Solar Geoengineering: Emergency Lever or Planetary Gamble?

Core Concepts

What does "solar geoengineering" actually mean?

The terminology used in this space matters more than it might seem. Terms like "geoengineering," "climate intervention," "solar radiation management" (SRM), "stratospheric aerosol injection" (SAI), and "carbon dioxide removal" (CDR) are not interchangeable. Scholars emphasize that conflating them leads to policy confusion, because SRM and CDR carry different technical profiles, different reversibility characteristics, and raise different governance challenges.

This module focuses on solar radiation management — specifically on two leading approaches:

• Stratospheric Aerosol Injection (SAI): releasing reflective particles (typically sulfur dioxide) into the stratosphere to scatter incoming solar radiation back to space.
• Marine Cloud Brightening (MCB): seeding low-lying maritime clouds with sea salt or other particles to increase their reflectivity.

Neither approach reduces atmospheric CO2. Both reflect sunlight. Beyond that, they work quite differently and carry distinct risk profiles.

---

How SAI works — and why volcanoes are only a partial guide

SAI cools the Earth by increasing the stratosphere's reflectivity, reducing the amount of solar energy reaching the surface. The physical mechanism is analogous to volcanic eruptions, which inject sulfate aerosols into the stratosphere. The 1991 eruption of Mount Pinatubo is the canonical natural analog: it cooled the global average temperature by approximately 0.4–0.5°C for 18–36 months, demonstrating the scale of cooling that stratospheric aerosols can produce.

Proposed deployment platforms include high-altitude aircraft, tethered balloons, and rockets. Research suggests aircraft-based delivery systems are the most technically feasible, and developing purpose-built high-altitude tankers is neither particularly difficult nor prohibitively expensive. This low barrier to entry is, as we will see, central to the governance problem.

But volcanic eruptions are imperfect models for sustained SAI. Volcanic events are transitory, lasting months to a few years before aerosols settle out of the stratosphere. Sustained SAI would continuously inject aerosols, forcing the atmosphere toward a new chemical equilibrium — a fundamentally different scenario with effects that transient volcanic analogs cannot fully predict. Recent sectional aerosol-chemistry-climate models have identified microphysical and chemical processes that were missing from earlier modeling work, indicating that existing guidance may have underestimated SAI's full climate effects.

SAI also differs from simply "dimming the sun." Injecting aerosols causes heating of the stratosphere itself, which alters energy balance and atmospheric circulation in ways that would not occur if you simply reduced incoming solar radiation without adding aerosols. The location, altitude, season, and magnitude of injection all affect how aerosols behave and what regional climate patterns result.

---

How MCB works

Marine cloud brightening operates at much lower altitudes. Ships or aircraft seed low marine clouds with fine sea-salt particles, increasing cloud droplet number, which makes clouds brighter and more reflective. The cooling effect is regional rather than global in the first instance, and highly dependent on deployment strategy.

Research documents that the regional climate impacts of MCB show extreme sensitivity to the geographic location, areal coverage, intensity, and timing of cloud-brightening interventions. The same technology applied in different ocean regions or at different intensities could produce vastly different outcomes for precipitation, sea-level patterns, and circulation elsewhere in the world.

---

Compare & Contrast

SAI vs. MCB: Two different tools with different risk profiles

Dimension	Stratospheric Aerosol Injection (SAI)	Marine Cloud Brightening (MCB)
Altitude	Stratosphere (~15–25 km)	Low troposphere (near-surface clouds)
Reach	Globally distributed	More regional, depends heavily on deployment
Reversibility	Aerosols settle out in 1–3 years if stopped	Effects dissipate relatively quickly
Ozone risk	Significant — sulfate aerosols catalyze ozone destruction	Lower direct ozone risk
Monsoon risk	Moderate to significant, especially South Asian	High, including ENSO suppression
Deployment barrier	Low — adapted high-altitude aircraft	Moderate — requires fleet of ships or aircraft
Research maturity	More studied; still large uncertainties	Less studied; viability unresolved

SAI vs. "just reducing solar output"

It might seem intuitive that SAI is equivalent to "turning down the sun." It is not. Because aerosols absorb infrared radiation, they heat the lower tropical stratosphere as an unavoidable physical consequence — a heating that does not occur if solar forcing is reduced by other means. This stratospheric heating, in turn, drives ozone chemistry impacts and alters atmospheric circulation in ways that create distinct regional climate patterns.

