[Home] [Lectures] [Previous] [Next]

Controversy: Global temperature trends over the last 2000 years

The plots of global temperature changes prior to 1860 (i.e., before the period of instrumental records) provided in the previous lecture about the Holocene Epoch should not be interpreted as a literal and exact reconstruction of global average temperature changes. There is a great deal of difficulty and uncertainty in trying to reconstruct global average temperature changes based on proxy data. One thing we can say with some certainty is that global average temperatures have fluctuated all throughout Earth's history, including the last 2000 years. Prior to 1900, we can safely say that the changes were not caused by anthropogenic additions of greenhouse gases, but were natural fluctuations.

The plots previously provided to you are only one estimation for global average temperature change and do not give you any indication about the uncertainty. Other research groups have come up with different reconstructions based on their interpretation of available proxy evidence for global average temperature changes. The figure below gives some indication of the unceratainty as it shows 10 different published global temperature reconstructions for the last 2,000 years (note that most of the reconstructions shown only go back about 1,000 years). Please understand that there is not one single correct reconstruction. Please disregard the black line and asterisk labeled "2004" in the plot below. This was someone's attempt to splice the recent measured changes in global average temperature (after 1850) onto the graph. The problem is that while relative temperature changes can be estimated, there is no way to tell where today's global average temperature fits in an absolute sense. If proxy records were used to construct the past changes in temperature, then those same proxy records should be used up to the present time. It is misleading to try to tack on higher resolution observations of temperature changes at the end of a low resolution proxy estimated record. The point of the figure is to indicate the uncertainty in reconstructing past temperature changes based on proxy evidence. You should not use this plot to conclude that the year 2004 was warmer than any time over the last 2000 years.

[Temperatures of the Last 2,000 years]

One of the most controversial (and often referenced) reconstructions was first released by Michael Mann and co-authors in the 1998 article, Global-scale temperature patterns and climate forcing over the past six centuries, which appeared in the journal Nature. This particular reconstruction has been labeled the "Hockey stick" (see image below) because of the hockey stick shape of the trend lines: gradual cooling from 1000 - 1900 AD, followed by a sharp warming after 1900. This reconstruction makes the medieval warm period and little ice age look insignificant compared with the rapid increase in global average temperature since 1900. (NOTE: based on written historical evidence, we can be fairly certain that these periods were relatively warm and cold throught much of Europe. Mann's study indicates that global average temperature did not behave like that of Europe.) Although the authors of this study tried to properly merge proxy-record information (tree rings, corals, ice cores, boreholes, etc) prior to 1900 with the measured temperature record after 1900, this is quite challenging and uncertain. It is very difficult to try to compare temperatures from proxy records with observed temperature from today in a comparative sense. Again, one cannot scientifically conclude that the global average temperature in the year 1998 (based on thermometer measurements) was 0.7°C warmer than the global average temperature in the year 1000 based on proxy inferences.

Hockey stick 1998
"Hockey stick" reconstruction of global average temperature changes over 1000 years (ending in 1998). Reconstruction performed using proxy data before 1900 and instrumental data after 1900. Yellow and red shading show author's estimation of uncertainty in the reconstruction.

It is difficult to decide which, if any, of the reconstructions of global mean temperature is the most accurate reflection of temperature fluctuations during the last 1000 years. However, it is likely that such reconstructions, accurate or not, will play a significant role in the ongoing discussions of global climate change and global warming.

If you "believe" the temperature reconstruction presented on the previous reading page, then global average temperature seemed to be relatively warm during the medieval warm period, generally cooling into the little ice age period, then warming after the little ice age. This may lead some to conclude that the present warm conditions are not extraordinary when viewed in the context of the last 1000 years. In other words there are recent natural climate fluctations in global average temperature very close to what we are observing today, and the human influence on global aveage temperature changes may be small.

If you "believe" the hockey stick reconstruction, then the current global average temperature is much warmer than at any time over the last 1000 years and much of the warming began after 1900, which coincides with the increase in human emissions of greenhouse gases. This can lead to the conclusion that the conditions today are extraordinarily warm compared with the recent past and more likely can be attributed to anthropogenic greenhouse gas increases.

