A subpolar-focused stratospheric aerosol injection deployment scenario

The three Working Group reports issued by the Intergovernmental Panel on Climate Change (IPCC) as a part of the Sixth Assessment Report present a sobering picture of the status of the changing climate and humanity's response to date. The average global surface temperature in 2011–2020 was 1.09 °C higher than that in 1850–1900 whereas by 2018, the global mean sea level had already risen by 0.20 m above the 1901 average (IPCC 2021). Under all shared socioeconomic pathways that serve as a basis for climate projections assessed by the IPCC, global surface temperatures continue to rise until at least mid-century (IPCC 2021). Perhaps most concerning, many changes caused by past and future greenhouse gas emissions are irreversible for centuries to millennia (IPCC 2021, 2022). The Arctic faces a particularly dire threat from climate change, warming at roughly twice the global average (IPCC 2021). This enhanced warming of the Arctic results from a combination of processes including: reduction in snow- and sea ice-albedo; increased downward longwave heating due to increased Arctic cloud cover and water vapor content; increased transport of energy from lower latitudes to the Arctic from changes in oceanic and atmospheric heat flux convergence; and enhanced heat absorption by the increase in soot and black carbon aerosols (Hansen and Nazarenko 2004, Gillett et al 2008, Graversen and Wang 2009, Shindell and Faluvegi 2009, Screen and Simmonds 2010, Serreze and Barry 2011). In fact, due to this 'Arctic amplification', the Arctic annual mean surface temperature had already increased by over 3 °C between 1971 and 2019 (AMAP 2021). In addition, the average September sea ice extent in 2010–2019 was 40 percent lower than that in 1979–1988 (IPCC 2021). By mid-century, if not earlier, summer Arctic sea ice will likely have effectively disappeared, with potentially catastrophic climate consequences for both the Arctic and the planet as a whole (AMAP 2017). Though polar amplification in the Antarctic is less pronounced, it too is warming faster than the planetary average and there remain concerns about Antarctic ice sheet melt as a climate change tipping point (Clem et al 2020, DeConto et al 2021, IPCC 2021).

Stratospheric aerosol injection (SAI) is a prospective climate intervention that would seek to abate global warming by slightly increasing the reflectiveness of the Earth's upper atmosphere. SAI is a potential supplement to (but not a replacement for) other climate strategies including mitigation, adaptation, and carbon dioxide removal. However, it remains controversial, and research on SAI technology and its governance are still at very early stages. The vast majority of SAI simulations involve deploying aerosols (or their precursors) globally in order to lower temperatures worldwide. For example, the Stratospheric Aerosol Geoengineering Large Ensemble (GLENS) project and the Geoengineering Model Intercomparison Project's (GeoMIP) G6Sulfur experiment both involve injecting SO₂ at low latitudes (30°S-30°N for GLENS, above the equator for GeoMIP G6) to offset climate change-driven increases in global mean temperature (Kravitz et al 2015, Tilmes et al 2018).

In contrast to global solar geoengineering, subpolar geoengineering would involve geographically limited deployments at latitudes of roughly 60°N/S. Because the tropopause is considerably lower at high latitudes, aerosols or their precursors would not need to be lofted as high, reducing the engineering challenges relative to a global deployment. Only a few existing studies consider Arctic and/or polar SAI deployment. In one of the earliest simulations, (Robock et al 2008) modelled both tropical and Arctic SAI deployments, finding that Arctic injection was more effective per unit of SO₂ at preserving sea ice than equatorial injection. (Jackson et al 2015) also modelled the sea ice impacts of Arctic injection of SO₂, finding that injection masses in excess of 10 Tg-SO₂/yr would be required indefinitely to recover sea ice. A recent study by (Lee et al 2021) finds that, per teragram of SO₂ injected, spring-only injection at 60°N restores approximately twice as much summer sea ice and achieves approximately 50% more Arctic and global mean temperature reductions than year-round injection at that latitude. While Arctic-only SAI deployment has been found to be highly effective, there are also potential concerns. (MacCracken et al 2013) and (Nalam et al 2018) find that deployment of SAI at higher latitudes in the Northern Hemisphere moves the Inter-Tropical Convergence Zone (ITCZ) southward, affecting global precipitation patterns. However, both papers also find that if counterbalancing SAI is deployed in the Southern Hemisphere, the position of the ITCZ can remain relatively unchanged (MacCracken et al 2013, Nalam et al 2018). These early findings motivate our focus on not only Arctic SAI deployment, but also on Antarctic SAI deployment.

