Final report

The Hockey Stick

A Survival Guide for Climate Conversation

DYN-CLIMATE VAR AND CLIMATE CHG, Columbia University New York

Daniela Domeisen, Joshua Gellers and Heather Raven

Dec 10th, 2006 1 INTRODUCTION 2

1 Introduction

When the new report of the IPCC (Intergovernmental Panel on ) was published in 2001, it led to extensive debate among scientists and in politics. The IPCC report had put a lot of emphasis on the hockey stick, a diagram shown within the report, specifically in the ’Summary for Policymakers’, where it was the cause for much discussion. From the last IPCC report and the news, the public still believed that scientists were not too sure about climate change, but then the hockey stick was published. The hockey stick is a diagram representing global warming for the last thousand years. It got its name due to its long, constant shape with a sudden upward slope in the end which looks similar to a hockey stick.

2 Overview

The graph shows the diagram published in the IPCC report which caused a lot of discussion about global warming.

Figure 1: The Hockey Stick Model as it was presented in the ’Summary for Policy Makers’ within the IPCC report 2001

The diagram shows a significant increase in temperature for the last few decades, indicating that since industrialization, global temperatures have been rising much faster than ever before. Therefore, it was concluded in the report that the warming had an anthropogenic influence which caused temperatures around the earth to rise and thereby impact humans as well as entire ecosystems. The hockey stick represents the average global temperature to many scientists who work on the IPCC reports. The task of the IPCC, which was founded in 1988, is the presentation of the actual state of in the various fields of climate change, the estimation of the consequences of climate variation for envi- ronment and society, and the formulation of realistic mitigation strategies. The report is published every few years and the next report is set for publication in 2007. The consensus of the IPCC consists of the following main points: 1) Global average temperature on the Earth’s surface has been rising by about 0.6 ± 0.2◦C until 2001, compared to the 30-year from 1961 to 1990, and by about 0.17◦C per decade during the last 30 years. 2) Most of the warming during the past 50 years can be attributed to humans, especially the emission of greenhouse gases such as CO2 and methane. 3 AFFIRMATIVE VIEW (JOSH GELLERS) 3

3) Depending on the future development of the emissions of greenhouse gases, Earth will be warming by 1.4◦C to 5.8◦C by the end of the 21st century. The publication of the hockey stick curve led to major controversies among scientists as well as in political arenas. Scientists who do not believe that climate change will be as dramatic as predicted by the IPCC report published papers contradicting the report. Various opinions exist about how much the earth will warm and in what time. However, most scientists agree that there is a warming trend and that it is due to anthropogenic influence, but since it is impossible to do an accurate prediction for the future, due to many unknown factors such as future emissions and political/policy change as well as positive and negative feedbacks of climate change, there are large differences between the future projections for warming. This disagreement has misleadingly been interpreted as a debate about whether climate change is happening at all.

