Skip to content

The Geological Perspective Of Global Warming: The Debate Continues

Dr Colin P. Summerhayes, Professor Robert Carter and Professor Vincent Courtillot

Editor’s Note: The following letter by Dr Colin Summerhayes and the response by Professors Bob Carter and Vincent Courtillot are a continuation of their debate on The Geological Perspective of Global Warming which the GWPF published on 14 February. Dr Summerhayes’ letters have also been published by the Geological Society. We welcome this scientific exchange and hope that readers will find it both enlightening and encouraging. 


Dr Colin P.  Summerhayes, Scott Polar Research Institute, Cambridge

Dear Dr Peiser,

Thank you for the opportunity to respond to the critique by Drs Carter and Courtillot of my note of 14/2/13 on “The Geological Perspective of Global Warming”. I initially wrote to you to draw attention to Geological Society of London’s statement on this topic, because the geological perspective is usually overlooked in discussions about climate change, and it should not be. But, because Drs Carter and Courtillot moved the debate out of just the geological arena, I am responding in my own capacity, not as a representative of the GSL.

Drs Carter and Courtillot took exception to my use of the phrase “The cooling [of the past 50 million years] was directly associated with a decline in the amount of CO2 in the atmosphere”, saying that correlation was not causation. True. What I should have said was “The cooling of the past 50 million years was driven by a decline in CO2 in the atmosphere.” Prior to the Ice Age of the last 2.6 million years the amount of CO2 in the atmosphere resulted from the interplay between the emission of CO2 by volcanoes and its absorption by the weathering of rocks, especially in mountainous areas, as well as by sequestration in sediments. Methods to determine the likely concentration of CO2 in the atmosphere in the geological past have improved in recent years. They include the numbers of pores (stomata) on leaves, the abundance of the mineral nahcolite (stable above concentrations of 1000 ppm CO2), and the carbon isotopic composition of alkenones from marine plankton. Methods for determining global temperature through time have also improved. We now know that the Eocene was a time of greater volcanic output of CO2, and that the rise of major mountain chains after that time pulled CO2 out of the atmosphere. Geochemical models of the carbon cycle simulate the decline in CO2 after the middle Eocene. Convergence between the CO2 data and the output from those models provide confidence that we understand the process. There is no geologically plausible alternative. We are not talking about a loose association where there is uncertainty about cause as Drs Carter and Courtillot imply. Indeed, even Drs Carter and Courtillot accept that CO2 is a greenhouse gas, and that accumulation of greenhouse gas in the atmosphere warms it. Likewise, its loss will cool the atmosphere.

Besides that, the GSL statement regards the massive injection of carbon into the atmosphere that took place over a short period 55 million years ago, raising temperature, raising sea level, and causing ocean acidity, as a case history that we can draw upon to tell us what may happen in the future if we continue to pump CO2 into the atmosphere at rapid rates. It was not alone; there was another such event in the Toarcian, for example, some 180 million years ago.

Moving on to the Ice Age of the past 2.6 million years, by this time the levels of CO2 in the atmosphere were so low that other drivers of the climate system had more effect. The primary drivers of change in the Ice Age were the tiny changes in solar radiation received at the Earth’s surface due to regular and predictable changes in the Earth-Sun distance and in the tilt of the Earth’s axis. These made the climate of the Ice Age fluctuate between cold periods – glacials- and warm ones – interglacials – in one of which we now live. The role of CO2 in this system was to provide positive feedback to the rises in temperature that took us from glacials into interglacials. In this narrow context, Drs Carter and Courtillot are correct – CO2 increased during the interglacials mainly by outgassing from the ocean. But that was not the main source of CO2 during the Cenozoic era.

We should all reflect on the fact that the past 4 interglacials were warmer than today, and sea levels then were higher than today. Drs Carter and Courtillot wonder if we know enough about the behaviour of the climate system during the Ice Age to be confident in our analysis. Yes we do. The uncertainties are minor. Given what we know from the link between CO2 and temperature with time from the geological record, it would be foolish to imagine that if we warm our planet to the same extent as it warmed in previous interglacials, we will not also see similar rises in sea level to those that occurred in them.  In any case, waiting until all small uncertainties are resolved is not a reasonable option.

