ReportForum on U.S. SolarGeoengineering ResearchMarch 24, 2017The Conference Center at theCarnegie Endowment for International Peace1779 Massachusetts Ave NW, Washington, DC 20036

This document was originally prepared asbackground for participants at the Forumon U.S. Solar Geoengineering Research.It now also includes reflections writtenafter the Forum. Moving forward, wehope it proves useful for anyone seekingto gain background information on solargeoengineering.For a full video of the event,

Table of ContentsForum Agenda . 1Background Paper . 3Scott Barrett . 15Holly Jean Buck . 18Rose Cairns . 21Kerry Emanuel . 24Peter Fiekowsky . 26Peter C. Frumhoff and Jennie C. Stephens . 28Anna-Maria Hubert . 32Jane C. S. Long . 35Douglas G. MacMartin . 39Joseph Majkut . 41Oliver Morton . 44Janos Pasztor . 47Joyce E. Penner . 50Jesse Reynolds . 54Kate Ricke . 56Alan Robock . 60Daniel P. Schrag . 65Kelly Wanser . 68Janie Wise Thompson . 72

Forum AgendaMarch 24, 20178:45 – 9:00 a.m.Setting the Stage Lizzie Burns — Fellow, Harvard John A. Paulson School of Engineering and Applied SciencesEdward A. (Ted) Parson — Dan and Rae Emmett Professor of Environmental Law; FacultyCo-Director, Emmett Center on Climate Change and the Environment, UCLA School of LawGernot Wagner — Research Associate, Harvard John A. Paulson School of Engineering andApplied Sciences; Lecturer, Environmental Science and Public Policy; Associate, HarvardUniversity Center for the EnvironmentPart I: The Science9:00 – 10:15 a.mSocial Science: What we know, and what we ought to know Edward A. (Ted) Parson [Moderator] — Dan and Rae Emmett Professor of EnvironmentalLaw; Faculty Co-Director, Emmett Center on Climate Change and the Environment, UCLASchool of Law Scott Barrett — Lenfest-Earth Institute Professor of Natural Resource Economics, ColumbiaUniversity Holly Buck — Doctoral Candidate, Development Sociology, Cornell University; FacultyFellow, Forum for Climate Engineering Assessment, American University Rose Cairns — Research Fellow, SPRU – Science Policy Research Unit, University of Sussex Kate Ricke — Assistant Professor, Scripps Institution of Oceanography and the School ofGlobal Policy and Strategy at University of California San Diego10:30 a.m. – 11:45 a.m.Natural Science: What we know, and what we ought to know Doug MacMartin [Moderator] — Senior Research Associate, Cornell University Thomas Ackerman — Professor of Atmospheric Sciences and Director of the Joint Institutefor the Study of the Atmosphere and Ocean (JISAO), University of Washington David Keith — Gordon McKay Professor of Applied Physics, Harvard John A. Paulson Schoolof Engineering and Applied Sciences; Professor of Public Policy, Harvard Kennedy School Joyce Penner — Ralph J. Cicerone Distinguished University Professor of AtmosphericScience, University of Michigan Alan Robock — Distinguished Professor, Department of Environmental Sciences, RutgersUniversity1

