In all events, these circumstances emphasize how imprudent it is to rely upon emissions cuts as our sole means of defense against tipping elements. Mitigation remains a plausible defense against triggering tipping elements whose thresholds lie in the latter part of this century or the next, but if thresholds lurk in the next several decades and before the world’s net zero emission point, mitigation is likely to be an ineffective means of avoiding them. We would remain on a collision course with climate disasters that could unfold in a non-linear fashion and evade any subsequent efforts to control them. That would portend runaway climate change.
2. SAI and tipping elements
A growing body of literature supports the assertion that a global stratospheric aerosol injection program could be effective in delaying or avoiding many tipping points10–21. This would suggest a distinct rationale for the commencement of global SAI – not for the direct purpose of reducing global mean surface temperatures to avoid heat-related damages, but rather to avoid tipping thresholds until mitigation or carbon capture could obviate the need for further solar geoengineering. This is a twist on the common conception that SAI may be a means to “buy time” before other interventions can kick in.
However, this too may prove too late. While global SAI is often portrayed as a possible “emergency” climate intervention, humanity should anticipate a roughly two-decade interval between a funded launch decision and the achievement of substantial global cooling on the order of 0.5ºC or 1ºC22. Such a program would require a fleet of several hundred large-payload, high-altitude jets of a type that does not yet exist23. Developing such a prototype aircraft along with sufficiently powerful engines would consume roughly a decade, and then manufacturing the fleet would consume a second decade22. Limited and ramping deployment could begin at the beginning of the second decade, but only when the fleet is complete would the target level of cooling be reached.
Moreover, this does not imply that we can expect global SAI capability ~20 years from today. What would commence this countdown to achieving a cooling target would be a “funded launch” – a firm contractual commitment by a financially capable actor to spend tens of billions of dollars to purchase a large fleet of new jets from one of the world’s major airframers. A roughly equivalent sum would also be required for ground infrastructure – principally airports and aerosol management facilities. Such a spending spree can only be underwritten by major governments24, whose approval or consent would also be required to authorize such a program. No government of any size is known to be even considering embarking on such a program, so a funded launch of a global SAI endeavor is not a near-term prospect.
Given the immensity of the technological, scientific and governance obstacles that would need to be overcome before undertaking such a launch decision, it is our subjective view that a funded launch decision seems all but impossible in the coming decade and improbable in the subsequent one. To draw an admittedly arbitrary but no longer flatly implausible line in the sand, let us imagine that it is the crossing of the 2ºC GMST threshold in ~2050 that finally galvanizes a funded launch decision. That would enable the commencement of a gradually ramping deployment program starting in ~2060 and the achievement of the cooling target by ~2070. So long as no thresholds had been transgressed before 2070, humanity would thereafter have the ability not only to cool the planet generally, but to ward off tipping elements thereafter. We would have successfully crossed the abyss.
But this would leave a task undone. If there are any tipping points lying in our path before 2070, we would be destined for a head-on collision. This begs the question as to whether there are any tools that could reduce our risk of triggering tips before 2070. Fortunately, the answer appears to be – perhaps.
An alternative to the global “peak-shaving” conception of SAI is a program that would target just the poles. Aerosols injected into the stratosphere equatorward of the Arctic and Antarctic circles would be carried poleward by the Brewer Dobson Circulation, fitting a parasol over the top and bottom of the earth. Such a sunshade appears likely to slow or prevent the deterioration of high-latitude tipping elements such as the AMOC10,12, the SPG12, and the ice sheets covering the Greenland13,14,20,25 and the West Antarctic17,18,26. It could ward off abrupt permafrost melt11,27–29 and prevent Arctic winter sea ice loss15,30–32. Though not tipping systems, gradual permafrost melt11,27–29 and September Arctic sea ice loss15,30–32 can also be ameliorated by a polar SAI program. This would provide humanity with a direct defense against colliding with some, though not all, of the tipping thresholds that may be encountered in the middle of this century.
