Climate change is among the most talked about and investigated global risks. No other environmental issue receives quite as much attention in the popular press, even though the impacts of pandemics and asteroid strikes, for instance, may be much more severe. Since the first Intergovernmental Panel on Climate Change (IPCC) report in 1990, significant progress has been made in terms of (1) establishing the reality of anthropogenic climate change and (2) understanding enough about the scale of the problem to establish that it warrants a public policy response. However, considerable scientific uncertainty remains. In particular scientists have been unable to narrow the range of the uncertainty in the global mean temperature response to a doubling of carbon dioxide from pre-industrial levels, although we do have a better understanding of why this is the case. Advances in science have, in some ways, made us more uncertain, or at least aware of the uncertainties generated by previously unexamined processes. To a considerable extent these new processes, as well as familiar processes that will be stressed in new ways by the speed of twenty-first century climate change, underpin recent heightened concerns about the possibility of catastrophic climate change.
Discussion of ‘tipping points’ in the Earth system (for instance Kemp, 2005; Lenton, 2007) has raised awareness of the possibility that climate change might be considerably worse than we have previously thought, and that some of the worst impacts might be triggered well before they come to pass, essentially suggesting the alarming image of the current generation having lit the very long, slow-burning fuse on a climate bomb that will cause great devastation to future generations. Possible mechanisms through which such catastrophes could play out have been developed by scientists in the last 15 years, as a natural output of increased scientific interest in Earth system science and, in particular, further investigation of the deep history of climate. Although scientific discussion of such possibilities has usually been characteristically guarded and responsible, the same probably cannot be said for the public debate around such notions. Indeed, many scientists regard these hypotheses and images as premature, even alarming. Mike Hulme, the Director of the United Kingdom’s Tyndall Centre recently complained that:
The IPCC scenarios of future climate change – warming somewhere between 1.4 and 5.8° Celsius by 2100 – are significant enough without invoking catastrophe and chaos as unguided weapons with which forlornly to threaten society into behavioural change. […] The discourse of catastrophe is in danger of tipping society onto a negative, depressive and reactionary trajectory.
This chapter aims to explain the issue of catastrophic climate change by first explaining the mainstream scientific (and policy) position: that twenty-first century climate change is likely to be essentially linear, though with the possibility of some non-linearity towards the top end of the possible temperature range. This chapter begins with a brief introduction to climate modelling, along with the concept of climate forcing. Possible ways in which things could be considerably more alarming are discussed in a section on limits to our current knowledge, which concludes with a discussion of uncertainty in the context of palaeoclimate studies. We then discuss impacts, defining the concept of ‘dangerous anthropogenic interference’ in the climate system, and some regional impacts are discussed alongside possible adaptation strategies. The chapter then addresses mitigation policy-policies that seek to reduce the atmospheric loading of greenhouse gases (GHG) – in light of the preceeding treatment of linear and non-linear climate change. We conclude with a brief discussion of some of the main points and problems.
Scientists attempting to understand climate try to represent the underlying physics of the climate system by building various sorts of climate models. These can either be top down, as in the case of simple models that treat the Earth essentially as a closed system possessing certain simple thermodynamic properties, or they can be bottom up, as in the case of general circulation models (GCMs), which attempt to mimic climate processes (such as cloud formation, radiative transfer, and weather system dynamics). The range of models, and the range of processes they contain, is large: the model we use to discuss climate change below can be written in one line, and contains two physical parameters; the latest generation of GCMs comprise well over a million lines of computer code, and contain thousands of physical variables and parameters. In between there lies a range of Earth system models of intermediate complexity (EMICs) (Claussen et al., 2002; Lenton et al., 2006) which aim at resolving some range of physical processes between the global scale represented by EBMs and the more comprehensive scales represented by GCMs. EMICs are often used to investigate long-term phenomena, such as the millennial scale response to Milankovitch cycles, Dansger-Oeschgaard events or other such episodes and periods that it would be prohibitively expensive to investigate with a full GCM. In the following section we introduce and use a simple EBM to illustrate the global response to various sorts of forcings, and discuss the source of current and past climate forcings.
A very simple model for the response of the global mean temperature to a specified climate forcing is given in the equation below. This model uses a single physical constraint – energy balance – to consider the effects of various drivers on global mean temperature. Though this is a very simple and impressionistic model of climate change, it does a reasonable job of capturing the aggregate climate response to fluctuations in forcing. Perturbations to the Earth’s energy budget can be approximated by the following equation (Hansen et al., 1985):
in which ceff is the effective heat capacity of the system, governed mainly (Levitus et al., 2005) by the ocean, X is a feedback parameter, and AT is a global temperature anomaly. The rate of change is governed by the thermal inertia of the system, while the equilibrium response is governed by the feedback parameter alone (since the term on the left hand side of the equation tends to zero as the system equilibrates). The forcing, F, is essentially the perturbation to the Earth’s energy budget (in W/m2) which is driving the temperature response (ΔT). Climate forcing can arise from various sources, such as changes in composition of the atmosphere (volcanic aerosols; GHG) or changes in insolation. An estimate of current forcings is displayed in Fig. 13.1, and an estimate of a possible range of twenty-first century responses is shown in Fig.13.2. It canbe seen that the bulk of current forcing comes from elevated levels of carbondioxide (CO2), though other agents are also significant. Historically, three main forcing mechanisms are evident in the temperature record: solar forcing, volcanic forcing and, more recently, GHG forcing. The range of responses in Fig. 13.2 is mainly governed by (1) choice of future GHG scenario and (2) uncertainty in the climate response, governed in our simple model (which can essentially replicate Fig. 13.3) by uncertainty in the parameters ceff and X. The scenarios considered are listed in the figure, along with corresponding grey bands representing climate response uncertainty for each scenario (at right in grey bands).