Daniel Schrag — Sturgis Hooper Professor of Geology, Professor of Environmental Scienceand Engineering, Harvard University; Director, Harvard University Center for theEnvironment; Director, Harvard Kennedy School Program on Science, Technology, andPublic Policy11:45 a.m. – 12:45 p.mLunchPart II: Policy and Politics12:45 – 2:00 p.m.State of Play Jesse Reynolds [Moderator] — Postdoctoral Researcher, Faculty of Law, Economics andGovernance, Utrecht University, The Netherlands Peter C. Frumhoff — Director of Science and Policy, Union of Concerned Scientists Steven P. Hamburg — Chief Scientist, Environmental Defense Fund Joseph Majkut — Director of Climate Science, Niskanen Center Janos Pasztor — Senior Fellow, Carnegie Council for Ethics in International Affairs;Executive Director, Carnegie Climate Geoengineering Governance Initiative (C2G2) Janie Wise Thompson — Vice President, Cassidy & Associates2:15 – 3:30 p.m.The Path Forward Oliver Morton [Moderator] — Senior Editor, Essays and Briefings, The Economist Anna-Maria Hubert — Assistant Professor, Faculty of Law, University of Calgary; AssociateFellow, Institute for Science, Innovation and Society (InSIS), University of Oxford Peter Kareiva — Director, Institute of the Environment and Sustainability, UCLA; FormerChief Scientist and Vice President, The Nature Conservancy Andrew Light — Distinguished Senior Fellow in the Climate Program, World ResourcesInstitute; University Professor, George Mason University Jane C. S. Long — Lawrence Livermore National Lab (ret) Kelly Wanser — Principal Director, Marine Cloud Brightening Project3:30 – 3:45 p.m.Conclusion & Next Steps Gernot Wagner — Research Associate, Harvard John A. Paulson School of Engineering andApplied Sciences; Lecturer, Environmental Science and Public Policy; Associate, HarvardUniversity Center for the Environment2

Background PaperForum on U.S. Solar Geoengineering ResearchEdward A. Parsoni, Lizzie Burnsii, John Dykemaii,Peter Irvineii, David Keithii, and Gernot WagneriiThis paper was prepared as background for the Forum on U.S. Solar Geoengineering Research,which was held at the Conference Center of the Carnegie Endowment for International Peace inWashington, DC on March 24, 2017. The Forum was co-hosted by the Solar GeoengineeringResearch Program at Harvard University and the Emmett Center on Climate Change and theEnvironment at the University of California, Los Angeles, and funded through the generoussupport of the Alfred P. Sloan Foundation. The paper provided background information ongeoengineering and associated debates for Forum participants unfamiliar with these issues, andframed several key questions to be addressed at the Forum. In addition, since the context fordiscussing solar geoengineering research changed substantially in the year between when theForum was planned and when it took place, the paper briefly discussed the current context andits implications.Importantly, while this background paper was initially intended for Forum participants, we hope itproves useful for those who are seeking to gain background information on solar geoengineeringmore broadly.Background: Geoengineering Methods, Effects, and ConcernsGeoengineering – also called climate engineering, climate intervention, or climate remediation –is a third class of potential responses to global climate change, additional to mitigation (cuttinggreenhouse-gas emissions) and adaptation (reducing vulnerability to climate change). Whilegeoengineering responses have been recognized for decades and periodically discussed inscientific assessments – going as far back as the first official report to a U.S. President on globalwarming, President Johnson in 1965 – they received little attention until the past ten years.Geoengineering is defined by intentionality and scale: intentional intervention to alter the climateat global scale. Of the two broad types of geoengineering – modifications of the global carboncycle, or of Earth’s radiative balance – the Forum mainly addressed the latter, which we call solargeoengineering. Solar geoengineering alters the energy balance of the Earth, either by slightlyincreasing the fraction of incoming sunlight that is reflected from the Earth rather than absorbed;or by increasing the Earth’s ability to cool by emitting thermal (infrared) radiation.iiiUCLA School of Law; [email protected] John A. Paulson School of Engineering and Applied Sciences.3