3 Affirmative view (Josh Gellers)

The so-called ’hockey stick’ model developed by Mann et al (hereafter ’MBH98’) has been under fire for its methodology and statistical analysis ever since it was first published in Geophysical Research Letters. Due to the high visibility of the study thanks to its prominent inclusion in the United Nations Intergovernmental Panel on Climate Change Third Assessment Report in 2001, MBH98 has been subject to an almost unprecedented amount of criticism, mostly by a few determined naysayers. However, the condemnation brought onto the model itself is largely unwarranted and yet the scientific community remains largely in accord with the findings of MBH98. First, the degree to which the hockey stick model is at the very least valid or at the most alarming has been egregiously overstated. The title of the initial paper from which this controversy derives is quite telling; Northern Hemisphere Temperatures During the Past Millennium: Inferences, Uncertainties, and Limitations. This unabashedly modest headline speaks to the conviction with which the original researchers assert their results- not much. The authors even begin their seminal work with a significant admonition; We focus not just on the reconstructions, but the uncertainties therein, and important caveats our results suggest that the latter 20th century is anomalous in the context of at least the past millennium (MBH98; emphasis added). Furthermore, doubts involving the quality of the paleoproxies used have circulated unjustly. Only a careless observer who has not read the literature would castigate the model for neglecting to address issues surrounding the validity of the proxies as MBH98 clearly states, a more widespread network of quality millennial climate indicators will be required for more confident inferences. Considering that the overarching extension of the hockey stick model was not initiated by its designers, one cannot place blame the questionable accuracy of the model solely on MBH98 by virtue of its unintended overuse. Second, despite its flaws the hockey stick model has earned the lukewarm endorsement of the U.S. National Academy of Sciences (About the NAS 2006). The NAS did not wholly adopt the model, however, contending that the has been highly politicized, often extending the fledgling research well beyond its deserved territory. According to Gerald North, chair of the NAS committee, the organization roughly agree[s] with the substance of [MBH98s] findings. North went on to state that the committee has a high level of confidence that the second half of the twentieth century was warmer than any other period in the past four centuries (Brumfiel 2006). Quite simply, the brouhaha about the hockey stick model is mostly overblown and exaggerated. It has never been the intention of the authors to provide a fail-safe graphic depicting a definitive warming trend in the Earths climate, despite the claims made by the models opponents. Ignorant of MBH98s unpretentious title, hockey stick contrarians have senselessly drawn and quartered Mann et al for their efforts. McIntyre and McKitrick, two of the most notorious hockey stick antagonists, have mounted quite a challenge to the model, though their endeavor has been almost entirely misdirected as the authors have never attested to the infallibility of their product. However, the attention given to them in the mainstream media has suffocated more important hockey stick-related revelations, such as the models moderate acceptance by the highly reputable National Academy of Sciences. To this day, Mann et al makes no attempt to disguise the uncertainties inherent in the hockey stick model and maintains their conclusion that both the past decade and past year [1998] are likely the warmest for the Northern Hemisphere this millennium all while 4 HOCKEY STICK MODEL SKEPTICAL VIEW (HEATHER RAVEN) 4 working to continuously refine the methods of their research (MBH98; authors emphasis).

