Scientific uncertainties

Scientific uncertainties

Uncertainty, scepticism, debate and independent verification are fundamental components of the scientific process. In science, it is important to be sceptical of a controversial or unusual theory or claim, especially when the ramifications of that theory, if true, are potentially grave. For this reason, scientists are sceptical of many of the claims put forward about climate change. Evidence, and continual re-assessment in the light of new evidence, is always required. Scientists rely on maintaining a healthy level of scepticism to allow them to objectively assess theories, and to minimise bias in the collection and interpretation of the evidence. It is true that publication in a peer-reviewed scientific journal does not guarantee that a theory or interpretation is correct and complete. However, subsequent independent studies that may be based on the same evidence but use an alternative perspective or, as is often the case, incorporate new evidence, provide a process of assessment and verification of prior work. Often, one team's work will be replicated by another group, just to be sure. In this way, science is inherently self-correcting. Incorrect or incomplete theories tend to be weeded out or clarified through this process of incremental knowledge accumulation, with the majority of self-correction occurring within only a few years.

The IPCC Fourth Assessment Report (AR4) suggests that these basic attributes of the scientific process can be used to assess contrasting claims about climate change. If any of the following questions cannot be answered in the positive, then less credence should be given to the claim until it can be further tested and independently verified:

  • Can the statement under consideration, in principle, be proven false?
  • Has it been rigorously tested?
  • Did it appear in the peer-reviewed literature?
  • Did it build on the existing research record where appropriate?

The majority of well-informed scientists and non-scientists familiar with the scientific process, or those with relevant expertise, accept that climate change is occurring and is mostly a result of human activities. In many cases, the arguments put forward by people who are disinclined to agree with the mainstream science rely on opinions or work that has not undergone rigorous scientific review. Such work is also often authored by non-scientists or scientists with no relevant expertise. In such cases the above questions cannot all be answered in the affirmative, and the argument cannot be given the same credence as arguments that draw on work that has been subject to scientific review. These claims may be classed as 'non-scientific', and are addressed in our pages dealing with common misconceptions and unscientific arguments.

Since scepticism is an integral part of science, we do not use the term 'sceptic' here to refer to those who challenge the reality of climate change or human influence on the climate. Instead, those individuals are referred to as 'doubters'. The information presented on this page, however, addresses both genuine uncertainties in climate change science as well as interpretations of the science that oppose the mainstream view.

As a general strategy, climate change doubters often tend to focus on individual scientific studies or publications that present evidence to challenge one or more aspects of the more generally accepted view. This is a selective way of viewing the information (often referred to as 'cherry picking'). This evidence often contradicts the doubters' own arguments in other ways, or sometimes the studies referred to are mutually contradictory. One should bear in mind that apparently contradictory evidence within scientific research is not in itself an unusual situation—ultimately, the scientific process relies on the accumulation of sufficient such evidence to allow the most likely explanation or 'truth' to emerge. For this to happen, one must view all the evidence and not just selected pieces of it.

Such a process, for example, was undertaken by the more than 500 lead authors and 2000 expert reviewers of the IPCC AR4 and the governments of more than 100 participating nations who agreed to its content word by word. The report examined the full body of work on past and present climate to determine whether climate change is occurring and if so, what the causes are (it found that global warming is certainly occurring and most of it is very likely due to human activities). Climate change doubters' apparent preference to focus on particular aspects of individual studies bypasses this more powerful synthesis path to understanding.

It should be acknowledged that the study of climate change is a particularly difficult discipline within science. This is because, in many relevant areas, there is no 'control' case available for comparison. There is no duplicate earth to observe in which humans have had no influence. Climate change is effectively the biggest manipulative experiment ever undertaken, but the experiment is unplanned and lacks the usual replicates and controls that scientists like to use for statistical assessment of the effect of the manipulative treatments. As well, there are elements of chaos in the climate system and this chaos, by definition, defies analysis and prediction. However, this does not invalidate the process of science because we can use other means to validate our observations and conclusions. For example, data from other planets; data from different time periods; the use of validated models; and experimentation on isolated components of the system, whilst being mindful of the fact that a whole system may behave differently.


