1. introduction and context
In global debates on climate change policy, the familiar dichotomy of two types of response – mitigation (reducing emissions of carbon dioxide (CO2) and other greenhouse gases (GHGs) that are causing climate change) and adaptation (reducing the harmful impacts of realized climate changes) – is being disrupted by the appearance of a third form of response: climate engineering. Climate engineering (CE), also called geoengineering, consists of intentional, engineered measures to actively change the global climate system and so reduce the realized climate changes that result from elevated GHGs.Footnote 1 CE is not a new idea: it was first proposed as a response to anthropogenic climate change in the 1960s, and has been mentioned in multiple assessments over subsequent decades.Footnote 2 But CE has reappeared in policy debates over the past several years, triggered by several sources of concern – including the growth of evident climate change impacts, the continued failure of mitigation efforts, and continuing scientific uncertainties that suggest even a shift to extreme mitigation would only reduce, not eliminate, the risks of severe climate change impacts. CE is highly controversial, and is not yet being explicitly addressed in international climate negotiations, but probably soon will be, and should be, on the international policy agenda. In its effect on climate response and policy debates, CE is a disruptive technology, presenting risks and opportunities that are large, novel, and deeply challenging to international law and governance.
Other contributions have started to investigate the general governance challenges posed by CE.Footnote 3 In this article, I address the previously unexamined question of how CE will affect issues related to the configuration of international cooperation on climate change. To date, the main focus of both diplomatic efforts and academic studies has been full global cooperation. While this focus makes sense given the global scale of the causes and consequences of climate change, alternative configurations of cooperation and participation must be considered if full global cooperation is not achievable, or not achievable initially – as the failure of more than two decades of diplomatic effort thus far suggests.
What specific alternatives to global cooperation must be considered? The theme of the conference from which this paper originates, ‘Global Climate Change without the United States’, examined one instance of a major class of alternatives: partial cooperation approaches in which some states take coordinated action but others stand aside. Such partial cooperation alternatives, including the specific case of cooperation without the United States (US), have been periodically discussed since shortly after the emergence of climate on policy agendas in the late 1980s.Footnote 4 They are the main alternatives to global cooperation that must be considered if climate policy is taken exclusively or predominantly to mean mitigation, but the situation changes with a broader policy agenda that includes CE. With this broader agenda, global cooperation still commands attention as a preferred approach. But if it is not attainable and alternatives must be considered, this broader substantive policy agenda requires consideration of a different set of alternative configurations of participation and non-participation. This article makes a preliminary, admittedly speculative, examination of these issues:
• How does the addition of CE to the climate policy agenda change salient configurations of participation and non-participation?
• How might these configurations develop, and what novel risks or opportunities do they present?
• What priorities for research and analysis follow from this new perspective?
Section 2 introduces the major technical approaches to CE, and outlines the three basic characteristics that shape the nature and severity of the challenges they pose to international law and governance. Section 3 examines the effects of CE on questions of participation, initially treating CE as separate from other elements of climate response. Under this rather artificial assumption, the most prominent issues concern the potential for, the risks of, and control of the unilateral pursuit of CE by major states. Section 4 considers CE in the context of a complete response to climate change, focusing on potential ways to build constructive bargaining linkages between CE and mitigation. It proposes four speculative linkage scenarios by which CE might enhance, rather than undermine, mitigation incentives. Section 5 draws tentative conclusions and identifies research priorities suggested by this preliminary investigation.
2. climate engineering technologies and their policy – relevant characteristics
Climate engineering (CE) entails interventions that modify global-scale properties of the Earth’s environment in order to counteract the heating and climate disruption caused by elevated atmospheric concentrations of GHGs.Footnote 5 Many specific forms of CE intervention have been proposed, which fall into two broad classes: (i) interventions in the global carbon cycle that reduce the atmospheric concentration of CO2; and (ii) interventions in the Earth’s radiation balance that reduce the amount of sunlight absorbed at the Earth’s surface, thereby offsetting the aggregate heating caused by elevated GHGs.
The approach that now seems the most promising and is receiving most attention is stratospheric aerosol injection: spraying a fine mist of lightly coloured or reflective particles (sulphate aerosols, for example) into the stratosphere. Viewed from Earth, this would make the sun appear a little dimmer (by about 1%), and the sky a little brighter and whiter. Although research may identify other approaches that are preferred, stratospheric aerosol injection has certain characteristics that clearly illustrate the policy and strategic challenges likely to be posed by any radiation-based CE. Its underlying scientific principles are well understood, as are the basic engineering approaches by which it would be implemented. Consequently, it could be done today, albeit crudely, with current knowledge and technology. Nature provides clear analogues for how such interventions would work in the occasional explosive volcanic eruptions that inject large quantities of sulphur into the stratosphere – most recently the 1991 eruption of Mount Pinatubo in the Philippines, which cooled the Earth by about half a degree Celsius over the following year or two.Footnote 6
Research is needed to study the many uncertainties about how specific CE interventions would work, their effects and risks – including, crucially, the regional and seasonal distribution of effects. Preliminary studies of these issues are under way – mostly laboratory and computer model studies, but there are also a few small field experiments of atmospheric aerosols and other proposed approaches, such as ocean fertilization. Early efforts to create explicit research programmes are also under way in a few jurisdictions, as are various ‘dual-use’ studies to investigate CE capabilities and effects, but which equally address other scientific questions. Since much of the field research to develop and inform CE capabilities can be carried out with small-scale interventions that are essentially riskless – indeed, many proposed experiments would resemble existing projects in small-scale weather modification, or the inadvertent impacts of normal commercial activities such as aviation and shipping – small-scale CE research would be hard to detect from a distance; it is therefore possible that other experimental interventions have already been undertaken.Footnote 7
For the purposes of understanding their role in societal responses to climate change, CE technologies have three salient characteristics: they are fast, cheap, and imperfect.Footnote 8 Climate engineering is fast. A manageable scale of intervention by means already known, involving one or two hundred transport aircraft in continuous operation, could cool the Earth by 1–2°C within a few years.Footnote 9 Consequently, an effective intervention could be deployed to arrest or reverse global heating even after it was known that rapid change or severe impacts were under way. Radiation-based CE is the only known response capable of such rapid effect: achieving a similar effect through even an extreme programme of emissions cuts, or by removing CO2 from the atmosphere, would take decades. This capability for rapid action is the principal way, although not the only way, in which CE offers a large expansion in human capability to limit the risks of climate change.
