Uncertainty and the basis for confidence in solar geoengineering research


Solar geoengineering is an emerging topic in climate-change discussions. To support future decisions on the deployment of this technology, society requires better estimates of its environmental impacts and limitations. As solar geoengineering has never been deployed, conclusions about its climatic effects are primarily obtained through models and natural analogues. As such, our confidence in projections of solar geoengineering, the basis for that confidence and how our confidence can be improved is limited. In this Perspective, we review our current understanding of uncertainty and risk in solar geoengineering via stratospheric aerosols. Using a risk-register framework, we illustrate key uncertainties, such as sub-grid-scale mixing or effects of stratospheric heating, investigations of which should be prioritized to transition the field to a mission-driven research agenda. We conclude with recommendations for possible avenues of research, including targeted model intercomparisons and appropriately governed small-scale field experiments.


Solar geoengineering — which describes technologies that deliberately modify the Earth’s climate — was recognized as a possible method of climate control as early as 1955 (ref.1). Since then, increasing confidence in the impacts and severity of anthropogenic warming has accelerated scientific research and public interest in geoengineering2, which now encompasses a portfolio of proposals to offset the effects of anthropogenic climate change3,4,5. Among the numerous proposed methods4,5, the most commonly discussed include stratospheric aerosols, brightening marine low clouds and thinning cirrus clouds. Stratospheric aerosol geoengineering (SAG) has received the most attention, largely as it is perceived to be more feasible than other methods. Inspired by the cooling seen after large volcanic eruptions6, SAG involves placing reflective aerosols such as sulfate in the stratosphere, which would reflect incoming solar radiation and cool the surface2,7 (Box 1; Fig. 1).

Fig. 1: Climatic effects and uncertainties associated with SAG.

Schematic of some of the conclusions and risks involved in stratospheric aerosol engineering. SAG, stratospheric aerosol geoengineering; UV, ultraviolet.

SAG is virtually certain to reduce global mean temperature, offsetting, at least partially, changes associated with rising CO2 concentrations. A wide range of modelled temperature reductions have been explored, including fully offsetting temperature change8, partially offsetting temperature change9, or slowing the rate of global warming10. Model-based evidence further indicates that SAG will alleviate many other impacts of climate change8: these include offsetting projected acceleration of the hydrologic cycle11,12, ice melt13,14, increased intensity or frequency of extreme events15,16 and tropical cyclone intensity9,17. Indeed, under SAG, nearly all regions are predicted to experience a climate closer to the historical baseline9,18. However, there are almost certainly trade-offs, leading to concerns about winners and losers18,19,20; for example, no model results indicate that SAG can offset both temperature and precipitation changes in all regions of the globe.

Different SAG strategies are also known to produce different regional effects. SAG injection at higher latitudes, for example, preferentially cools the polar regions (with subsequent impacts on sea and land ice)14,21, whereas injection in only one hemisphere preferentially cools that hemisphere (with concomitant shifts in tropical precipitation)22. SAG could also pose additional physical climate risks, such as depletion of stratospheric ozone and subsequent ultraviolet radiation changes23,24, interactions with cirrus clouds and possible effects on the radiative balance25, increased acid rain26, changes to ecosystems27, effects on the ocean28, agricultural impacts29,30 and the potential for climate rebound from the sudden termination of SAG31,32. Figure 1 summarizes some of the conclusions and risks involved in SAG, which have also been previously reviewed4,5,33,34,35,36,37,38,39.

Despite numerous efforts to estimate the impacts of SAG, a systematic assessment of uncertainty (defined here as anything that is currently unknown) and confidence is absent; as such, there is little information to bound conclusions about the range and impacts of possible SAG effects. Thus, to support future decisions, there is a need to clearly articulate how confident the research community is in the potential effects of solar geoengineering, the basis for that confidence and what needs to be done to improve that confidence; the limits on what we can know also need to be determined.