4 Hockey Stick Model Skeptical View (Heather Raven)

Even though the ’hockey stick’ model developed by Mann et al. (1999) has been used and referenced by well-renowned organizations such as the Intergovernmental Panel on Climate Change (IPCC, Third Assessment Report, 2001) and the National Academy of Sciences (NAS), there are a few significant errors in the production of the model and its use. The general public with little or no information on climate change could look at this graph as an indisputable indication that the climate is warming rapidly and that it has never happened before, at least not in the past thousand years, at this unprecedented rate. They also might point their fingers at anthropogenic causes for the warming, yet this is not what the graph is telling us. The hockey stick model is simply showing us how much the Northern Hemisphere (NH) temperatures depart from the 1961-1990 AD average (i.e. temperature anomalies in the NH). It is important to remember that this graph and the resultant anomalies are deviations from only a 30-year record of temperatures rather than a long-term record of what is ’average’ for the NH. The anomalies in the hockey stick from 1900 to 2000 AD seem to be experiencing an even greater positive deviation from average which is not seen in an earlier time period in the graph; this positive deviation happens to roughly correspond to the time when instrumental records were available for use in this graph, the prior time periods reconstructed using proxy data. Proxy data is gathered from tree rings, corals, ice cores, and historical records (Mann et al., 1999). Difficulties arise when using proxy data since there are limits to each method: ice cores can only be taken from a small fraction of the global surface, corals may not be reliable for use on long timescales due to changing temperature and salinity of ocean waters, and large samples of trees from many sites are needed to analyze tree rings for a change in climate (Jones and Mann, 2004). The use of proxy data from 1900 to 1998 AD seems to correspond with the instrumental record, but the proxy data from any earlier timeframe contains a large amount of error, error that could put past temperature deviations possibly on par with the anomalies experienced in the NH today. Prior to 1600 AD, the error can be seen to increase even further as proxy reconstructions become more uncertain. Mann even admitted that uncertainties prevent decisive conclusions for data collected prior to 1400 AD (, Corr., 2006), hence the large error in the ’stick’ of the hockey stick graph. Uncertainties arise from the use of ’insufficient data and flawed statistical analysis’ (McKitrick and McIntyre, as referenced in Nature, News) of the proxy data temperature and climate reconstructions. Wegman’s (statistician, ) analysis concludes that Mann et al. (1999) cannot claim that the 1990’s were the hottest decade in the past millenium, given their methods of analysis (suppression of low-frequency variation, excluding large regional warm events that could have affected overall anomaly results) [Figure 2]. Wegman also points out that the results by Mann et al. (1999) are not wholly independently verified since they were peer-reviewed within the paleoclimate community [a statistician would have been a good choice for reviewing the work by Mann et al. (1999)]. Indeed, Mann himself concedes that higher resolution data is needed before ’more confident conclusions can be reached’ about their study’s results (Mann et al., 1999), ultimately decreasing the error in the past proxy reconstructions and statistical methods. A skeptic’s point of view is easy to find in the literature, especially given the topic of the hockey stick model, a model peppered with political, scientific, and societal implications for the future; however, maintaining an objective view of the hockey stick, its benefits and flaws, will be valuable in all conver- sations related to climate change and specifically, regional/global warming. The hockey stick is not a ’global’ model, nor is it quite representative of the Southern Hemisphere (SH) [Figure 3, model outputs of SH and Globe in form of ’hockey stick’] but it can be used as a tool for understanding how temperature anomalies are changing in the NH. It has been redeveloped to include projections of future anomalies in temperature [Figure 4] but these projections contain great amounts of error and predict a warming somewhere between 1◦C to 5.7◦C. The future of the hockey stick model, whether the positive anomaly trend will grow to unprecedented values, will be further contested as improved data is used for the model and human impacts on climate are resolved. (Mann et al., 1999), ultimately decreasing the error in the past proxy reconstructions and statistical methods. A skeptic’s point of view is easy to find in the literature, especially given the topic of the hockey stick model, a model peppered with political, scientific, and societal implications for the future; however, maintaining an objective view of the hockey stick, its benefits and flaws, will be valuable in all conversations related to climate change and specifically, regional/global warming. The hockey stick is not a ‘global’ model, nor is it quite representative of the Southern Hemisphere (SH) [Figure 2, model outputs of SH and Globe in form of ‘hockey stick’] but it can be used as a tool for understanding how temperature anomalies are changing in the NH. It has been redeveloped to include projections of future anomalies in temperature [Figure 3] but these projections contain great amounts of error and predict a warming somewhere between 1 ºC to 5.7 ºC. The 4 HOCKEY STICKfuture of MODEL the hockey SKEPTICAL stick model, whether VIEW the (HEATHERpositive anomaly RAVEN) trend will grow to 5 unprecedented values, will be further contested as improved data is used for the model and human impacts on climate are resolved.

Figure 1. In this figure from Jones and Mann, 2004, they show regional temperature Figure 2: In thisanomaly figure from data from Jones about and 1660 Mann, to 2000. 2004, Note they how showregional regional change varies temperature greatly from anomaly the data from about 1660 to 2000.Northern Note Hemisphere how regional average changeanomaly (instrumental varies greatly data). from Jones the and NorthernMann state that Hemisphere average there is “limited agreement between the NH and the European regions” and that the NH anomaly (instrumentaldata cannot data). be inferred Jones from and a single Mann region. state However, that there Mann is et ’limited al. (2003a) agreement [referenced betweenin the NH and the EuropeanJones regions’ and Mann, and that2004] thethenNH try “tentative data cannot extensions” be inferred of the NH from mean a singleusing data region. that However, Mann et al. (2003a) [referencedthey acknowledge in Jones provides and Mann, a “highly 2004] uncertain” then estimate try ’tentative of NH temperature extensions’ change. of the NH mean using data that they acknowledge provides a ’highly uncertain’ estimate of NH temperature change.

