Does science require experiments with our climate?

Human carbon emissions have been described as a planet-wide experiment with a sample size of one. Are there ways for science to accept uncertainties and just point at what could happen, while not testing the theory? This article is adapted from a talk at the NZ Skeptics Conference, in Wellington, 7 September 2013.

In 2011, France was running the G20 group of countries and their Minister of Agriculture, Bruno Le Maire, started a new way of exchanging information on the current status of crop production globally. His motivation was driven by the extreme drought that had affected Russia in 2010 to the extent that they had decided to ban all food exports, causing international prices to skyrocket and leading to riots in some countries dependent on importing food from Russia.

Le Maire said that the 2010 drought was caused by climate change and so there now needed to be a system to deal with this type of impact. The G20 countries agreed and established an Agricultural Market Information System to get better control of food prices when extreme events occurred.

At the same time several science papers came out debating whether the extreme droughts in Russia, which occurred at the same time as record flooding over large areas of Pakistan and parts of China, could be attributed to human-induced climate change. This also led to debate about whether analyses of such climate events should always be based on testing the null hypothesis of 'no change' in order to avoid what are called the Type I (false positive) errors1-3.

When Hurricane Sandy hit the east coast of the US in 2012, New York's mayor, Michael Bloomberg, believed that this was worse because of climate change caused by fossil fuels and he accelerated plans to radically improve the state's energy efficiency and make their infrastructure more resilient to this type of extreme event. But scientists have not proved that Sandy was caused by increasing greenhouse gases, so that was another decision based on judgment.

For about 20 years the reinsurance industry has reported an increasing trend in weather-related damages caused by climate change. Similarly, the Institutional Investors Group on Climate Change, which represents companies that manage about 20 trillion dollars of long-term investments for pension schemes, is pushing on governments to move faster to reduce the risks that they see climate change causing for their asset base.

These are just a few examples showing that responses to climate change so far tend to be driven by those involved in professional risk management. Risks are directly related to uncertainties and we deal with them all the time, both in insurance schemes and investment strategies.

But some are using the fact that climate scientists admit to uncertainties as reason for not making changes, or at least for delaying responses until we can be sure.

The science basis

In science we work with uncertainties, because they define the challenge for new research that should lead to better understanding backed up by good data. Many see Galileo as the father of modern science because of his emphasis on careful observations to prove a theory. So while Copernicus had postulated that the Earth goes round the Sun, Galileo provided the evidence for this from detailed observations. And was then excommunicated from the Catholic Church by the power base of cardinals who did not want to believe what he said.

Similarly, in the 19th century, Svante Arrhenius had estimated how much the Earth would warm if we doubled CO2 in the atmosphere but it was seen as still being a complex issue involving water vapour feedbacks and clouds. It was also thought that nearly all of the CO2 from our use of fossil fuels was dissolved into the oceans, so that this doubling would take about 3000 years.

But from the 1950s on there has been a steadily accumulating body of evidence based on much more detailed observations. We now know why CO2 is accumulating in the atmosphere much more rapidly than Arrhenius had expected. Paleoclimatic records for the last 800,000 years show temperature changes correlated with greenhouse gas concentrations rather than with changes in the Earth's orbit around the Sun, which are just expected to be a trigger for climate change.

Establishment of the Intergovernmental Panel on Climate Change (IPCC) in 1988 was recognition of the need to link policy decisions to what is still a developing field of science. Recently the fifth in a series of detailed assessment reports has started to come out4 and some changes can be seen from the early assessments, but the basics are still very much the same.

Much of the IPCC focus has been on attribution of causes for climate change by considering all the known factors such as changes in solar radiation, effects of deforestation, the cooling effect of aerosols that scatter sunlight away, as well as increases in greenhouse gases. Comparing all these effects shows that while an increase in energy coming from the Sun would have contributed to warming prior to 1950, the predominant cause since then has been increases in greenhouse gases.

More than 90 percent of the extra heat that is being trapped by greenhouse gases goes into the ocean, some of the rest melts glaciers and ice sheets, and there are increases in latent heat in the form of water vapour in the atmosphere. So only a small part of this heating is going into the surface temperature and while that has slowed down over the last 15 years it can be due to the accelerating loss of ice sheets and to periodic changes in ocean circulation, such as the Pacific Decadal Oscillation. Scientists still want to know more about the details of change in the Earth's heat distribution, but production of more total heat near the surface by an increasing greenhouse effect is well defined.

Projections of future climate raise several different types of uncertainty. Estimates for the change in global average temperature caused by doubling CO2 have been sitting in the range 1.5 - 4.5°C for several decades. Much of this range comes from how cloud cover might change and a recent analysis has shown that the upper half of this range is now more likely5, but this remains an active area for research.

A quite different type of uncertainty applies to the amount of greenhouse gases that will be emitted into the atmosphere during the rest of this century. That is treated by considering a range of scenarios for future social and technological development and the corresponding emissions, or concentrations, of the greenhouse gases. These now produce a range for global average temperatures in 2100 increasing by less than 2°C or by nearly 5°C from the preindustrial value.