Fig. 13.1 Global average radiative forcing (RF) estimates and ranges in 2005 for anthropogenic carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O), and other important agents and mechanisms, together with the typical geographical extent (spatial scale) of the forcing and the assessed level of scientific understanding (LOSU). The net anthropogenic radiative forcing and its range are also shown. These require summing asymmetric uncertainty estimates from the component terms and cannot be obtained by simple addition. Additional forcing factors not included here are considered to have a very low LOSU. Volcanic aerosols contribute an additional natural forcing but are not included in this figure due to their episodic nature. The range for linear contrails does not include other possible effects of aviation on cloudiness. Reprinted with permission from Solomon et al. (2007). Climatic Change 2007: The physical Science Basis.
Among the most obvious ways in which the energy balance can be altered is if the amount of solar forcing changes. This is in fact what happens as the earth wobbles on its axis in three separate ways on times cales of tens to hundreds of thousands of years. The earth’s axis precesses with a period of 15,000 years, the obliquity of the Earth’s orbit oscillates with a period of around 41,000 years, and its eccentricity varies on multiple times cales (95,000, 125,000, and 400,000 years).
These wobbles provide the source of the earth’s periodic ice ages, by varying the amount of insolation at the Earth’s surface. In particular, the Earth’s climate is highly sensitive to the amount of solar radiation reaching latitudes north of about 60°N. Essentially, if the amount of summer radiation is insufficient to melt the ice that accumulates over winter, the depth of ice thickens, reflecting more back. Eventually, this slow accrual of highly reflective and very cold ice acts to reduce the Earth’s global mean temperature, and the Earth enters an ice age.
Fig. 13.2 Forcing and temperature response curves using the simple models presented in this chapter. In the top panel the thin line corresponds to historical forcings over the twentieth century; the solid line corresponds to GHG-only forcings, and these are projected forward under an extended version of the IPCC B1 scenario (SRES, 1992). In the bottom panel is the base model [Equation (13.1)] response to the B1 scenario with best guess parameters from Frame et al. (2006). Also shown (dashed line) is the response with the arbitrary disruption represented by Equation (13.1b). SRES: N. Nakicenovic et al., IPCC Special Report on Emission Scenarios, IPCC, Cambridge University Press, 2000. Frame, D.J., D.A. Stone, P.A. Stoll, M.R. Allen, Alternatives to stabilization scenarios, Geophysical Research Letters, 33, L14707.
Current solar forcing is thought to be +0.12 Wm−2, around one- eighth of the total current forcing (above pre-industrial levels). There has been some advocacy for a solar role in current climate change, but it seems unlikely that this is the case, since solar trends over some of the strongest parts of the climate change signal appear to be either insignificant or even out of phase with the climate response (Lockwood and Frohlich, 2007). Detection and attribution studies of recent climate change have consistently been able to detect a climate change signal attributable to elevated levels of GHG, but have been unable to detect a signal attributable to solar activity (Hegerl et al., 2007). Though dominant historically over millennial and geological timescales, the sun apparently is not a significant contributor to current climate change.
Another form of climate forcing comes from the explosive release of aerosols into the atmosphere through volcanoes. This is a significant forcing when aerosols can make it into the stratosphere where the residence time for aerosols is several years. During this time, they act to cool the planet by back-scattering incoming solar radiation. Though peak forcings due to large volcanoes are massive, the stratospheric aerosol residence time is short (~4 years) in comparison to that of the principal anthropogenic GHG, CO2 (~150 years) so volcanic forcings tend to be much more dramatic and spikier (Fig. 13.3) than those associated with GHG. Peak forcing from the recent El Chichon (1982) and Mt Pinatubo (1991) volcanoes was on the order of – 3Wm−2. Though it is harder to infer the forcing from volcanoes that pre-date the satellite record, estimates suggest that volcanic forcing from Karakatau and Tambora in the nineteenth century are likely to have been larger.
Fig. 13.3 Solid lines are multi-model global averages of surface warming (relative to 1980–1999) for the scenarios A2, A1B, and B1, shown as continuations of the twentieth century simulations. Shading denotes the ±1 standard deviation range of individual model annual averages. The orange line is for the experiment where concentrations were held constant at year 2000 values. The grey bars at right indicate the best estimate (solid line within each bar) and the likely range assessed for the six SRES marker scenarios. The assessment of the best estimate and likely ranges in the grey bars includes the AOGCMs in the left part of the figure, as well as results from a hierarchy of independent models and observational constraints. Reprinted with permission from Solomonet al. (2007). Climatic Change 2007: The physical Science Basis.
Even bigger are supervolcanoes, the most extreme class of volcanic events seen on Earth. Supervolcanoes eject over a 1000 km3 of material into the atmosphere (compared with around 25 km3 for Pinatubo). They are highly uncommon, occurring less than once every 10,000 years, with some of the largest events being increasingly rare. Examples include the formation of the La Garita Caldera, possibly the largest eruption in history, which occurred some 28 million years ago, ejecting around 5000 km3 of matter and, more recently, the Toba explosion, which occurred about 71,000 years ago and ejected around 2800 km3 of material into the atmosphere. These very large events occur with a frequency in the vicinity of 1.4-22 events per million years (Mason et al., 2004), and the fact that the frequency of the most extreme events tends to fall away quickly with the magnitude of the event suggests that there is some sort of upper limit to the size of volcanic events (Mason et al., 2004).
In the last 1000 years, massive volcanic eruptions, responsible for climate forcings on the order of 10, occurred around 1258, probably in the tropics (Crowley, 2000). This medieval explosion was enormous, releasing around eight times the sulphate of the Krakatau explosion. However, correlations between the amount of sulphate released and observed global mean temperatures show that the temperature response for extremely large volcanoes does not scale simply from Krakatau-sized volcanoes (and smaller). Pinto et al. (1989) suggest that this is because beyond a certain level, increases in sulphate loading act to increase the size of stratospheric aerosols, rather than increasing their number. In essence, the radiative forcing per unit mass decreases at very large stratospheric sulphate loadings, and this moderates the cooling effects of very large volcanoes.
