Article

The Technical Feasibility and Costs of SAI

Stratospheric aerosol injection (SAI), an idea to add millions of tonnes of reflective particles to the upper atmosphere, seems to offer enormous leverage over the climate, as a relatively small amount of material could produce a substantial cooling effect. But what are the technical and economic challenges of deploying SAI at a large scale, and who might be able to do it?

Key takeaways

  • A large planetary-scale deployment to slow or halt global warming would require a fleet of specialised aircraft, which would take many years to develop and cost tens of billions of dollars per year to build and operate.
  • A lower altitude polar deployment over the Arctic and Antarctic regions could be carried out using modified versions of existing aircraft, and so have a lower up-front cost but would be less efficient.
  • While such large deployments could only be implemented by powerful states, a broad range of actors, including rich individuals, could carry out much smaller deployments which would have little effect on the climate.

SAI has been described by some as “easy” – a quick, simple technical fix that can reduce climate risks at low cost. In this vein, researchers have speculated about small island states threatened by extreme weather and sea level rise – Tuvalu,1 for instance – deploying SAI on their own, or even billionaires – “Greenfingers”2 in the mould of the James Bond villain Goldfinger – taking it upon themselves to “save the world” by implementing the technology. 

But how easy would it be to deploy SAI in technical and economic terms?3 

Implementation: methods and scenarios 

One reason that SAI has the potential for a substantial cooling effect is that the aerosols – tiny reflective particles – that would be released would have a much longer lifetime in the stratosphere compared to the lower atmosphere. The dry conditions in the stratosphere mean that clouds and rain would not clean out the particles, and strong winds in the stratosphere would help to spread the particles around the world and towards the poles.

Stratospheric aerosol injection (SAI)

Tiny particles released in the stratosphere would directly reflect a small fraction of sunlight.

Sunlight

It would be possible to create a global layer of particles with a lifetime of 1–2 years by releasing sulphur dioxide (SO2) or other particles in the tropical or subtropical stratosphere at around 20 km altitude.4 However, this is far above the flight ceiling of most aircraft – around 13 km – and so would require special means of delivering the particles. Outside of the sub-tropics, the height of the stratosphere is much lower, dipping down as low as 7 km at the poles, but particles introduced at these lower altitudes would have shorter lifetimes and so a smaller cooling effect.5 

Researchers have looked at several possible mechanisms for delivering aerosols to the stratosphere to implement SAI, including high-altitude balloons, tethered hoses, and rockets.4 However, engineering assessments consistently suggest that a fleet of specially designed aircraft would be the best way to deliver aerosols to the stratosphere.6,7

The technical requirements for SAI implementation depend on the scale of the deployment. Here, three illustrative possibilities are discussed: a micro-scale deployment or “stunt” with no significant climate impact, a low-altitude polar deployment aiming to achieve 0.1°C of global cooling, and a high-altitude planetary-scale deployment intended to achieve 1°C of global cooling.

Micro-scale SAI deployment 

Imagine a scenario in which an actor wanted to release a planeload of SO2 – say, 5 tonnes – into the stratosphere to capture the world’s attention and force SAI onto the international agenda. Such a “micro-scale” release would be around one millionth the scale needed to produce a substantial global cooling effect and would have no meaningful effect on the climate system. This kind of one-off deployment could be achieved by buying a commercially available business jet and modifying it to carry and release SO2 for around $30 million.8

A G650 business jet in flight.

A stunt of this nature could be carried out today. A wide variety of actors, including large and small countries, corporations, and even individuals could pursue it. Again, however, a micro-scale release like this would be primarily symbolic and scaling up a deployment with business jets would not be a practical route to achieving a substantial cooling.4

Low-altitude polar SAI deployment 

The lower height of the stratosphere over the poles compared to the tropics and subtropics puts it within reach of existing aircraft, making larger polar deployments based on current technology achievable. A deployment limited to the Arctic but aimed at reducing global temperatures by 0.1°C would be possible using a fleet of modified commercial aircraft (such as the Boeing 777)9 injecting roughly 2 million tonnes of SO2 into the stratosphere per year.10 

Such a deployment would not necessarily be advisable since it would be inefficient and so carry greater side effects per unit cooling. Furthermore, if such an asymmetric deployment of SAI were scaled up to offset a large amount of warming, it would produce large shifts in tropical rainfall patterns.11 

Preparing for such a deployment would take at least a decade, however.9 Planes would need to be modified – including by installing tanks, nozzles, and a plumbing system – and airbases would need to be built.9 Costs would be in the billions of dollars per year.12 

Planetary SAI deployment 

Now imagine a scenario in which a deployment is intended to lower global temperatures by 1°C. Unlike a polar deployment, a full planetary deployment would require stratospheric injection in the tropics.5 Implementing this would therefore necessitate manufacturing a fleet of hundreds of specially designed high-altitude, high-payload aircraft capable of reaching the stratosphere at locations closer to the equator.4 Such a fleet would need to inject roughly 12 million tonnes of SO2 per year.13 

A diagram of a plane from four angles with a thin fuselage, wide wings, and four engines.

A schematic for a hypothetical stratospheric aerosol delivery aircraft (Diagram: Climatic Change)

 

A fleet of this type could be developed by major aerospace companies like Boeing and Airbus, since they are uniquely capable of designing, developing, and manufacturing the types of airframes and engines that would be required for a large-scale planetary deployment based on existing technology.14 The primary technical challenge for such companies would be designing and building engines capable of sustained combustion at high altitudes.

