The Hockey Stick Goes to Congress

Posted on November 5, 2012


The so-called “hockey stick” graph, displaying dramatic rises in the Earth’s temperatures over the last century, has been a major bugaboo for climate change deniers ever since it took a starring role in Al Gore’s “An Inconvenient Truth.”  As Mann points out in the following testimony before Congress, the focus on the original 1999 graph has been to the exclusion of Mann’s subsequent work, and the work of many other scientists which has duplicated the graph again and again.  In 2006, Mann was summoned to testify before a hostile panel of climate-denying Republicans, anxious to discredit him.  To bolster the denier side, they brought in Edward Wegman, a statistician with no climate understanding at all, who made an entirely statistical critique of Mann’s work on the 1999 graph, which Mann refutes here.  

Wegman has since been thoroughly discredited, and a published paper of his based on his congressional testimony was withdrawn by its publisher for plagiarism.  


The following represents a somewhat condensed version of Mann’s testimony to congress that day.  His full testimony, with graphs and exhibits, may be found in the accompanying video by Peter Sinclair.

Partial Transcript of Michael Mann’s Testimony to Congress, July 27th, 2006

….I became a climate scientist because the Earth’s climate is a fascinating and complex system and understanding how it works is so important. Part of my research has involved examining pre-industrial climate history in order to learn about the natural variations in the Earth’s climate….

Of course, we have accurate thermometer measurements only back about a 100 years and so we estimate climate prior to that period from indirect sources called climate proxies such as tree rings, corals and ice cores.  This work involves many uncertainties, and there are numerous judgment calls that must be made.  For that reason, we are rarely categorical in the conclusions that we reach.

What is important, however, is that the scientific community has reached consensus that recent Northern Hemisphere average warmth appears to be unprecedented over at least the past 1000 years, and that this warmth can only be explained by anthropogenic or human influences on the climate.

This conclusion is not based on single studies or isolated research, but is confirmed by many studies using different sets of data and independent statistical methods, and indeed this conclusion was just echoed weeks ago by a report of the National Academy of Sciences, the most prestigious non-partisan scientific body in the nation.

So where does my research fit into this? Taken as a whole, my own research is in accord with the scientific mainstream reflected in the National Academy report and elsewhere…

But this committee is not looking at my work on the whole or on the larger body of science on this issue.  It is instead focusing on the first study of this type my colleagues and I published and undertook in 1996 when I was still a graduate student.

While there were previous reconstructions based on proxy data, our study was the first to estimate global patterns of past temperature change and the first to estimate uncertainties.  Our initial study published in the journal Nature in 1998 was followed by an additional study in the journal Geophysical Research Letters in 1999.

The main conclusion of the 1998 study was that there had been unprecedented warming in the Northern Hemisphere in recent decades.  The 1999 study reinforced this conclusion but also reassessed and expanded the uncertainties and added a tentative conclusion that it was likely the 1990s were the warmest decade over that 1000 year time period and that 1998 was the warmest year.

The 1999 study included a graphic depiction of the temperature history over the last millennium which demonstrated an unprecedented rise during the twentieth century.  Some have dubbed this graphic “the hockey stick.”

If the question this committee seeks to answer is whether, knowing what I know today, a decade after starting the original study, my colleagues and I would conduct it in exactly the same way, the answer is plainly no.  The field of paleoclimate reconstruction has evolved tremendously over the past decade.  Important new proxy data have been developed.  Reconstructions have been compared with independent estimates from climate model simulations and confirmed by those simulations.  Statistical methods for reconstructing climate from proxy data have been refined and rigorously tested.  And I have been actively working in each of these areas.

This is important because all the focused criticism on our work in the late 1990s has been on the statistical conventions we used.  My colleagues and I have not used those conventions in our later work.

The critique goes only to our first reconstruction effort.  It does not apply to our more recent studies.  All of which indicate the same basic “hockey stick” result.

Now our critics do not confront the fact that our basic conclusion is not an isolated or aberrational finding reached only in one study.  Every climate scientist who’s performed a detailed reconstruction of the climate of the past 1000 years, using different proxy data, using different statistical methods, has come up with same basic “hockey stick” pattern.  That is to say, a reconstruction that agrees with our original reconstruction within its estimated uncertainties.

My critics also fail to recognize that even if their criticisms are accepted it has no bearing on the outcome.  Dr. Wegman’s report argues that the “hockey stick” pattern derives from the statistical conventions used in our 1998, 1999 studies.  However, using alternative statistical conventions yields the same “hockey stick” pattern.  The “hockey stick” pattern is intrinsic to the data.

That was the conclusion of the National Academy.  Page 116 of the National Academy report says the statistical convention my colleagues and I used, quote, “does not appear to unduly influence reconstructions of hemispheric mean temperature; reconstructions performed without using principal component analysis are qualitatively similar to the original curves presented by Mann et all,” end quote.

This was also the conclusion reached by Doctors Hunt and Stork who testified here last week and by four independent teams of scientists who published peer-reviewed articles considering and rejecting the conclusion that the statistical methods used in our early studies were responsible for the “hockey stick” result.

Finally, my critics ignore the fact that other scientists have repeated our original results using the centred [principal component] analysis that Dr. Wegman favours, and have concluded that the result is basically the same as we originally reported.

So even if one accepts as valid the criticisms about the statistical conventions used in our early work, our results are essentially unaffected…. …the two curves are barely distinguishable within the width of the lines that are shown.

And, as I have said before, our key conclusion that recent hemispheric warmth appears unprecedented over at least the past millennium has been confirmed by every study that has examined the same question.

Finally, it’s worth pressing again that paleoclimate reconstructions represent just one of many independent lines of evidence that support the conclusion that human activity is already having a substantial impact on global climate.  –Dr. Michael Mann, 27 July 2006