The hockey stick

Why now really is different

It was a seductive image. So seductive that the IPCC put it right at the front of its thousand-page assessment of climate change, published in 2001. The panel hoped that it would become as talked about as the Keeling curve. And scientists gave it a snappy caption: this was the graph they called the "hockey stick." As I don't play hockey, I was initially left wondering why. But if you lay a hockey stick on the ground and look at its shape as a graph, you will see that the long, flat shaft has at the end of it a short but sharply upturned blade, the bit you hit the puck with. And that, according to the IPCC authors, is the shape of the world's temperatures over the past thousand years: about 900 years of little or no change, followed by a century with a short, sharp upturn.

The assembly of the data behind the hockey stick graph has become a political cause célèbre. It began with high hopes: it was to be the first serious attempt to piece together a global picture of climate over the past millennium from a wide variety of different kinds of sources. Rather than carrying on the well-established work of reconstructing past temperatures from analysis of tree rings, it sought to add in other proxy data from ice cores, coral growth rings, and lake sediments. The idea was to lose the built-in bias of tree-ring chronologies, which must rely on trees from Northern Hemisphere regions outside the tropics, because those are the trees with well-defined annual growth rings.

The hockey stick graph was first put together in 1998. The politics soon got going. That year turned out to be the warmest in the instrumental record. So it wasn't much of a stretch to argue that the hockey stick revealed 1998 to be the warmest year in the warmest century of the past millennium. That got headlines. And brought trouble—not least for the voluble, self-confident, and likable collator of the hockey stick data, Mike Mann. Even though the IPCC published other data sets showing much the same, Mann was accused of concocting a spurious case that late-twentieth- century warming was exceptional and therefore, presumably, a result of man-made pollution.

It probably didn't help that at the time, Mann was based at the University of Virginia, home of the biggest voice among the climate skeptics: Pat Michaels. Soon Mann was fraud-of-the-month on the Web sites of the climate skeptics. But the criticism went beyond the normal community of climate skeptics: some serious climate researchers expressed misgivings about Mann's methods.

When I finally met Mann, he had moved from Virginia to Penn State University, where he is now director of the Earth Science Systems Center. But the flak had followed. Some was fair; some was unfair; some was deployed as political hand grenades; some was just a part of the normal adversarial flow of scientific debate; and some was just plain personal —like Wally Broecker's bad-mouthing of Mann, quoted at the start of Chapter 23. Mann was even damned in Washington, where Senator James Inhofe of Oklahoma accused him of playing fast and loose with the data, and Representative Joe Barton of Texas summoned him to provide his committee with voluminous details about working procedures and funding. Some called it a McCarthyite vendetta. But Mann seemed up for it, dismissing Inhofe as "the single largest Senate recipient of oil industry money."

I will now entertain some of the criticisms that have rained on Mann, because they matter. But it is worth saying first that nothing I have heard impugns Mann's scientific integrity, credentials, or motives. He is just braver than some, and more willing to have his debates in public—and to fight back when the brickbats start flying. (You can read him in action on the Web site he started with scientific colleagues at www.realclimate.org.) Some researchers have suffered real personal and psychological damage from attacks by skeptics. I hope that won't happen to Mann. I wish more scientists were like him.

First, does the hockey stick fairly represent the temperature record? Does Mann's take-home conclusion, that the last century warmed faster and further than any other in the past thousand years, stand up to scrutiny? The short answer is yes—but only just.

The world of proxy data trends is a statistical minefield. This is partly because the physical material that shows past climate loses detail with time. Tree rings, for instance, get smaller as the tree gets older, so annual and even decadal detail gets lost. "You lose roughly 40 percent of the amplitude of changes," says the tree ring specialist Gordon Jacoby, of Lamont-Doherty. But it goes far beyond that. To make any sense, analysis of a single data set—for instance, from the tree rings in a forest—involves smoothing out the data from individual trees to reveal a "signal" behind the "noise" of short-term and random change. The kind of analysis pioneered by Mann, in which a series of different data sets are merged, involves further sorting and aggregating these independently derived signals, and smoothing the result. And Mann's work involves a further stage: meshing that proxy synthesis with the current instrumental record.

Some, including Jacoby, complain that by combining smoothed-out proxy data from past centuries with the recent instrumental record, which preserves many more short-term trends, Mann created a false impression of anomalous recent change. "You just can't do that if you are losing so much of the amplitude of change in the rest of the data," Jacoby told me. Mann argues the contrary—that in fact he was one of the first analysts in the field to include error bars on his graph. "The error bars represent how much variance is lost due to the smoothing," he says.

But the accusation that he has somehow fixed the data analysis continues to dog him. The most persistent line of criticism, and the one most widely championed by anti-IPCC lobbyists, came from two Canadians: Stephen Mclntyre, a mathematician and oil industry consultant, and Ross McKitrick, an economist at the University of Guelph. They claimed to have found a fundamental flaw in Mann's statistical methodology that biased the temperature reconstruction toward producing the hockey stick shape.