We posit here another line in the sand, though we perceive this one to be less arbitrary. It would be prudent for humanity to arm itself to defend directly against tipping high latitude tipping elements by 2040. We choose this temporal target not because no protection is required sooner, but rather because we judge this timescale to be as soon as is reasonably possible. Later is certainly plausible (much more so, in fact), but sooner is not. Therefore, from here, we proceed from a simple question, which is “if humanity were to seek by 2040 the capability to respond within a year to a high-latitude tipping point emergency, what would we need to do between now and then?”
We will first describe in greater detail the deployment scenarios that might be responsive to the sorts of climate emergencies we can envision at the poles. We will next consider the air infrastructure that would be required to mount such responses, and subsequently, the required ground infrastructure.
3. Polar solar geoengineering deployment scenario
In describing a deployment scenario that might prevent or delay crossing tipping thresholds at the poles, we start with the sub-polar focused program described in Smith et al. (2022), which was characterized as follows:
a) Temperature anomaly target: Annual average surface temperature reduction of 2°C for the area between 60ºN and the North Pole
b) North/south symmetry: Any deployment at one pole must be countervailed by a roughly equivalent deployment at the opposite pole
c) Injection seasonality: March through June in the north, and September through December in the south
d) Injection locations: At latitudes of roughly 60ºN and as close to 60ºS as logistics allow
e) Injection altitude: 13 km
f) Deployed material: Sulfur dioxide, vented as a gas
g) Deployed masses: 6.7 Tg-SO2/yr in each hemisphere, or a global aggregate of 13.3 Tg
Several elements of the above program merit further discussion or modification here.
a) While an average annual surface temperature reduction of 2°C was chosen for Smith et al. (2022), there was no suggestion that this level of cooling was in any way optimal or fit for purpose. Moreover, in the current paper, we note many different cryospheric hazards to which we might potentially seek to respond, and the level of cooling that may be optimal to suppress one hazard is unlikely to be the same as might be warranted in respect of another. Further modeling beyond our intent here will be required to calibrate cooling objectives to particular hazards. Nonetheless, a -2°C annual temperature change in the area north of 60°N would be a very substantial cooling, particularly since the seasonal imbalance would concentrate most of the cooling in the summer months
b) As conceived here, the trigger for an SAI response would be convincing data suggesting that a particular high-latitude target tipping element was approaching its threshold, but any large-scale SAI response at one pole must be countervailed by a roughly equivalent one in the opposite hemisphere. This is because a program in just one hemisphere would substantially shift the location of the Intertropical Convergence Zone (ITCZ), which in turn materially impacts rainfall patterns in the tropics. So as not to disturb agriculture and ecosystems in the most densely populated parts of the world, any polar SAI program must be bi-polar and roughly symmetrical. Symmetry could be conceived as meaning the same deployed mass in each hemisphere, or the same cooling impact, or such ratio of hemispheric distribution as creates the least disturbance to the ITCZ. Noting these prospectively different definitions, we will assume herein that equivalent masses are deployed in each hemisphere.
c) Given the extraordinary seasonal imbalance of insolation at the poles, one would intervene only in the sunny months and not in the dark ones. Building on conclusions reached in Lee et al. (2021), one would inject only in the spring/early summer months, such that the aerosols are aloft when the days are long and the insolation is at its maximum. While aerosols vented above the tropics would have an atmospheric lifetime on the order of a year or more, aerosols injected at 60ºN/S would endure for less than half that time30, meaning the particles injected in the spring would have mostly fallen out of the air by autumn.
 d) In respect of injection locations, latitude matters greatly but longitude seems not to substantially impact deployed aerosol efficacy, given the efficient east/west mixing provided by the earth’s rotation. Since the hazards we seek to manage reside in the high latitudes and since material vented into the lower stratosphere flows mostly poleward, high-latitude injection locations are sensible for this purpose. In the Arctic, a sunshade from 60ºN northward would shield all of Greenland, the Arctic Ocean, and most of the permafrost regions from the most intense summer sunlight. Likewise, all of Antarctica lies south of 60ºS. Bases near 60ºN are plentiful, with each of Anchorage, Oslo, Stockholm, Helsinki, and St. Petersburg being proximate. On the other hand, the southernmost widebody-capable airports in the world are at the tip of Tierra del Fuego, lying only 54 degrees south of the equator. These are less ideal than their northern counterparts, but will have to do.