Three proposed methods of solar geoengineering are most prominent in current debate.Stratospheric aerosol injection would involve the distribution of reflective aerosols in the upperatmosphere. Marine cloud brightening would involve modifying the properties of low-altitudemarine clouds to make them more reflective. Cirrus thinning would reduce the density of iceparticles in high-altitude cirrus clouds, which would increase the emission of thermal radiation tospace. Several other approaches have been proposed but have fallen out of serious considerationdue to early indications of limited effectiveness, high cost, or risk. It is likely that new forms ofintervention, or improvements to existing forms, will be identified.The most promising of these methods offer the prospect of modifying global-scale characteristicswith extremely high leverage. Several of their prominent characteristics are related to this highleverage. First, solar geoengineering could act fast. Like large volcanic eruptions, some methodscould, if deployed at large enough scale, significantly cool global temperatures within months. Theaerosol particles that figure in these methods have lifetimes of only a few days in the loweratmosphere and a few years in the upper atmosphere, the stratosphere. The benefit of these shortlifetimes is that aggregate cooling can be started, modified, or – if some unfavorable consequenceis discovered – stopped, quickly. A corresponding risk arising from these short lifetimes is that, ifa large program of solar geoengineering were suddenly terminated, the heating being offset bythe program would occur rapidly, which would bring even more severe risks than if geoengineeringhad not been done and the same heating had occurred more slowly. This is because many of therisks of climate change arise from the rate of change, not simply the change itself.Additionally, most solar geoengineering techniques would have a global, not local, impact. Onecountry, for example, could not deploy solar geoengineering to slow global warming over its ownborders without affecting other nations and ecosystems around the globe. This fact poses manygovernance challenges, particularly when combined with a second fact: that solar geoengineeringis inexpensive, at least by comparison of the direct cost of making the interventions to the cost ofachieving the same total cooling by carbon removal or mitigation.1 Indeed, the direct costs of solargeoengineering are likely to be trivial relative to other risks and benefits. As a result, the capabilityto deploy solar geoengineering and change the global climate may be within reach of manynations. This makes the problem of governing geoengineering the inverse of that posed bymitigation – a strategic structure that has been called a “free-driver” problem, in contrast to the“free-rider” problem of mitigation.All solar geoengineering methods, however, offer only imperfect corrections for the harms causedby elevated greenhouse gases. They target only certain climate effects of elevated CO2, not itseffects on the chemistry of the oceans that makes them more acidic, its alteration of competitiverelationships among plants, which depends on how they use CO2 in photosynthesis, or other keyfactors.2 All identified methods also have environmental side effects in addition to their targetedclimate effects. For different methods, these side-effects may include alterations of stratospheric1Mitigation can only reduce future heating, not cool the climate relative to heating already realized or committed.Solar geoengineering will affect the carbon cycle indirectly via temperature-carbon feedbacks, for example byreducing the thawing of permafrost and associated emissions of methane and CO224

chemistry, in particular stratospheric ozone (noting that ozone decreases for some methods, butincreases for other recently proposed methods); changes in the appearance of the sky; and theeffects of any material injected into the atmosphere when it is deposited on the ground surface.Moreover, no solar geoengineering method could perfectly offset the climate effects of elevatedgreenhouse gases. This is mainly because the effects of reducing absorption of light at the Earth’ssurface are quite different from the greenhouse effect, which occurs aloft. As a result, comparedto greenhouse heating, solar geoengineering reduces precipitation and evaporation more stronglythan temperature. In addition, some methods – marine cloud brightening and cirrus thinning –operate by modifying naturally occurring phenomena (in this case, clouds), so their potentialimpact is limited by the spatial distribution of those phenomena. These methods may thus havepatchy effects, or quantitative limits to their global effect. Still, model studies show mismatches ofeffect that are smaller than was initially expected. Compared to climate conditions with projectedincreases in greenhouse gases, model studies suggest that solar geoengineering interventions maybe able to move both temperature and precipitation closer to pre-industrial values over a largefraction of world land surface. As the Intergovernmental Panel on Climate Change (IPCC) stated,“Models consistently suggest that [solar geoengineering] would generally reduce climatedifferences compared to a world with elevated GHG concentrations and no [solar geoengineering];however, there would also be residual regional differences in climate (e.g., temperature andrainfall) when compared to a climate without elevated GHGs.”3These basic properties of solar geoengineering interventions – fast effect and controllability, crossborder impacts, low cost, and imperfect correction for the effects of elevated greenhouse gases –define the large-scale nature of the governance problem they pose. They also explain why we arediscussing these interventions, and why now.Debate on solar geoengineering became prominent ten years ago. The trigger for the debate wasa widely-noted essay by eminent atmospheric scientist Paul Crutzen, who argued theseinterventions merited investigation because their risks might be less severe than those ofcontinuing climate change. But beyond the specific triggering event of Crutzen’s essay, thebroader cause of renewed attention to solar geoengineering lay in the underlying realities hedescribed: increasingly severe risks from projected climate change, continued uncertainty aboutthe character and timing of these risks, and increased recognition that mitigation and adaptationmay be inadequate to manage the risks. Adaptation may fall short due to limited knowledge orexperience of how to do it, resource constraints, political conflict, or institutional failure – as wellas the possibility that climate changes may overwhelm adaptation capability. Mitigation may fallshort because it cannot reverse realized or committed climate change, due in part to the hugeamount of installed plant and equipment that must be changed – and the high costs oftransitioning this infrastructure – and due to the slow response of the climate system to changesin forcing. With the exception of measures that target short-lived gases, even intense andsuccessful mitigation efforts will only significantly deflect climate risks after a few decades.3IPCC Assessment Report 5, Working Group 1, Chapter 7. (Different terminology: “solar geoengineering” replaces“Solar Radiation Management (SRM)” in the original.)5