In this Perspective, we do not attempt to address any of these questions directly but, rather, map out the processes by which these questions could be addressed. To allow for a more thorough discussion of uncertainty, we restrict our focus to SAG; other methods, such as marine cloud brightening, encompass fundamentally different uncertainties. We also restrict our attention to uncertainties in natural science, rather than also discussing those in the societal response; this is not to imply that any set of disciplines is more important than others, but some bounding of scope is necessary to make the subsequent discussions tractable. As the SAG method can influence which uncertainties matter, we first describe a design perspective for SAG, before introducing an SAG risk register as a basis for comparing and prioritizing different uncertainties. We then consider the epistemology and basis for confidence in our conclusions about SAG and discuss several specific categories of uncertainty. Finally, we discuss the implications of these uncertainties in supporting future decisions.

Scenario and strategy dependence

Before being able to assess confidence in the effects of SAG, the motivation of which is to inform future decisions, there needs to be increased understanding of where and why uncertainties arise. Here, using a few examples, we illustrate that the background climatic scenario and strategy in which SAG is deployed affect the relative importance of different risks and uncertainties.

The scenario — which describes the severity of background climate change — exerts a strong control on SAG risks. Under climate change, uncertainty in surface-climate effects increases with time, in part due to temperature-dependent climate feedbacks40,41. Under solar geoengineering, temperature rise and, hence, feedback strength are suppressed, leading to reduced model spread in projections of future change12,42,43. As another example, ozone depletion is a modelled effect of SAG deployment23, attributed to an increase in aerosol surfaces where chemical reactions involving chlorofluorocarbons (CFCs) can occur44,45, as observed after large volcanic eruptions. The concentration of atmospheric CFCs, however, has declined since the 1990s and will continue to do so. Thus, if SAG were to be deployed late in the 21st century instead of in the near future, lower CFC concentrations would result in substantially smaller stratospheric ozone depletion46. Similarly, the severity of potential climate rebound following rapid termination of SAG is dependent on the level of background greenhouse gas (GHG) emissions and the magnitude of SAG deployment; for instance, a high GHG scenario coupled with strong cooling achieved through SAG would increase the consequences of a pronounced rebound effect. Many of the conclusions about scenario dependence are based on representing solar geoengineering via solar reduction, and they have not yet been comprehensively explored for SAG. It may be expected that stratospheric circulation and surface-climate responses will be different for SAG47, but changes in the net terrestrial carbon cycle may be similar between the two representations, despite the enhanced diffuse radiative flux under SAG48.

The effects of SAG, as well as the risks and uncertainties, are further dependent upon the deployment strategy, including: the objectives (what SAG is trying to achieve) and the way in which SAG is deployed (latitude, altitude, magnitude and time of year of injection, as well as aerosol or precursor composition)49. A similar list could be made for other solar geoengineering methods; for example, for marine cloud brightening, the effects might depend upon the location and timing of injection, marine boundary-layer stability, particle composition and particle size50,51,52. We illustrate different risks associated with different deployment strategies through a case study of the quasi-biennial oscillation (QBO), a mode of stratospheric variability that describes the equatorial zonal winds as being easterly or westerly, impacting the spread of injected aerosols and, thereby, surface impacts53. Under tropical SAG injection, the QBO transitions to a persistent westerly phase (Fig. 2), with dynamical effects that can alter tropospheric winds and precipitation patterns54,55,56,57. However, for off-equatorial injection that still results in similar levels of global mean cooling, the phase and magnitude of the QBO remain relatively unchanged (Fig. 2), and some of the side effects experienced under equatorial injection do not materialize58.

Fig. 2: Contrast between QBO strength for SAG via equatorial SO2 injection and GLENS.

Stratospheric zonal mean wind speed (m s−1) averaged over 2°S–2°N in Geoengineering Large Ensemble (GLENS)86, where SO2 is injected at four off-equatorial locations (30°N, 30°S, 15°N and 30°S) (part a), and equatorial stratospheric aerosol engineering (part b). Positive wind values indicate a westerly phase of the quasi-biennial oscillation (QBO) and negative values an easterly phase of the QBO. SAG, stratospheric aerosol geoengineering. Adapted with permission from ref.58, Wiley-VCH.