e) An injection height of 13 km is chosen based on the height of the tropopause at deployment latitudes, noting that aerosols must be injected above the tropopause to retain sufficiently high residence times in the stratosphere. Compared to our proposed program, the principal source of difficulty in respect of tropical deployment is that an altitude of roughly 20 km is required to access the lower stratosphere in these regions, requiring the development of novel aircraft. Latitude is the primary determinant of tropopause height, and fluctuations due to seasonal33,34, longitudinal35,36, and interannual variability36,37 are relatively minor. During our proposed injection periods, the tropopause sits at roughly 10 km altitude at 60ºN and 9 km altitude at 60ºS36. Tropopause heights at a deployment latitude of 54º in the Southern Hemisphere rise by only about 1 km compared to 60ºS36, suggesting that deployment from 54ºS is an acceptable alternative. To all of these altitudes, we add a 3 km buffer to account for tropopause abnormalities and minor seasonal, longitudinal, and interannual variation, arriving at an injection height of 13 km. However, tropopause height variability may mean that on certain days in the deployment window, an abnormally high tropopause may render deployment ineffectual. We have reduced dispatch reliability by 5% herein to account for this, noting that that is simply a plug number intended to take account of this issue, subject to further refinement.
f) While other forms of sulfur are possible and perhaps preferable in deployment schemes, and non-sulfur aerosols would avoid interfering with the recovery of the ozone layer particularly in the Antarctic, we have for simplicity retained the most common assumption, which is that our program’s aerosol is SO2 vented as a gas. After a few weeks, SO2 would oxidize to H2SO4, in which state it would be effective at deflecting the intense summer sunlight at the poles.
g) Sourcing from Lee et al. (2021) and assuming linear scaling, 6.7 Tg-SO2/yr was determined to be required in the Arctic in order to reduce temperatures by 2°C15,38. An equivalent deployed mass was assumed in the Antarctic. We reiterate that the resulting aggregate deployed mass of 13.3 Tg/yr should be perceived as an order-of-magnitude first guess, subject to substantial further refinement.
4. Air infrastructure
The lower altitude required for high-latitude aerosol injections is a game-changer in respect of aeronautical infrastructure. Unlike tropical injections that would require novel high-altitude aircraft with minimal payloads23, sub-polar injections could be performed with existing aircraft designs, because the 13 km altitude requirement (roughly 43,000 feet) is at or near the service ceilings certified for many modern jetliners. Moreover, contrary to the conclusions reached in Smith et al. (2022), we conclude here that some existing aircraft would be well-suited to this mission and that purpose-built novel aircraft would not be required. This removes one substantial task from the infrastructural build-out necessary to undertake such an intervention.
Smith et al. (2022) examined the capabilities of existing air-to-air refueling tankers for polar SAI deployment given the mission similarities, since in both cases the planes are tasked with lofting dense loads of fluids into the air. However, the aging aircraft designs used for air-to-air refueling caused these aircraft to perform poorly at 13 km, suggesting that purpose-built SAI aircraft would be substantially more economical despite the required upfront investment in a new design. In this study, we have analyzed the capabilities of the most modern freighters either in production or on the drawing boards at Boeing and Airbus and find those to be much more capable than the tanker platforms and well-suited to a 13 km deployment mission. In particular, our analysis suggests that a modified variant of the 777F (a “777 Special Tanker”) could haul a whopping 110.6 metric tonnes (243,750 lbs) to the target altitude, which is roughly 3 times the payload of the tankers Boeing and Airbus are currently delivering to air forces around the world and over 1/3 more than the payload of the “SAIL-43k” bespoke deployment aircraft conceived in Smith et al. (2022). For the task of developing an emergency response capability to tipping point phenomena at the poles, the 777F appears to be an excellent starting point.