Moreover, even if emissions are reduced rapidly to zero, there are still risks of global warmingbecause of carbon’s long atmospheric lifetime and the presence of carbon-climate feedbacks, suchas the release of carbon dioxide and methane from melting permafrost, which could acceleratewarming. As a practical matter, mitigation may also fall short because, despite nearly threedecades of attempts, nations’ mitigation efforts have not been intense – or even, in many cases,serious.The Prospect of Future Operational Use: Benefits, Risks, ConditionsIn this context, solar geoengineering offers a high-stakes, two-sided prospect. On the one hand, itmay, under some conditions, be able to substantially reduce climate-change risks and harms inways that mitigation and adaptation alone cannot. On the other hand, it could be ignorantly,incompetently, dangerously, and illegitimately used in ways that cause severe harms to humansin the environment – greater than those posed by climate change. The conditions for the potentialbenefits to dominate, broadly, are that interventions are identified that work well with limitedharmful side-effects, and that they are developed and used competently, prudently, andlegitimately.Assuming these conditions, three broad ways have been proposed that solar geoengineeringmight be beneficially used – each of them subject to various scientific and socio-politicallimitations and concerns.First, it might be used in response to some future severe climate-change impacts beingexperienced or imminently anticipated. This mode of use has been described as “emergencyresponse,” or “Plan B.” Used this way, solar geoengineering deployment would be delayed, rapidand strong – not deployed at all in the near term, but then deployed quickly and intensely at somefuture time.Second, it might be used in conjunction with aggressive mitigation, adaptation, and carbonremoval, as part of a strategic, integrated, multi-decade climate response. This mode of use hasbeen described as “buying time,” or “shaving the peak” (reducing the 50 to 100 year period ofheating that even extreme mitigation and carbon removal are too slow to avoid). Used this way,deployment would be immediate, incremental, and temporary – ramping up, then down, as theother responses grow to full scale. Even if carbon capture were not included in such an integratedresponse – which would imply that climate change could only be stopped, not reversed – such atemporary program of solar geoengineering could still reduce risks by slowing the rate of heatingtoward whatever hotter climate the given level of mitigation effort is moving the world toward.Third, deployments at less than global scale have been proposed, to target large-scale regionalprocesses of global concern, such as summer loss of Arctic sea ice or tropical cyclone formation.Early research suggests that such proposals could have the potential to bring certain benefits, suchas reduced sea level rise, but there are also many uncertainties as well as several potential risks,including changes to regional hydrology. Like all methods, more research would need to be doneto increase our understanding of the potential benefits and risks.6