Climate engineering is cheap. Estimates of the direct cost of offsetting projected 21st century global-average heating are in the order of a few billion dollars per year,Footnote 10 and are likely to decrease with further research into and development of approaches. Various commentators have proposed that, for considering the strategic implications of these technologies, it is a useful approximation to consider their cost as zero.Footnote 11 While normally it is an advantage if a potentially desired option is cheap, in this case low cost is a double-edged sword, with two potentially destructive consequences. Firstly, it has deluded some observers into a stance of naïve cheerleading for the technologies.Footnote 12 This, in turn, has raised concerns about excessive reliance on CE as a complete response to climate change – which it emphatically cannot be, for reasons noted below – further weakening the already inadequate support for cutting emissions. Secondly, the low cost of CE raises problems of control by putting it within reach of more actors. Although I argue below that the prospects of unilateral CE by small states or non-state actors have been overstated, CE is still more widely available than past examples of potentially destabilizing technologies, of which the most relevant parallels are novel weapons capabilities.
Finally, CE offers only a highly imperfect corrective for the environmental effects of elevated GHGs. Their correction is imperfect even if only their global-average climate effect is considered, because CE counteracts a heating that occurs aloft by a cooling at the Earth’s surface, where the blocked sunlight would otherwise have been absorbed. The result is that CE controls precipitation more strongly than temperature, so a world in which CE fully offsets average greenhouse heating would have a climate drier than the starting climate.Footnote 13 These global average differences cascade to diverse, albeit uncertain, differences in regional and seasonal climate effects.Footnote 14 In addition, CE does nothing to counteract the non-climate (that is, the chemical and biological) effects of elevated CO2, which include making the oceans more acidic, and disrupting competitive relationships between different types of plants with different responses to increased CO2.Footnote 15
These three characteristics – fast, cheap, and imperfect – outline the basic governance and policy challenges posed by CE. Considered together, they present an acute tension: in common with all technological expansions of human capabilities, CE may offer the prospect of either large benefits – reducing the climate change risks we otherwise face – or large harms, depending on how it is used and how it influences related choices. Used prudently and benevolently, it may bring large benefits in multiple forms. It can provide a contingency response to a future climate emergency, as discussed above; it can also be used earlier and less intensely, to shave the peak off projected near-term heating while a serious mitigation effort is ramped up, thereby reducing the cost of a global transition to climate-safe energy sources; or it can be targeted to reduce specific high-priority regional or seasonal risks, such as cooling Arctic summers to slow the loss of sea ice, or cooling tropical oceans to block formation of the highest-energy hurricanes.Footnote 16 But used incompetently, negligently, or destructively, CE technologies may make matters much worse. They thus present new needs, and new challenges, for governance and control, to pursue the benefits and minimize the harms they hold.
3. unilateralism and multilateralism in climate engineering
When CE is added to the set of potential responses to climate change, the aspiration for global cooperation still exerts powerful attraction, perhaps even more so than when policy is just mitigation. Early discussions suggest that every group that takes the prospect of CE seriously asserts the importance of broad consultation and participation in decision-making.Footnote 17 But if global cooperation appears unattainable, CE requires consideration of a different set of alternative configurations of participation and non-participation than is the case when climate policy is simply mitigation. Partial cooperation approaches are still relevant, but various configurations of unilateral action, by the US or other states, must also be considered. So, too, must scenarios of involuntary non-participation by, or the exclusion of, some states. This section begins to explore these possibilities, initially and somewhat artificially treating governance of CE as separate from other elements of climate policy. The next section adds more realism by considering complete climate responses that include CE with linkages with other response elements, particularly with mitigation. The discussion is unavoidably speculative, but it aims to use the speculation to identify key uncertainties that require investigation, and to discipline the speculation by anchoring it to current knowledge, particularly about characteristics of CE technologies relevant to state capabilities and interests.
Considering CE governance separately from other elements of climate policy, the most prominent alternative to global cooperation that must be considered is unilateral action. This section considers unilateral action from the perspectives of international law, state capabilities, and state interests.