Below, we address the following specific issues or arguments that are used to challenge the mainstream interpretation of the evidence of climate change:

  • There is so much uncertainty and disagreement among climate models that they cannot be believed.
  • The observed temperature increase is due to spurious data or faulty reconstructions of past temperature.
  • Current global temperature is no warmer than the medieval warm period.
  • Recent global warming is due to natural causes.
  • Glacial transitions show that warming causes CO2 increase, not vice versa.
  • Global warming has ceased in the last decade.
  • The climate system is self-stabilising, and feedback mechanisms will counteract the effects of greenhouse gases.

Why is there so much uncertainty and disagreement among climate models?

A variety of factors contribute to uncertainties and bias in climate change modelling. These may arise from, for example, incomplete knowledge of the processes being simulated; insufficient representation of these processes, e.g. over-simplification to accommodate limitations in computing resources; uncertainty in assumptions about future socioeconomic scenarios and emission trajectories; inherent limitations to the representation of many of these processes due to their stochastic nature, e.g. the chaotic nature of the climate system and indeterminacy of human behaviour; errors in observations due to instrumental limitations; or errors in observations due to insufficient sampling.

The climate system is complex. Mass and energy exchange and transport processes occur over a wide range of space and time scales and through a variety of mechanisms that are constantly interacting and responding to changing conditions. Climate models have undergone a dramatic evolution in recent years, with ever-increasing sophistication (inclusion of more processes and feedback mechanisms), higher resolution in both space and time, and increasing length of simulations. This has been facilitated by an increase in supercomputing processing speeds of a factor of about a million over the past three decades.

Many of the processes driving the climate system are not easily characterised or discerned through observations, and they must be formulated in models through use of laws of physics, fluid dynamics and thermodynamics. For example, how fast does air within the atmosphere circulate between the tropics and the poles? Or between the lower and upper atmosphere? How does the amount of water vapour in the air change with changes in ocean temperature, and how does this affect cloud formation and rainfall distribution? To what extent do aerosols or small particles in the atmosphere encourage the condensation of water vapour and formation of clouds, and the size of water droplets within clouds?

There do exist, however, measurable quantities within the climate system that change in response to processes such as these, which can be used to indirectly observe their effect. Such quantities include, for example, the distribution in time and space of concentrations of carbon dioxide and other gases, wind speeds, temperature, rainfall, and cloud cover. These data may be used at the outset for model calibration, to adjust parameters within the model that define the relative strength or influence of various mechanisms contributing to the state of the climate system at a given point in time and space. Observations may also be used for model validation—the accuracy of process formulation can be assessed by comparing model simulations of measurable quantities with observations.

Climate models developed by different scientists and different institutes do not all produce the same model results. Due to the complex nature of the climate system, the limited coverage of observational data available, and the inherent error associated with such data, it is impossible to fine-tune the process formulations within any model such that all observations are perfectly simulated at all times. Therefore climate modellers must make informed judgements about matters such as the quality of different kinds of data and the relative importance and strength of various processes, and sacrifice model-observation agreement in some areas to better simulate others. Furthermore, when models are run outside the range of the observations (e.g. for future projections), the assumptions underlying the evolution of interacting mechanisms are no longer verifiable and thus become rather speculative, with the level of uncertainty increasing the further the simulations stray from the range of experience.

Other than testing models through their ability to reproduce historical and current observations, another valuable means of quantifying model uncertainty is through model intercomparisons. A range of alternative interpretations of the science have been realised in the development of a range of well-established and rigorously reviewed climate models. These models may also be based on different mathematical frameworks, which affect the treatment of input data and estimation of model parameters. Comparison of simulations from these models allows an assessment of the range of plausible responses of the climate system to various forcings (for example human influences such as increasing greenhouse gas concentrations, or natural factors such as changes in solar output or volcanic eruptions). Several such model intercomparisons have been undertaken to establish the level of confidence that we can attribute to our simulation of various components of the climate system and its response to these externally imposed changes. Furthermore, each individual model produces a range of results depending on the assumptions put in to the model, for example corresponding to different emissions scenarios.