Moreover, for any of these modes of use to be beneficial, certain conditions must hold. Some ofthese conditions are matters of knowledge and technical capability: are feasible methodsidentified that would confidently have the intended effect and not carry severe side effects? Otherconditions are matters of the social, ethical, institutional, legal, and political setting in whichinterventions would be considered, decided upon, and (if adopted) implemented and managed: isthere basis for confidence that these decisions would in fact be competent, prudent, andlegitimate? Of these two groups of conditions, the first are fundamentally about research –research into proposed methods, and the natural systems with which they would interact. Thesecond are fundamentally about governance, mainly at the international level because of theinternational scope and impacts of these interventions.The governance requirements posed by potential future proposals for operational use of solargeoengineering are novel and severe. Many serious governance-related risks have been identifiedrelated to geoengineering being used incompetently, recklessly, rivalrously, or relied on too much.Examples of such dangerous conditions of geoengineering use are easy to imagine: for example,use in a crisis with inadequate risk assessment; uncoordinated or opposing interventions bymultiple states or other actors; relying on these imperfect interventions too much, making theneglect of essential mitigation and adaptation measures even more severe; use in ways thatundermine or destabilize institutions for international cooperation, on climate or related issues;or use in ways that generate international destabilization and conflict, particularly in the event ofwide differences in severity of climate impacts, or interventions that suggest the prospect ofregional climate control.While we recognize the novelty and high stakes of these governance challenges raised by theprospect of future operational interventions, these were not the focus of the Forum. Rather, theForum focused on the first class of conditions identified above: the need for solar geoengineeringrelated research, the risks and challenges associated with research, and the governance needsposed by research. In contrast to the larger but more distant governance challenges raised by theprospect of future operational deployment, these research issues are immediate and concrete.Longer-term questions related to operational governance were on the table at the Forum, butonly insofar as they were implicated by, or likely to be influenced by, near-term decisions relatedto research.Solar Geoengineering Research: Arguments in Favor, Experience, ProposalsThe basic argument for expanded research is straightforward. If it is likely that future decisions willhave to be made regarding proposals, demands, or charges about operational geoengineeringdeployment – whatever the outcome of such decisions, whether to authorize, prohibit, or regulateand control proposed interventions – then providing any basis to inform those decisions requiresresearch. Research is needed to identify and characterize methods and capabilities, to designpossible implementation scenarios, to identify and characterize efficacy and associated risks, andto understand the social, ethical, institutional, legal, and political setting within which they mightbe used.7

Beyond informing such future decisions, there are also additional reasons research is needed,including developing the ability to detect, identify, and monitor interventions (for example, toprotect against clandestine interventions, or to improve our understanding of legitimateinterventions by observing their risks and efficacy in the natural environment); and informing theparticulars of future governance needs, since these will be strongly influenced by specific technicalcapabilities and anticipated risks.Certain research into solar geoengineering has already been conducted. There have been manycomputer-modeling studies, including comprehensive inter-comparisons of climate-modelprojections driven by standardized scenarios of future greenhouse-gas emissions and solargeoengineering interventions conducted under the Geoengineering Model IntercomparisonProject (GeoMIP). There have also been many observational studies of natural or already existinganthropogenic (human-influenced) processes relevant to likely effects of solar geoengineeringmethods, e.g., atmospheric aerosols, volcanic plumes, and tracks left by ships and aircraft.Moreover, there have been lab-bench studies of related processes and a few preliminaryengineering studies of potential methods to estimate performance, effectiveness, technicalrequirements, and cost.One of the key sites of controversy over solar geoengineering research – and the key margin ofnear-term decision that made the Forum timely – concerns active outdoor perturbationexperiments. These would involve intentional introduction of materials into the open environmentor some other active manipulation of environmental conditions in ways that aim to informunderstanding of the efficacy and risks of potential future interventions.The case for doing such studies that are small in scale appears strong. All such studies thus farproposed (and the few that have been attempted or done) are of tiny scale, posing negligibleenvironmental risk, yet offer to substantially advance knowledge on atmospheric processes crucialto understanding what potential future interventions might do. The proposed studies would addknowledge to that gained by laboratory or computer-model studies, which cannot fully replicateconditions in the open environment of potential relevance to the effects of interventions. Theyare thus likely to inform understanding of whether and how potential geoengineeringinterventions can be done, what their effects are likely to be (both intended and unintended), howinterventions can be detected, what risks they may carry, and how these risks can be managed.But few to no such active perturbation studies have been done. Two small-scale interventions thatare known to have been done, both using existing funding on related topics, include one in theUnited States (the 2011 E-PEACE experiment sprayed smoke and salt particles from a barge offthe California coast to study effects on cloud formation), and one in Russia (a 2009 experimentsprayed smoke from a helicopter and a truck and observed resultant radiative effects). In addition,one study, a proof-of-concept experiment to spray water from a tethered balloon (the SPICEexperiment), was proposed, funded, and partly implemented in the United Kingdom, then delayeddue to objections that the associated public consultation process was inadequate, and cancelled8

