on other aspects of the earth system, including ecosystems on land and in the oceans. Because the coupled human-environment system is large and complex, it is impossible to fully anticipate all consequences of a geoengineering intervention in advance, or any other type of intervention for that matter. Nevertheless, it is possible to predict and anticipate some of these consequences through a combination of analysis; small-scale de minimis experiments; and climate, Earth system, and integrated assessment modeling. Again, in the case of stratospheric sulfur aerosol injection options, experiments that evaluate how increases in diffuse solar radiation would affect ecosystem productivity or how stratospheric particles might affect the ozone layer could be carried out. Similarly, modeling studies and analysis of observations around volcanic eruptions may provide insight into the changes to be expected in the hydrologic cycle from SRM.

Develop metrics and methods for informing discussions and decisions related to “climate emergencies.” There are at least two components to this research need. For use of SRM as a potential “backstop option” in the case of an emerging “climate emergency,” improved observations and understanding of climate system thresholds, reversibility, and abrupt changes (see Chapter 6)—for example, observations to let us know when an ice sheet or methane hydrate field may become unstable (e.g., Khvorostyanov et al., 2008; Shakhova et al., 2010)—could inform societal debate and decision making about needs for deployment of a climate intervention system. Second, there is no consensus on what constitutes a “climate emergency,” nor is there a consensus regarding when an SRM deployment might be warranted. The notion of an “emergency” is not simply a scientific concept, but one that involves both scientific facts and human values—quite similar to discussions about “dangerous interference in the climate system” (e.g., Dessai et al., 2004; Gupta and van Asselt, 2006; Hansen, 2005; Lorenzoni et al., 2005; Oppenheimer, 2005; Smith et al., 2009). To some people, losing Arctic ecosystems constitutes a climate emergency, whereas to others the declaration of an “emergency” might require widespread loss of human life. Therefore, to inform a broader discussion of how society wants to address issues of risk, climate intervention cannot be studied in isolation but must be placed in a broader context considering, for example, drivers of climate change, climate consequences, sociopolitical systems, and human values.

Develop and evaluate systems of governance that provide models for decision making about whether, when, and how to intentionally intervene in the climate system. Because decisions about intentional alteration in the climate system will have widespread consequences, options for governance, including different types of institutions, assigned decision makers, procedures, norms, and rules and regulations, will be needed and can be provided through analysis. Much can be learned, for exam-