| 英文摘要: | Current climate engineering proposals do not come close to addressing the complex and contested nature of conceivable 'climate emergencies' resulting from unabated greenhouse-gas emissions. Continuing business-as-usual with regards to greenhouse-gas emissions will increase the likelihood of 'dangerous' climate changes. In response to this risk, Crutzen1 argued in 2006 that a 5 °C warmer world will probably have catastrophic consequences and that the only way out may be to engineer the Earth's climate by injecting aerosols into the stratosphere. The possibility of a future 'climate emergency' has subsequently been used to justify research on climate engineering2 — the deliberate modification of the Earth's climate. Over time, the emergency framing has evolved to become a central argument for why we should consider investigating solar radiation management (SRM) techniques, which reduce the amount of sunlight absorbed at the Earth's surface. But whether SRM can possibly prevent or counteract a climate emergency raises the more fundamental question of what a climate emergency actually is.
Crossing a tipping point in the Earth system has often been used as an example of a potential climate emergency2. Several 'policy-relevant' tipping elements have been identified that could conceivably be tipped by anthropogenic activities this century3. Among these are the Atlantic thermohaline circulation, the West Antarctic ice sheet, the Amazon rainforest and the West African monsoon4. But whether SRM intervention could actually prevent these elements from tipping, or counteract tipping that was underway, depends on: (1) their predictability, (2) their timescale of tipping and (3) their reversibility. A proactive 'emergency' response is only conceivable if a tipping point can be convincingly forecast in advance. Although early warning signals have been found for some tipping points4, the methods do not precisely forecast the time of tipping, and only work if a system is forced slowly relative to the internal timescale of its dynamics4, 5, 6. Under relatively rapid climate change, this can prevent 'slow' systems such as ice sheets, ocean circulation or major forest biomes from giving a reliable early warning signal of approaching tipping. This restricts climate engineering to being a reactive response to tipping that is already underway. 'Slow' tipping elements such as ice sheets7, 7 or the Amazon rainforest9, 10 tend to exhibit hysteresis and a high degree of irreversibility. They also tend to lag climate forcing such that by the time tipping is perceived, their original state may have long since lost its stability. This means that excessive climate engineering — that is, over-cooling the planet — is likely to be required to recover their original state (and even then it may not work). The steadily accumulating consequences of slow tipping are also not obvious triggers for a rapid 'emergency' response. Notably, evidence suggests11 that the West Antarctic ice sheet has been tipped by oceanic warming during the past 20 years, yet no climate emergency has been declared thus far. If it were, it is unlikely that SRM would be able to reverse the ice discharge from West Antarctica. 'Fast' tipping elements that could trigger an 'emergency' situation, such as an abrupt shift in a monsoon, are generally related to regional changes in climate. As SRM, for instance by stratospheric aerosol injection, has effects over a much larger scale, it is not an obvious response to such a regional emergency and, owing to spatially heterogeneous hydrological responses, may pose more of an additional threat than offer a remedy12. Thus, the potential for SRM to respond effectively to tipping-point 'emergencies' is very restricted. Even if there were a case where it could be a logical response, there is one final problem: decisions on how much SRM to implement would have to be based on experiments with the same global climate models that had failed to predict the occurrence of a tipping point in the first place. These models would by definition be insufficiently sensitive to climate forcing, and therefore run the risk of recommending an excessive SRM intervention.
© IMAGEGALLERY2 / ALAMY
The devastating consequences for human lives and properties from Typhoon Haiyan hitting Southeast Asia, and particularly the Philippines, in early November 2013. Socio-economic dynamics add a new dimension of complexity to the climate emergency problem. Whereas a purely environmental climate emergency might not even have detectable socio-economic impacts, an event regarded as a socio-economic climate emergency might be based on very few tangible environmental observations. For example, through complex global supply chains, the effects of extreme local weather events might spread fast14 and have global impacts on critical socio-economic variables such as food prices, commodity prices, trade flows and migration. A cascade of such damages could lead to a more general socio-economic emergency. Indeed, the perception of a single extreme event as a potential threat for a strategic region might itself lead to considerable political instability. In this and any sense, an emergency can only be 'declared' rather than be 'discovered'. Whether a given phenomenon is regarded as an emergency is ultimately based on shared societal understandings of what constitutes an emergency and when it is appropriate and legitimate to declare one19. Emergencies are not just pure facts, but a combination of facts and values, perceptions and interests. This socio-political character of a climate emergency leads ultimately to a number of critical questions21 such as who will be affected, at what scale, and who is authorized to declare the emergency. On top of this complication, a fundamental scientific question remains: can SRM counteract the climatic root of such a socio-economic emergency? The evidence suggests not, as it is difficult to envisage how SRM could be used effectively to address, for instance, interruptions in global supply chains or outbreaks of social unrest. Instead, SRM interventions are likely to result in changes in regional climate patterns22, and these will carry regional to global socio-economic and political implications of their own. Furthermore, early warning signals for such social tipping points are even more difficult to determine23.