after the (unrelated) discovery of previously undisclosed financial conflict-of-interest of one of theresearchers.Several additional solar geoengineering experiments have been proposed in scientific literatureand some have taken various degrees toward technology development and implementation. Noneof these has yet been implemented or fully funded, however, and all have raised controversy andopposition. Leading examples of these include:1. A stratospheric controlled perturbation experiment (SCoPEx), which would use a balloonto release 100 g to 1 kg of aerosol material in the stratosphere, to study resultant aerosolsize distribution, radiative forcing, and chemical effects;2. An experiment in marine cloud brightening, which would loft sea salt spray into the marineboundary layer to study resultant effects on cloud formation and properties;3. A study, described in the scientific literature, which would seed high-latitude cirrus cloudswith aerosols to reduce their optical thickness and so increase infrared radiation from thetop of the atmosphere.Solar Geoengineering Research: Concerns and Arguments AgainstThe slow pace of developing, funding, and implementing solar geoengineering field researchreflects not just constrained resources and bureaucratic inertia, but also widespread nervousnessabout the endeavor, based on existence of significant concerns and some political opposition. Webriefly outline these concerns and objections, grouped in five categories:First, some objections to expanded research originate in concerns that pertain, reasonably, topotential future deployment – e.g., difficulties of control, disruption to the climate-policy agenda,risk of excessive reliance (potentially leading to future termination-shock scenarios), orinternational conflict – but extend these concerns to assert that they also support opposition toresearch. But opposing future deployment does not necessarily imply opposing research,particularly since research can inform and benefit any future decisions about deployment,including the decision to reject it. One way to make this inference from opposing deployment toopposing research valid would be by making an extreme prior assumption: that operationaldeployment is certain never to be warranted, at any time or under any circumstances. Suchcategorical opposition to deployment might be based on prior certainty that the consequences offuture deployment can only be worse than the consequences of the climate change that it mightreduce or delay. But such opposition is more typically based on some non-consequential moralstance: geoengineering deployment would be intrinsically wrong – e.g., because it is messing withnature, is hubris, or is an impermissible step to the Anthropocene – and cannot be redeemed byany evidence that it might bring benefits relative to the available alternative. We find this extremepremise implausible, but if you accept it, the argument for research to inform future decisions isgreatly weakened – although not completely: even under this assumption, research could still bewarranted to build capacity to monitor and detect unauthorized or clandestine use.9

Second, some related objections state that there is no way to distinguish between small-scale fieldresearch and global-scale operational deployment. In part these objections rely on the continuityof intervention scale, from tiny to global. They thus reduce to a “where-to-draw-the-line”argument, and are vulnerable to the normal rejoinder to arguments of this type: we might notknow precisely where to draw the line, yet still be confident that these things lie on one side, thosethings on the other. A subtler form of this argument relies on the complexity of atmosphericprocesses, which makes the effects of any intervention always uncertain until it is actually done.Tiny-scale experiments give information about atmospheric processes that can help understandlikely responses to larger interventions. But some uncertaint

1 Forum Agenda March 24, 2017 8:45 - 9:00 a.m. Setting the Stage Lizzie Burns — Fellow, Harvard John A. Paulson School of Engineering and Applied Sciences Edward A. (Ted) Parson — Dan and Rae Emmett Professor of Environmental Law; Faculty Co-Director, Emmett Center on Climate Change and the Environment, UCLA School of Law