Solar Geoengineering: What Each Position Is Protecting

A climate scientist at Harvard's Solar Geoengineering Research Program has spent a decade studying stratospheric aerosol injection — the idea that aircraft could release fine reflective particles into the upper atmosphere, mimicking the cooling effect of a large volcanic eruption. The models are consistent: even a modest SAI program could reduce global average temperature by half a degree or more within years. In 2023, the world crossed 1.5°C of warming above pre-industrial levels for the first time. She is not proposing deployment. She is proposing research — field experiments that would answer basic questions about stratospheric chemistry, regional precipitation effects, and governance requirements before anyone is in a position to make a deployment decision. The world, she believes, does not have the luxury of an ignorance-based precautionary principle when the alternative is uncontrolled warming.

A climate policy researcher at a civil society organization in Mozambique has heard this argument many times. She knows what volcanic cooling looks like in the Southern Hemisphere: Pinatubo's 1991 eruption disrupted the Asian and African monsoons for two years. Mozambique already faces intensifying cyclones, erratic rainfall, and coastal flooding caused by emissions from countries with which it has almost no negotiating leverage. An SAI program controlled by the United States, China, or a coalition of wealthy nations would set a new global thermostat — one calibrated to the preferences of whoever deployed it. She is not denying that her country needs cooling. She is asking who would decide how much cooling, where the rain would fall, and what recourse her government would have if the answer was wrong.

A legal scholar at a Swiss environmental governance institute is watching this debate with a different concern. The current international framework governing atmospheric interventions is, essentially, a void: the 1978 Environmental Modification Convention addresses weaponized weather modification, not emergency climate management. The UN Environment Assembly passed a non-binding resolution calling for an expert panel in 2023 — it was rejected. What governs solar geoengineering research today is a patchwork of national regulations and voluntary scientific norms. She is not opposed to research. She is terrified of the governance gap: that by the time the world's governments are ready to seriously negotiate a framework, several actors will have moved from research to unilateral deployment, creating facts that governance must catch up to.

These three people are not disagreeing about whether the climate is warming or whether the crisis is serious. They are disagreeing about something harder: whether the scale of the emergency justifies technologies whose side effects are global, whose governance is nonexistent, and whose deployment cannot be undone once begun — and what the failure to answer that question in advance will cost.

What the climate emergency research position is protecting

People who advocate for solar geoengineering research — and potentially, eventually, for deployment — are protecting something specific: the option to prevent warming that mitigation has already failed to prevent. Their core argument is temporal. The emissions reductions required to stay below 1.5°C did not happen in the 1990s, did not happen in the 2000s, and are not happening now at the required pace. The IPCC's modeled pathways that keep warming below 2°C all involve negative emissions in the second half of the century — removing CO₂ from the atmosphere at enormous scale. Carbon removal is real and important, but at current deployment rates, it will not arrive in time to prevent the feedback loops — permafrost thaw, ice-albedo loss, forest dieback — that lock in further warming regardless of what humans do after the fact.

From this view, refusing to research solar geoengineering is not neutrality. It is a choice to remain ignorant about a potentially necessary tool. The Harvard SAI research program's proposed stratospheric balloon experiments — measuring particle dispersion and stratospheric chemistry at very small scale, releasing kilograms of calcium carbonate rather than megatons of sulfur dioxide — would provide data that does not currently exist. Without that data, any deployment decision, for or against, is made blind. The research position is not a deployment recommendation; it is an argument that the knowledge required to make a deployment decision responsibly should be accumulated before the decision is forced.

Proponents also note that solar geoengineering research is already happening — in China, in Russia, in private sector contexts — and that the choice facing Western science institutions is not between a world with and without SAI research, but between research conducted with open publication, international scientific scrutiny, and governance norms, or research conducted in opacity. The Harvard program's approach explicitly includes governance design as a core research component: they are trying to build the international oversight framework alongside the technical understanding, not after it.

What this position costs is harder to acknowledge: it normalizes a technological intervention of planetary scale, and normalization tends to move toward deployment. Each research step answers a question and generates the next question, building constituencies and expertise and sunk costs in a direction that has its own momentum. The history of large-scale technological systems — nuclear power, fossil fuels, industrial agriculture — suggests that "research does not imply deployment" is a harder promise to keep than it sounds.

What the climate justice opposition is protecting

People who oppose solar geoengineering — research and certainly deployment — are protecting something different: the right of vulnerable communities not to have unilateral decisions about planetary systems imposed on them by the actors most responsible for creating the emergency in the first place. Their argument is structural. The countries and communities most threatened by climate change — Bangladesh, the Sahel, Pacific Island states, low-lying coastal cities in Asia and Africa — had almost no role in generating the emissions that caused warming, almost no role in designing the international climate regime, and, under any plausible governance scenario, almost no ability to veto a solar geoengineering deployment decision made by major powers.

The regional climate effects of SAI are not uniformly beneficial. Volcanic analogues — the 1815 Tambora eruption, the 1991 Pinatubo event — show that stratospheric aerosols disrupt the Asian and African monsoon systems. A 2013 study in the journal Nature Climate Change found that SAI optimized to reduce global mean temperature would simultaneously over-cool some regions and under-cool others, with the distributional burden falling disproportionately on the Global South. The history of large-scale technological interventions by wealthy countries in the global commons — atmospheric nuclear testing, industrial CFC emissions, fossil fuel combustion itself — is a history of harms distributed to those with the least power to resist them.

