On September 9th, 2025, we published the article Safeguarding the polar regions from dangerous geoengineering: a critical assessment of proposed concepts and future prospects in the scientific journal Frontiers in Science. The authors are 42 experts with knowledge of polar and climate sciences, policy and governance. Over the course of two years, we reviewed five polar geoengineering concepts that have been proposed to reduce the adverse effects of climate change and are gaining increasing attention in the public and policy spheres. These five concepts are labeled as “polar” as they intend to reduce climate change impacts in polar regions and/or would be deployed in the polar regions. One of the five concepts we examined is Stratospheric Aerosol Injection, or SAI, a type of Solar Radiation Modification (SRM for short).
SAI is the concept of injecting aerosols into the stratosphere to reflect incoming solar radiation, resulting in a net cooling effect. The concept is based on observations during and after past volcanic eruptions. For instance, when Mount Pinatubo erupted in 1991, it emitted 17 megatons of sulfate aerosols into the stratosphere, resulting in a cooling of global mean temperatures by 0.5 °C over the course of two years (McCormick et al., 1995). On the face of it, this seems like a simple enough idea; repeat artificially what a volcano does naturally, and we have got an easy cooling thermostat for our overheating planet, right?
Except, this is a vast oversimplification. What we know from the Mount Pinatubo eruption, for instance, is that this event resulted in severe weather disruption, including the disturbance of the South Asian monsoon rains and moderate to severe drought (Trenberth and Dai, 2007) and reduced crop yields (Proctor et al., 2018). A further example is that coal-fired power plants, releasing a lot of sulfate pollution into the atmosphere, resulted in acid rain damage to forests in Europe and North America until scrubber technology was introduced in the 1980s. We therefore already know that manipulating the chemical and physical properties of our atmosphere comes with a wide range of negative environmental impacts.
In our published assessment, we evaluated SAI using six criteria: could it be achieved (feasibility)? What are the negative impacts (unintended consequences)? Could it be scaled to global levels (scaling)? Could it be deployed in time to offset the immediate climate crisis (timing)? Who has responsibility and accountability (governance)? And crucially, is this concept introducing a promise of a solution without being able to deliver (false hope)?
Let us dig into the first five criteria a little:
1. Feasibility
- To be effective in reducing global temperatures, SAI would require sustained global-scale injections of sulfates into the stratosphere. Any cessation of activity (for example, from war or geopolitical conflict) would result in “termination shock” as temperatures rapidly return to the levels corresponding to the existing GHG concentrations in the atmosphere at the time over the course of 1 to 2 decades. This rapid temperature jump would be too fast for many human communities and ecosystems to adapt to.
2. Timing
- SAI would require a fleet of specialized delivery planes that exceed current technology. The technique would not be ready to deploy in the polar regions (where the stratosphere is at a lower altitude) for at least another 10 years, whilst estimates for global deployment are twice as long. According to the latest assessments (Forster et al., 2025), 1.5 °C would be crossed in ~2 years.
3. Costs and upscaling
- Estimated costs conservatively reach $1.7 billion (USD) per year, excluding the costs needed for monitoring or liability for damages, and would need to be sustained over decades to centuries to avoid termination shock.
4. Unintended consequences
- SAI deployment would induce highly unpredictable and non-uniform changes in atmospheric circulation, regional precipitation and regional temperatures.
- For the same global temperature, a world with SAI and a world without SAI are fundamentally different. SAI creates an artificial climate with new regional risks and extremes. Certain regions will show enhanced warming, or drying as a result of the induced atmospheric circulation changes. A recent study using Earth System Models shows that a world at 2 °C without SAI is more desirable than a world at 1.5 °C with SAI deployment (Hwong et al., in prep), although neither alternative is desirable, or for that matter necessary, given other available pathways to 1.5 °C this century without SAI.
- SAI deployment would interfere with seasonality and disrupt water availability for agriculture, biodiversity and food security.
- Reduced incoming solar radiation would induce crop stress, resulting in reduced crop growth and yield.
- Reduced insolation would impact solar energy production (Baur et al., 2024).