There is a growing body of literature that considers the cost and logistics of SAI deployment (The Royal Society 2009, McClellan et al 2012, Moriyama et al 2017, Smith and Wagner 2018, Smith 2020, Smith et al 2022). All of these contemplate deployments intended to have global impact and which would therefore take place in the tropics or sub-tropics at altitudes of 20 km or higher. Existing studies of high latitude deployment limit their scope simply to climate impacts. No existing study builds a complete bi-hemispheric polar or subpolar SAI deployment scenario including logistical and cost considerations, nor do existing studies clarify whether existing aircraft would be suitable for this mission. In section 2 of this paper, we establish a subpolar SAI deployment scenario. In section 3, we lay out the logistics and costs of the scenario. In section 4, we consider the scenario's climate impacts both regionally and globally.

To clarify the feasibility of subpolar SAI, we seek here to articulate a plausible deployment scenario for which we can thereafter assemble a logistical plan. The key parameters of our deployment scenario are as follows:

• Temperature anomaly target: As mentioned earlier, the polar amplification has caused a substantially greater warming in the high latitudes compared to the global average warming over the last several decades. With global greenhouse gas emissions still rising, this additional warming means that the conditions at the poles are likely to be substantially warmer on the threshold of a prospective deployment than they are today. If the objective of such a deployment were to arrest ice and permafrost melt and therefore constrain global sea level rise, a substantial temperature anomaly would seem warranted rather than a barely detectable one. With these considerations in mind, we propose that a plausible temperature anomaly target for a Polar SAI program might be a 2 °C cooling in the Arctic (calculated as the area-weighted average of surface temperatures between 60°N and the pole); this allows us to estimate costs independently of the emissions scenario. We do not argue that this is an optimal or likely target, but as the impacts of deployments in the mass range discussed herein are reasonably linear, readers seeking to estimate the logistics and costs of a smaller or larger deployment may reasonably interpolate or extrapolate from our figures below.
• North/south symmetry: In order to minimize disturbances to distant weather and circulation patterns, any deployment in the Northern Hemisphere must take account of its impact on the Southern Hemisphere. Previous studies have done so by countervailing Northern deployments with roughly similar southern deployments, thereby reducing any shift in the Intertropical Convergence Zone (Ban-Weiss and Caldeira 2010, Kravitz et al 2016). To simplify the calculations and optimize aircraft usage, we define 'symmetry' here as calling for an equivalent aerosol mass deployment in the Antarctic rather than an equivalent temperature anomaly.
• Injection seasonality: Building on the conclusions reached in (Lee et al 2021), we propose to inject only in the spring and early summer months, which is to say March—June in the Northern Hemisphere and September—December in the Southern. Since the intended effect of deployment is to deflect incoming sunlight, deployment in the local winter would have limited impact, as there is little sunlight in the region (Peixoto and Oort 1992). Spring deployment takes advantage of the waxing days and resulting solar intensity, remaining aloft through the polar summer. Aerosols deployed in the very high latitudes have considerably reduced stratospheric endurance relative to that deployed at low latitudes, but since the effective season for deflecting sunlight is merely six months long, the earlier sedimentation of this material at the poles is of limited impact on its efficacy.