3.1. Current International Law and Climate Engineering
Present international law imposes virtually no control on any state’s conduct of most forms of CE, whether conducted for the purposes of research or operational climate modification. Multiple regimes are relevant but none meaningfully constrain CE, with the result that any state may legally conduct CE, on or over its own territory, or that of other consenting states, or over the high seas.Footnote 18
The reasons for this lack of legal control are unique to each treaty and institution, but generally lie in the narrowness and specificity of obligations imposed by environmental treaties.Footnote 19 The regimes of greatest relevance are those on stratospheric ozone depletion, climate change, and long-range air pollution. Yet the concrete obligations of the Montreal Protocol on the ozone layer are limited to controls on the production and consumption of listed chemicals, and do not include comprehensive controls on other activities that affect ozone.Footnote 20 Similarly, the Kyoto Protocol on climate change only limits national emissions of six listed GHGs, and only for Parties listed in Annex B.Footnote 21 None of the sulphur-based species now considered promising candidates for stratospheric aerosol injection appear on the list of controlled substances in either of these treaties. National emissions of sulphur dioxide (SO2) are controlled under the 1999 Gothenburg ProtocolFootnote 22 to the Convention on Long-Range Transboundary Air Pollution.Footnote 23 But this Convention is a regional treaty the membership of which includes only European nations plus the US and Canada, and the way in which the 1999 Protocol specifies national emissions limits appears likely to seriously constrain participation in a CE programme for only the smaller European states.Footnote 24 Another treaty of apparent relevance, the 1977 Environmental Modification Convention (ENMOD), prohibits large-scale environmental modification, but only if undertaken for military or other hostile purposes, and includes an explicit exemption for activities carried out for peaceful purposes.Footnote 25 The result is that none of these treaties impose concrete obligations that would be violated by proposed CE interventions.
Pushed by vigorous advocacy by a few non-governmental organizations (NGOs), two treaties have taken explicit steps to limit or discourage CE activities: (i) the London Convention and Protocol under the International Maritime Organization,Footnote 26 and (ii) the Convention on Biological Diversity (CBD).Footnote 27 Even these initiatives, however, currently impose no binding legal restrictions on CE. The initiatives in the CBD, despite claims by their NGO proponents that they comprise a moratorium, impose no binding controls.Footnote 28 At most, they express a generalized disapproval for CE, in language that is remarkable for its weakness, opacity, and multiple escape clauses.Footnote 29 Action within the London Convention and Protocol has been more focused, but is limited to ocean fertilization, a CE method that appears increasingly unlikely to be effective.Footnote 30 London Parties have asserted that ocean fertilization falls within the scope of these treaties and have expressed concern about its potential adverse impacts,Footnote 31 but have constructively decided that it does not comprise ‘dumping’ and so falls outside the Protocol’s general prohibition. Further, Parties have drawn an even stronger distinction between dumping and ‘legitimate scientific research’ into ocean fertilization, and have developed an ‘assessment framework’ to which such research should be subject – a rather generic set of procedures for environmental impact and risk assessment – and are now developing legal measures to implement this framework.Footnote 32 The upshot is that ocean fertilization is at present subject only to generalized normative statements of concern urging caution, not yet to any legally binding control, while other forms of CE, including stratospheric aerosol injection, are under even less international legal control. In the specific case of controlling US conduct, the legal situation is even weaker because the US is not a party to either the CBD or the London Protocol. Consequently, even if binding controls were adopted under one of these treaties, the US as a non-party would not be bound by them.Footnote 33
In the absence of specific treaty provisions that would constrain national CE activities, the points of existing international law of potential relevance to CE fall into two classes: (i) general obligations to protect and preserve the environment that appear in many treaties, such as the Vienna Convention for the Protection of the Ozone LayerFootnote 34 and the United Nations (UN) Convention on the Law of the Sea (UNCLOS);Footnote 35 and (ii) relevant principles of customary international law, such as the duty to avoid transboundary harm.Footnote 36 Although any of these could be elaborated or interpreted to apply to CE, they lack the specificity to provide operational guidance on what CE interventions, and under what conditions, would be permissible or impermissible – particularly in view of the tension between the potential of CE both to reduce climate change risks and to introduce new risks. Other provisions of customary international law, such as the duty to undertake environmental impact assessment, would not limit CE itself but may create procedural obligations related to how it is conducted.Footnote 37
3.2. Distribution of State Capabilities
The present lack of any controlling international law, however, does not necessarily imply a serious threat of unilateral action to develop or deploy CE technologies. The severity of this risk will depend additionally, indeed primarily, on the distribution of relevant state capabilities and interests. Focusing on these, one common way to express the strategic novelty and challenge of CE has been to contrast its basic structure to that of cutting emissions. Cutting emissions is generally understood as a problem of collective action in which the basic strategic challenge is to motivate and enforce costly contributions to a shared goal, while for CE the basic problem is to bring a widely distributed capability under competent and legitimate collective control. One recent discussion used the vivid ‘free-rider versus free-driver’ image to illustrate this distinction: for effective global policy, the basic problem of emissions control is to overcome free-rider incentives, while the basic problem of CE is to corral multiple potential drivers, each able to act alone, into a collective decision process.Footnote 38
Taken to an extreme, this logic would suggest that virtually anyone can do CE – as has been proposed in various colourful scenarios of CE conducted by terrorist groups, apocalyptic cults, or wealthy individuals.Footnote 39 But these scenarios overstate the distribution of capabilities and thus the risk of unilateral action, because they focus too narrowly on financial cost as the determinant of capability and neglect other, non-financial, requirements and constraints. To assess these other constraints, it is crucial to note that achieving a non-trivial, sustained alteration of global climate requires continued large-scale material inputs. These, in turn, depend upon delivery equipment and supporting infrastructure – for example, balloons, tethered pipes, aircraft or ships, backed up by airports, bases and ports – that is visible, hard to conceal, and vulnerable to military attack. This is not to claim that even powerful states would take such military action lightly, in view of the substantial associated costs and risks; yet such action will clearly be a feasible response for some states under some conditions, if they judge another state’s CE actions to threaten their vital interests and have been unable to stop it through other means.