The debate about the validity of the hockey stick reconstruction still continues more than 15 years after its initial publication. The graph has been used almost as an icon of global warming due to human emissions of greenhouse gases. It was published in the 2001 IPCC report as "evidence" that the increase in global average temperature since 1900 was unprecendented over the last 1000 years . However, I believe that we can now say that the methodology used to create the original hockey stick graph was flawed. The biggest problem is due to mathematical and statistical errors in Mann's method that were discoverd and proven to exist. This is discussed in a non-technal manner in this article, A Global Warming Bombshell: A prime piece of evidence linking human activity to climate change turns out to be an artifact of poor mathematics written by Richard Muller, who says he is concerned about global warming and thinks that human-added carbon dioxide has largely contributed to the recently measured warming (after 1900). In the article, Muller makes the point that a blind belief about the implications of the hockey stick could actually weaken the case of human-caused gloabal warming if the Earth were to begin a period of global cooling. If one were to claim, based on the hockey stick graph, that greenhouse gas increases are the main driver of recent climate changes, it would be would be difficult to explain a period of global cooling (or even an extended period with no global warming, such as the recent period from about 2000 to 2013) during a time when greenhouse gas concentrations in the atmosphere are increasing. A brief history of the hockey stick controversy is provided in Unscientific hockey sticks and hidden data and The rise and fall of the Hockey Stick. It is interesting how quickly the hockey stick graph spread through the IPCC, world governments, and the media, even though some had reasonable doubts about the data and methods. The journal Nature, which published the paper, did not do a good peer review of the manuscript and along with Michael Mann did not provide other scientists access to the data or methods. This is generally required for all scientific papers ... once a paper is published, other scientists are able to verify or disprove the results. In the instance of the hockey stick, it tooks years via legal action to get access to information that should have been made readily available. Once access to the data and methods were granted, it took only a matter of weeks for the hockey stick graph to be scientifically and mathematically invalidated. The hockey stick shows how personal beliefs and politics can blind people with respect to this issue of possible global warming due to human emissions of greenhouse gases. Since many wanted to believe the implications of the hockey stick, they failed to question the validity of the study. This group included scientists who should know better. The fact that the hockey stick graph has been invalidated, though, does not mean significant global warming and climate changes have not happened or will not happen due to human emissions of greenhouse gases. However, the best scientific assessment of available proxy data does not support the temperature changes indicated by the hockey stick no matter how much some want it to be true.

Another criticism of Mann's work is that the results are highly influenced by the use of tree ring proxy data. It can be difficult and misleading to interpret past temperatures based on tree ring data. For one, the yearly growth of a tree depends on more than just temperature changes, e.g., rainfall, availablity of nutrients, etc., making it easy to mistake a change in rainfall for a change in temperature for example. Another reason is that for each tree there is likely some optimum growing temperature. If it is warmer or colder than the optimum less growth will occur, which can make it difficult to interpret the growth ring in terms of temperature changes. Note that the hockey stick graph was not included in the 2007 IPCC report or the latest 2013 report, which further indicates that even those concerned about climate change have realized the graph cannot be scientifically justified. Just for comparison, another group of reasearchers in the late 1970s had a different and interesting interpretation of recent climate change cycles based on tree ring data, Prediction: Warming trend until year 2000, then very cold from the St. Petersburg Times. It will be interesting to see if these predictions bear out. So far it has not gotten cold in the first part of the 21st century, but this is not enough years to rule out the prediction. It is also possible that cooling would have been the natural cycle of climate change, but that is being altered by human activity mainly by the emission of greenhouse gases into the atmosphere. For another point of comparison, the temperature reconstruction shown below was constructed based on many sources of proxy data, but not including tree rings. The recent, measured changes in global average surface temperature have not been appended to the end of the proxy record.

loehle 2008
A Temperature reconstruction with 95% confidence intervals shown, from Loehle and McCullough (published in Energy and Environment, 2008): Correction to: A 2000 year global temperature record based on non-tree ring proxies. As this reconstruction is based soley on proxy data, the recent, measured global average temperature was not added to the end of the record as was done by Mann. The last data point represents a 29 year average centered around 1935.

We certainly cannot resolve this issue here. There is lively debate on these reconstructions even among the "experts". In any case, all reconstructions of global average temperature based on proxy data are so uncertain that one cannot be used to disprove another. As with many other issues in the global warming debate, we are forced to decide how to act based on uncertain information.

[Home] [Lectures] [Previous] [Next]