• Deployed material: (Lee et al 2021) assumes injections of SO₂, which will oxidize into H₂SO₄ (the sulfur species that is effective for radiative forcing) and coagulate into liquid super cooled aerosols after a month in the stratosphere. While recent studies have explored the direct injection of accumulation mode-H₂SO₄ as an alternative to SO₂ (Vattioni et al 2019, Weisenstein et al 2021), neither the aeronautical tradeoffs associated with carrying this heavier substance nor the mechanics of venting it at the optimal particle size have been convincingly explored. Therefore, despite the prospective advantages of deploying other species of sulfur, we have retained the selection of SO₂ as made in (Lee et al 2021).
• Injection locations: We propose target injection latitudes of 60°N and 60°S, which delimit zones that include the entirety of both Greenland and Antarctica. In the Northern Hemisphere, this is roughly the latitude of Oslo, Helsinki, Homer Alaska, and Magadan in eastern Siberia. In the Southern Hemisphere, the 60th parallel lies entirely in the Southern Ocean well south of the tip of Patagonia. It should be noted that these latitudinal bands delimit an area roughly twice as large as either the Arctic or the Antarctic, each of which are properly defined as the areas poleward of 66.30°. Our descriptions of deployment in the 'Arctic' and 'Antarctic' should be understood in all cases herein to refer to these greater subpolar regions rather than merely to the areas poleward of 66.30°. Given the efficient East/West mixing, particularly at these high latitudes, we assume that injection longitudes are irrelevant to the design of the injection program and should instead be determined by the location of capable air bases proximate to the intended injection latitudes.
• Deployed masses: (Lee et al 2021) estimates that a 12 Tg-SO₂/yr spring deployment at 60°N would force a −3.7 °C annualized average surface temperature anomaly in the region north of 60°N. This estimate was made with a background RCP 8.5 °C scenario but should not be strongly dependent upon the specific scenario. Assuming for simplicity a linear radiative forcing in these mass ranges, this would suggest that each Tg of SO₂ begets approximately −0.3 °C of temperature response in the target zone. Given Arctic temperature forcing targets of −2.0 °C and the assumption of an equivalent mass deployment in the Antarctic, this calls for 6.7 Tg-SO₂/yr in each hemisphere or a total annual deployed mass of 13.4 Tg.
• Injection altitude: (Lee et al 2021) assumed an injection altitude of 14.8 km. Seeking to balance radiative forcing efficacy of deployed aerosols with operational efficiency, we reduce the assumed injection altitude herein to 13 km. Average tropopause altitudes in June (the final month of the northern deployment season) exceed 10 km, and to allow for longitudinal and diurnal tropopause height variability among other factors, 13 km is considered here to be a feasible and prudent deployment altitude, but we do not plan for higher deployments. While we acknowledge that there may be a small difference in the radiative forcing that may result from this lower deployment, we have assumed that the results would be broadly similar and have therefore applied no decrement to the radiative forcing efficacy assumed in (Lee et al 2021).

With the deployment scenario described in section 2 as the objective, we pivot to the matter of how it might be fulfilled. We will assume for this exercise that deployment is undertaken in what would, from an operational standpoint, be idealized conditions, wherein a single global monopolist deployer is able to operate continuously and consistently across multiple national airspace regimes without local interference. We do not seek here to address how such a legitimate global mandate might be secured other than to note that it would be very difficult. Alternatively, deployment plans that instead assume multiple uncoordinated actors, funding challenges, airspace sovereignty disputes, and other routine complications could only be less efficient than what is described below.