Our geological knowledge of past climate change is independent of the numerical models used by climatologists to tell us what may happen if we add more CO2 to the atmosphere. The geological data, recalibrated in 2012 by the PALAEOSENS team led by Eelco Rohling (Nature 491, 683-691, 29 November 2012), tell us that the sensitivity of the climate in the past to a doubling of CO2 was 2.2-4.8°C, about the same as calculated for the modern climate by the climate modellers who feed data to the IPCC. This match is unlikely to be a coincidence. Indeed, it suggests that the climate modellers may well be on the right track, and that Dr Lindzen and others may be wrong in suggesting that the sensitivity is 1°C or less. However, Drs Carter and Courtillot are right to point out that some recent studies suggest that the climate sensitivity to a doubling of CO2 may be closer to the low than to the high end of the IPCC range. While that may appear comforting, it only postpones the inevitable.

Drs Carter and Courtillot took me to task over the relationship between CO2, temperature and sea level. However, their sea level calculations are simplistic. The 20 cm rise that we have seen since 1900 is not an equilibrium response – it is instead a transient response to a rise in temperature of 0.8°C occasioned by a rise in CO2 of 40%, or 100 ppm. The sea level will go on rising even if we stop putting CO2 into the atmosphere, as the ocean equilibrates with the atmosphere over decades to centuries, and as ice sheets slowly decay. Models suggest that the equilibrium position may be 0.5m/1°C due to thermal expansion alone. Currently thermal expansion accounts for around 1/3 of sea level rise, and glaciers and ice sheets for another 1/3 each. It is not difficult to see how a further rise in CO2 could by 2100 lead to a rise in sea level of perhaps as much as 1.4 m as estimated by Stefan Rahmstorf and colleagues.

Drs Carter and Courtillot took exception to my statement that the Earth should have been cooling over the past 10,000 years. Indeed it should because that’s what we calculate from known phenomena like changes in the Earth-Sun distance and tilt of the Earth’s axis. Other shorter-term changes will of course be superimposed upon that trend. Drs Carter and Courtillot emphasize them by providing a graph of Greenland temperatures, but as they point out those were regional. Even so, that graph too shows underlying cooling for the past 5000 years. The small divergences from the mean on the Greenland graph were caused by short term climate changes like those of the Medieval Warm Period and the cooling of the Little Ice Age, which coincided with the Maunder Minimum in sunspot activity between around 1645 and 1715. Both events seem to have been most intense in the North Atlantic and European region, not globally. There is no evidence that the Medieval Warm Period was warmer than today globally. Nor is there any evidence to suggest that we are now living through a similar event.

Drs Carter and Courtillot would like us to believe that the current rapid global warming event is purely natural. This seems odd given that they also accept that carbon dioxide is a greenhouse gas that warms the lower atmosphere and that a portion of human emissions of CO2 is now accumulating in the atmosphere. Moreover one of their key references (Ring et al 2012) makes it clear that human activity has caused the warming since 1900. All our attempts to relate the post 1970 warming to natural sources of heat have failed. Our burning of fossil fuels is detectable in the atmosphere from a reduction in oxygen as well as from an increase in CO2 and from the carbon isotopic signature typical of the burning source materials. Since the 1970s, warming has been taking place while the suns output has not been increasing. Nobody has yet come up with a better explanation of this recent warming than that it is caused by the known increases in CO2 and related greenhouse gases, much as we might expect from what we know of the effect of CO2 in the climates of the past, and from the basic physics of radiation.

The warming of the recent past up to and including 2012 is shown in the attached graph by Hansen, J., Sato M., and Ruedy, R., 2013 “Global Temperature Update Through 2012” (available from The reader will notice that the rise has not proceeded smoothly, but in a series of steps like the one that started in 2002. It was inaccurate of Drs Carter and Courtillot to suggest that this flat spot started in 1998, which was a prominent El Niño year. During El Niño years, shown in the lower graph, the emission of heat from the Pacific Ocean warms the world. Temperatures drop during the subsequent cool La Niña events. They also drop during volcanic eruptions large enough to eject fine particulates and acid gases into the stratosphere. Thus the 1998 El Niño effect visible in the graph was not the start of a flat step; it was followed by a cooling due to a large La Niña. Other large-scale oscillations within the climate system will also have had an effect, one such being the Interdecadal Pacific Oscillation, which shifted to a positive phase in the 1970s and led to a warmer Pacific. In the 2000s that Oscillation reversed, cooling the Pacific and likely thereby contributing to masking the rise in global temperature (EOS, v.94, No.6, 5 February 2013).