Figure 2. Jones and Mann, 2004, show the average of temperature anomalies (proxy and Figure 3: Jonesinstrumental) and Mann, over 2004, the show NH, Southern the average Hemisphere of temperature and the estimated anomalies Global (proxy average. and instrumental) It is over the NH, Southerninteresting Hemisphere to note the error and bars the and estimated the anomalies Global above average. 0 ºC It in isthe interesting SH. These toare note all the error bars and the anomaliesvariations aboveof the 0‘hockey◦C in the stick’ SH. model, These although are all variations the original of model the ’hockey only depicts stick’ model,NH although the original modeltemperature only depicts anomalies. NH temperature anomalies. 5 REFERENCES 6

Figure 4: Nature, Vol. 441, 2006. The hockey stick graph by Mann et al., 1999, with a future projection of temperature anomalyFigure 3.increase Nature, Vol. from 441, average. 2006. The Note hockey error stick in futuregraph by projections Mann et al., as 1999, great with as a 2.5◦C. future projection of temperature anomaly increase from average. Note error in future projections as great as 2.5 ºC. 5 References

• ’About theReferences NAS.’ National Academy of Sciences. 2006. 4 Dec. 2006. http://www.nasonline.org/site/. • Brumfiel, Geoff.Appell, D.Academy (2005) Behind affirms the hockey-stickHockey Stick. Scientific graph Nature. American 4, Insights, Dec.2006. March, pp. 28 34- Jun. 2006. http://www.nature.com/news/2006/060626/full/4411032a.html35.

• Mann, ME,Bradley, RS Bradley R.S., Hughes, and MK M.K., Hughes. Mann, NorthernM.E. (2006) hemisphere Authors were temperatures clear about hockey-stick during the past mil- lennium: Inferences,uncertainties. uncertainties, Nature, Correspondence, and limitations Vol. 442,Geophysical pp.627. Research Letters, 1999. 26 (6): 759-762. Brumfiel, G. (2006) Academy affirms hockey-stick graph. Nature, News, Vol. 441, pp. • IPCC Third1032-1033. Assessment Report, Climate Change 2001: Synthesis Report, ’Summary for Policy- makers’, 2001 Jones, P.D., Mann, M.E. (2004) Climate Over Past Millennia. Reviews of , • Appell, D.Vol. (2005) 42, BehindNo. 43, pp. the 1-42. Hockey Stick. Scientific American, Insights, March, pp. 34-35.

• Bradley, R.S.,Tippee, Hughes, B. (2005) M.K., Questions Mann, haunt M.E. global (2006) warmingAuthors ‘hockey were stick’. clear aboutOil & Gas hockey-stick Journal, uncertain- ties. Nature,Vol. Correspondence, 103, No. 26, pp.68. Vol. [Used 442, in pp.627.Presentation] • Brumfiel, G. (2006) Academy affirms hockey-stick graph. Nature, News, Vol. 441, pp. 1032-1033. • Jones, P.D., Mann, M.E. (2004) Climate Over Past Millennia. Reviews of Geophysics, Vol. 42, No. 43, pp. 1-42. • Tippee, B. (2005) Questions haunt global warming ’hockey stick’. Oil and Gas Journal, Vol. 103, No. 26, pp.68. [Used in Presentation] • House Energy and Commerce Subcommittee on Oversight and Investigations. Accessed 07/22/06 at http://energycommerce.house.gov/. Subcommittee to Hold Global Warming Hearing, Witnesses to Testify on Climate Change Assessment Report. Dr. Edward Wegman Report.