In view of the possibility of such military interdiction, unilaterally achieving a climate alteration that matters would require not just the money, technological capability, and delivery assets, but also the command of territory, global stature, and ability to deploy and project force necessary to protect a continuing operation against opposition from other states, including deterring their threats of stopping it through military action. These requirements exclude the nightmare scenarios of climate alteration by megalomaniac billionaires, terrorist groups, or apocalyptic cults, and also exclude the prospect of unilateral action by most states. Rather, the capability is likely to be limited to a few major world powers, probably numbering fewer than a dozen. Precisely which states could act unilaterally is indeterminate, not just because the feasibility, the precise requirements, and the effects of particular CE interventions are uncertain, but also because who can do it (or, more precisely, who can do what) will also depend on the intensity of other states’ interests in who does what. The more strongly others care – in particular, the more other powerful states are intensely opposed – the fewer states will be able to conduct unilateral CE in the face of that opposition.
However, while the distribution of unilateral capability is narrower than the most apocalyptic commentaries have suggested, it is still broad enough to be significantly destabilizing. The potential for destabilization and conflict arises, on the one hand, from the virtual certainty that any intervention by one state strong enough to alter its own climate would also exert similarly large effects on other states and, on the other hand, from the presumption that no state would undertake a major CE programme lightly. That is, in the event that any state seriously pursues a CE programme, its leaders and polity must perceive the intervention as a matter of high-order national interest because of the realized and impending harm from climate change they see and their expectation that CE can reduce these. Given the likely inability to limit the effects of a CE intervention to one state, if the state making the intervention perceives such acute interest then so also will other states.
3.3. Distribution of State Interests
Characterizing the severity of the resultant risks of international conflict requires a closer examination of states’ likely interests, in particular the degree to which the interests of major states over the available set of CE choices are aligned or opposed. State interests in CE will depend on (i) what specific capabilities are available, including how controllable they are; (ii) the specific projected regional effects of available capabilities; and (iii) how these effects are expected to interact with ongoing greenhouse heating and natural climate variability. Although these factors are uncertain and likely to change with further research, the range of possibilities can instructively be clustered into three alternative degrees of interest alignment.
At one extreme, the interests of states over available CE options might be closely aligned. This situation would be most likely to arise under three conditions:
• Experienced and anticipated climate change harms are widely distributed worldwide, so states perceive a broadly shared peril.
• The projected effect of CE is to limit these harms in a manner that is also roughly consistent across world regions. This would require that even the imperfect joint correction of temperature and precipitation noted above is roughly consistent across regions, facilitating agreement among states on some preferred compromise between restoring prior precipitation and prior temperature.
• Crucially, the available CE capabilities remain rather crude, with little ability to control effects beyond choosing the aggregate intensity of intervention. In particular, there is no ability to tune interventions to achieve differential control of climate effects in different regions.
At the opposite extreme from this first possibility, states’ interests in CE might be strongly opposed. The conditions favouring this situation would be the opposite of those above – that is, the severity of climate disruptions varies strongly among regions, so some states perceive a climate change crisis requiring urgent response while others do not. The anticipated effects of CE in limiting harmful climate impacts also show large regional differences, in their effects or how they are valued, so states disagree over whether and, if so, how to use CE even if they agree that a climate crisis is occurring. Finally, scientific and technological advances have brought a strong ability to control regional effects of CE interventions. In the extreme, plausible advances in information technology and nanotechnology over several decades might lead to individually controllable, optimally asymmetric stratospheric particles, allowing some degree of real-time, regional controllability of climate and weather.Footnote 40 Like CE itself, such extreme advances in regional control would present sharp double-edged possibilities. They would greatly advance the ability to reduce harm from climate change and manage climate and weather for global benefit. However, precisely the same advances would enable control of CE to distribute large regional benefits and harm on command, even on a time-scale of days or hours, starkly raising the stakes in how the capability is used and who controls it.Footnote 41 If states even suspected others of pursuing such capabilities, or of withholding information about capabilities and effects, opportunities for international mistrust and tension would be substantial.
Perhaps the most likely is the intermediate possibility in which states’ interests are mixed and variable, aligned for some choices and opposed for others. Early model studies have already cast doubt on the strongest commonality of interests from shared benefits of CE. Even if the only available dimension of CE control is the aggregate intensity of an intervention, it appears that different regions most closely approach their prior climate at different levels of intervention.Footnote 42 As interventions come to vary in respect of more dimensions – they have more dials to turn – the possibility of trade-offs among different regional harm-reduction objectives increases.Footnote 43 Current early study of these questions is highly tentative, of course: it has only identified the possibility of trade-offs, not characterized the intensity or degree of opposition of the resultant interests, which will depend on currently unknown future advances in capabilities. Advancing knowledge may perhaps reveal large-scale regional patterns in joint effects of climate change and CE that create foreseeable common or opposed interests, for example, by latitude band (low-latitude versus mid-latitude countries) or position relative to major mountain ranges (China versus India, for example, over how precipitation and storms are distributed over the Himalayas). Although the details of these interests are unknown, it appears generally likely that increased CE controllability, including more dimensions for control of interventions, will increase the potential for opposing interests.Footnote 44
Moreover, the discussion thus far may understate the prospects for opposition, because it assumes some rational process of forming nationally aggregated interests, based on realized or projected climate effects, with each region viewing its recent climate as ideal; but any of these assumptions might not hold. State interests could be driven by smaller-scale patchiness of climate effects within countries and resultant domestic political conflict. Alternatively, climate preferences might shift in response to realized climate changes or to recognition of the possibility of intentional climate control, such that a region’s present climate is no longer judged to be ideal. State interests in CE might also be dominated by non-consequential or non-rational processes, such as religious or symbolic commitments, general technological optimism or pessimism, or generalized suspicion about other states’ intentions. To the extent that these other processes show strong regional variation, they could further increase the possibility of interstate conflict over CE.
