Engineering
the Sky

Pinatubo proved the physics. Nobody asked it to.

The mechanism is straightforward: sulfur dioxide injected into the stratosphere oxidizes into sulfate aerosols — tiny particles that scatter incoming solar radiation back into space before it can warm the surface. The stratosphere sits above weather, above rain, above the scrubbing action of the lower atmosphere. Particles that reach it linger for one to two years. Pinatubo demonstrated that a sufficiently large aerosol injection could measurably cool the entire planet in months.

The 1991 eruption cut global average temperature by approximately 0.5°C over the following twelve months. Crop yields fell in some regions. The ozone layer thinned. Rainfall patterns shifted. The summer of 1992 was the coldest in the Northern Hemisphere in decades. It also prevented — by rough calculation — perhaps 0.1°C of the warming that would otherwise have accumulated that year. Climate scientists noted all of this very carefully.

The phrase stratospheric aerosol injection — SAI — entered the technical literature before Pinatubo, but the eruption made it real. Here was proof that aerosols at altitude could alter the planet's energy balance. The question the eruption posed, and has never stopped posing, is: what happens when humans decide to do this deliberately?

The technology is simpler than you think. That is part of the problem.

The proposed method is almost mundane: high-altitude aircraft — modified versions of existing military tankers — would fly continuous missions into the lower stratosphere, roughly 15 to 25 kilometers above the surface, and release sulfur dioxide or calcium carbonate. The aerosols disperse on stratospheric wind currents, spread globally within weeks, and form a diffuse reflective veil. The engineering challenge is not exotic. The logistics are comparable to a mid-size airline operation. A 2018 Harvard study estimated that a mature SAI program could deliver meaningful cooling for approximately $2.25 billion per year — cheap enough for a mid-income nation to fund unilaterally. That is not a reassuring number.

The science has sharpened considerably since Pinatubo. Climate models can now simulate the aerosol chemistry with reasonable fidelity. The rough consensus: injecting 1–5 million tonnes of SO₂ per year into the stratosphere could offset roughly 1–2°C of global warming. The distribution of that cooling, however, would not be uniform. The tropics would benefit most. Higher latitudes — especially the Arctic and Antarctic — would see less effect. The global temperature average would fall while regional climates shifted in ways that are hard to model precisely and impossible to negotiate in advance.

Field experiments are already underway. Make Sunsets, a California-based startup, began releasing sulfur-bearing balloons into the stratosphere in 2022 without regulatory approval — and without meaningful scientific controls. Harvard's SCoPEx project, a more rigorous balloon experiment that would have tested aerosol dispersion chemistry at small scale, was cancelled in 2024 after objections from Indigenous Sámi communities in Sweden, where the launch was planned. The science is advancing. The governance frameworks governing it are not.

The risks are not science fiction. They are in the physics.

The most discussed risk has a name: termination shock. If a large-scale SAI program is abruptly halted — due to geopolitical breakdown, war, funding collapse, or technical failure — the aerosols that have been suppressing warming would settle out of the stratosphere within 12 to 18 months. The underlying warming that has accumulated, masked but not removed, would then be released in a sudden surge. Models suggest a termination shock from a decade-long SAI program could expose the world to decades of warming compressed into two or three years — faster than ecosystems or agriculture can adapt.

A second risk sits in the monsoon systems. The Indian and West African monsoons are among the most sensitive and consequential regional climate systems on Earth, supplying water to roughly three billion people. Stratospheric aerosol injection alters the temperature differential between land and sea that drives monsoon circulation. Model studies are not in agreement on the direction or magnitude of the effect — some show strengthening, some weakening — but the uncertainty itself is the point. A program calibrated to benefit one hemisphere could disrupt rainfall on which billions of subsistence farmers depend.

Ozone chemistry is a third concern. Sulfate aerosols provide reactive surfaces that accelerate the destruction of stratospheric ozone, the layer that shields the surface from ultraviolet radiation. Pinatubo produced measurable ozone thinning at high latitudes. A sustained multi-year SAI program would do the same, potentially delaying the recovery of the Antarctic ozone hole that international action under the Montreal Protocol has been slowly achieving since 1987.

And then there is the question of what SAI doesn't fix. Aerosol injection cools the planet by reducing incoming sunlight. It does nothing about ocean acidification, which is driven by the concentration of CO₂ in the water regardless of temperature. A cooler ocean with the same atmospheric carbon load would still be acidifying at the rate that is dissolving the base of the marine food chain. SAI addresses the fever. It does not address the infection.

No one is in charge. Anyone could start.

There is no international treaty that governs stratospheric aerosol injection. There is no institution with the authority to approve, prohibit, or regulate it. The UN Environment Programme has issued statements. The IPCC has flagged it as a research area requiring governance frameworks. Various academic bodies have published principles and proposed oversight mechanisms. None of this has legal force. And the technology does not require international cooperation to deploy.

This creates a scenario that political scientists call the "free driver" problem — the mirror image of the free rider problem that bedevils climate cooperation. In conventional climate negotiations, every country has an incentive to let others reduce emissions while continuing to emit freely. In SAI governance, the dynamic inverts: a single nation or even a wealthy private actor could unilaterally deploy a program that cools the entire planet, reaping the benefits of their preferred climate outcome while imposing costs — altered monsoons, disrupted agriculture, ozone loss — on nations that had no vote.

The scenario is not hypothetical in the abstract. Make Sunsets has already demonstrated that a startup with minimal capital can begin stratospheric experiments without regulatory oversight. As warming intensifies and its costs — floods, droughts, crop failures, heat mortality — become politically unbearable, the pressure on governments of vulnerable nations to act unilaterally will increase. The 2020s and 2030s are the period during which the governance gap either gets closed or becomes permanent.

The scientific community is increasingly divided not about whether SAI could work — the physics is not seriously disputed — but about whether research should proceed at all, given that demonstrating efficacy accelerates deployment pressure. Some researchers argue that a world hurtling past 2°C will need every option. Others argue that the mere existence of a cheap technofix licenses continued fossil fuel use and removes pressure to decarbonize. What both sides agree on is that this decision — whether to begin engineering the sky — may be the most consequential governance choice of the 21st century, and it is currently being made by default, in the absence of institutions capable of making it deliberately.