Fig. 1. From Hansen, Sato and Ruedy, 2013. Global surface temperature anomalies relative to 1951-1980. The Nino index is based on the detrended temperature in the Nino 3.4 area in the eastern tropical Pacific. Green triangles mark volcanic eruptions that produced an extensive stratospheric aerosol layer. Blue vertical bars are estimates of the 95% confidence interval for comparisons of nearby years.

In conclusion, I consider that the data from the geological record are consistent with the data from the modern environment, and with projections made on the basis of those modern data as to how our climate may change in the future. Anyone who accepts that CO2 is a greenhouse gas, as Drs Carter and Courtillot do, must expect that a large increase in its concentration is bound to have a warming effect, and observations show a warming that is consistent with this effect. Remarkably few climate scientists dispute that fact. The world is indeed exposed to real short-term climate related events, as Drs Carter and Courtillot point out, but what we face in human-made global warming is an insidious underlying upward trend that will exacerbate those short term events unless action is taken to deal now with the causes of that trend.

C. P. Summerhayes, Scott Polar Research Institute, Cambridge

Revised 3.3.2013



Professor Robert Carter and Professor Vincent Courtillot respond:

We thank Dr Summerhayes for his further comments, and agree with him that the geological perspective on climate change is an important one that is often underappreciated.

However, the science of geology is a holistic one, and includes an understanding of modern earth processes such as those involved in meteorology and climatology. We therefore do not see that we have moved the debate out of the geological arena. Rather, our approach to the global warming issue is to undertake an assessment of all the major and relevant scientific considerations within proper geological context.

Turning to the main points made by Dr Summerhayes, we are aware of the Eocene as a time of great output of CO2. One of us (VC) has worked for the past decade on geological evidence of huge and rapid volcanic pulses in large igneous provinces (the Deccan traps at the KT boundary and the Karoo traps at the Toarcian extinction), and has suggested that the Late Paleocene Thermal Maximum could have been caused by one phase of the eruption of the Greenland traps. Volcanic flows exceeding 10,000 km3 in volume and erupted in less than a century have been documented, and their climatic consequences, due to release of SO2 and CO2, have been studied and modelled. We are therefore alert to the fact that the release of CO2 by volcanism, and its subsequent withdrawal by alteration and sequestration in sediments, have played roles in influencing ancient climate.

Nonetheless, and as we stated in our earlier letter, the key questions that remain concern:

(i) the accuracy (or uncertainties) of the quantitative estimates of climate sensitivity to CO2 increase; and

(ii) the precise, quantitative nature of important processes of cloud microphysics and solar forcings.

For example, though total solar irradiance (TSI) varies by only 0.1 percent over the decadal time scale of instrumental observations, ultra-violet (UV) and extreme ultra-violet (EUV) radiation varies by tens of percent. Depending on whether the forcing mechanism depends only on TSI or also on UV and EUV (through atmospheric electricity and induced cloud cover variations), the climate response will be totally different. And the degree to which such variations occur also on longer, geological timescales remains unknown.

For these reasons, applying what is seen on geological time scales to the present, or vice versa, remains fraught with uncertainty. It is these uncertainties that are often forgotten, that remain as topics for much further research, and that we wish to emphasize.

As Dr Summerhayes recalls, some past interglacials were warmer and sea levels higher than today due to the Sun and not CO2. Nonetheless, the unattentive reader here may be led to forget that the warming that caused these earlier sea level rises was due to the Sun, through Milankovic orbital changes.

Further on sea level rise: the steady 20th century increase in sea level displayed by the tide gauge record starts around 1900, when the flat slope in previous decades changes to a linear, regular increase of about 20 cm/century. This slope was established before the major post-WWII CO2 increases, and continues unchanged after them. How then can it be considered as a transient response to CO2 rise?