3.1. Platforms

For the sort of globally effective SAI deployment in the tropics and sub-tropics envisioned in (Smith and Wagner 2018) and (Smith 2020), a deployment altitude of 20 km is commonly assumed in order to remain well above the tropopause, which can often appear as high as 17 km in the tropics. Injection of large masses of aerosols at 20 km is not judged to be feasible with existing aircraft, requiring the development of new lofting platforms designed for this mission as envisioned in (Bingaman et al 2020). Alternative lofting technologies such as guns, rockets, and balloons were considered in prior studies (McClellan et al 2012, Smith and Wagner 2018) but were determined to be more expensive than aircraft on a cost-per-lofted-tonne basis. And while fixed hoses lofted by tethered balloons could have lower unit costs than aircraft (Davidson et al 2012), their technological immaturity renders them unreliable as lofting options for SAI (McClellan et al 2012, Kuo and Hunt 2015, Lockley et al 2020).

A threshold question arising in respect of the lower 13 km deployment altitude sufficient for a polar program is whether existing aircraft platforms can serve in this instance. After consideration, the simple but surprising answer is—only poorly, and therefore, likely not at all. Experimental sub-scale initial deployments could potentially reuse existing tanker designs, but to implement a program of the scale considered here, a much larger fleet would be required than could be assembled from used aircraft, and the reduced capabilities of existing designs would clearly justify a new, purpose-built platform.

In surveying existing platforms that would seem most likely to serve, the obvious starting point is the large air-to-air refueling tankers used to extend the operational range of military aircraft. In common with our prospective polar SAI deployment platform, these tankers are designed to haul a dense, heavy load of liquid (in their case, jet fuel) into the heavens and transfer it to other aircraft at altitude. By far the most numerous large tanker is the aged but still capable KC-135, which is still aiding US military efforts more than 60 years after its entry into service (U.S. Air Force). These are projected to remain operational at low utilization levels (U.S. Government Accountability Office 2020) through 2040 (U.S. Defense Science Board 2004), but will remain in service until each encounters its firm structural fatigue limits. This means there is not and likely will not be a substantial fleet of retired but operable KC-135s that can be drafted into service for SAI. Their replacements are two current-production tankers: the Boeing KC-46 and the Airbus A330 MRTT (Tegler 2022). An earlier but discontinued replacement tanker is the KC-10. For completeness, we have also considered a theoretical replacement tanker modified from the A340, whose four engines give it an advantage over its twin-engine challengers (KC-46, A330) in this competition. All five of these aircraft are capable of hauling fuel loads of at least 200,000 pounds to altitudes of at least 30,000 feet (roughly 9 km) and would therefore seem ideally suited to the SAI deployment mission.

However, none of these aircraft is capable of ascending with that full payload the additional 4 km necessary to get to our minimum target altitude of 13 km. To sustain a substantial rate-of-climb above their optimal cruise altitude in the 9–10 km range, each of these aircraft would need to get lighter by leaving payload on the ground. Flying reduced loads would enable these aircraft to reach as high as 12 km, but only the KC-135 has a service ceiling enabling it to get comfortably to 13 km. Nonetheless, to facilitate cost comparisons between all of these platforms, we will assume (perhaps unreasonably) that all can be stretched incrementally above their current service ceilings to attain 13 km, albeit with reduced payloads, which in turn increases fleet requirements and costs.

Since each of these pre-existing platforms achieves a dismal payload fraction (net payload/ maximum take-off weight) at 13 km (roughly 43,000 feet), we have added to the platform set a version of the SAIL-01 (Bingaman et al 2020) reconfigured specifically for the subpolar deployment mission. The 'SAIL-43K' could loft a payload nearly five times as great as its predecessor given that the air density at 13 km is so much greater than that for which the SAIL-01 was designed. Even with this huge payload increase, the SAIL-01's six engines are overkill for the 13 km mission, so SAIL-43K has merely four (see figure 1 below for a diagram of the SAIL-43K).

Zoom In Zoom Out Reset image size

The much greater payload on the SAIL-43K required more robust structure and landing gears than SAIL-01, leading to a roughly 18% increase in operating empty weight despite the two fewer engines. Structural augmentation was also required to bring the ultimate load factor up to 4.5 g (from 3.0 g previously), such that it presents an apples-to-apples comparison with the former airliners being alternatively considered. Given these changes, the SAIL-43K could achieve a payload fraction of 56%, making it vastly more efficient for the subpolar mission than the alternatives (see table 1 below).