The expectations for the coming decades are that the odds of a volcanically induced radiative forcing of – 1.0 Wm – 2 or larger occurring in a single decade are in the range 35–40%. The odds of two such eruptions in a single decade are around 15%, while the odds of three or more such eruptions is something like 5%. The odds of getting lucky volcanically (though perhaps not climatically) and experiencing no such eruptions is approximately 44%. Considering larger volcanic events, there is a 20% likelihood of a Pinatubo-scale eruption in the next decade and a 25% chance for an El Chichón-scale eruption (Hyde and Crowley, 2000). Although events of this magnitude would certainly act to cool the climate, and lead to some deleterious impacts (especially in the vicinity of the volcano) they would represent essentially temporary perturbations to the expected warming trend.
A simple model for the accumulation of GHG (simplified here to carbon dioxide), its conversion to radiative forcing, and the subsequent effect on global mean temperatures, can be given by the following two equations:
Here, CO2 (t) represents the atmospheric concentration of carbon dioxide over time, which is taken to be the sum of emissions from natural (enat) and anthropogenic (eant) sources. We can further specify a relationship between carbon dioxide concentration and forcing:
in which F2XCO2 is the forcing corresponding to a doubling of CO2, CO2 refers to the atmospheric concentration of CO2 and CO2pre is the (constant) pre-industrial concentration of CO2. There is some uncertainty surrounding F2x CO2, but most general circulation models (GCMs) diagnose it to be in the range 3.5−4.0 W/m2. Current forcings are displayed in Fig. 13.1. This shows that the dominant forcing is due to elevated levels (above pre-industrial) of carbon dioxide, with the heating effects of other GHG and the cooling effects of anthropogenic sulphates happening to largely cancel each other out.
It suffers from certain limitations but mimics the global mean temperature response of much more sophisticated state of the art GCMs adequately enough, especially with regard to relatively smooth forcing series. Both the simple model described above and the climate research community’s best-guess GCMs predict twenty-first century warmings, in response to the kinds of elevated levels of GHG we expect to see this century, of between about 1.5°C and 5.8°C.
The key point to emphasize in the context of catastrophic risk is that this warming is expected to be quite smooth and stable. No existing best guess state-of-the-art model predicts any sort of surprising non-linear disruption in terms of global mean temperature. In recent years, climate scientists have attempted to quantify the parameters in the above model by utilizing combinations of detuned models, simple (Forest et al., 2002) or complex (Murphy et al., 2004) constrained by recent observations (and sometimes using additional information from longer term climate data (Hegerl et al., 2006; Schneider von Deimling et al., 2006). These studies have reported a range of distributions for climate sensitivity, and while the lower bound and median values tend to be similar across studies there is considerable disagreement regarding the upper bound.
Such probabilistic estimates are dependent on the choice of data used to constrain the model, the weightings given to different data streams, the sampling strategies and statistical details of the experiment, and even the functional forms of the models themselves. Although estimation of climate sensitivity remains a contested arena (Knutti et al., 2007) the global mean temperature response of the models used in probabilistic climate forecasting can be reasonably well approximated by manipulation of the parameters ceff and, especially, X, in Equation(13.1) of this chapter. Essentially, Equation (13.1) can be tuned to mimic virtually any climate model in which atmospheric feedbacks scale linearly with surface warming and in which effective oceanic heat capacity is approximately constant under twentieth century climate forcing. This includes the atmospheric and oceanic components of virtually all current best-guess GCMs.
However, we know there are plenty of ways in which our best-guess models may be wrong. Feedbacks within the carbon cycle, in particular, have been the source of much recent scientific discussion. Amazon dieback, release of methane hydrates from the ocean floor, and release of methane from melting permafrost are among mechanisms thought to have the potential to add nasty surprises to the linear picture of anthropogenic climate change we see in our best models. There are essentially two ways in which we could be wrong about the evolution of the system towards the target, even if we knew the climate sensitivity S:
• Forcing could be more temperature-dependent than we thought. That is: 
.
• Feedbacks could be more temperature-dependent than we thought. In this case: 
.
In Fig. 13.2, we plot the latter sort of variant: in the dotted line in the bottom panel of Fig. 13.2 the forcing remains the same as in the base case, but the response is different, reflecting an example of a disruptive temperature dependence of X, which is triggered as the global mean temperature anomaly passes 2.5°C. This is a purely illustrative example, to show how the smooth rise in global mean temperature could receive a kick from some as yet un-triggered feedback in the system. In the last decade, scientists have investigated several mechanisms that might make our base equation more like (1) or (2). These include the following:
• Methane hydrate release (F) in which methane, currently locked up in the ocean in the form of hydrates can be released rapidly if ocean temperatures rise beyond some threshold value (Kennett et al., 2000). Since methane sublimes (goes straight from a solid phase to a gaseous phase) this would lead to a very rapid rise in atmospheric GHG loadings. Maslin et al. (2004) claim that there is evidence for the hypothesis from Dansgaard-Oeschger interstadials; Benton and Twitchett (2003) claim that the sudden release of methane gas from methane hydrates (sometimes called the ‘clathrate gun hypothesis’) may have been responsible for the mass extinction at the end of the Permian, the ‘biggest crisis on Earth in the past 500 Myr’. If this is so, they claim with considerable understatement, ‘it is […] worth exploring further’.
• Methane release from melting permafrost (F) is related to the oceanic clathrate gun hypothesis. In this case, methane is trapped in permafrost through the annual cycle of summer thawing and winter freezing. Vegetable matter and methane gas from the rotting vegetable material get sequestered into the top layers of permafrost; the reaction with water forms methane hydrates (Dallimore and Collet, 1995). Increased temperatures in permafrost regions may give rise to sufficient melting to unlock the methane currently sequestered, providing an amplification of current forcing (Sazonova et al., 2004).
• Tropical forest (especially Amazonian) dieback (F) is perhaps the best known amplification to expected climate forcing. Unlike the previous two examples which are additional sources of GHG, tropical forest dieback is a reduction in the strength of a carbon sink (Cox et al., 2004). As temperatures rise, Amazonia dries and warms, initiating the loss of forest, leading to a reduction in the uptake of carbondioxide (Cox et al., 2000).