- SAI could worsen stratospheric ozone depletion, with adverse effects on human health and ecosystems.
- Inhalation of sulfuric acid aerosols can create respiratory health issues and increase global health costs.
- Ocean acidification would carry on unabated (without emissions reductions), and would be exacerbated through acid rains, beyond critical levels for shelled organisms, especially in polar oceans, where ocean acidification is already causing shell damage. Ocean acidification is incredibly long-term, taking 30-50,000 years to return to today’s levels.
5. Governance challenges
- SAI will create challenging situations in terms of liability and compensation. It is unclear who would be responsible for fisheries loss from ocean acidification, forestry losses or human respiratory health impacts due to acid rains, loss in crop productivity due to reduced rainfall or solar radiation.
- Linked to the issue of liability, the impacts due to climate extremes will be difficult to attribute. In the case of an extreme drought event, for example, would this drought have been intensified by SAI, or would it have been worse without SAI? There is an interesting parallel we could make with an existing case of weather intervention in Hungary, which deploys cloud seeding systems to avert hail destruction of plantation. Droughts in the eastern part of the country resulted in threats of violent repercussions towards system operators (without proof) who were blamed for these droughts, resulting in a halt of all operations (article in Hungarian press). Tensions such as these would likely arise, at the global scale, with international actors.
- Existing environmental agreements are very complex, and global cooperation among signatories (e.g., the Paris Agreement under UNFCCC) is, at most times, a challenge (a good example is the lengthy and difficult negotiations each year at the climate COPs). Creation of an international regulatory agreement for the management and monitoring of possible SAI deployment would be far more complicated; it would require knowledge of which actors would be responsible for deployment, which actors would be responsible for adverse impacts, and so on.
- Due to the risk of termination shock, SAI deployment would have to be carried out over multiple generations, and until greenhouse gas concentrations are reduced to an appropriate and commensurate level with atmospheric temperature. The timescales required for effective governance can be compared to the debates that have arisen from the need to dispose of nuclear waste.
- SAI global deployment would require a level of geopolitical stability over a significant (potentially multi-decadal) period that does not exist today.
From our in-depth assessment of the existing scientific literature, we concluded that we know enough about SAI to demonstrate how it does not pass any of the criteria listed above and that the risks of SAI far outweigh the hypothetical benefits to be had. SAI is not effective, nor desirable, due to its damaging environmental impacts, its cost (both direct and indirect), its negative and unpredictable unintended consequences, and its insurmountable global governance. More information can be found in the article and the accompanying policy brief.
The risk of false hopes regarding SAI research and eventual development is real. Many of the arguments that are brought up by proponents to pursue further research fall into that “false hope” category, and we debunk a few of the most common ones below.
False premise #1: SAI is a quick fix to our warming climate
Both sides of the SAI argument agree that we are nowhere close to deployment. Research is immature and currently relies on simplistic modeling experiments. Let us imagine that, for the sake of argument, we agreed with the premise that SAI could be desirable. A great amount of additional research would be needed to understand whether the benefits of SAI could outweigh the negative impacts of SAI, if possible at all. Such work would largely be computer-based, given the inherent global risks of field trials, making it unlikely that we can ever arrive at a satisfactory science-based position to proceed (Opinion on solar radiation modification – Ethical perspectives, Publications Office of the European Union, 2024). In addition to this research lead time, a minimum of 10 years would be needed to build a fleet of planes capable of injecting enough sulfates high enough in the stratosphere. SAI is inherently not a quick fix.
False premise #2: SAI is cheap
As stated above, costly specialized fleets of airplanes would have to be developed and maintained over decades. Furthermore, SAI can only attenuate one impact – temperature – not non-temperature-related impacts (e.g., ocean acidification, committed sea-level rise). And since studies show that deployment would likely result in non-uniform warming in certain parts of the world, there would be regional adverse impacts. Costs of SAI should be compared to the cost of mitigation, not the cost of damages from climate change. Indeed, SAI can only reduce costs for damages that could potentially be reduced/prevented through deployment. SAI, and SRM as a whole, cannot reduce past damages, and cannot reduce all future damages (like those caused by acidification or unavoidable sea level rise mentioned before). Furthermore, additional costs arising from damages to fisheries, crops, etc, would be incurred. SAI will inherently not be cheap.