By modifying several of these degrees of freedom — that is, factors related to the deployment strategy — it may be possible to design SAG to achieve climate outcomes beyond solely reducing global mean temperature19,38,49,59. For example, if SAG was to be deployed only in the equatorial regions, this would prevent the rise in global mean temperature58 but have the side effect of residual polar warming8 (Fig. 3). However, by injecting SO2 in four independent locations, it is possible to meet simultaneous temperature objectives (such as offsetting changes in global mean temperature, the interhemispheric temperature gradient and the equator-to-pole temperature gradient; Fig. 3)21,60. By considering the design of SAG, it may be possible to further predictably modify related climate features, such as the position of the intertropical convergence zone or Arctic sea ice extent49.

Fig. 3: Effectiveness of SAG strategies in achieving temperature targets.

Comparison of stratospheric aerosol engineering strategy on various temperature metrics49: global mean temperature (part a), the interhemispheric temperature gradient (part b) and the equator-to-pole temperature gradient (part c). The blue lines show Representative Concentration Pathway 8.5 (RCP8.5), the background against which stratospheric aerosol engineering is being performed, the black lines show results for Geoengineering Large Ensemble (GLENS)86 and the orange lines show results for equatorial-only injection of SO2 (ref.58). The faint lines show different simulation ensemble members and the darker lines show the ensemble mean. The dashed grey zero line shows the objective (no change from 2020 values). SAG, stratospheric aerosol geoengineering. Adapted with permission from ref.58, Wiley-VCH.

In addition to choosing the injection strategy based on the projected response, the deployment strategy can be selected to reduce uncertainty. Off-equatorial injection, for example, could be chosen to reduce dependence on unclear aerosol-coagulation rates (so aerosols are quickly transported away from the injection site)61 or sulfate aerosols may be chosen over calcite, despite projected detrimental ozone effects, to avoid poorly bounded uncertainties39.

However, there are fundamental trade-offs in the climate system and, hence, in what SAG can achieve. For example, solar geoengineering cannot completely offset simultaneous CO2-related changes in global mean temperature and global mean precipitation11. Other trade-offs need to be further understood; even though sea level would continue to rise if global mean temperature change was arrested by solar geoengineering62, conclusions about the amount of resulting sea level rise should be revisited63. There is still substantial uncertainty in the range of climate features that can be effectively managed with SAG38 and whether that space can be expanded by complementing SAG with, for example, marine cloud brightening or cirrus thinning64,65. In the absence of complete knowledge, the climate community has adopted proxies that encapsulate many different climate objectives, typified by the 1.5 °C or 2 °C warming targets to prevent dangerous anthropogenic interference in the climate system66. Similar high-level heuristics for solar geoengineering may also be useful for encompassing a wide range of effects; as described in the following section, a risk register could offer a path towards developing such heuristics for comparing risks under a range of possible future scenarios and strategies.

Prioritizing uncertainties

Although there have been attempts to articulate physical-science uncertainties in SAG39, there has not yet been the prioritization that is necessary for a transition to mission-driven research67, that is, research aimed at identifying and reducing the most important uncertainties. One method of accomplishing this is a risk register68 (Fig. 4). In a risk register, each uncertainty, such as what effect would termination have or how does stratospheric transport influence SAG, is assessed on two axes, representing probability (of occurrence or of being wrong) and consequences. The combination of these parameters describes the overall risk associated with that uncertainty, categorized as low, low–medium, medium, medium–high or high risk. Some uncertainties might never have any basis for assigning an objective likelihood69 or an absolute ranking of all uncertainties, yet some qualitative and subjective assessment may still be possible. Any item that sits in the top-right corner of a risk register immediately becomes a high research priority; actions might aim to reduce the likelihood or degree of uncertainty (through additional observations or experiments) or to find ways to reduce the consequences (choosing strategies that are less dependent on knowledge of a particular parameter or process). However, prioritizing uncertainties is independent of efforts to address them. Research is a primary means of reducing uncertainty but some uncertainties are empirically or practically irreducible, regardless of priority; for example, precise prediction of regional temperature and precipitation outcomes for any given solar geoengineering strategy.

Fig. 4: Schematic of a risk register.

Risks are placed in terms of their probability of occurrence (x-axis), for example, being wrong about some process or parameter, and the consequence (y-axis). The risk register serves as a heuristic for comparing risks and identifying the highest priorities (items towards the top-right of the register). Adapted with permission from ref.67, PNAS.