---

Common Misconceptions

Misconception 1: "The Pinatubo eruption proves SAI works and is safe."

What the evidence shows: Pinatubo does demonstrate the cooling power of stratospheric aerosols. But volcanic eruptions are transitory events. Sustained SAI would force the atmosphere into a new chemical equilibrium that differs fundamentally from temporary volcanic perturbations. The analogy provides useful physical intuition, but cannot substitute for dedicated SAI climate modeling.

---

Misconception 2: "Talking about geoengineering undermines public support for cutting emissions."

What the evidence shows: This is the "moral hazard" concern — the idea that a perceived backup plan reduces urgency. Multiple large-scale empirical studies — including a pre-registered, money-incentivized experiment with 2,500 US participants — have found little to no evidence that exposure to geoengineering information actually reduces support for emissions reductions or mitigation policies. Some studies have found the opposite effect: information about SRM can increase stated support for carbon taxes.

The more significant moral hazard risk may not be citizens reducing their climate concern — it may be policymakers preemptively blocking beneficial options because they *assume* citizens will react badly. In experimental economics games, policymakers declined to use geoengineering even when it would benefit everyone, because they incorrectly anticipated public moral hazard that did not materialize.

This distinction between individual-level and institutional-level moral hazard matters for governance design. Scholars identify regulatory drift — the gradual weakening of oversight frameworks through complacency or deliberate deregulation — as a more durable institutional risk than citizens reducing their personal climate action.

---

Misconception 3: "SAI would at worst reduce rainfall slightly, which we can adapt to."

What the evidence shows: The precipitation effects of SAI are poorly understood, regionally variable, and in some models, severe. Climate models frequently disagree on the sign (increase vs. decrease) of precipitation changes over critical regions including India and the Tibetan Plateau. This is not a measurement precision problem — it reflects genuine gaps in understanding how SAI affects atmospheric circulation and moisture transport.

The pattern that emerges across models is a counterintuitive redistribution: wet regions tend to become drier and dry regions wetter under SAI relative to greenhouse gas-only scenarios. For regions that depend on monsoon rainfall, this asymmetry is not a minor adaptation challenge.

---

Misconception 4: "We can always just stop if something goes wrong."

What the evidence shows: Stopping is the hard part. See: Termination Shock, below.

---

Misconception 5: "Sulfate aerosols are the only option, so ozone risk is unavoidable."

What the evidence shows: Calcium carbonate (CaCO3) particles have been proposed as an alternative to sulfate aerosols that could reduce or even reverse ozone depletion. Alkaline CaCO3 particles chemically convert ozone-depleting species into stable salts, theoretically enabling SAI while simultaneously increasing column ozone. Experimental measurements have been conducted to constrain ozone-response modeling. However, this remains largely theoretical and would require further investigation into potential unintended consequences.

---

Boundary Conditions

Where do these approaches break down?

Termination shock: the dependency trap

The single most consequential structural risk of SAI is what happens if deployment stops. If SAI is masking a high level of accumulated greenhouse gas warming and is abruptly terminated, the global climate would rapidly warm toward the temperatures that would have existed without SRM deployment. Research indicates the rate of warming following abrupt termination could be approximately four times larger than the rate caused by ongoing anthropogenic climate change alone.

This is not just a temperature spike. Termination triggers simultaneous rapid changes in precipitation, sea level rise, land drying, and weakened carbon sinks — a multi-variable climate shock compressed into years rather than decades.

We have already observed a real-world analog. When international maritime regulations reduced shipping sulfur dioxide emissions by approximately 80 percent in 2020, this inadvertent termination of accidental aerosol forcing doubled the warming rate in the 2020s compared to the warming rate since 1980. Termination shock is not theoretical.