The last sentence of Dr Summerhayes’ comments on sea level concludes that “In any case, waiting until all small uncertainties are resolved is not a reasonable option”. This is an affirmation of the “precautionary principle”, which is a sociological and not a scientific concept. Apart from the fact that the remaining uncertainties in climate science are not small but large and many, the doctrinaire application of the precautionary principle often leads to expensive or unreasonable consequences, to the degree that a 2006 report by the House of Commons Select Committee on Science and Technology recommended against the use of the concept in public policy formulation.[1]

On climate sensitivity, Dr Summerhayes acknowledges that some recent studies suggest that climate sensitivity is significantly smaller than previously emphasized in many reports, and by the IPCC. Dr Lindzen indeed “may be wrong” in suggesting that sensitivity is less than 1°C; alternatively, he may be right. The key point is that the debate is clearly not over and therefore the situation cannot be considered as being scientifically settled to the degree that public policy formulation requires. Most climate modellers, who have achieved significant advances over the past decades, acknowledge that a number of ill-understood physical processes that bear on climate sensitivity remain to be modelled accurately, such as details of water phase changes and cloud microphysics.

Dr Summerhayes asserts that we took exception to his statement that the Earth should have been cooling over the past 10,000 years. In fact, we queried the comment that “the Earth should be cooling slightly. Evidently it is not”, and gave reasons why the statement is, at best, ambiguous. We also made the specific point that the Earth had indeed cooled through the Holocene, consistent with a Milankovic control.

On the Medieval Warm Period (MWP) being global and warmer than today, there are many papers arguing for that and the debate is certainly not closed. Whether the MWP was more intense in some locations than others, as claimed by Dr Summerhayes, is largely irrelevant for the same reason that it is irrelevant that the climatic signature of the last glaciation was much reduced in lower, and enhanced in higher, latitudes. “Global climate” is an abstraction within which great local and regional variation can and does occur. In any case, an excellent and comprehensive database of papers on the MWP from locations around the world has been compiled by Dr Craig Idso.[2]

On the temperature rise of just under a degree Celsius that has occurred over the past 150 years, the idea that it reflects a millennial cycle of solar activity remains as valid as (and we believe more likely than) the idea that CO2 rise is responsible. Again, the issue is one of climate sensitivity and the disentanglement of the solar and CO2 responses, all of which are topics of ongoing research.

On the global temperature plateau since 1997, this has now been acknowledged by even the head of the IPCC, Dr Rajendra Pachauri, who has previously strongly denied its reality. However, the temperature pause is a short-term feature, and its significance within any longer-term trends or multi-decadal rhythms that may be present remains a challenge, not a settled point.

Dr Summerhayes asserts that we wish people to believe “that the current rapid global warming event is purely natural”. In actuality, we wrote regarding the mild, not rapid, 20th century warming that “our null hypothesis is that the global climate changes that have occurred over the last 150 years (and continue to occur today) are mainly natural in origin”, and pointed to the abundant evidence that favours this hypothesis. Given that the warming response to increasing carbon dioxide decreases at a logarithmic rate, the facts that CO2 is a greenhouse gas and is accumulating in the atmosphere do not necessarily imply that dangerous or even measurable warming will occur – as may already be indicated by the lack of warming over the last 17 years in face of an 8% increase in carbon dioxide.

In conclusion, climate change is as much a geological matter as it is a meteorological one. For that reason we agree with Dr Summerhayes that studying the geological past, trying to make observations at always increasing time resolution and extracting constraints on past climate behaviour can greatly enlighten our understanding of recent and current climate change.

Yet all of that accepted, the most important point to be reiterated here is that many disputed scientific matters need to be better understood before climate science can be viewed as mature enough to be used to inform public policy.

Vincent Courtillot
Bob Carter

March 4, 2013

[1] “We can confirm our initial view that the term ‘precautionary principle’ should not be used, and recommend that it cease to be included in policy guidance…In our view, the terms ‘precautionary principle’ and ‘precautionary approach’ in isolation from any such clarification have been the subject of such confusion and different interpretations as to be devalued and of little practical help, particularly in public debate.” House of Commons Science and Technology Committee, 2006 Scientific Advice, Risk and Evidence Based Policy Making. Seventh Report of Session 2005-06.