• Ocean acidification(F)is an oceanic analogue to forest dieback from a carbon cycle perspective in that it is the reduction in the efficacy of a sink. Carbon dioxide is scrubbed out of the system by phytoplankton. As carbon dioxide builds up the ocean acidifies, rendering the phytoplankton less able to perform this role, weakening the ocean carbon cycle (Orr et al., 2005).
• Thermohaline circulation disruption (λ), while perhaps less famous than Amazonian dieback, can at least claim the dubious credit of having inspired a fully fledged disaster movie. In this scenario, cool fresh water flowing from rapid disintegration of the Greenland ice sheet acts to reduce the strength of the (density-driven) vertical circulation in the North Atlantic (sometimes known as the ‘thermohaline circulation’, but better described as the Atlantic meridional overturning circulation. There is strong evidence that this sort of process has occurred in the Earth’s recent past, the effect of which was to plunge Europe into an ice age for another 1000 years.
• The iris effect (λ) would be a happy surprise. This is a probable mechanism through which the Earth could self-regulate its temperature. The argument, developed in Lindzen et al. (2001) is that tropical cloud cover could conceivably increase in such a way as to reduce the shortwave radiation received at the surface. On this view, the expected (and worrying warming) will not happen or will largely be offset by reductions in insolation. Thus, while the Earth becomes less efficient at getting rid of thermal radiation, it also becomes better at reflecting sunlight. Unlike the other mechanisms discussed here, it would cause us to undershoot, rather than overshoot, our target (the ball breaks left, rather than right).
• Unexpected enhancements to water vapour or cloud feedbacks (λ). In this case warming triggers new cloud or water vapour feedbacks that amplify the expected warming.
Those marked (F) are disruptions to the earth’s carbon cycle, the others, marked (λ), can be seen as temperature-dependent disruptions to feedbacks. In reality, many of the processes that affect feedbacks would likely trigger at least some changes in carbon sources and sinks. In practice, the separation between forcing surprises and feedback surprises is a little artificial, but it is a convenient first order way to organize potential surprises. All are temperature-dependent, and this temperature-dependence goes beyond the simple picture sketched out in the section above. Some could be triggered – committed to – well in advance of their actual coming to pass. Most disturbingly, none of these surprises are well-quantified. We regard them as unlikely, but find it hard to know just how unlikely. We are also unsure of the amplitude of these effects. Given the current state of the climate system, the current global mean temperature and the expected temperature rise, we do not know much about just how big an effect they will have. In spite of possibly being responsible for some dramatic effects in previous climates, it is hard to know how big an effect surprises like these will have in the future. Quantifying these remote and in many cases troubling possibilities is extremely difficult. Scientists write down models that they think give a justifiable and physically meaningful description of the process under study, and then attempt to use (usually palaeoclimate) data to constrain the magnitude and likelihood of the event. This is a difficult process: our knowledge of past climates essentially involves an inverse problem, in which we build up a picture of three-dimensional global climates from surface data proxies, which are often either biological data (tree-ring patterns, animal middens, etc.) or cryospheric (ice cores). Climate models are often run in under palaeo conditions in order to provide an independent check on the adequacy of the models, which have generally been built to simulate today’s conditions. Several sources of uncertainty are present in this process:
1. Data uncertainty. Generally, the deeper we look into the past, the less sure we are that our data are a good representation of global climate. The oceans, in particular, are something of a data void, and there are also issues about the behaviour of clouds in previous climates.
2. Forcing uncertainty. Uncertainty in forcing is quite large for the pre-satellite era generally, and again generally grows as one moves back in time. In very different climate regimes, such as ice ages, there are difficult questions regarding the size and reflectivity of ice sheets, for instance. Knowing that we have the right inputs when we model past climates is a significant challenge.
3. Model uncertainty. Climate models, while often providing reasonable simulations over continental scales, struggle to do a good job of modelling smaller-scale processes (such as cloud formation). Model imperfections thus add another layer of complexity to the problem, and thus far at least, GCMs simulating previous climates such as the last glacial maximum have had little success in constraining even global mean properties (Crucifix, 2006).
In the presence of multiple scientific uncertainties, it is hard to know how to reliably quantify possibilities such as methane hydrate release or even meridional overturning circulation disruption. These things remain possibilities, and live areas of research, but in the absence of reliable quantification it is difficult to know how to incorporate such possibilities into our thinking about the more immediate and obvious problem of twenty-first century climate change.
While this book has drawn the distinction between terminal and endurable risks (see Chapter 1), limits to acceptable climate change risk have usually been couched in terms of ‘dangerous’ risk. The United Nations Framework Convention on Climate Change (UNFCCC), outlines a framework through which nations might address climate change. Under the Convention, governments share data and try to develop national strategies for reducing GHG emissions while also adapting to expected impacts. Central to the convention, in article 2, is the intention to stabilize GHG concentrations (and hence climate):
The ultimate objective of this Convention and any related legal instruments that the Conference of the Parties may adopt is to achieve, in accordance with the relevant provisions of the Convention, stabilization of greenhouse gas concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with the climate system.
Article 2 introduced the problematic but perhaps inevitable concept of ‘dangerous anthropogenic interference’ (DAI) in the climate system into the scientific and policy communities. DAI has been much discussed, with many quite simple studies presenting it, to first order, in terms of possible global mean temperature thresholds (O’Neill and Oppenheimer, 2002; Mastrandrea and Schneider, 2004), or equilibrium responses to some indefinite atmospheric concentration of CO2 (Nordahus, 2005), although some commentators provide more complete discussions of regional or not-always-temperature based metrics of DAI (Keller et al., 2005; Oppenheimer, 2005).
It has become usual to distinguish between two, ideally complementary ways of dealing with the threats posed by climate change: mitigation and adaptation.