From this sketch of potential state capabilities and interests in CE, two large-scale implications can be drawn about unilateralism in CE. On the one hand, major powers such as the US are likely to face significant temptations to unilateralism – that is, to develop CE capabilities unilaterally; to conceal information about plans, research results and capabilities; and to act diplomatically to preserve a unilateral right of action. On the other hand, such unilateral actions are likely to be dangerous and disruptive to international stability.
Temptations to unilateralism may arise from several factors. The scientific and technical challenges of doing CE well – developing high-benefit, low-risk interventions – are sufficiently large that rich, scientifically advanced nations are likely to have substantial advantages in developing them. Scientific and government elites in such nations may be confident of these advantages, and may also be confident – perhaps over-confident – of their ability to persuade others to their view of CE. Temptations to unilateralism may be exacerbated by anticipation of economic benefits if CE research produces private intellectual property. They may also be exacerbated by the polarization of early debates on CE governance, in which widespread hostility to CE and calls for bans may lead those who favour developing CE capability to judge that doing it unilaterally may be the best way to ensure that it gets done. These temptations are already evident in US policy debate, both in a few explicit calls to preserve US freedom of action and in more widespread scepticism about international consultation over early-stage research.Footnote 45
Yet unilateral pursuit of CE is likely to carry serious risks, which also follow from the same observations about the likely distribution of state capabilities and interests. The ability to develop CE capability, and even to deploy it, will not be limited to the US or to any single state. Other world powers can do it, possibly just as well; and even if some leading state achieves a technological breakthrough – for example, with an approach that is cheaper, safer, or more controllable – less advanced approaches can make similarly large climate perturbations, albeit more crudely. Other states can also assert the same legal arguments for a unilateral right of action. Indeed, states with programmes of regional weather modification may be favoured in advancing these arguments because of the blurry line between these activities, which clearly lie within their sovereign authority, and early CE development. With both capabilities and potential justifications broadly distributed, at least among major powers, unilateral pursuit of CE by any world power, including the US, would risk others deciding to do the same. Once any major power decides to pursue this course, attempting to stop them would be difficult and risky.
Moreover, states are likely to perceive strong interests in whether and how other states pursue CE, not just at the deployment stage but also from early unilateral steps to develop capabilities that might make future deployment more likely. As discussed above, the severity of these risks will depend on how states’ future interests in CE are aligned or opposed. But, given the current uncertainties about CE capabilities and effects, these interests might be subject to some degree of influence. In particular, states’ perceived interests may in part form reactively in response to early acts by other states that signal either anticipated rivalry or cooperation over CE. Thus, early unilateral acts by a major state – including development of capabilities, secrecy about intentions, or aggressive declaration of rights of action – may induce others to perceive CE as predominantly rivalrous and to pursue similar acts, either because they interpret these acts as indicating hostile or rivalrous intent or because they infer from these acts that it is valuable to have an independent CE capability. Conversely, early signals of cooperation and openness may have the opposite effect, steering others’ perceptions and choices towards cooperation. Given the uncertain and labile nature of future CE capabilities, such cooperative early moves may even influence the direction in which future capabilities are developed towards those that pose less risk of conflict.
In sum, following a unilateral approach to CE – including not just eventual deployment but also early steps to pursue research and development alone, maintain secrecy about capabilities and results, and reserve unilateral legal rights – is a superficially tempting but dangerous course of action, for the US and other major powers. States should anticipate and resist these temptations and instead pursue a cooperative approach to CE. Such an approach could start immediately, with informal consultations on research programmes, agreement on common standards for transparency, and joint development of assessment frameworks.Footnote 46 A cooperative approach need not involve universal participation, but could start with only the dozen or so nations likely to be most interested in developing CE and most able to pursue it unilaterally. It also need not await a comprehensive climate regime. By building cooperation and transparency on CE while the stakes are relatively low, such early cooperation may help to build norms for cooperative management of CE, which would then be available to help in resolving the more challenging governance problems raised by future proposals for operational interventions.
4. climate engineering within an integrated climate policy: participation, unilateralism, and exclusion
The discussion thus far has considered CE and its implications for global climate cooperation separately from other elements of climate policy. This approach is consistent with the present literature on CE governance, which has given limited consideration to interactions with other elements of climate policy. There has yet been no examination of interactions or trade-offs between CE and adaptation, although these may represent important future decisions over alternative ways to reduce harm from climate impacts on relatively fast time-scales. Discussions of CE mitigation interactions have been more extensive, but thus far fall into two classes: (i) analyses of how CE interacts with mitigation in a global dynamically optimal climate response, neglecting all politics and negotiation;Footnote 47 and (ii) discussions of the potentially destructive implicit interaction known as the ‘moral hazard’ effect of CE, whereby its perceived availability may undermine already inadequate political support for needed mitigation.Footnote 48
But if and when CE comes onto the policy agenda, there are likely to be large and explicit interactions with other elements of climate response, including attempts by policy-makers to link them in ways that favour their objectives. Such intentional linkages may expand the space for effective responses. In this section, I begin to consider specific possibilities for interaction and linkage, focusing in particular on how these linkages may affect incentives for, and consequences of, less than full participation. While interactions of CE with adaptation may be important, particularly for future responses to realized or impending climate changes, in this preliminary exploration I consider only potential linkages between CE and mitigation. The discussion is based on the strategically relevant characteristics of CE discussed above, but is otherwise entirely speculative.