Table 1. Candidate aircraft (The methodology behind the calculation of the TOW for 43,000 ft has been explained in appendix I).

Aircraft	Service ceiling (ft)	Maximum takeoff weight (lb)	Takeoff weight for 43K ft (lb)	Net payload AT 43K ft (lb)	Net payload AS % OF PAYLOAD	Derivation
KC-135R	50,000	322,500	219,400	105,200	33%	Converted B707
A330 MRTT	42,700	514,000	357,105	88,300	17%	Converted A330
KC-46A PEGASUS	40,100	415,000	256,500	74,870	18%	Converted 767
KC-10	42,000	590,000	376,000	128,801	22%	Converted DC-10
A340F	40,100	840,000	492,000	102,761	12%	Converted A340
SAIL-43K	47,000	299,397	299,397	167,971	56%	Purpose built

Another platform category often casually considered for high altitude flight is top-of-the-line business jets such as the Bombardier Global Express 6000 and the Gulfstream G650, both of which have service ceilings above 15 km. However, these aircraft can achieve such high altitudes in part because they are designed to carry essentially nothing—a handful of well-tailored passengers and their suitcases. Were these same aircraft to be freighted down with their full fuel capacity and maximum structural payloads, they too would be forced to remain at much lower altitudes, with payloads substantially smaller than those of the medium widebody tanker platforms noted above. For a tanker or freighter, the operative question is not how high the plane can get empty and out of fuel, but rather how high it can climb with a full payload at the commencement of cruise, which is a very different matter.

Despite the dramatically lower target altitude required in the polar deployment scheme relative to the global scheme, a purpose-built aircraft would still be warranted for this mission.

3.2. Fleet and activity

3.3. Bases

3.4. Speed to launch and governance

3.5. Costs

The subpolar SAI intervention described herein is calibrated to reduce the average surface temperatures north of 60°N by a year-round average of 2 °C. For simplicity, we have assumed an identical amount of deployed aerosols in the Southern Hemisphere. While the Antarctic is not heating as fast as the Arctic and may have a similarly muted response to SAI, we assume that such an intervention might offset a similar proportion of local warming. Despite the general poleward flow of the Brewer-Dobson Circulation into which the SO₂ would be injected, some of the aerosols would actually flow towards the equator rather than towards the poles. Both because of this, and because of changes in atmospheric and oceanic heat transport, the resulting cooling would not be confined northward of 60°N and would instead be detectable throughout most of the Northern Hemisphere (Lee et al 2021). Similar results north of the deployment zone can be expected with injections in the high latitudes of the Southern Hemisphere (Nalam et al 2018).

Several studies have shown that SAI at low- to mid-latitudes could be effective in reducing and reversing the losses of sea ice and permafrost brought on by global warming, since the injected aerosols would eventually flow poleward, enveloping the entire earth (Moore et al 2019, Chen et al 2020, Lee et al 2020). Furthermore, due to increased Arctic cooling per unit of aerosol optical depth, strategic injections at higher latitudes are more effective at reversing sea ice loss than global or mid- to low-latitude injections (Caldeira and Wood 2008, MacCracken et al 2013, Kravitz et al 2016). (Lee et al 2021) show that spring injection of 12 Tg of SO₂ annually restores the September sea ice extent in a climate model simulation by 5.0 million km². Annual Northern Hemisphere sea ice extent also increases considerably in the Arctic injection scenario.

Arctic SAI has been expected to shift the ITCZ southward, with potentially serious implications for the distribution of tropical precipitation (Robock et al 2008, MacCracken et al 2013, Nalam et al 2018, Lee et al 2020). Balancing the Arctic injection with an Antarctic injection is expected to nearly nullify such a shift (Nalam et al 2018).