Focussing on global average temperatures has been motivated by ensuring that climate models are reliable. But it can be deceptive because much of the land warms by 50 percent more and annual extreme daily temperatures by twice as much. Also the latest IPCC assessment has significantly increased the estimates of future sea level rise and provided likely ranges, but did not set an upper bound for how much may occur by 2100.

There are still major issues for uncertainty analysis in climate change because every few months some new result seems to shift the range for what can happen in the future. For example, as I am writing this, a paper has just come out showing that what was seen in climate models as a stable part of the Greenland ice sheet has actually started to break up and slide into the ocean6. The question is now to what extent can that continue, because the drainage basin it covers has enough ice to raise sea level globally by about one metre.

Can science live with uncertainty?

When development of radiocarbon dating led to the discovery that carbon in the atmosphere was getting older7, this was rapidly followed by others discovering why all the CO2 from fossil fuels was not being taken up in the oceans8, which led them to say: "Thus human beings are now carrying out a large scale geophysical experiment of a kind that could not have happened in the past nor be reproduced in the future".

Science has always been seen as based on experiments that test hypotheses and this can be traced back to the philosophy for developing our understanding that was established by Socrates and Plato over 2000 years ago. Socrates was critical of people who often seemed to think that they knew everything that was relevant, whereas his approach was to focus on addressing our limits in knowledge.

While a significant evolution in the philosophy of science was started by Karl Popper, a recent review has shown that it is still heavily based on carrying out tests of our understanding, whether we are considering Einstein's general relativity theory or Darwin's theory of evolution9.

So can science only make progress by experiments that test theories, even if this would create major problems for humanity? Or are there ways for science to accept uncertainties and just point at what could happen, while not testing the theory?

The last 70 years has seen significant advances in the way that limits to current understanding are considered in science. Quite different types of uncertainty have become recognised as forms of both aleatory uncertainty, due to limits in the data for processes that we are trying to quantify, and epistemic uncertainty, due to an incomplete knowledge of the key processes themselves. The difference was noted by David Hawkins, who was part of the Manhattan Project developing atom bombs, pointing out that it was quite misleading to treat all forms of uncertainty in the same way as one would treat rolling dice10.

A recent major review of formalisms for describing uncertainty has set out four levels of understanding, with use of Bayesian statistics and probability distributions being the most advanced of these11. Ways of dealing objectively with more limited knowledge can be based on defining possibility distributions or using fuzzy sets of values and this could become a way of addressing issues such as the threshold for sustainability of the major ice sheets.

But there are still the policy questions as to whether this could be done in ways that would set a stronger basis for decisions that curtailed fossil fuel emissions globally. In several different contexts, governments have adopted forms of a precautionary principle to deal with these deeper types of epistemic uncertainty, but that is not yet being applied in the case of climate change.

Some have argued that a precautionary principle tends to just create legalistic debates, such as whether precaution should focus on environmental or economic values. However, Cass Sunstein has suggested that for issues such as climate change it would be better to adopt an anti-catastrophe principle12.

How should this issue evolve?

Well, during a meeting of authors for the IPCC's Third Assessment Report, 14 years ago, there was a discussion about the key issues that needed to be addressed. While many were raising questions about cloud properties, or the sustainability of the Amazon forests, a social scientist stood up and said that we were all wrong because the real question is how society actually responds to major issues. There was a bit of a stunned silence in the room and the chairman changed the subject, but many of us have never forgotten that point, and the person who said it became a lead author for the first chapter in our synthesis report.

References

1. Trenberth, K. E. 2011. Wiley Interdisciplinary Reviews: Climate Change 2, 925-930.
2. Curry, J. 2011. Wiley Interdisciplinary Reviews: Climate Change 2, 919-924.
3. Allen, M. 2011. Wiley Interdisciplinary Reviews: Climate Change 2, 931-934.
4. IPCC 2014. in Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds. Stocker, T. F. et al.) Cambridge University Press.
5. Sherwood, S. C., Bony, S. & Dufresne, J.-L. 2014. Nature 505, 37-42.
6. Khan, S. A. et al. 2014. Nature Climate Change online 16 March 2014.
7. Rafter, T. A. 1955. NZ J. Sci. Tech. B37, 20-38.
8. Revelle, R. & Suess, H. E. 1957. Tellus 9, 18-27.
9. Losee, J. 2005. Theories on the scrap heap: scientists and philosophers on the falsification, rejection and replacement of theories. University of Pittsburgh Press.
10. Hawkins, D. 1943. Philosophy of Science 10, 255-261.
11. Helton, J. C., Johnson, J. D., Oberkampf, W. L. & Sallaberry, C. J. 2010. Int. J. General Systems 39, 605-646.
12. Sunstein, C. R. 2006. Pace Environmental Law Review 23, 3-17.