False premise #3: SAI just requires that we develop the right governance
Global cooperation for mitigation has been tried for years, and it has been painstakingly difficult and slow. Why would global cooperation be any easier for SAI? Indeed, some have argued that SRM is effectively ungovernable (Biermann et al., 2022). To avert termination shock, we would need a governance framework that works globally, across geopolitical turmoil and lasts for many generations. Without a global agreement on banning SAI, there is a danger that rogue states or actors begin experimenting with our atmosphere in ways that are unaccountable. The development of an effective governance framework for SAI is unlikely to be achievable.
False premise #4: Mitigation is not happening fast enough
Although current policies are insufficient to guarantee 1.5 °C compliance, these policies do provide a 1 in 5 chance of limiting long-term warming to 1.5 °C. And they substantially increase the likelihood of limiting warming to below 2 °C (4 in 5 chance). This adheres to the goals of the Paris Agreement to “holding the increase in the global average temperature to well below 2°C above pre-industrial levels and pursuing efforts to limit the temperature increase to 1.5°C above pre-industrial levels”. We need to focus on accelerating mitigation efforts, and not add to climate risks by deploying unpredictable SAI (or SRM as a whole) that fails to address the root cause of climate change. Many proven approaches exist that can bring us to net-zero, necessary to stabilize our climate, followed by necessary net-negative emissions, or removals, as outlined in IPCC assessments. These include the adoption and deployment of clean tech in many sectors (transport, energy), already entering a phase of rapid adoption where costs are cheap.
False premise #5: It is our moral duty to look at all options
At a time when carbon emissions need to be reduced quickly and deeply, political focus, funding, expertise, and innovation should concentrate on achieving this. Geoengineering concepts such as SAI constitute a moral hazard. They can induce complacency as they are portrayed as a quick fix (which, as we have already discussed, is not true) that avoids difficult political trade-offs. (Well-intentioned) researchers may emphasize that SRM should complement rather than replace mitigation, yet this does not prevent other actors from deciding that (the prospect of) geoengineering options (is) are enough to reduce climate risks, using this to justify delaying emission reductions, and resulting in mitigation deterrence. Another type of moral hazard that SAI deployment can induce is “predatory delay” by powerful institutions, which promote seemingly easier options to slow implementation of actions that address the root causes of problems to preserve their own financial and political power. We must look at options that tackle the root cause of climate change effectively and on the long term.
False premise #6: Geoengineering will prevent tipping points from being crossed
Certain climate tipping points are close to being breached with 1.5 °C of warming, such as the collapse of the Greenland and West Antarctic ice sheets, abrupt permafrost thaw, or collapse of the global ocean circulation. The recent Global Tipping Points Report indicates that the loss of warm-water corals may have already passed such a point. While SRM is often framed as a means to delay the crossing of such tipping points, many studies outline that SRM deployment is less effective than greenhouse gas mitigation strategies due to the risks and uncertainties involved with SRM. According to the latest Indicators of Global Climate Change (IGCC) report, we are only ~2 years from crossing 1.5 °C, at least temporarily, with global efforts to then focus on coming back to 1.5 °C “from above” (Forster et al., 2025). No SRM deployment scenario would be ready in time to avert this. Furthermore, the unpredictable changes in atmospheric circulation due to SAI deployment could induce regional temperature and rainfall changes in regions already fragile and thus induce their tipping (e.g., the Amazon rainforest). Many studies that advocate for SAI rely on simplified scenarios that look only into physical-chemical impacts and neglect biological impacts, as well as sociopolitical factors. Such impacts would cause real harm or raise ethical dilemmas, which would further restrict the potential of geoengineering as an emergency solution (e.g., governance of SAI would have to be global, with a certain degree of geopolitical stability).
The false premises and plethora of feasibility, governance, and ethical challenges of implementing SAI mean that we already know enough to categorize this technique as a dangerous distraction from our need to reduce greenhouse gas emissions.