Termination does not require a voluntary decision to stop. Triggers can include political instability, economic collapse, war, pandemic disruption, or geopolitical conflict. There are no globally agreed termination protocols, and existing international law is unlikely to constrain use of force in response to unilateral deployment decisions.

Technological lock-in

The termination shock dynamic creates an intergenerational governance problem. Once SAI is deployed at scale, continuous operation becomes mandatory. Future generations inherit the obligation to maintain a planetary intervention they had no role in initiating. Future decision-making becomes structurally biased toward programmatic expansion rather than voluntary termination.

Ecosystem adaptation limits

Rapid warming from termination shock exceeds the migration and adaptation capacity of many ecosystems. When climate velocities — the speed at which climate conditions shift spatially — exceed the migration rates of foundational species, ecosystems face disruption. Sudden termination following decades of artificial cooling creates unprecedented climate velocity that overwhelms evolutionary adaptation timescales.

MCB viability is unresolved

An international consensus of leading MCB researchers concludes that MCB's scientific viability depends entirely on whether observations and models can demonstrate that equitable geographic distribution of benefits and risks is achievable. That question is currently unanswered. Without a solution to the equitable deployment problem, MCB viability remains scientifically open.

MCB also risks suppressing the El Niño-Southern Oscillation (ENSO), producing a permanent La Niña-like climate state. ENSO affects global weather patterns, fisheries, agriculture, and water availability across multiple continents. A persistent shift toward La Niña conditions would redistribute rainfall globally, creating winners and losers in ways that no governance body currently has the authority to arbitrate.

Monsoon disruption

Both SAI and MCB risk disrupting Asian and African summer monsoons. The South Asian monsoon contributes approximately 80% of total annual rainfall in India. SAI-induced precipitation reductions would affect soil moisture, agricultural productivity, freshwater availability, and ecosystem services for billions of people in already water-stressed regions. Many of those regions would have little voice in decisions made by the states capable of deploying these technologies.

Ozone depletion

Sulfate aerosols provide surfaces in the cold polar stratosphere on which dormant chlorine gases can be activated. Research indicates sulfate SAI could destroy 25–75% of the ozone layer above the Arctic and significantly delay Antarctic ozone hole recovery. The sulfate heating of the lower tropical stratosphere also increases stratospheric water vapor, creating a positive feedback loop that further accelerates ozone loss. This is not a peripheral side-effect — it is a fundamental chemical consequence of using sulfate aerosols that cannot be eliminated by adjusting injection parameters within the tropics.

Injection latitude matters

The latitude of aerosol injection significantly affects the distribution of side effects. Tropical injections produce the strongest undesirable circulation and precipitation changes, while injections at higher latitudes or near the poles substantially reduce these adverse regional effects. This finding has important design implications but does not eliminate the underlying risks.

---

The governance boundary: who decides?

The free-driver problem

In climate negotiations, the central collective action challenge is the "free rider" — the temptation to benefit from others' mitigation without contributing. Solar geoengineering creates the opposite problem: a "free driver." A single nation — or even a non-state actor — could unilaterally deploy SAI at relatively low cost, imposing climate and environmental risks on other nations without their consent.

SAI is both technically feasible and financially within the means of many individual states today. The development of a purpose-built high-altitude tanker fleet is within the technical and budgetary reach of numerous governments. Existing international law provides only weak constraints on unilateral SAI efforts. Unlike a nuclear weapon, you cannot sanction your way to stopping a state from flying aerosol-laden aircraft in its own airspace.

The geopolitical consequences of unilateral deployment could be severe. States would likely view unilateral SAI as a national security threat, and the literature raises the possibility that unilateral deployment could, in extreme cases, trigger armed conflict. Uncoordinated deployment creates international tensions even short of armed conflict, particularly when some regions benefit while others experience adverse agricultural impacts from altered precipitation.

The governance gap

No comprehensive, legally binding international governance framework exists to regulate geoengineering research or deployment at a global level. Existing instruments cover partial terrain:

• The Convention on Biological Diversity (CBD) established a moratorium on geoengineering, but it is non-binding. The CBD COP lacks authority to issue legally enforceable rules. The moratorium uses qualified language ("invites" states to "consider guidance"), not mandatory language. The 2024 COP16 reaffirmed the moratorium, underscoring its ongoing relevance — and ongoing toothlessness.