The UNFCCC’s glossary 1 describes mitigationas follows:
In the context of climate change, a human intervention to reduce the sources or enhance the sinks of greenhouse gases. Examples include using fossil fuels more efficiently for industrial processes or electricity generation, switching to solar energy or wind power, improving the insulation of buildings, and expanding forests and other ‘sinks’ to remove greater amounts of carbon dioxide from the atmosphere.
While adaptationis:
Adjustment in natural or human systems in response to actual or expected climatic stimuli or their effects, which moderates harm or exploits beneficial opportunities.
Internationally, the focus of most climate policy has been on establishing targets for mitigation. The Kyoto Protocol, for instance, set ‘binding’ targets for emissions reductions for those industrialized countries which chose to ratify it. This is changing as nations, communities, and businesses (1) come to appreciate the difficulty of the mitigation coordination problem and (2) realize that they are beginning to need to adapt their practices to a changing climate.
In the general case, what is dangerous depends on who one is, and what one is concerned about. Local thresholds in temperature (for coral reefs, say), integrated warming (polar bears), or sea level (for small island states) rise, to take a few examples, may constitute compelling metrics for regional or highly localized DAI, but it is not obvious that they warrant a great deal of global consideration. Moreover, the inertia in the system response implies that we are already committed to a certain level of warming: the climate of the decade between 2020 and 2030 is essentially determined by the recent accumulation of GHG and no realistic scenario for the mitigation will alter that. Thus, as well as mitigating against the worst aspects of centennial timescale temperature rise, we also need to learn to adapt to climate change. Adaptation is inherently a regional or localized response, since the effects of climate change – though both most discernable and best constrained at the global level – actually affect people’s lives at regional scales. Although mitigation is usually thought of as a genuinely global problem requiring a global scale response, adaptation is offers more opportunity for nation states and smaller communities to exercise effective climate policy. Adaptation alone is not expected to be sufficient to deal with climate change in the next century, and is best seen as part of a combined policy approach (along with mitigation) that allows us to avoid the very worst impacts of climate change, while adapting to those aspects that we are either already committed to, or for which the costs of mitigation would be too high.
The IPCC’s Fourth Assessment Report highlights a number of expected changes at regional scales. These include the following:
• Extra-tropical storm tracks are expected to move polewards.
• Hot extremes are expected to increase in frequency.
• Hurricanes are expected to become more frequent, with more intense precipitation.
• Warming is expected over land areas, especially over the continents, and is also expected to be strong in the Arctic.
• Snow cover and sea-ice are expected to decrease.
• Patterns of precipitation are generally expected to intensify.
These findings are qualitatively similar to those highlighted in the IPCC’s Third Assessment Report, though some of the details have changed, and in many cases scientists are more confident of the predicted impacts.
Impacts generally become more uncertain as one moves to smaller spatio-temporal scales. However, many large-scale processes, such as El Nino Southern Oscillation and the North Atlantic Oscillation, are crucially dependent on small-scale features such as convective systems over the maritime continent, and patterns Atlantic sea-surface temperature, respectively. Given that the El Nino Southern Oscillation and the North Atlantic Oscillation are key determinants of variability and climate-related impacts on seasonal-to-decadal timescales, good representation of these features will become an essential part of many regions’ adaptation strategies.
Adaptation strategies can take many forms. They can be based on strategies that take different sorts of stakeholder relationships with the impacts into account. Depending on the nature of one’s exposure, risk profile, and preferences, one may opt to base one’s strategy on one, or a mix, of several different principles, including: some reading of the precautionary principle, traditional cost-benefit analysis (usually where one is confident of the numbers underpinning the analysis) or robustness against vulnerability. Other adaptive principles have also been discussed, including novel solutions such as legal remedies through climate detection and attribution studies, which is like cost-benefit analysis, though backward looking (at actual attributable harm done) and done via the legal system rather than through economics (Allen et al., 2007). As the choice of principle, tools and policies tend to be highly situation specific, there tends to be a great diversity in adaptation strategies, which contrasts with the surprisingly narrow band of mitigation policies that have entered the public domain.
Apart from a range of possible principles guiding the adaptation strategy, the range of possible responses is also very large. Technological responses include improved sea and flood defences, as well as improved air-conditioning systems; policy responses include changes in zoning and planning practices (to avoid building in increasingly fire- or flood-prone areas, for instance); managerial responses include changes in firm behaviour – such as planting new crops or shifts in production- in anticipation of the impacts associated with climate change; and behavioural responses include changes in patterns of work and leisure and changes in food consumption.
In traditional risk analysis, the risk of an event is defined as the costs associated with its happening multiplied by the probability of the event. However, identifying the costs associated with adapting to climate change is deeply problematic, since (1) the damages associated with climate change are extremely hard to quantify in the absence of compelling regional models and (2) the economic future of the community implementing the adaptive strategy is highly uncertain. As the IPCC AR4 states:
At present we do not have a clear picture of the limits to adaptation, or the cost, partly because effective adaptation measures are highly dependent on specific, geographical and climate risk factors as well as institutional, political and financial constraints. p. 19
An alternative strategy, and one that seems to be gaining popularity with the adaptation community, is the development of ‘robust’ strategies that seek to minimize the vulnerability of a community to either climate change, or even natural climate variability. Because of the difficulty in quantifying the costs of adaptation and likely development pathways, strategies emphasizing resilience are gaining in popularity.
The default science position from which to think about climate mitigation policy has been the simple linear model of anthropogenic warming along with the expected solar and volcanic forcings, which remains basically the same (in the long-run) even in the presence of large volcanoes that we might expect only once in every 1000 years. This is usually coupled with a cost-benefit analysis of some kind to think about what sort of policy we should adopt. However, many of the hard to quantify possibilities that would disrupt this linear picture of climate change are often thought to lead to dramatic changes in our preferred policy responses to twenty-first century climate change.