The risk register is not intended for quantitative accuracy or to replace human judgment. Indeed, positioning on a risk register is a somewhat subjective process shaped by expert opinion and is dependent on scenario and strategy. Using the previous example of the QBO, the consequences (the y-axis of the risk register) are much lower for off-equatorial injection than equatorial injection58, so the probability of being wrong about the QBO (an appropriate interpretation of the x-axis in this case) is low. Instead, a risk register serves as a means for ensuring a conscious, explicit conversation about relative priorities and the basis for that prioritization, as well as a way to track progress on reducing risk over time. Since the purpose of a risk register is comparability, the scope of any risk register needs to be carefully chosen and any major project (such as global-scale SAG) would likely have numerous risk registers that cover different aspects of the problem; for example, the risk of technological lock-in due to private financial interests in maintaining a solar geoengineering programme70 is not directly comparable to uncertainties in climate effects due to aerosol coagulation53, so those two risks are unlikely to appear on the same risk register. However, uncertainties in stratospheric water vapour changes and stratospheric ozone chemical effects71 due to SAG are likely to appear on the same register. The overall project would also involve a high-level synthesis, where all of the risk registers are evaluated simultaneously to look for overlaps, blind spots and potentially compounding risks.

Basis for confidence

Positioning on a risk register becomes less subjective with more confidence in conclusions about the effects of SAG, which, in turn, requires reducing (or at least bounding) uncertainties. Sources of evidence available to reduce geoengineering uncertainties are few and, as such, there is a limited basis for confidence in any conclusions.

Since geoengineering has never been deployed, bounds on potential uncertainties are reliant on information from natural analogues, such as large volcanic eruptions. These analogues, however, often provide limited information72; differences in aerosol microphysics73 and climate responses74, for example, will vary markedly when comparing pulse volcanic eruptions and sustained injections through SAG. In addition, there is sometimes no natural analogue that can be used. Calcite, for instance, has been proposed as a potential SAG aerosol, as it is hypothesized to be less harmful to stratospheric ozone than sulfate75. However, as there has never been a large amount of calcite in the stratosphere, the risks of ‘unknown unknowns’ are higher.

Given the limitations of observing natural analogues, experiments to reduce uncertainties have been proposed. Observational experiments, such as aircraft measurements of the next major volcanic eruption, could provide process-level information about aerosol microphysical growth or stratospheric transport76. Laboratory experiments are constrained in the types of problems to which they can be applied but can be useful in narrowing down specific uncertainties, such as highly accurate measurements of aerosol refractive indices77. Similarly, limited-scope field experiments — either observational campaigns or perturbation experiments — may provide insight into particular processes78,79. For example, the Eastern Pacific Emitted Aerosol Cloud Experiment (E-PEACE) deliberately introduced aerosols into marine clouds to learn how cloud properties changed80, providing information relevant to marine cloud brightening. Similarly, small experiments have been proposed for measuring chemistry in the stratosphere78, while other uncertainties, such as local surface-climate effects, may not be possible to directly validate, even after deployment has started67. In all instances, experiments would require appropriate governance mechanisms to assess whether the science justifies any environmental risk, as could be informed through the risk register.

To date, models have been the primary source of natural science evidence regarding solar geoengineering81, allowing a variety of situations and uncertainties to be explored with minimal environmental impact. Process-level studies can reveal mechanistic understanding; for example, simulations that prescribe or turn off longwave radiative heating47,82 can explore the effects of stratospheric heating on surface climate, revealing and isolating mechanisms of change. Idealized simulations investigating how the altitude of injection influences stratospheric heating may further inform the importance of these effects83. Model intercomparisons (for example, the Geoengineering Model Intercomparison Project (GeoMIP)84) also offer useful insight, particularly with understanding the similarities and differences between responses to standardized modelling experiments. Furthermore, large ensembles of simulations with slightly different initial conditions, such as the Community Earth System Model (CESM) large ensemble85 or the Geoengineering Large Ensemble (GLENS)86 can reveal the influence of internal variability; while these large ensembles can be computationally expensive, emulators offer a useful compromise, generating numerous ensemble members cheaply but at reduced fidelity or granularity87.