Key decisions on CE will be made by states – most likely through some negotiation process in which the strongest influence is exercised by the dozen or so states with credible capacity to act unilaterally, even if the group nominally participating is larger. These states are roughly the same group of major economies that account for the bulk of global emissions, among whom it is often suggested that a serious agreement on emissions cuts might also be crafted.Footnote 49 Several such forums of major states have been proposed, with membership ranging from the G8+5 at the smaller end, through the Major Economies Forum, to the G20 at the larger end.Footnote 50 In the event that such a group of major states emerges to organize a decision-making forum for both mitigation and CE policy, this group would have the capability to explicitly link decisions on the two issues.
The most prominent near-term question concerning these linkages is how (and whether) decisions about CE can be linked with emissions cuts so as to make the two approaches complementary, rather than acting as substitutes or competitors, as has most often been presumed. Although this question has been previously identified, little progress has yet been achieved beyond posing it and noting its importance.Footnote 51 This section attempts to sharpen the question and advance its investigation, by proposing four alternative speculative scenarios of how mitigation and CE could be linked, with a preliminary discussion of the plausibility, likely effects, and challenges of each scenario.
4.1. Linking Climate Engineering with Mitigation: Four Possible Forms of Linkage
The first scenario, Plan B Linkage, represents the simplest, most minimal, form of linkage between CE and mitigation. Effectively, it makes explicit the relationship often presumed to exist between the two responses even in the absence of intentional linkage. In this scenario, states keep trying to pursue serious emissions cuts. At the same time, they conduct research to develop CE capability for future interventions and agree that these will be used if and as needed to limit future severe climate changes and impacts.
For this minimal form of linkage to promote an effective climate response that includes increased near-term mitigation efforts, it must present future CE deployment as a threat – an outcome so abhorrent that it motivates increased mitigation efforts to avoid it. In effect, the scenario presumes that the normal ‘moral hazard’ concern about CE can be reversed, even with no specific measures to change incentives. This presumption seems odd if we assume rationality in future decision-making, since any future decision to deploy CE would presumably be made only if CE promised to reduce climate harms otherwise anticipated – when the prospect of these worse climate harms at present is failing to provide adequate motivation for mitigation.
This scenario is not completely implausible, however, and could come about under various assumptions, related to uncertain CE effects or non-rational decision-making. For example, future CE use could be perceived as a gamble carrying the risk of outcomes worse than uncontrolled climate change. If future decision-makers regard CE as likely to improve matters on average, but have not learned enough to be fully confident that it will not increase harms, they might still favour deploying it as a desperate measure in the face of severe climate change. Looking ahead to this possibility, current decision-makers might be motivated to greater mitigation efforts to avoid this awful future choice. Alternatively, the prospect of deploying CE might somehow gain more saliency or mobilize more horror about the severity of human disruption of the global environment than severe climate change alone. At first glance, these eventualities appear barely plausible – suggesting that this scenario is unlikely to motivate much strengthening of near-term mitigation – but cannot be completely dismissed.
The second scenario, Reverse Linkage, would reverse the contingency relationship between mitigation and future CE use from that in the Plan B scenario. In this scenario, states would jointly agree to withhold CE, no matter how severe the climate impacts occurring or anticipated, unless states had achieved some agreed level of acceptable performance on cutting emissions. This scenario admittedly requires some suspension of disbelief, yet is still instructive to explore.
The linkage in this scenario would aim to motivate states to cut emissions early, to avoid the prospect of facing severe future climate change without access to CE to moderate the impacts. The most obvious difficulty with the scenario is credibility: how could a threat to refuse CE in response to some future climate emergency be credible? As preposterous as this may first appear, such refusal could be plausible in a political setting marked by intense domestic opposition to CE in the states otherwise able to deploy it. In such a setting, it would be difficult to achieve agreement to use CE under any conditions; but achieving this agreement would be easier if states had made serious contributions to reducing emissions, in part because moral hazard concerns would then be less severe.
In this regard, it is important to recall that future climate change risks come from two distinct routes – either continued failure to cut emissions, or unfavourable resolution of major uncertainties. In this scenario CE would be available to respond to a climate crisis arising from the second cause – unlucky resolution of uncertainties – because under this condition it would be easier to overcome general opposition to CE. CE would not be available, however, to respond to climate harms caused or confounded by failure to make agreed emissions cuts. Given sufficiently strong opposition to CE, the threat to withhold it under these conditions might be credible – or at least, credible enough to provide stronger motivation for near-term mitigation than the first scenario. And, of course, if it does motivate stronger mitigation, the credibility of the threat to refuse future CE in the event of inadequate mitigation would not be tested.Footnote 52
The scenario poses two further difficulties, however. Firstly, while one can construct plausible accounts in each case of how the required future threats could be made credible, both the first and second scenarios involve long time lags between the present mitigation decisions to be influenced, and the future use or non-use of CE that provide the incentive. Both thus suffer from the special credibility problems that afflict any attempt to link acts widely separated in time, as changes in conditions, actors, preferences, or capabilities can easily intervene to prevent the promised (or threatened) future actions. Second, both scenarios make future CE decisions depend on aggregate, collective mitigation performance in the interim. They thus suffer from the same free-rider, collective action problems that hinder current attempts at mitigation. Even if the inter-temporal linkage can be made credible in aggregate, each state will still have incentives to weaken its mitigation efforts in the hope that others do enough to trigger the desired future condition, as long as mitigation remains costly or difficult.