The basic problem of anthropogenic climate change is that we are releasing too much carbondioxide 2 into the atmosphere. The principal aspect of the system we can control is the amount of CO2 we emit.3 This is related to the temperature response (a reasonable proxy for the climate change we are interested in) by the chain emissions concentrations forcings temperature response, or in terms of the equations we introduced above, (13.2) – (13.3) – (13.1). Various aspects of this chainare subject to considerable scientific uncertainty. The mapping between emissions and concentrations is subject to carbon cycle uncertainties on the order of 30% of the emissions themselves, for a doubling of CO2 (Huntingford and Jones, 2007); and though the concentrations to forcings step seems reasonably well behaved the forcing-response step is subject to large uncertainties surrounding the values of the parameters in Equation (13.1) (Hegerl et al., 2006). So eventhough decision makers are interested in avoiding certain aspects of the response, they are stuck with a series of highly uncertain inferences in order to calculate the response for a given emissions path.
In order to control CO2 emissions, policy makers can prefer price controls (taxes) or quantity controls (permits). Quantity controls directly control the amount of a pollutant released into a system, whereas price controls do not. In the presence of uncertainty, if one fixes the price of a pollutant, one is uncertain about the quantity released; whereas if one fixes the quantity, one is uncertain about how to price the pollutant. Weitzman (1974) showed how, in the presence of uncertainty surrounding the costs and benefits of a good, a flat marginal benefit curve, compared to the marginal cost curve, leads to a preference for price regulation over quantity regulation. If anthropogenic climate change is as linear as it seems under the simple model mentioned above, as most studies have quite reasonably assumed to be the case, then a classical optimal cost-benefit strategy suggests that we ought to prefer price controls (Nordhaus, 1994; Roughgarden and Schneider, 1997, for instance). Intuitively, this makes sense: if the marginal benefits of abatement are a weaker function of quantity than the marginal costs are, then it suggests getting the price right is more important than specifying the quantity, since there is no reason to try to hit some given quantity regardless of the cost. If, however, the marginal benefits of abatement are a steep function of quantity, then it suggests that specifying the quantity is more important, since we would rather pay a sub-optimal amount than miss our environmental target.
This sort of intuition has implied that our best-guess, linear model of climate change leads to a preference for price controls. At the same time, if it were shown that catastrophic disruptions to the climate system were a real threat, it would lead us to opt instead for quantity controls. Ideally, we would like to work out exactly how likely catastrophic processes are, factor these into our climate model, and thus get a better idea of what sort of mitigation policy instrument to choose. However, catastrophic surprises are difficult to incorporate into our thinking about climate change policy in the absence of strong scientific evidence for the reliable calculation of their likelihood. Our most comprehensive models provide scant evidence to support the idea that anthropogenic climate change will provide dramatic disruptions to the reasonably linear picture of climate change represented by the simple EBM presented above (Meehl et al., 2007). At the same time, we know that there are plenty of parts of the Earth System that could, at least in theory, present catastrophic risks to human kind. Scientists attempt to model these possibilities, but this task is not always well-posed, because (1) in a system as complex as the Earth system it is can be hard to know if the model structure is representative of the phenomena one is purporting to describe and (2) data constraints are often extremely weak. Subjective estimates of the reduction in strength of the Atlantic meridional overturning circulation in response to a 4° C global mean temperature rise put the probability of an overturning circulation collapse at anywhere between about 0% and 60% (Zickfeld et al., 2007). Worse, the ocean circulation is reasonably well understood in comparison to many of the other processes we considered above in Section 13.4: we have very few constraints to help us quantify the odds of a clathrate gun event, or to know just how much tropical forest dieback or permafrost methane release to expect. Attempting to quantify the expected frequency or magnitude of some of the more extreme possibilities even to several magnitudes is difficult, in the presence of multiple uncertainties.
Therefore, while the science behind our best expectations of anthropogenic climate change suggests a comparatively linear response, warranting a traditional sort of cost-benefit, price control-based strategy, catastrophe avoidance suggests a quite different, more cautious approach. However, we find it hard to work out just how effective our catastrophe avoidance policy will be, since we know so little about the relative likelihood of catastrophes with or without our catastrophe avoidance policy.
Different policy positions are suggested depending on whether twenty-first century climate change continues to be essentially linear, in line with recent temperature change and our best-guess expectations, or disruptive in some new and possibly catastrophic way. In the first case, as we have seen, price controls are preferable since there’s no particular reason to need to avoid any given quantity of atmospheric GHG; in the second, we are trying to remain under some threshold, and allow prices to adjust in response. Ideally, science would tell us what that threshold is, but as we have seen, this is a difficult, and elusive, goal. Furthermore, quantity controls remain indirect. The thing we wish to avoid is dangerous anthropogenic interference in the climate system. Yet we cannot control temperature, we can control only anthropogenic emissions of GHG, and these map to temperature increases through lagged and uncertain processes. This matters because it affects the way we construct policy. European policy makers, who have been at the forefront of designing policy responses to the threat of anthropogenic warming, generally back-infer acceptable ‘emissions pathways’ by inverting, starting with an arbitrarily chosen temperature target, and then attempting to invert the equations in this chapter. This bases short- and medium-term climate policy on the inversion of a series of long-term relationships, each of which is uncertain.
Dangerous, possibly even catastrophic, climate change is a risk, but it is exceptionally hard to quantify. There is evidence that the climate has changed radically and disruptively in the past. While its true that a global mean temperature rise of 4°C, would probably constitute dangerous climate change on most metrics, it would not necessarily be catastrophic in the sense of being a terminal threat. The challenges they suggest our descendents will face look tough, but endurable.
In terms of global climate, potential catastrophes and putative tipping points retain a sort of mythic aspect: they are part of a useful way of thinking about potentially rapid surprises in a system we do not fully understand, but we find it hard to know just how to incorporate these suspicions into our strategies. We cannot reliably factor them into our decision algorithms because we have such a poorly quantified understanding of their probability. We need to investigate the mechanisms more thoroughly to get a better handle on their amplitude, triggering mechanisms, and likelihood. There is no unique and uncontroversial way to bring these concerns within the realm of climate forecasting, nor, especially, our decision-making processes.