The climate change projections presented by the IPCC AR4 are not single specific projections of temperature increases or sea level rise for a given time in the future. Rather, they represent an estimate of the range of possible temperature increases or sea level rise that result from different greenhouse gas emissions scenarios corresponding, for example, to different assumptions about the rate of substitution of fossil fuels with renewable energy sources. Furthermore, the report gives a confidence in the projections by providing a 'likely range' for each scenario, including a 'best estimate' as well as a 'low' and 'high' estimate. These confidence ranges embody the range and variability of estimates from the suite of climate models synthesised in the report, which incorporate various different assumptions about the relevant processes and feedbacks. In this manner, the IPCC Report presents the best possible estimate of climate change projections based on the current state of knowledge, and assigns a level of confidence in those projections. The models that are incorporated into these projections and confidence estimates have all undergone testing and review through the scientific process.

It is possible, even likely, that future observations and advances in understanding will reveal processes or assumptions that are poorly represented or faulty in some of the models currently in use, or even throughout the whole suite of models. This may significantly change model simulations of the evolution of the climate system under external forcing and hence projections of climate change impacts. However, this should not prevent us from accepting the current best estimate of the likely range of projections as the basis on which we should make our policy decisions at this time. Our response to climate change will necessarily need to be able to adapt to ongoing analysis and assessment not just from evolving scientific knowledge, but also changes and feedbacks in our social and economic systems.

Are the data and methods that provide evidence of global warming valid?

A small number of sites demonstrate a clear urban heat island (UHI) effect. Dark tarmac absorbs sunlight and radiates heat in the immediate vicinity; windows and bright concrete surfaces of buildings reflect sunlight towards the ground; and industrial and urban effluent from external air conditioning units, generators and vehicles send hot air flow into the surrounds. Influences such as these can cause the temperature within parts or all of a city or urban area to be significantly higher than would otherwise occur in that geographic location. Studies that look at hemispheric and global temperature trends, however, conclude that any UHI trend is less than one tenth smaller than decadal and longer term trends that are evident in the data. In addition, other studies have corrected for this effect and demonstrated that the warming trend is still clearly evident.

There has been much debate over the 'hockey stick' graph of increasing northern hemisphere temperature anomalies (departures from the long-term mean) that was published in the IPCC Third Assessment Report in 2001, which showed a significant sudden rise in temperature from the 1950s. The dispute centred on the methodology and data used to construct the graph, with other scientists being unable to reproduce the results. It was demonstrated that this was due to differences in the way that the methodology was implemented, but there remained criticisms of the methodology and selection of data that led to the hockey stick graph. However, regardless of the controversy over that graph, several independent studies have since confirmed the incontrovertible existence of a substantial rise in global surface temperature over the past few decades compared to the historical average. The authors of the original hockey stick graph recently published a new analysis utilising a greatly expanded data set, which demonstrated that recent warming is unusual in the context of the record over the past 1300 years, with no previous periods in this timeframe being as warm as the current conditions. If tree-ring data are included as a proxy measure of temperature (the accuracy of which had been questioned), the conclusion can be extended to the past 1700 years. These results are consistent with other studies presented in the IPCC Fourth Assessment Report.

Is the current warming unusual compared to historical warm periods?