Luckily, we have existing paths forward to mitigate climate harm and address the climate crisis. And technology is part of our arsenal of effective solutions. We must accelerate the deployment of clean technologies in many sectors and push for innovation. But rather than direct limited resources towards unproven and risky “dead-end” geoengineering technologies, we should prioritize new low-carbon technologies and carbon dioxide removals. These will be needed for hard-to-abate emissions to reach carbon neutrality. IPCC assessments have highlighted these approaches as necessary to bring us to net zero. In addition, we must also strengthen (our understanding of) carbon sinks through Nature-based Solutions that carry many economic, social, health and environmental co-benefits, including restoration and protection of biodiversity.
We need to invest in research into our fundamental natural systems. There are still a lot of fundamental processes that are poorly understood and we need a lot more research into our Earth system to improve our understanding of it. One of the largest uncertainties in IPCC climate assessments, for example, is in the radiative balance, and in particular, cloud-aerosol interactions. Investing further research into this is fundamental. However, we advocate that there is no need to add a geoengineering component in research that examines systems that are still poorly understood.
And finally, we wanted to end on two words of caution. The first is that with funding for climate science, and science as a whole, shrinking worldwide, well-intentioned scientists are increasingly pressured to accept funding that is ostensibly for geoengineering research, even if what they do with it is actually valuable, fundamental research. The resulting risk is that, by doing research that is promoted as being an enabling study for a large geoengineering concept, it increases the credibility of the concept from the perspective of the media. And if they report such research positively as building towards deployment of the concept, this influences how the general public, policy makers and politicians view it. It falls once again in the “moral hazard” argument. Similarly, when major national funding agencies such as UKRI (£11M made available in a call for SRM modelling studies last year) and ARIA announce funds for geoengineering research, this increases the credibility of the concepts from the perspective of the media, policy makers and politicians. This is a concerning issue in a time when decision-makers are under pressure for stronger and faster climate action.
The second word of caution is that we must watch out for false equivalences. SAI is often presented, particularly true in the media, with both sides of the argument equally represented (and this is true of geoengineering concepts as a whole). Only a few decades ago, a handful of climate skeptics would be given media exposure equal to that of the thousands of scientists raising the alarm about climate change and its consequences. We are seeing the same trap happening here with SRM. There is growing scientific and political support for a non-use agreement on SRM. Almost 600 academics and 2,000 Civil Society Organizations have supported the call for an International Non-Use Agreement on Solar Geoengineering, which clearly dwarfs other petitions (e.g., an open letter calling for more research with just over 100 signatures). Yet, many well-respected media outlets will include quotes from both sides of the argument, not representing the balance in the academic world.
Marie Cavitte, on behalf of the authors of the article Safeguarding the polar regions from dangerous geoengineering: a critical assessment of proposed concepts and future prospects, published in Frontiers in Science.
Resources
Siegert M, Sevestre H, Bentley MJ, Brigham-Grette J, Burgess H, Buzzard S, Cavitte M, Chown SL, Colleoni F, DeConto RM, Fricker HA, Gasson E, Grant SM, Gulisano AM, Hancock S, Hendry KR, Henley SF, Hock R, Hughes KA, Karentz D, Kirkham JD, Kulessa B, Larter RD, Mackintosh A, Masson-Delmotte V, McCormack FS, Millman H, Mottram R, Moon TA, Naish T, Nath C, Orlove B, Pearson P, Rogelj J, Rumble J, Seabrook S, Silvano A, Sommerkorn M, Stearns LA, Stokes CR, Stroeve J and Truffer M (2025) Safeguarding the polar regions from dangerous geoengineering: a critical assessment of proposed concepts and future prospects. Front Sci 3:1527393. doi: 10.3389/fsci.2025.1527393
Policy Brief: Safeguarding the polar regions from dangerous geoengineering, Cavitte, M., Kirkham, J., & Pearson, P. (2025). Policy Brief: Safeguarding the polar regions from dangerous geoengineering. Zenodo. https://doi.org/10.5281/zenodo.17488507
Disclaimer: The views expressed in this blog are personal.