manner with which it had been prepared and circulated. Few other incidents in recent history have done more to damage the credibility of global warming skeptics and Far-Right antienvironmentalism. Yet to this day, many Far-Right and Astroturf forums still cite the petition project as “proof” that a majority of “scientists and engineers” do not believe human caused global warming is happening!

Energy & Environment is another interesting story. It is not even a scientific journal. It’s a social science journal devoted mainly to issues of interest to industry, and not surprisingly receives most of its funding from Astroturf sources. In the wake of the SB controversy at Climate Research, editor Sonja Boehmer-Christensen defended EE’s publication of the Soon et al study in an interview in the Chronicle of Higher Education by stating that she was "…following my political agenda—a bit, anyway. But isn't that the right of the editor?". Today the two papers are not taken seriously anywhere but in Astroturf and Far-Right forums that are, in general, unacquainted with the relevant science.

Another assault on the hockey stick was published in Geophysical Research Letters in 2005 by industry consultant Stephen McIntyre and economist Ross McKitrick. McIntyre and McKitrick (MM) performed their own PCA analysis of the proxies used by MBH98 and argued that their “building blocks” had overemphasized one particular proxy (bristlecone pine tree rings). As a result, they claimed, the MBH98 analysis was bound to produce a “hockey stick” shape no matter what data were used and did not prove that 20th century temperatures were historically unprecedented. Unlike the works of Soon and Baliunas, the MM paper managed to survive an acceptable peer-review process because it lacked its predecessors’ obvious blunders and amateurishness (Geophysical Research Letters is a widely respected journal that has a far better editorial record than Climate Research or Energy and Environment). It was received with much fanfare in skeptic circles and hailed as the war club that had broken the hockey stick for good.

But alas, MM’s analysis also fell to inadequate data processing methods. The one common thread running through all skeptic criticisms of the hockey stick is that MBH98 did not “sand the paleoclimate block” fine enough to prove that 20th century temperatures were unprecedented. But this claim has little meaning unless we clarify what is meant by “fine enough” and “unprecedented”. It is not enough to show that some historical dataset (or time series) has uncertainties—all do. We must show that its uncertainties are large enough to invalidate the underlying trend that has been ascribed to it. To date, no skeptic criticism of the hockey stick has done this. Virtually all have been generalized accusations of “uncertainty” and/or “variability” in the proxy datasets—nothing specific enough or demonstrably large enough to invalidate the Mann et al results.

In this case MM used a different set of components that they felt did not over-represent the tree ring data but failed to recognize that changing the components used changes the number required for an accurate answer (smaller, or differently shaped bricks will change the number of bricks needed for a foundation, and how they should be placed). They did their reconstruction using the same number of components that Mann et al used, but it has since been shown that their method requires more components to be reliable. When this problem is corrected their analysis does not significantly differ from that of MBH98. A statistically significant hockey stick shape emphasizing 20th century warming is inherent in the resulting time series.

Furthermore, other paleoclimate studies that used more complete datasets and different methodologies altogether (for instance, Moberg et al, 2005 and Rutherford et al, 2005). These analyses are not subject to the criticisms of MM05 yet yield results similar to those of MBH98. In June of 2006 the National Academy of Sciences published the results of its investigation of MBH98 and the hockey stick. They concluded that while there were certain problems with the MBH98 treatment of tree ring proxies, their analysis had been adequately complete given the data available at the time, and the resulting errors did not significantly alter the study’s conclusions. Despite this affirmation, many skeptics decided that the NAS study “vindicated” their complaints about MBH98… simply because it acknowledged the presence of some errors. Once again, a touting of “errors” and “uncertainties” without consideration of their impact (one is reminded of the methods of fundamentalist creationists and their claims of alleged “problems” with evolution—problems that somehow never end up being quantified or even clarified adequately!).

Finally, even if we were to concede to MM that 20th century warming was not “unprecedented” in the last 600-1000 years, this is still a far cry from proving that it is commonplace and not due to anthropogenic activity. At best, on the few occasions that climate skeptics have gotten specific on this point they’ve managed to come up with no more than one or two instances during the last 1000 or more years when they believe temperatures were higher. This does not relegate the last century to business-as-usual that we needn’t worry about. It may be that the guy up the block from me ended up with a broken basement window during a storm 10 years ago. It doesn’t follow from this that I shouldn’t be concerned about my house being broken into if I awake tonight at 3 AM to the sound of my ground floor windows shattering!

In the end, skeptic criticisms of the hockey stick are no different than their treatment of so many other climate change datasets. Highly generalized accusations of alleged problems and/or uncertainties for which defensible specifics are never provided. It comes as little surprise that the large majority of climate scientists today recognize that 20th century temperatures are in fact disturbingly unusual, and a direct consequence of human caused activity.