It has been suggested that global scale historical climate irregularities have occurred in the past, and the current situation is not unusual. Various authors throughout the 20th century compiled anecdotal and environmental evidence pointing to a 'MEDIEVAL WARM PERIOD' at around 900–1200 A.D. Most of this evidence, which includes the northern extent of cultivable land in Iceland and Greenland and ice-free ocean routes, comes from Western Europe. Analysis of tree stumps in lakes also reveals that parts of the western United States experienced severe and prolonged drought during this period. More recent studies demonstrate that the medieval climate probably varied heterogeneously in different regions. The evidence suggests that on the whole the Northern Hemisphere probably experienced temperatures warmer than average in a 2000-year context, and certainly warmer than the subsequent 'LITTLE ICE AGE' from ~1500 to ~1850 A.D. However, while some regions may have experienced warmer conditions than those that have prevailed in recent decades, the evidence does not point to a warming as pronounced or as spatially extensive as the global warming we are currently experiencing.

Is the current global warming caused by natural factors?

Natural factors such as changes in solar output, volcanic activity, and variations in the earth's orbit are thought to significantly influence the earth's climate. While the extent of these influences is not known with great certainty, scientists are confident that these natural factors cannot account for the observed warming in recent decades. The figure below demonstrates that when climate models incorporate only these natural factors, they are able to reproduce the general pattern of global temperature up until about 1960, but not the pronounced rise in temperature observed since then (see bottom panel). Note that the high variability in observed temperature from year to year is not well simulated, because this arises from stochastic processes in the climate system that cannot be effectively modelled (see model uncertainties above). The models are more concerned with accurate simulation of the trend in temperature over several years. As shown in the figure (see top panel), in order to reproduce the warming of the past several decades, anthropogenic forcing (the change in radiative balance due to greenhouse gas emissions from human activities) must be included. This result presents some of the most convincing evidence that human activities are causing global warming—there are no known natural processes that can account for the observed warming, and the increase in greenhouse gas concentrations due to human activities is necessary and sufficient to explain the observations.

Comparison between global mean surface temperature anomalies (°C) from observations (black) and AOGCM simulations forced with (a) both anthropogenic and natural forcings and (b) natural forcings only.

Comparison between global mean surface temperature anomalies (°C) from observations (black) and AOGCM simulations forced with (a) both anthropogenic and natural forcings and (b) natural forcings only. All data are shown as global mean temperature anomalies relative to the period 1901 to 1950, as observed (black, Hadley Centre/Climatic Research Unit gridded surface temperature data set (HadCRUT3); Brohan et al., 2006) and, in (a) as obtained from 58 simulations produced by 14 models with both anthropogenic and natural forcings.

Source: Intergovernmental Panel on Climate Change, Contribution of Working Group I to the Fourth Assessment Report, Climate change 2007—the physical science basis, Chapter 9 Understanding and attributing climate change, Figure 9.5, p. 684.

Does CO2 cause warming, or vice versa?

Bubbles of air trapped in deep layers of ice in the thick ice sheets on Greenland and Antarctica reveal the composition of the atmosphere over the last several hundred thousand years, including the concentration of carbon dioxide in the atmosphere. Furthermore, proxy measures of temperature can be obtained at the same time, using the isotopic composition of the ice itself (the amount of deuterium, or heavy hydrogen, in the ice, which varies with temperature). The data show that carbon dioxide concentrations and temperature both increase fairly rapidly and dramatically in interglacial (warm) periods, and both decline more slowly in glacial (cold) periods. The underling cause of these glacial-interglacial transitions is thought to be changes in the earth's orbit, which change the amount of solar energy arriving at the surface and in different regions of the planet.

The transitions between glacial and interglacial periods are enhanced by feedbacks in the earth-climate system. The extent and timing of various mechanisms contributing to these transitions is uncertain, but it is thought that the changes in orbit at glacial 'termination' trigger an initial warming of the oceans. This in turn releases carbon dioxide (because CO2 is less soluble in warm water) and the extra carbon dioxide in the atmosphere causes further warming. Decrease in albedo (reflectivity) resulting from melting of ice also acts to increase warming. These mechanisms amount to a positive feedback that magnifies the initial tendency and leads to the warm interglacial periods. Concurrent changes in ocean biogeochemistry and in wetlands causing changes in the amount of methane released to the atmosphere may also play a role in the glacial/interglacial transitions.

