Science and regulation. End the deadlock on governance of geoengineering research.

P roposals for research on geoengineer-ing methods to offset greenhouse-gas– driven climate change have attracted controversy (1– 6). Multiple methods have been proposed (7), but attention and controversy have centered on methods to reduce incoming sunlight—for example, spreading refl ective aerosols in the stratosphere or spraying condensation nuclei to increase low ocean clouds (1, 2). Such high-leverage interventions offer the dual prospect of large benefi ts and harms. They may reduce climate-change risks faster than any other response. Yet they may also cause environmental harm or worsen policy failures— for example, undermining emissions cuts or triggering international conflict. Research is needed to develop capabilities and assess effectiveness and risks (fi eld research as well as model and laboratory studies), but geo-engineering requires competent, prudent, and legitimate governance (1, 2, 8). We propose specifi c steps to advance progress on research governance. No such governance now exists beyond normal scientifi c review processes and national laws, so geoengineering outside national ter-ritory—from small fi eld research to operational deployment—falls under no international legal control (9). Recognizing this void, several projects have tried to develop guidelines on governance of geoengineer-ing research (10– 13). These projects have achieved agreement on the need for research, the need for governance of research, and the principle that as the scale and anticipated risk of interventions increase, so does the need for assessment, scrutiny, and control. But these consensus statements have been at high levels of abstraction, lacking the specificity needed to help any body— governmental or scientifi c—enact operational governance and assessment procedures. In particular, no progress has been made on two questions that are basic to designing a governance system. First, if large interventions need more control than small ones, how is the boundary between " small " and " large " defi ned? Second, can scientifi c self-regulation adequately control small-scale research, or is government regulation needed—and, if it is, what should be the relation between regulatory and scientifi c processes? Debate on these questions is increasingly polarized. One view, advanced by some non-governmental organizations and a few scientists , invokes direct environmental risks (often exaggerated) and a slippery slope from research to deployment to seek strict control on a broad set of activities—for example, all geoengineering research, all fi eld research, or all active environmental perturbation, no matter how small. Practical obstacles to this approach are considerable, because impacts of proposed research can be …