The argument is a technical one that hangs on how Mann used well-established mathematical techniques for classifying data called principal component analysis. Mclntyre and McKitrick claimed that Mann's method had the effect of damping down unwanted natural variability, straightening the shaft of the hockey stick and accentuating twentieth-century warming. Mann agrees there was some truth in this charge. He analyzed the data in terms of their divergence from twentieth-century levels, and this had the inevitable effect of giving greater significance to data showing the biggest differences from that period.

But the critical charge was that he had somehow created the hockey stick out of nothing—"mining" the data for hockey-stick-shaped trends, as his critics put it. Mclntyre and McKitrick produced their own analysis, showing an apparent rise in temperatures in the fifteenth century, which, they claimed, may have been as warm as the twentieth century. The shaft of the hockey stick had a big kink in it. When it was published, in 2005, this analysis was hailed by some as a refutation of Mann's study.

But while Mann was open to attack, so were Mclntyre and McKitrick. Would their "refutation" of Mann stand up to critical attention? During 2005, three different research groups concluded that Mann's findings bear scrutiny much better than do those of his critics. They had bent the statistics more than he had, arbitrarily leaving out certain sets of data to reach their conclusion. Remove all the biases, and the real data looked more like Mann's—a conclusion underlined in early 2006, when Keith Briffa, a respected British tree-ring analyst at the University of East Anglia, published the most complete analysis to date, showing the twentieth century to have been the warmest era for at least the past 1,200 years. Briffa's take was confirmed in June 2006 by the U.S. National Academy of Sciences, which, in a long-awaited review of the hockey-stick debate, endorsed Mann's work. The analysts expressed a "high degree of confidence" that the second half of the twentieth century was warmer than any other period in the previous four centuries. But they said that although many places were clearly warmer now than at any other time since 900, there was simply not enough data to be quite so sure about the period before 1600.

If the key to successful science is producing findings that can be replicated by other groups using different methodologies, then Mann is on a winning streak. Upward of a dozen studies, using both different collections of proxy data and different analytical techniques, have now produced graphs similar to Mann's original hockey stick. Not identical, for sure, but with the same basic features of unremarkable variability for 900 years followed by a sharp upturn in temperatures in the final decades.

The one unexplained factor is that most of these studies show paltry evidence for the medieval warm period and the little ice age. But an answer to that conundrum now seems at hand. There is growing agreement that the most substantial evidence for the existence of both a medieval warm period and a little ice age comes from the northern latitudes. "What we know about the cold in the little-ice-age era is primarily a European and North Atlantic phenomenon," says Keith Briffa. Most interesting, there is growing evidence from a range of new proxy data that other parts of the world were seeing climate trends opposite to those going on in Europe. The tropical Pacific appears to have cooled during the medieval warm period and warmed during the little ice age. One ice core from the Antarctic shows temperatures during the medieval warm period that were 5°F colder than those in the little ice age. Under the circumstances, says Mann, it is not surprising that his more global assessment of temperatures does not spot much difference during these earlier climatic shifts. They undoubtedly had major influences on regional climates, but the cumulative effect on global temperature was small.

It is no part of this book's case that climate didn't change in the past. Parts of the world clearly saw substantial warming and cooling during the medieval warm period and the little ice age. Other parts saw other changes. In the American West, there were huge, century-long droughts during the medieval warm period. Even Broecker, who holds that the little ice age was global, admits that the evidence of a global medieval warm period is "spotty and circumstantial." But there is a good case for saying that over the millennium until the mid-twentieth century, most climate change concerned the redistribution of heat and moisture around the globe rather than big changes in overall heating. Only recently has there been a major additional "forcing," caused by the introduction of hundreds of billions of tons of greenhouse gases into the atmosphere. Recent warming may be the first global warming since the closure of the ice age itself.

The argument over the hockey stick is an interesting sideshow in the debate about climate change. But it remains a sideshow. Right now, it matters little for the planet as a whole whether the medieval warm period was or was not warmer than temperatures today—or, indeed, whether it was a warm period at all. The subtext of the climate skeptics' assault on Mann's hockey stick has always been that if the current warming is shown not to be unique, then somehow the case that man-made global warming is happening evaporates. But this is a spurious argument. Briffa is not alone in arguing precisely the opposite. If it was indeed very warm globally in the medieval warm period, that is truly worrying, he says. "Greater long-term [natural] climatic variability implies a greater sensitivity of climate to forcing, whether from the sun or greenhouse gases. So greater past climate variations imply greater future climate change."

Was this article helpful?

0 0
The Basic Survival Guide

The Basic Survival Guide

Disasters: Why No ones Really 100 Safe. This is common knowledgethat disaster is everywhere. Its in the streets, its inside your campuses, and it can even be found inside your home. The question is not whether we are safe because no one is really THAT secure anymore but whether we can do something to lessen the odds of ever becoming a victim.

Get My Free Ebook


Post a comment