High-resolution ice core data during deglaciation indicate that Antarctic temperature starts to rise several hundred years before CO2. However, the evidence suggests that a range of interacting mechanisms are at play in glacial-interglacial transition periods, and the enhancement of the greenhouse effect as CO2 concentrations increase is one of the most important.

The issue of whether warming during past glacial terminations led or lagged CO2 increases is not relevant to the current situation. There is no question that the rapid increase in CO2 concentrations since 1850 is a direct result of human activities (mainly burning of coal and oil). We also know that CO2 and other greenhouse gases change the energy balance of the planet and cause the surface temperature to be warmer than it would be in the absence of those gases. The evidence surrounding the warming we have observed over recent decades indicates that most of this warming is very likely due to the increase in greenhouse gas concentrations in the atmosphere due to human activities.

Does the apparent slowing or halting of global warming over the last decade mean that climate change has stopped?

Globally, 12 of the last 13 years are the hottest years on record. Natural climate variability causes variations in temperature from year to year, particularly in response to the El Niño cycle whereby atmospheric and ocean circulation patterns change and cause regional climates to change for up to several years at a time. The year of 1998 experienced a strong El Niño event, which underpinned very warm conditions on average over the earth. In addition, solar sunspots affect the brightness of the sun on an 11-year cycle and may also change global temperatures by a few tenths of a degree. The effect of warming and cooling on the earth lags peaks and troughs in the cycle. The sunspot cycle is currently at a minimum, which may be contributing a slight cooling effect. The Hadley Centre data show that no year since 1998 has been as hot as that year, but most have been much hotter than the long-term average.

It has recently been suggested that we may be entering a period of up to two decades of global temperature stasis or even slight cooling caused by reduced solar activity, though there is dispute over whether such a mechanism could reverse the warming effect of anthropogenic emissions. The current apparent slowing in the rate of global warming will no doubt continue to generate debate, and the next several years will see the issue evolve subject to inputs of new data and modelling studies and scrutiny of the scientific process, from which a more complete understanding should emerge. It is clear, however, that the current global surface temperature is unprecedented in recorded history. Though natural processes may moderate the rate of warming over timescales of a few years, the evidence suggests that human-induced warming will continue over the next several decades at a rate that can only effectively be reduced by substantial cuts in emissions.

To what extent is the climate system self-stabilising, and will feedback mechanisms counteract the effects of greenhouse gases?

The IPCC has reviewed all published data and research, and its projections account for all known feedbacks to the climate system. However, feedbacks to the climate system are difficult to predict and there are conflicting results from a number of modelling studies. Clouds and their interactions with aerosols in particular are one of the most difficult aspects of the climate system to assess and predict, and this issue is addressed in more detail here.

The IPCC AR4 notes that cloud feedbacks are the largest source of uncertainty in model predictions of climate sensitivity to greenhouse gas emissions, and the models do not even simulate present-climate clouds well. Depending on their altitude, thickness and droplet size, clouds can either have a net warming effect on the earth or a net cooling effect, with research suggesting that overall the cooling effect outweighs the warming effect.

It is hypothesised that aerosols (small particles suspended in the atmosphere, e.g. dust and urban and industrial pollutants) increase the depth and lifetime of clouds, as well as their brightness. They also make the atmosphere more reflective in the absence of clouds. These influences all increase the amount of solar radiation that is reflected back into space, thereby reducing the radiation reaching the earth's surface and providing a cooling effect. Since human industry emits aerosols to the atmosphere alongside greenhouse gases, this provides a possible negative feedback mechanism to alleviate the global warming that would otherwise occur. However, aerosols can also trap heat within the atmosphere, and it is difficult to determine where the balance between the competing and interacting mechanisms lies.