Similar to the results from global or tropical injections, subpolar SAI will also result in heating of the lower stratosphere (Niemeier et al 2013, Ferraro et al 2015) , although because the aerosols would be focused at higher latitudes, there would be less heating per unit injection. In addition, the introduction of aerosols into the stratosphere enhances the aerosol-induced surface area density which leads to an increase in heterogeneous reactions required for halogen activations (Solomon 1999, Tilmes et al 2021), though as the aerosols would primarily be present during the summer this may be less of an effect than for global SAI. Consequently, subpolar SAI may impact stratospheric ozone concentrations through a combination of dynamical and chemical effects, possibly slowing the recovery of the Antarctic ozone hole, or at higher deployed masses, reversing it (Pitari et al 2014, Lee et al 2021, Tilmes et al 2021). Further, both wet and dry depositions of the added sulfates pose a risk to humankind and ecosystem. Previous studies have shown that, under global injection, only a small fraction of the sulfate deposition takes place at high latitudes (Visioni et al 2020). Even under injections at 60°N, a significantly larger fraction of the injected aerosols can be expected to deposit southward of the injection latitude (Lee et al 2021). All of these effects would need more research to evaluate.

In addition to the direct climatic impacts resulting from the interaction of injected aerosols with the stratosphere, the carbon footprint associated with the deployment program also carries environmental risks. The pre-deployment emissions stem from the development and production of the deployment fleet as well as the retrofitting of the target airports with the infrastructure to enable SAI. We use the scope 1 and scope 2 carbon dioxide emission figures reported by Airbus for their commercial airliners as a proxy to calculate the total emissions associated developing the fleet for SAI (Airbus 2022). This category of emissions is directly dependent on the fleet size as shown in figure 2. A recent study (de Vries et al 2020) estimates that the CO₂eq associated with airport modifications are ∼2.5 million metric tonnes per airport.

On top of these one-time preliminary emissions, the program also entails recurring operating emissions. These derive from: the combustion of jet A fuel by the planes; the manufacture, transport, and handling of the sulfur dioxide; and the ground handling operations at airports. Jet A fuel upon combustion releases carbon dioxide at a constant rate of 3.16 kg for every kg of fuel (Penner et al 1999).

In addition to the CO₂, aircraft engine combustion also results in non-CO₂ climatic impacts, primarily via the formation of contrails and the release of nitrogen oxides (Azar and Johansson 2012). To account for these in the overall impacts of fuel combustion, (Azar and Johansson 2012) have calculated an emission weighting factor of 1.7 (ranging from 1.3 to 2.9)—a multiplier that allows for computation of CO₂ equivalency based on a hundred-year global warming potential. Figure 2 shows the CO₂eq effects associated with direct combustion of fuel in aircraft engines.

Zoom In Zoom Out Reset image size

While the calculation of life cycle emissions associated with the manufacture, handling, and transportation of sulfur dioxide is beyond the scope of this study, existing studies looking at the cradle-to-grave carbon dioxide emissions resulting from the manufacture of sulfur dioxide and sulfuric acid show that the emissions can vary by an order of magnitude (Veolia 2011, Adeniran et al 2017, Edwards et al 2017). This depends on the carbon intensity of the electricity, the means of sourcing the elemental sulfur, and transport distance required to ship the elemental sulfur to the destination of its use. Barring any significant deviation, over the lifetime of the program, the CO₂eq emissions from the combustion of fuel will likely be significantly larger than the emissions from preparing sulfur dioxide. Similarly, (de Vries et al 2020) found the emissions associated with operating and maintaining the airport to be negligible compared to that from fuel combustion.