The third scenario, Real-time Linkage, aims to address the first of these problems, the inter-temporal disconnect, by linking actions on mitigation and CE concurrently rather than through future commitments. This scenario thus diverges from the most widely proposed way to use CE, as a response held in reserve for use in some future climate crisis. Instead, in this scenario CE would be used in one or both of the two proposed near-term modes: incremental use to shave the peak off near-term heating coupled with a phased programme of steep emissions cuts, with CE gradually phased out thereafter; or use targeting regional processes (such as Arctic summer sea ice loss or tropical hurricane formation) that are strongly linked to global climate risks.
In this scenario, participating states would simultaneously pursue agreed programmes of steep emissions cuts and limited CE deployment in one or both of these modes. Such real-time linkage of mitigation and CE could ease several strategic and political problems, making both responses more politically feasible and effective by coupling them. Linkage would make mitigation easier by addressing the distinct inter-temporal disconnect that has obstructed efforts thus far, which may be an even more severe obstacle than the collective action character of mitigation.Footnote 53 Whereas emissions cuts carry immediate costs to reduce climate risks decades in the future, introducing some small, modulated level of CE concurrently with mitigation would reduce climate risks in the near-term when they are politically salient. In addition to this clear political benefit, mitigation linked to concurrent CE could be less costly, because reductions could build gradually to match capital turnover cycles and allow the orderly development and roll-out of new technologies.
At the same time, real-time linkage could make CE less politically explosive, both because its deployment would be limited in intensity or spatial extent (albeit also performed earlier, when it is arguably not ‘needed’ to manage an imminent climate crisis); and, crucially, because the parallel enactment of mitigation and CE would address the strongest concern about harmful effects of CE, that it may undermine mitigation incentives. Moreover, concurrent linkage would enhance the credibility of nations’ mitigation commitments, because ongoing agreement and authorization to undertake CE – which states would presumably want to continue because of their real-time risk-reduction benefits – would depend on continuing mitigation efforts, with performance verifiable year by year. In sum, this scenario would link the two responses in a ‘both or neither’ political bargain, under which opponents of both mitigation and CE each tolerate the response they oppose because its scale, cost, and risks are limited by parallel pursuit of the response they favour.
Like the first two scenarios, however, this one does not specify by what process or subject to what inducements these collective decisions are made. By implicitly treating all action and consequences as collective, all three scenarios assume a return to some form of full cooperation – if not globally, then at least among the participating major powers. They thus fail to address many strategic and operational difficulties with making real deals work. Most importantly, they do not specify the linkage between the actions of individual states and the consequences that shape their incentives. Since the benefits of both mitigation and CE are likely to be distributed globally, these scenarios thus all remain vulnerable to the collective action challenges that have hindered cooperation on mitigation thus far. In any scenario that links collective CE decisions to aggregate mitigation performance, states will still have incentives to underperform on mitigation, so long as they expect their free-riding will not unravel the global agreement.
The fourth scenario, Pay to Play Linkage, aims to address this problem. This scenario follows the third in proposing real-time coupling of a strong, agreed trajectory of mitigation with a parallel programme of incremental or targeted CE. However, this scenario also provides individual incentives to deter free-riding, by making each state’s mitigation performance a condition for its participation in decision-making on CE. Under the dual assumptions that (i) all participating states strongly desire a voice in CE decisions, and (ii) the threat to exclude them from such participation is credible, this approach would address the problem of providing effective incentives for states to accept and meet strong mitigation commitments. To explore whether this scenario could be plausible, we must thus examine these two assumptions.
Firstly, is the threatened consequence of non-performance – exclusion from participation in CE – sufficiently painful that states will be motivated to pursue strong emissions cuts in order to avoid it? This will turn, in part, on what it means to be excluded. A narrow interpretation of exclusion might be that non-performing states and their citizens and enterprises may not participate in implementing the agreed CE programme. Such narrow exclusion would mainly target commercial interests – for example, firms that want CE contracts – which are unlikely to be large components of overall national interest, although they could still be effective political motivators if sufficiently concentrated. More broadly, exclusion might mean that non-performing governments are also barred from participating in decisions on the design and implementation of the CE programme – what is done, where, and how. Oddly, the extent to which such exclusion is disagreeable to states is likely to vary with the alignment of their interests in CE, but in a manner opposite to the discussion of interest alignment above. If states’ interests in CE are strongly aligned – at least over the range of choices being considered – then the cost to any state of being excluded from decision-making is likely to be low. In this case, other states are likely to implement essentially the same programme as they would if the excluded state was present. As states’ interests diverge – as discussed above, a function of regional variation in climate impacts, how CE limits these, CE capabilities and their controllability, and states’ beliefs and suspicions about these – the cost of being excluded from decisions, and thus the motivating power of the threat of exclusion, would grow.