However, models have uncertainties, presenting their own challenges when assessing confidence in SAG. As described, model intercomparisons may be useful when diagnosing responses to standardized experiments, but as many of these models are related88,89, spread may provide a biased estimate of uncertainty. Moreover, no single model can span all relevant scales, meaning that any behaviour on sub-grid scales is parameterized. Models may further have incomplete representations of reality, often due to poor understanding of real-world behaviour.

Our understanding of the underlying mechanisms and projected effects of climate change is now well supported by models, observational platforms and underlying theory, providing robustness and confidence in our estimates of corresponding uncertainty90. However, solar geoengineering thus far lacks diversity in its sources of evidence, leading to an incomplete picture of uncertainty; model uncertainty is not a good proxy for all other sources of uncertainty. As such, it is presently difficult to determine the level of confidence in our conclusions about SAG, let alone build a high degree of confidence or how present confidence levels90 will change after further research.

Key uncertainties

Here, building on previous efforts39, we review key uncertainties in SAG in the context of the risk-register formulation (that is, probability and consequence), focusing on discussions around stratospheric processes (many of which determine the radiative forcing associated with the aerosols), the resulting climate response and their impacts. We also provide our personal assessment of where each uncertainty falls on the risk register, recognizing that there is not yet an adequate basis for objectivity and that the process will constantly evolve with new information (Table 1). Table 1 is not exhaustive and other potential uncertainties could include impacts on extreme events15, methane chemistry71, the carbon cycle91,92, water cycling93, vegetation94, agriculture29,30, human health95 and the cryosphere13,14. Nevertheless, Table 1 is a useful starting point to identify, prioritize and reduce uncertainties in SAG, as well as motivate discussion on how to reduce or manage these uncertainties.

Table 1 Summary of key risks and uncertainties associated with SAG

Stratospheric processes

Based on what occurs with a volcanic eruption, a typical assumption for SAG is that SO2 gas would be injected into the stratosphere and undergo transport and oxidation to sulfate aerosols. The direct condensation of H2SO4 into droplets96, or the use of calcite75, would bypass the oxidation step. Sulfate aerosols would then grow due to condensation onto existing particles and coagulation between particles97. Larger particles lead to decreased shortwave scattering efficiency, increased longwave absorption (and, hence, stratospheric heating) and increased sedimentation rate (and, thus, shorter lifetime)96. As the aerosols fall, they will likely interact with cirrus clouds; cirrus would also be affected by changes in ice-crystal size distribution98 and altered vertical velocities resulting from stratospheric heating99. Heating will also lead to increased stratospheric water vapour46,100. Each of these processes has uncertainties that would affect the necessary injection amount needed to meet the chosen objectives53, the spatial distribution of the effects and the severity of the side effects induced by stratospheric heating.

The mean modelled radiative forcing from equatorial injection of sulfate for SAG is approximately −0.23 ± 0.07 W m−2 per Tg SO2 year−1, with the range across models being −0.11 to −0.31 W m−2 per Tg53. There have not yet been sufficient inter-model studies to quantify a range of uncertainty for off-equatorial injection or geoengineering during only part of the year. Models can reproduce observed properties of the aerosol layer (for example, thickness and aerosol radius) after volcanic eruptions101, but these observations are insufficient to constrain aerosol microphysical uncertainties for SAG because of different aerosol growth rates between pulse and sustained injections, as well as uncertainties in eruption observations102,103. Changes in stratospheric water vapour are difficult to quantify on sub-decadal timescales104,105,106, so, although stratospheric aerosol loading is a known source of stratospheric water vapour100,107, it is difficult to quantitatively validate how well models reproduce this feature. The amount of ozone destruction from stratospheric aerosols depends strongly on the location and size of the aerosols108, as well as changes in stratospheric heating109,110,111. Moreover, models have trouble reproducing both the baseline and the perturbed state of the stratosphere112,113,114, a fact that is complicated by predicted but unknown changes in large-scale atmospheric flow under climate change61. Interactions of stratospheric sulfate aerosols with cirrus are poorly understood both in observations of volcanic eruptions115,116 and in model simulations of SAG25. Furthermore, many of the aerosol and transport processes happen on the sub-grid scale, which is subject to additional uncertainties.