Two papers based on the latest satellite data claim to show that the aerosol/cloud feedback effects on the climate system are more negative than was previously thought—that is, they counteract the warming effect. The authors state that the models on which the IPCC AR4's conclusions were based underestimate these effects, and overestimate the greenhouse warming effect. If true, this could have profound implications for our interpretation of climate change, because the observed rate of warming is larger than the rate that he calculates should occur from the greenhouse effect, implying that this warming must be mostly due to natural effects rather than human activities. Other recent independent research aligns with these findings. A NASA news release in December 2007 states that new data from the Aqua satellite shows the first global evidence that pollution of clouds by aerosols is making clouds brighter and more reflective, thus increasing the cooling effect of clouds. The research, currently in press, supports the idea that clouds present a more negative feedback than previously thought.

However, other scientists have challenged the validity of the model assumptions in the papers, and point out that the modelling has not been subject to the same scrutiny as those that underpin the IPCC's conclusions. Furthermore, the models that best replicate current conditions lead to the opposite conclusion, that the negative feedbacks are perhaps overestimated. Scientists at the Hadley Centre (UK) analysed new satellite data and found that current models may overestimate the extent to which anthropogenic aerosols increase cloud reflectivity. The researchers note, however, that the uncertainty in their estimates of the influence of aerosols is larger than the influence itself, i.e. they cannot say with certainty whether it is a warming or cooling influence.

These various results and interpretations indicate the high level of uncertainty surrounding the influence of feedbacks on the climate system and how they might respond as the climate warms further. However, the IPCC Assessment Reports review all the published literature to assess the state of knowledge, to try to reconcile contrasting evidence and document uncertainties. The uncertainty about the influence of aerosols on the climate system and the strength or sign of the feedback has been acknowledged and addressed for several years and the recent publications utilising the latest satellite data (that have been published after the latest IPCC report) add data to the debate but do not resolve it at this time. Based on the evidence currently available, there is no reason to suppose that the IPCC interpretations and projections are erroneous.

Further reading:

Climate change 2007—The physical science basis, Working Group I contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change.

H. H. Lamb, 'The early medieval warm epoch and its sequel', Palaeogeography, Palaeoclimatology, Palaeoecology, 1: 13–37, 1965.

R. S. Bradley, M. K. Hughes and H. F. Diaz, 'Climate in Medieval time', Science, 302: 404–5, 2003.

T. J. Osborn and K. R. Briffa, 'The spatial extent of 20th century warmth in the context of the past 1200 years', Science, 311: 841–4, 2006.

R. W. Spencer, W. D. Braswell, J. R. Christy and J. Hnilo, 'Cloud and radiation budget changes associated with tropical intraseasonal oscillations', Geophysical Research Letters, 34, L15707, doi:10.1029/2007GL029698, 2007.

R. W. Spencer and W. D. Braswell, 'Potential biases in feedback diagnosis from observational data—a simple model demonstration', Journal of Climate, 21, p. 5624, 2008.

'NASA satellites help lift cloud of uncertainty on climate change', NASA media release, 12 December 2007.

M. Lebsock,, G. L. Stephens and C. Kummerow, 'Multi-sensor satellite observations of aerosol effects on warm clouds', Journal of Geophysical Research, 113, D15205, doi:10.1029/2008JD009876, 2008.

N. Bellouin, A. Jones, J. Haywood and S. A. Christopher, 'Updated estimate of aerosol direct radiative forcing from satellite observations and comparison against the Hadley model', Journal of Geophysical Research, 113, D10205, doi:10.1029/2007JD009385, 2008.

J. Quaas, O. Boucher, N. Bellouin and S. Kinne, 'Satellite-based estimate of the direct and indirect aerosol climate forcing', Journal of Geophysical Research, 113, D05204, doi:10.1029/2007JD008962, 2008.

M. E. Mann et al, 'Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia', Proc. Natl. Acad. Sci., 105, pp. 13252–7, 2008.

22 October, 2010

Facebook LinkedIn Twitter Add | Email Print
Back to top