Negotiating solar radiation modification governance

Executive Summary

This report presents the findings of the year-long project, Exploring Geopolitical Tipping Points and Climate Cooling Strategies, a work package of the Advanced Research and Invention Agency (ARIA) funded Strategic Foresight on Climate and Geopolitics: Toward Governance of Solar Radiation Modification (SAFEGEOGOV) project. This project aimed to gain a better understanding of how geopolitical and sociopolitical factors influence the research, development, and governance of Solar Radiation Modification (SRM) technologies, a set of proposed deliberate large-scale interventions to alter global average surface temperatures.

To accomplish this aim, the project team created and ran three two-day crisis simulations in which a network of 35 international academics, civil society actors, and practitioners embodied their expertise in role-playing exercises on SRM governance. There is widespread agreement that some kind of governance framework for both research and potential deployment are essential for SRM but the path to achieving any framework – let alone a transparent, fair, and democratic one – is unclear. While there are ongoing discussions about what governance should look like and what the key challenges might be, having experts embody their positions in a series of simulated negotiations highlighted unexpected and underexplored sources of tension and potential deadlock that may not be obvious from more abstract stock-taking exercises.

Method

The project used the method of non-ideal, open-ended crisis simulations. Rather than pushing for the adoption of a particular set of recommendations or principles, the simulations focused on advancing an understanding of the potential points of failure or agreement of governance attempts. The project recruited a mix of natural scientists, social scientists, policy experts, and practitioners for an international research network that prioritised diversity in background, expertise, and experience. While many participants had specific expertise in SRM itself, others were recruited for their expertise in foreign policy, security, and international governance and diplomacy. Each participant was given a role profile with different aims and red lines based on their actor’s stated policies and historical background and asked to embody their expertise in these roles. In monthly online network meetings, all participants were given primers on a range of topics so that they shared a common baseline of knowledge.

A simulation pre-text formed the starting point for each exercise. Participants were given the freedom to pursue their goals and interests as they saw fit within the exercise. The first pre-text was a meeting set in September of 2025 organised by the United Nations Environment Assembly (UNEA) to negotiate a “rules of the road” framework for SRM research and potential deployment. The second was a consultative forum convened by the Arctic Council set in September 2027 to continue attempts to develop effective SRM governance in the context of increasing environmental and political instability in the region. The third moved forward in time to September 2028, with the starting pre-text that some actors had proceeded with SRM research and development without an effective international governance framework in place.

Key Findings

Network members engaged in over 40 hours of in-person exercise, participated in monthly online network meetings, and provided approximately 60 pages of reflections and analysis over the course of the year. At no point were the network members able to reach an agreement on the governance of SRM that could pass either a 2/3rds majority vote (Simulation 1 and 2) or consensus vote (Simulation 3). Most participants were motivated to pass a formal agreement but were unable to succeed.

Although participants were provided with baseline knowledge of all SRM methods, their negotiations and discussions consistently coalesced around Stratospheric Aerosol Injections (SAI), with some brief attention to space-based SRM methods. Marine Cloud Brightening Techniques (MCBT), Cirrus Cloud Thinning, and other albedo enhancement strategies received minimal engagement from participants. In the second simulation which focused on the Arctic region, there was some supplementary discussion of research into sea ice thickening and enhancing the reflectivity of polar ice.

Analysis of the simulation content and participant feedback further showed that:

• Issues of strategic competition and state sovereignty drove the simulations, not climate concerns. Despite sustained attempts by the project team and many network members to refocus negotiations and discussions on climate change, the simulations consistently focused on questions of control over the development and use of SRM technology and its political implications, rather than climate outcomes.
• Actors were unable to establish a shared foundation of knowledge and understanding, particularly around the purpose of SRM research and development. Distinctions between research and deployment and large and small scale experimentation could not be settled on the basis of scientific thresholds or technical assessments. These were instead explicitly political questions related to the risk tolerance of participants. While many network members approached SRM as an intervention to address the effects of climate change, others viewed SRM – primarily SAI – as an emerging technology with global implications and transboundary effects that would impinge on a range of political domains and governance regimes beyond climate.
• Multiple types of uncertainty hampered progress towards agreement. These were not limited to scientific uncertainties about the effects of SRM as a potential climate intervention. The most persistent and important of these uncertainties were: how to assess and trust information in a fraught political context; how to gauge the intentions of actors engaging with SRM; and how the continued emergence of SRM would influence existing relationships and dynamics between different actors, particularly states.
• SRM governance debates both reflected and showed the potential to exacerbate global inequalities. These inequalities reflect existing global hierarchies between wealthy, technologically-capable states and others, and between states and non-state actors. In broader SRM debates there is frequent discussion about the power to ‘set the global thermostat’ but a consistent finding was that controlling climate outcomes was less relevant than the capacity to develop and deploy SRM technologies to demonstrate geopolitical supremacy through the ability to produce global impacts, whether it would achieve climate stabilisation or not. Uneven access to scientific research infrastructure and capacity also influenced the willingness of different actors to engage or collaborate with each other.
• Participants and proceedings in the simulations were vulnerable to misinformation and misperceptions. Because the project team focused on designing an environment that reproduced current real-world geopolitical tensions and dynamics, participants experienced the effects of lack of trust between key actors which made it difficult to come to an agreement. Given this, efforts to deflect and mislead were often successful in breaking down consensus.

Key Takeaways

Any attempt to develop SRM governance should recognise both the nature and the scale of the obstacles involved and any further SRM technological development should be conducted with the awareness and acknowledgment of these governance obstacles.

Any governance proposals should recognise that there are sticking points for states that are not directly related to SRM or climate change.

Issues of capacity-building and knowledge-sharing are crucial if any agreement is to be meaningfully multilateral. Scientific instrumentation and infrastructure capacity is important in questions of monitoring and verification for any agreement but so is trust.

The context within which any negotiations happen is important. Whether or not a governance proposal is acceptable does not just rely on its content but also on who proposes it, who is supporting or not supporting it, and how it will affect non-SRM issue areas.

Full report