Pizer, W.A. http://www.ss.org.ss/Documents/RII-DP-03-31.pdg. Climate Change Catastrophes, Discussion Paper 03–31, Resources for the Future, Washington, DC, 2003. Provides a nice summary of the ways in which catastrophic risks can change the choice of climate mitigation policy instrument.
• Stern Review. The most comprehensive economic review yet undertaken into the http://www.hm-treasury.gov.uk/independant_reviews/stern_review_economics_climate_change/stern_review_report.cfm problem of climate change. Although controversial in some parts, the Stern review is probably the most important document on climate change produced in the last few years: it has moved the discourse surrounding climate change away from scientific issues towards policy responses.
• IPCC AR4 Summary for Policy Makers. WG1 provides a comprehensive review of the state of climate change science. To a large extent it echoes the previous assessment (2001).
• S.H. Schneider, A. Rosencranz, J.O. Niles 2002, climate change policy: A A survey Island Press. Schneider et al. on climate policy. Provides a nice one-volume synthesis of the physical and impacts science behind climate change alongside possible policy responses. Chapters solicited from a wide range of experts.
• Schellnhuber reference discusses the state of science and some policy thinking regarding our ability to meet the aims of the UNFCCC and avoid dangerous climate change. Provides a nice selection of papers and think pieces by many of the leading commentators in the field.
• Brooks. Something of an underappreciated gem, this paper combines the intriguing uncertainty of palaeoclimate research with a fascinating – if necessarily speculative – discussion of the developments of early civilizations around the globe. An interesting scientific and social look at how social responses to climate change can succeed or fail.
Andronova, N., Schlesinger, M., Hulme, M., Desai, S., and Li, B. (2006). The concept of climate sensitivity: history and development. In Schlesinger, M.E., Kheshgi, H.S., Joel Smith, dela Chesnaye, F.C., Reilly, J.R., Wilson, T, Kolstad, C., (eds.), Human-Induced Climate Change: An Interdisciplinary Assessment (Cambridge: Cambridge University Press).
Benton, M.J. and Twitchett, R.J. (2003). How to kill (almost) all life: the end-Permian extinction event. Trend Ecol. Evol., 18(7), 358–365.
Brooks, N. (2006). Cultural responses to aridity in the Middle Holocene and increased social complexity. Quat. Int., 151, 29–49.
Claussen, M., Mysak, L.A., Weaver, A.J., Crucifix, M., Fichefet, T., Loutre, M.-F., Weber, S.L., Alcamo, J., Alexeev, V.A., Berger, A., Calov, R., Ganopolski, A., Goosse, H., Lohman, G., Lunkeit, F., Mokhov, Petoukhov, V., Stone, P., and Wang, Zh. (2002). Earth system models of intermediate complexity: closing the gap in the spectrum of climate models. Clim. Dynam., 18, 579–586.
Cox, P.M., Richard A. Betts, Chris D. Jones, Steven A. Spall and Ian J. Totterdell (9 November, 2000). Acceleration of global warming due to carbon cycle feedbacks in a coupled climate model. Nature, 408(6809), 184–187.
Cox, P.M., Betts, R.A., Collins, M., Harris, P.P., Huntingford, C., and Jones, C.D. (2004). Amazonian forest dieback under climate-carbon cycle projections for the 21st century. Theor. Appl. Climatol., 78, 137–156.
Crowley, T.J. (2000). Causes of climate change over the past 1000 years. Science, 289, 270–277.
Crucifix, M. (2006). Does the last glacial maximum constrain climate sensitivity? Geophys. Res. Lett., 33(18), L18701-L18705.
Crutzen, P.J. (2006). Albedo enhancement by stratospheric sulphur injections: a contribution to resolve a policy dilemma? Clim. Change, 77, 211–220.
Dallimore, S.R. and Collett, T.S. (1995). Intrapermafrost gas hydrates from a deep corehole in the Mackenzie Delta, Northwest-Territories, Canada. Geology, 23(6), 527–530, 1995.
Diamond, J. (2005). Collapse (Penguin).
Flannery, T. (1994) The Future Eaters (Grove Press).
Forest, C.E., Stone, P.H., Sokolov, A.P., Allen, M.R., and Webster, M.D. (2002). Quantifying uncertainties in climate system properties with the use of recent climate observations. Science, 295, 113–117.
Govindasamy, B. and Caldeira, K. (2000). Geoengineering Earth’s Radiation Balance to Mitigate CO2-Induced Climate Change. Geophys. Res. Lett., 27, 2141–2144.
Hansen, J., Russell, G., Lacis, A., Fung, I., and Rind, D. (1985). Climate response times: dependence on climate sensitivity and ocean mixing. Science, 229, 857–859.
Hegerl, G.C., Zwiers, F.W., Braconnot, P., Gillett, N.P., Luo, Y., Marengo Orsini, J.A., Nicholls, N., Penner, J.E., and Stott, P.A. (2007): Understanding and Attributing Climate Change. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., and Miller, H.L. (eds.)], Cambridge University Press, United Kingdom and New York, NY, USA.
Hegerl, G.C., Crowley, T.J., Hyde, W.T. and Frame, D.J. (2006). Climate sensitivity constrained by temperature reconstructions over the past seven centuries. Nature, 440, 1029–1032.
Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K., and Johnson, C.A., (eds.). (2001). Climate Change 2001: The Science of Climate Change (Cambridge: Cambridge University Press).
Hyde, W.T. and Crowley, T.J. (2000). Probability of future climatically significant volcanic eruptions. J. Clim., 13, 1445–1450.
Keith, D.W. (2007). Engineering the planet. In Schneider, S. and Mastrandrea, M. (eds.), Climate Change Science and Policy (Island Press).
Keith, D.W. (2001). Geoengineering. Nature, 409, 420.
Kiehl, J.T. (2006). Geoengineering climate change: treating the symptom over the cause? Clim. Change, 77, 227–228. Kemp, M. (2005). Science in culture: inventing an icon. Nature, 437, 1238.