If we stipulate that exclusion is sufficiently disagreeable to states to be motivating, is the threat of exclusion credible? This will depend strongly on which states are in the participating group, and which state is the presumed target of the threat. Clearly the threat to exclude grows more credible if the group of states collectively binding themselves is large and strong, and the state considering free-riding is small and weak. The credibility of the threat will also depend on the alignment of states’ interests in CE, now once again in the direction that opposed interests pose a greater challenge. If states’ interests in specific uses of CE are strongly opposed, then the threat to exclude is constructively a threat to design a CE programme that takes no account of the excluded state’s interests. Under some conditions, this might inflict significant harm on the excluded state and thus risk serious conflict, raising the possibility that the threatened act may be too severe for the threat to be credible. This problem might be eased by the incremental and real-time nature of the CE interventions being made in this scenario, which are less intense or less global in scale than those that would be deployed in response to some future climate emergency. Over this more limited decision space, regional disparity of interests may be attenuated, allowing a balance between the disagreeability of exclusion and the credibility of threatening it. Such intermediate degrees of interest alignment may thus be close to a politically optimal configuration of interests for CE to be able to motivate contribution to mitigation.
5. conclusion
The discussion in this article suggests that the issues associated with less than full global cooperation in climate policy, and the associated risks and opportunities, are likely to change markedly as the agenda of climate policy is broadened to include CE.
Including CE on the policy agenda raises significant risks of unilateralism, which will be particularly tempting for major powers. Even small early steps towards separate or secret research, or aggressive claims to reserve unilateral rights of action, may trigger reciprocal actions, perceptions, and expectations that hinder the growth of cooperative decision-making norms that may be crucial for reducing risks of conflict under severe climate change futures. Conversely, there is high value to early cooperative steps such as informal consultation, collaboration in the design and management of research programmes, and risk assessment, which may build a foundation for future cooperation over higher stakes governance questions when these arise. In this early research cooperation, it may be valuable to declare an explicit moratorium on the largest-scale interventions, to assuage public concerns about a thoughtless slide from small-scale research to global operational interventions. Such a moratorium need not obstruct necessary, early small-scale research but may rather facilitate it, because addressing these concerns may be a necessary condition for allowing small-scale, low-risk research to proceed.Footnote 54
Although these first steps on CE research need not consider explicit linkages with other elements of climate policy, longer-term decisions on CE – including any that begin to consider proposals or concerns about deployment – must be integrated with decisions on mitigation to contribute to an effective total climate response. This article has examined four scenarios by which negotiated international decision-making could aim to advance effective climate response by explicitly linking CE and mitigation decisions. Although highly preliminary and speculative, this exploration of linkage scenarios identifies some possible conditions and requirements for effective linkage, and also highlights specific areas of research priority that have received little attention thus far.
Firstly, the discussion suggests the importance of which specific states participate or exercise most influence in this linked decision-making. At a minimum, participation must include those states that could plausibly pursue CE unilaterally if they valued it highly enough. Because this is a similar group of states to those which are crucial for global mitigation negotiations, joint decisions on mitigation and CE in such a group would create the possibility for explicit linkage between the two types of response.
Secondly, the discussion has identified the serious inter-temporal obstacle posed to effective mitigation-CE linkage when CE is considered only as a contingent response to some future climate emergency. Using CE earlier and at lower intensity, in one of its less widely considered modes, may help to craft the bargaining linkages needed for effective global climate response by making it possible to couple small, incremental deployment of CE with commitments to serious parallel emissions cuts, year by year. Although such early CE deployment may raise slippery slope concerns even more acutely than the prospect of using it in some future emergency, building a sufficiently strong linkage to concurrent mitigation may address these concerns.
However, while the strategic bargaining advantages of real-time linkage appear to be clear, this approach would pose daunting governance and management challenges. Participating states would need to make decisions similar in novelty and difficulty to those that would arise under a crisis-driven deployment of CE further into the future: for example, what specific interventions are undertaken, how they are monitored and their risks managed, what systems for liability and compensation are applied, and how are these decisions made? Just as in a later, crisis-driven deployment, controlling these smaller, near-term CE interventions would require international decision processes able to discharge three distinct functions: (i) to competently and fairly assimilate scientific knowledge about the effects and risks of proposed CE interventions; (ii) to make political decisions over what specific interventions to authorize; and (iii) to conduct competent real-time operational management and oversight of interventions under way, scan for unanticipated risks and modify or stop interventions as needed.Footnote 55 But, in this case, these decisions would have to be addressed earlier, under even more uncertainty about effects, and absent the potentially unifying factor of a widely perceived climate crisis.
The effectiveness and risks of these linkage-based strategies will depend on several points of uncertainty, suggesting priorities that are different from those targeted by CE research thus far. Firstly, in view of the apparent strategic and bargaining advantages of the alternative, near-term modes of CE use, research into methods, effects, risks, and management of these would be valuable in addition to research on the longer-term, global interventions that have received most attention thus far. Secondly, since so much about the geopolitical risks of CE turns on uncertainty about states’ interest alignment or opposition over available CE choices, research into such configurations of interests and associated conflicts and trade-offs should be a high priority. Although study of this question has begun, the extent to which these interests depend on the actual set of CE options available remains largely unexamined, which suggests a priority for joint examination of the science and technology of specific CE capabilities, the alternative ways in which such capabilities might be used, and their implications for regional-scale costs and benefits. In view of the high stakes, it may be especially valuable to conduct this research in open collaborative international groups, in the hope that potential threats associated with the development of specific destabilizing capabilities may be recognized and deflected in advance, rather than having to be managed after the fact.