There is also the moral hazard argument. If solar geoengineering is perceived as a technically feasible alternative to emissions reduction, it will reduce the political pressure on the actors most responsible for emissions to actually stop emitting. This is not speculative: the rhetoric of technological salvation has already functioned as a pressure-release mechanism for climate action across multiple cycles of CCS, bioenergy, and carbon offsetting. Every year that the fossil fuel economy continues generates additional warming that geoengineering would then need to mask, making the program larger and its termination more catastrophic.

What this position costs: opposing research does not prevent research. The countries and actors who would conduct solar geoengineering without adequate governance or international consent are not deterred by the opposition of civil society organizations in the Global South. A blanket opposition to research may foreclose the possibility of influencing how research is conducted — including its governance design — from the inside.

What the precautionary governance position is protecting

A third position — held by many international lawyers, climate governance scholars, and some environmental organizations — is not opposed to geoengineering research in principle but is alarmed by the governance vacuum in which it is currently proceeding. This position is protecting multilateral institutions and the principle that interventions at planetary scale require planetary-scale consent.

The central problem is what scholars call "termination shock." Solar geoengineering does not reduce atmospheric CO₂; it masks warming by reducing solar radiation. If an SAI program were deployed and then terminated — due to geopolitical conflict, state failure, funding collapse, or deliberate decision — global temperatures would rebound to where they would have been without the program, within years rather than decades. The rate of warming during termination shock would be far faster than anything in the climate record, far faster than most ecosystems and agricultural systems could adapt to. An SAI program of meaningful scale would therefore create an indefinite planetary dependency: once started, it could not be stopped without catastrophic consequences. This fact changes the governance requirement entirely. It is not enough to regulate deployment; it is necessary to ensure continuity of governance across political disruptions, leadership changes, and geopolitical realignments that no current international institution is designed to guarantee.

The governance position argues that the right sequence is: negotiate an international framework first, then conduct research within it. The legal tools exist — the UN Environment Assembly, the ENMOD Convention, the UNFCCC — but no state has been willing to prioritize a geoengineering governance negotiation. The 2023 UNEA resolution was blocked by a coalition including the United States. The precautionary multilateralist position holds that the failure to negotiate governance is itself a choice, and that proceeding with research under voluntary norms while governance is unresolved is not neutral — it is a choice to proceed in the absence of the consent that would make proceeding legitimate.

What this position costs: waiting for adequate multilateral governance before conducting research may mean waiting indefinitely. The international institutions capable of producing binding climate governance frameworks have, so far, not been adequate to the task. The Kyoto Protocol failed. The Paris Agreement's commitments are voluntary and insufficient. A geoengineering governance negotiation that attempts to give affected countries meaningful veto power faces exactly the same collective action and great-power competition dynamics that have blocked every other ambitious multilateral climate agreement.

What the carbon dioxide removal position is protecting

A fourth position — less visible in mainstream debate but increasingly prominent among climate scientists and some environmentalists — argues that the energy and resources being directed toward solar geoengineering research should instead be directed toward accelerating carbon dioxide removal. This position is protecting a path to genuine atmospheric restoration, rather than indefinite management of an unresolved crisis.

The logic is thermodynamic and governance-based simultaneously. CDR — whether direct air capture, enhanced weathering, ocean alkalinity enhancement, or bioenergy with carbon capture — actually removes the problem rather than masking it. An atmosphere with lower CO₂ concentration does not require permanent management; it does not create termination shock; it does not require international governance of a continuously operating deployment system. CDR is expensive and slow relative to the urgency of the crisis — current direct air capture technology costs around $300–$1000 per tonne of CO₂, and removing meaningful fractions of atmospheric CO₂ would require deployment at a scale that does not currently exist. But these are constraints on resources, technology, and political will — exactly the constraints that a serious international research and deployment program could address.

The CDR position also argues that solar geoengineering is being treated as a serious option partly because it seems cheap relative to the scale of the problem. SAI could reduce global average temperature at a cost of billions per year — compared to the trillions required for full decarbonization. But this cost comparison is misleading: it compares the operating cost of a system that must run indefinitely with the transition cost of a system that, once built, runs on renewable energy. The apparent cheapness of SAI is the cheapness of not solving the underlying problem.

What this position costs: CDR at scale does not exist yet, and the gap between current CDR deployment and the scale needed to make a meaningful difference in atmospheric CO₂ may be wider than the time available to close it. In the meantime, the people who are already suffering from warming cannot wait for a technological path that is decades from deployment at scale.

Where the debate actually is

Most participants in the solar geoengineering debate agree on more than the adversarial framing suggests. Almost everyone agrees that the climate emergency is real and that current emissions trajectories are inadequate. Almost everyone agrees that governance is necessary before any deployment could be legitimate. Almost everyone agrees that CDR is important and underfunded. The specific disputes are harder to navigate than they appear: they are disputes about sequencing (governance before or alongside research), about risk distribution (whose climate preferences would be satisfied by which deployment scenario), and about moral hazard (whether naming something a contingency plan reduces the urgency of preventing the condition that would make the contingency necessary).

The governance gap is the most tractable immediate problem and the one least being addressed. The international community's failure to develop even a framework for governing geoengineering research — not deployment, research — reflects the same collective action dynamics that have limited climate governance more broadly. The states with the most capacity to conduct research have the least incentive to submit to governance that would constrain their options; the states with the most at stake in how that research is governed have the least leverage to demand it.

What is genuinely contested — and not resolvable by more evidence — is the prior question: whether the scale of the climate emergency justifies taking seriously a technology that, at deployment scale, creates a permanent planetary dependency that no current institution is designed to manage. That question is not technical. It is a question about what kind of futures are acceptable, whose consent is required before a civilizational intervention, and whether the actors who failed to prevent the crisis are the ones who should be trusted to manage the response.