The costs of a large planetary deployment using a fleet produced by established airframe and engine manufacturers are estimated to be in the tens of billions of dollars per year.6 Developing and building aircraft and associated infrastructure would take at least two decades.15

It should be noted that very rapid developments in aerospace technologies are possible and have been achieved during the space race and in wartime. Whether new technology development pathways or sufficient political will for a rapid development program will materialise, however, is unclear. 

Feasible for whom? 

Thus, the larger the SAI intervention that is envisioned, the more difficult and costly it would be. Stunts can be performed today, but polar and especially global deployments depend on technical advances and infrastructure buildout occurring over decades. Yet, although implementing large-scale SAI would not be easy, neither would it face insurmountable technical obstacles. 

The technical and economic resources necessary to deploy SAI, particularly at larger scales, are not spread evenly around the world, rather they are concentrated in rich, powerful countries. At the same time, rich, powerful countries are generally less vulnerable to climate change than developing countries. This begs the question: what, if anything, can be done to ensure that those who would have the ability to implement SAI take into account the views and interests of those lacking such capabilities? 

SAI scenarios, technical feasibility, and costs

Scenario Goal Amount of SO2 Technical Requirements Cost
Micro-scale SAI deployment Provoke reaction Tonnes Modified business jet Around $30 million
Low-altitude polar deployment Reduce global temperatures by 0.1°C, deploying only in the Arctic About 2 million tonnes per year Modified commercial airliners Billions of dollars per year
Planetary deployment Reduce global temperatures by 1°C About 10 million tonnes per year New, specialised aircraft Tens of billions of dollars per year

Open questions

  • Which states would have the capacity to develop and deploy planetary SAI?
  • Which states could stop others from deploying SAI, and how?
  • If a state wanted to develop and deploy SAI as quickly as possible, how long would it take them to do so?

Ask us a question!

Ask a Question Form

Endnotes

  1. Millard-Ball A. (2012). The Tuvalu syndrome: can geoengineering solve climate’s collective action problem? Climatic Change. 110(3):1047-66. https://doi.org/10.1007/s10584-011-0102-0  
  2. Victor DG. (2008). On the regulation of geoengineering. Oxford Review of Economic Policy. 24(2):322-36. https://doi.org/10.1093/oxrep/grn018  
  3. The societal challenges and costs associated with SAI, such as mitigation displacement, are not considered here. 
  4. Smith W, Wagner G. (2018). Stratospheric Aerosol Injection Tactics and Costs in the First 15 Years of Deployment, Environmental Research Letters 13: 124001. https://doi.org/10.1088/1748-9326/aae98d 
  5. Zhang Y, MacMartin DG, Visioni D, et al. (2024). Hemispherically symmetric strategies for stratospheric aerosol injection. Earth System Dynamics. 15(2):191-213. https://doi.org/10.5194/esd-15-191-2024 
  6. Smith W. (2020). The Cost of Stratospheric Aerosol Injection Through 2100. Environmental Research Letters 15: 114004. https://doi.org/10.1088/1748-9326/aba7e7  
  7. Lockley A, MacMartin D, Hunt H. (2020). An update on engineering issues concerning stratospheric aerosol injection for geoengineering. Environmental Research Communications. 2(8):082001. https://doi.org/10.1088/2515-7620/aba944 
  8. Indeed, the company Make Sunsets is currently (and controversially) conducting similar micro-scale SAI interventions using weather balloons; this approach, however, cannot be feasibly scaled. 
  9. Smith W, Bartels MF, Boers JG, et al. (2024). On Thin Ice: Solar Geoengineering to Manage Tipping Element Risks in the Cryosphere by 2040. Earth’s Future 12. https://doi.org/10.1029/2024EF004797  
  10. Results of Lee et al. 202311 found a cooling efficiency of 0.1°C per million tonnes of SO2 for global, high-altitude SAI and 0.06°C per million tonnes for Arctic-focused SAI. 
  11. Lee WR, MacMartin DG, Visioni D, et al. (2023). High‐latitude stratospheric aerosol injection to preserve the Arctic. Earth’s Future. 11(1):e2022EF003052. https://doi.org/10.1029/2022EF003052 
  12. This estimate is approximate to one-tenth the cost of the $35 billion deployment described in Smith et al. 2024.9
  13. Haywood J, Tilmes S, Keutsch F, et al. (2022) Chapter 6: Stratospheric Aerosol injection and its Potential Effect on the Stratospheric Ozone Layer. In: Scientific Assessment of Ozone Depletion 2022. pp. 325–375. https://csl.noaa.gov/assessments/ozone/2022/downloads/Chapter6_2022OzoneAssessment.pdf 
  14. Horton JB, Smith W, Keith DW. (under review at Global Policy) Who Could Deploy Stratospheric Aerosol Injection? The US, China, and Large-Scale, Rapid Planetary Cooling. Preprint available at https://media.rff.org/documents/HORTON_paper2.pdf 
  15. Smith W. (2024). An assessment of the infrastructural and temporal barriers constraining a near-term implementation of a global stratospheric aerosol injection program. Environmental Research Communications. https://doi.org/10.1088/2515-7620/ad4f5c 

Citation

Josh Horton, Pete Irvine (2024) - "The Technical Feasibility and Costs of SAI" Published online at SRM360.org. Retrieved from: 'http://srm360.org/article/technical-feasibility-and-costs-of-sai/' [Online Resource]

Reuse this work freely

All visualisations, data, and code produced by SRM360 are open access under the Creative Commons BY license. You are free to use, distribute, and reproduce these in any medium, provided that SRM360 and the authors are credited.

Data produced by third parties and made available by SRM360 is subject to the licence terms of the original third-party authors. We will always indicate the original source of such data in our documentation, so please review the licence of any third-party data before use and redistribution.