Keller, K., Hall, M., Kim, S., Bradford, D.F., and Oppenheimer, M. (2005). Avoiding dangerous anthropogenic interference with the climate system. Clim. Change, 73, 227–238.
Kennett, J.P., Cannariato, K.G., Hendy, I.L., and Behl, R.J. (2000). Carbonisotopic evidence for methane hydrate instability during Quaternary interstadials. Science, 288, 128–133.
Levitus, S., Antonov, J., and Boyer, T. (2005). Warming of the world ocean, 1955–2003. Geophys. Res. Lett., 32, L02604.
Lenton, T.M. et al. (2006). The GENIE team, Millennial timescale carbon cycle and climate change inanefficient Earth system model. Clim. Dynam., 26, 687–711.
Lenton, T.M., Held, H., Hall, J.W., Kriegler, E., Ludd, W., Rahmstorf, S., and Schellnhuber, H.J. (2007). Tipping Elements in the Earth System. Proc. Natl. Acad. Sci., 1.
Lindzen, R.S., Chou, M.-D., and Hou, A.Y. (2001). Does the Earth have an adaptive iris? Bull. Am. Meteorol. Soc., 82, 417–432.
Lockwood, M. and Frôhlich, C. (2007). Recent oppositely directed trends in solar climate forcings and the global mean surface air temperature. Proc. R. Soc. A., 10.1098/rspa.2007.1880.
Maslin, M., Owen, M., Day, S., and Long, D. (2004). Linking continental-slope failures and climate change: testing the clathrate gun hypothesis. Geology, 32, 53–56.
Mason, B.G., Pyle, D.M., and Oppenheimer, C. (2004). The size and frequency of the largest explosive eruptions on Earth. Bull. Volcanol., 66(8), 735–748.
Mastrandrea, M. and Schneider, S. (2004). Probabilistic integrated assessment of “dangerous” climate change. Science, 304, 571–575.
Mastrandrea, M. and Schneider, S. (2005). Probabilistic assessment of dangerous climate change and emission scenarios. In Schellnhuber, H.-J. et al. (eds.), Avoiding Dangerous Climate Change (DEFRA).
Meehl, G.A., Stocker, T.F., Collins, W., Friedlingstein, P., Gaye, A., Gregory, J., Kitoh, A., Knutti, R., Murphy, J., Noda, A., Raper, S., Watterson, I., Weaver, A., and Zhao, Z.-C. (2007). Global climate projections Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, S. Solomon, D. Qin, and M. Manning, Eds., Cambridge University Press.
Meinshausen, M. (2005). On the risk of overshooting 2° C. In Schelln huber, H.-J. et al. (eds.), Avoiding Dangerous Climate Change (DEFRA).
Murphy, J., David M.H. Sexton, David N. Barnett, Gareth S. Jones, Mark J. Webb, Matthew Collins &David A. Stainforth, (2004). Quantification of modelling uncertainties in a large ensemble of climate change simulations. Nature, 430, 768–772.
Nordhaus, W.D. (1994). Managing the Global Commons. The Economics of Climate Change, MIT Press, Cambridge, MA.
Nordhaus, W. (2005). Life After Kyoto: Alternative Approaches to Global Warming Policies, National Bureau of Economic Research, Working Paper 11889.
O’Neill, B.C. and Oppenheimer, M. (2004). Climate change impacts are sensitive to the concentration stabilization path. Proc. Natl. Acad. Sci., 101, 16411–16416.
Oppenheimer, M. (2005). Defining dangerous anthropogenic interference: the role of science, the limits of science. Risk Anal., 25, 1–9.
Orr, J.C., Victoria J. Fabry, Olivier Aumont, Laurent Bopp, Scott C. Doney, Richard A. Feely, Anand Gnanadesikan, Nicolas Gruber, Akio Ishida, Fortunat Joos, Robert M. Key, Keith Lindsay, Ernst Maier-Reimer, Richard Matear, Patrick Monfray, Anne Mouchet, Raymond G. Najjar, Gian-Kasper Plattner, Keith B. Rodgers, Christopher L. Sabine, Jorge L. Sarmiento, Reiner Schlitzer, Richard D. Slater, Ian J. Totterdell, Marie-France Weirig, Yasuhiro Yamanaka and Andrew Yool, (2005). Anthropogenic ocean acidification over the twenty-first century and its impact on calcifying organisms. Nature, 437, 681–686.
Pinto, J.P., Turco, R.P., and Toon, O.B. (1989) Self-limiting physical and chemical effects in volcanic eruption clouds. J. Geophys. Res., 94, 11165–11174.
Pizer, W.A. (2003). Climate change catastrophes. Discussion Paper 03–31, Resources for the Future, Washington, DC.
Roughgarden, T. and Schneider, S.H. (1997). Climate change policy: quantifying uncertainties for damage and optimal carbon taxes. Energy Policy, 27(7), 371 -434.
Sazonova, T.S., Romanovsky, V.E., Walsh, J.E., and Sergueev, D.O. (2004). Permafrost dynamics in the 20th and 21st centuries along the East Siberian transect. J. Geophys. Res.-Atmos., 109, D1.
Schellnhuber, H.-J. (ed.). (2006). Avoiding Dangerous Climate Change (Cambridge, UK: Cambridge University Press).
Schneider von Deimling, T., Held, H., Ganopolski, A., and Rahmstorf, S. (2006). Climate sensitivity estimated from ensemble simulations of glacial climate. Clim. Dynam., 27, 149.
Solomon, S. et al. (2007). IPCC: Summary for Policymakers. In Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., and Miller, H.L. (eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (Cambridge, UK: Cambridge University Press and New York, NY, USA).
Weitzman, M.L., (1974). Prices vs. Quantities in: Review of Economic Studies, 41(4), 477–491.
Zickfeld, K., Levermann, A., Morgan, M.G., Kuhlbrodt, T., Rahmstorf, S., and Keith, D. Expert judgements on the response of the Atlantic meridional overturning circulation to climate change. Clim. Change, accepted.