• The London Protocol (2013 amendment) is the only international instrument specifically focused on regulating geoengineering — but only marine geoengineering. As of the most recent data, only 9 of the 55 parties required for entry into force have ratified it, meaning even this limited tool remains non-operational.

• The World Meteorological Organization (WMO) has responsibility to evaluate impacts of SRM on weather and climate, but lacks any regulatory mandate or governance capacity beyond assessment. There is a mismatch between the institution best positioned to understand atmospheric impacts and the authority to constrain them.

The IPCC AR6 characterizes SRM governance as facing "substantial risks and institutional and social constraints to deployment related to governance, ethics, and impacts on sustainable development." The fragmented institutional landscape means no single body can coordinate across environmental, health, security, and equity dimensions simultaneously.

Outdoor experiments are expanding without sufficient oversight

Outdoor geoengineering experiments are proliferating — particularly marine geoengineering field tests — with insufficient international regulatory coordination. The repeated reaffirmation of the CBD moratorium at successive COP meetings (2016, 2024) signals that governance gaps are not being closed.

The most prominent stratospheric experiment, the SCoPEx project at Harvard, designed to measure stratospheric aerosol behavior in plumes, created an external advisory committee in 2019 to develop governance norms covering transparency, risk management, and stakeholder engagement. The project was ultimately cancelled in 2024 when the principal investigator announced he would no longer pursue it. The governance model it developed remains a relevant precedent even so.

The risk-risk framework

Geoengineering is sometimes framed as a straightforward "lesser evil" choice: solar geoengineering risks vs. unchecked climate change risks. The evidence does not support this framing. Risk-risk assessments show that SAI could reduce some climate extremes (like heat waves) while simultaneously creating or worsening others (precipitation extremes, regional droughts). The distribution of who bears the harms and who gains the benefits differs substantially across regions and time periods. Governance frameworks must address distributional justice explicitly, not assume that global average temperature reduction constitutes a global benefit.

---

Governance principles that have emerged

Two non-binding frameworks have shaped the academic governance discussion:

• The Oxford Principles (2012) establish that geoengineering should be governed as a public good, with transparent research, public participation, independent risk assessment, and explicit ethical review before deployment.

• The Tollgate Principles (2018) refine the Oxford Principles by emphasizing democratic participation, distributive justice, and consultation with affected parties — particularly Global South nations that face asymmetric impacts but have been underrepresented in governance discussions.

Both remain non-binding academic proposals. Governance activity and expertise remain disproportionately concentrated in the Global North, even as the populations most exposed to transboundary harms are predominantly in the Global South.

---

Thought Experiment

It is 2038. Global temperatures have risen 1.9°C above pre-industrial levels. Major coral reef systems have effectively collapsed. A coalition of small island nations and low-lying coastal states — facing existential flooding and repeated crop failures — announces that it will begin stratospheric aerosol injection within six months. The coalition has acquired a fleet of adapted high-altitude aircraft. Their stated plan is to deploy enough SO2 to offset 0.5°C of warming over a ten-year period.

Three months before the announced start date, satellite imagery confirms the aircraft are operational.

Consider:

1. Is the coalition's action morally legitimate? Does the severity of what they face change your answer? Does the absence of any international governance framework change it?

2. What would the main objections be from nations whose monsoon rainfall might be affected? From nations that have not yet ratified any related international treaties? From future generations who would inherit the termination shock risk?

3. If you were designing an emergency governance framework in the next 90 days — knowing that full treaty negotiation is impossible in that timeframe — what would the minimum viable components of such a framework need to include? Who would need to be at the table for it to have any legitimacy?

4. The "arm the future" argument holds that geoengineering research should proceed as insurance against exactly this scenario. Does the scenario validate that argument, complicate it, or both?

There is no clean answer. The point is to reason through the tradeoffs with what you now know — including what you know about what we do not know.