Based on the foregoing, several conclusions emerge. While it has yet to be established that the physical or societal impacts of any SAI program would prove to be net positive, it seems clear that a program focused on substantially cooling the world's polar and subpolar regions would be logistically feasible. This could arrest and likely reverse the melting of sea ice, land ice, and permafrost in the most vulnerable regions of the Earth's cryosphere. This in turn would substantially slow sea level rise globally. Spring-only seasonal deployment would achieve substantially higher radiative efficacy per unit of mass deployed and would therefore minimize other negative environmental impacts relative to a year-round program. Despite the fact that Arctic warming is outpacing Antarctic warming, any deployment in one hemisphere should be countervailed by a roughly equivalent injection in the opposite hemisphere.

On the other hand, effective subpolar SAI could not be achieved with a small fleet of hand-me-down tankers or other pre-existing aircraft. Despite the roughly one-third reduction in deployment altitudes compared to a globally-focused program, operational economics would still call for a purpose-built platform, and the required fleet size would be large enough to justify such a new developmental effort if it were backstopped by government guarantees. If pre-existing tanker designs were employed instead, this would roughly double the required fleet size without shortening the time required to stand up such a program.

It is not merely the flight assets but the ground infrastructure that would need to be greatly enhanced in order to accommodate such a program. Appropriately located bases exist in multiple locations in the Northern Hemisphere, whereas in the Southern Hemisphere, the tip of Patagonia is the only plausible option and even it is not ideally proximate to the proposed deployment latitude. A single fleet of aircraft could be feasibly deployed to serve in both hemispheres. The design and build-out of both the flight and ground infrastructure would require more than a decade, such that a large subpolar SAI program is not a feasible emergency response to acute climate stress.

Nonetheless, as with alternative SAI applications, this would be extraordinarily cheap compared to other climate responses such as mitigation, adaptation, or carbon capture and sequestration. However, these are apple/orange comparisons since SAI would merely ameliorate a key symptom of climate change without curing the underlying disease. A subpolar SAI program would also be much cheaper than a program intended to cool the entire globe by the same −2 °C target.

While the cooling would be most pronounced poleward of the deployment latitudes, it would also be expressed in temperate latitudes. Hemispherically symmetrical deployments could likely minimize substantial shifts in the ITCZ, but other artifacts of SAI such as increased sulfur deposition, retarded ozone layer recovery, and increased stratospheric heating would remain. The deployment effort itself would add marginally to the CO₂ and non-CO₂ radiative forcings resulting from aviation via fuel combustion, supply chain-related emissions, and increased contrails.

Though deployment at or near 60°N/S would take place over the airspace of no more than a dozen countries and require bases in even fewer, it is unclear that this would substantially ease the governance and legitimacy challenges that would confront such a program. Therefore, a subpolar deployment seems unlikely to bypass the awesome governance challenges that would confront any SAI program, though this would seem to be a crucial avenue for subsequent social science research.

Nonetheless, an SAI program with global benefits that would entail deployment directly overhead of far less than 1% of the world's population and nearly none of its agriculture may prove an easier sell to a skeptical world than a full-on global deployment. Given its apparent feasibility and low cost, this scenario deserves further attention.

The authors would like to thank Donald C Bingaman of VPE Aerospace Consulting LLC and Jonathan Berger, Adam Guthorn, and Bradley Dailey of Alton Aviation Consultancy LTD for their inputs and Claire Henly for her contribution in the initial literature search.

The data that support the findings of this study are available upon reasonable request from the authors.

Support for BK was provided in part by the National Science Foundation through agreement CBET-1931641, the Indiana University Environmental Resilience Institute, and the Prepared for Environmental Change Grand Challenge initiative. For all the other authors, the work was unfunded.

The authors declare no competing interests.

WS and DGM designed and conceptualized the study. Material preparation, data collection, and analysis were performed by CVR, WS and UB. DV, WRL, and BK provided inputs on aerosol efficacy in subpolar regions. A substantial portion of the first draft of the manuscript was written by WS, with contributions from UB and all authors commented on previous versions of the manuscript. CVR prepared appendix I. All authors read and approved the final manuscript.

Not applicable.

Not applicable.

Not applicable.