It may not be possible to recognize a climate emergency before it takes the form of a declared socio-economic and political emergency, for which SRM seems obviously ill-suited as a remedy. Because emergencies are combinations of facts and values, they can be ignited by political strategies. They can also, like scandals, be triggered by the mass media or by politicians. The declaration of an emergency situation is ultimately a political act, and thus will inevitably be used for political purposes. By definition, declaring an emergency invokes a state of exception that carries many inherent risks24: the suspension of normal governance, the use of coercive rhetoric, calls for 'desperate measures', shallow thinking and deliberation, and even militarization. By definition, emergency situations are extraordinary and exceptional. To declare an emergency becomes an act of high moral and political significance, as it replaces the framework of ordinary politics with one of extraordinary politics25. In cases of humanitarian emergencies, for instance, foreign armies might be permitted to operate within a country's territory. In cases of epidemic diseases, civil liberties might be restricted. If these potential violations of the principles of international law are to be policed, then we need to avoid casual declarations of climate emergencies, even with the best of intentions. Further, if SRM is to be conceived and declared as a pre-emptive strike against putative future emergencies, the analogy to pre-emptive warfare is hard to avoid. The climate emergency narrative as an argument for SRM implementation must therefore be constantly scrutinized, especially when it is claimed to make scientific sense. There are many tragic examples of where normal politics has been suspended in the name of science and 'objective evidence'.
Solar radiation management may allow the control of one characteristic of the climate system, for example the global mean temperature. At the same time it changes many other characteristics of the system. Although a specific class of extreme climatic events might potentially be reduced under SRM, it remains completely unclear whether SRM increases or decreases other categories of weather extremes, such as those associated with jet-stream dynamics or monsoon systems. Currently, our models and techniques are insufficient to predict the tipping of climate subsystems, and these systems are sufficiently complex to prevent human-induced repair after tipping has occurred. Consequently, one can ask whether a climate emergency can ever be prevented by SRM, unless it is declared pre-emptively on the sole basis of unabated greenhouse gas emissions. In this case, an unprecedented amount of risk would have to be taken without knowing which emergencies would actually be avoided or even be provoked. The danger of declaring a climate emergency is further exacerbated when one considers the political stakes of doing so. Emergencies are by no means simple geophysical occurrences, but rather the outcome of highly complex interactions between the natural environment, political interests and social norms. In the context of considerable scientific uncertainty — and hence the multiple possible interpretations of scientific results and arguments — climate emergencies will be declared on largely political grounds. This interlinking of scientific uncertainty and political opportunism should caution against implementing SRM as a climate emergency measure, a conclusion that we reach on the basis of sound scientific arguments, good governance and ethical principles.
- Crutzen, P. Climatic Change 77, 2–220 (2006).
- Geoengineering the Climate: Science, Governance, and Uncertainty (The Royal Society, 2009).
- Lenton, T. M. et al. Proc. Natl Acad. Sci. USA 105, 1786–1793 (2008).
- Lenton, T. M. Nature Clim. Change 1, 201–209 (2011).
- Boettinger, C. & Hastings, A. Nature 493, 157–158 (2013).
Tipping pointsFrom patterns to predictions&rfr_id=info:sid/nature.com:Nature.com&id=doi:10.1038/493157a&id=pmid:23302842&genre=article&aulast=Boettinger&aufirst=C.&title=Nature&volume=493&spage=157&epage=158&date=2013&atitle=Tipping pointsFrom patterns to predictions&sid=nature:Nature">
- Scheffer, M. et al. Nature 461, 53–59 (2009).
- Mengel, M. & Levermann, A. Nature Clim. Change 4, 451–455 (2014).
- Robinson, A., Calov, R. & Ganopolski, A. Nature Clim. Change 2, 429–432 (2012).
- Jones, C. et al. Nature Geosci. 2, 484–487 (2009).
- Boulton, C. A., Good, P. & Lenton, T. M. Theor. Ecol. 6, 373–384 (2013).
- Favier, L. et al. Nature Clim. Change 4, 117–121 (2014).
- Robock, A., Oman, L. & Stenchikov, G. L. J. Geophys. Res. 113, D16101 (2008).
- Blackstock, J. J.
|