There are ongoing activities to reduce several of these uncertainties. The Model Intercomparison Project on the climatic response to Volcanic forcing (VolMIP) is aimed at validating modelled representations of volcanic eruptions by minimizing differences in the applied volcanic forcing117. The Interactive Stratospheric Aerosol Model Intercomparison Project (ISA-MIP) aims to understand the range of modelled stratospheric responses to quiescent and changing conditions over measured history, with the purpose of improving and validating models118. This project also includes a protocol to explicitly quantify the impact of aerosol microphysical uncertainties on climate outcomes under geoengineering. Preparing an observation platform that could be rapidly deployed during the next volcanic eruption would substantially improve our understanding of aerosol formation and microphysical growth76. In addition, during a hypothetical deployment, aerosol optical depth and surface air temperature could be measured regularly to adjust the injection amounts119, compensating for uncertainty in the injection required to achieve the desired cooling, even in the presence of uncertainties in size distribution, deposition and stratospheric heating.

Climate response

Due to temperature-dependent feedbacks, many of the uncertainties in the response to SAG are also uncertainties in the response to CO2 (ref.43). Focusing only on how the two forcings affect the climate differently120 thus simplifies the assessment of uncertainty. Whereas many broad features of the different climate responses to CO2 and SAG are robust across models121, details of regional changes can differ substantially between models. For example, all models predict that some locations may see increased departures from preindustrial precipitation due to solar dimming, but there is little agreement between models on where this might occur18,122.

It is useful to explicitly separate differential climate responses from SAG and CO2 into categories arising from shortwave versus longwave forcing, the spatial and seasonal pattern of forcing and processes unique to stratospheric aerosols, particularly those owing to stratospheric heating or ozone depletion. This subdivision illustrates the appropriate level of granularity in constructing a risk register. Also, because each of these categories is associated with their own uncertainties, subdividing indicates how to design dedicated simulations to isolate individual mechanisms. For example, offsetting CO2-induced global mean temperature change with reduction in the solar constants of models overcools the tropics and undercools the poles8,123. This differential latitudinal cooling is due in large part to different spatial patterns of forcing between CO2 and insolation and differences in shortwave versus longwave forcing124. Furthermore, geographically varying patterns of solar reduction result in different surface-temperature patterns49. Finally, regardless of the spatial pattern of forcing, offsetting increased GHG forcing with a reduction in shortwave radiation will reduce the strength of the hydrological cycle, as the shortwave reduction compensates for the temperature-dependent precipitation response but not for the ‘fast’ response to increased CO2, which is a radiative effect that acts to reduce evaporation and, hence, upward moisture flux into the free troposphere125. To determine which of these responses is most important for understanding the latitudinal distribution of cooling, one could design simulations with different spatial patterns of solar geoengineering and different characters (longwave versus shortwave) of forcing47,49,124.

The sources of uncertainty for differential regional responses to SAG are not yet fully understood. Many of the surface-climate effects are tied to changes in large-scale circulation61,126, which often have poorly represented processes in Earth system models127, leading to model spread. For example, Fig. 5 shows the maximum and minimum (on a grid-cell level) across models of regional responses to offsetting high CO2 with solar dimming; not even the sign of the regional changes is consistent across all models in most places. Similar results do not yet exist for simulations of SAG as there are not yet any inter-model comparisons using models that can simultaneously include enough of the relevant aerosol processes. However, the uncertainty in the surface response to SAG would presumably be even larger, as it would include both the uncertainties associated with compensating increased CO2 with a solar reduction, as well as further uncertainties associated with the stratospheric aerosol processes, including the effects of stratospheric heating on the surface climate. Moreover, different models have different representations of CO2 fertilization to accelerate plant growth, as well as different nutrient-limitation cycles, leading to different magnitudes of change in surface fluxes, temperature changes and hydrological responses32,128. A common approach to reducing spread or selecting models that accurately reproduce real-world phenomena is through observational constraints129, but such information is lacking for solar geoengineering. SAG field experiments directly aimed at eliciting a climate-system response are effectively deployment-scale130,131,132 and are best categorized as operational ‘pre-deployment’ activities67, rather than tools to reduce uncertainty in climate response.

Fig. 5: Uncertainty in climate responses to SAG.

Model spread from 12 models in which the solar reduction fully offsets the global mean temperature increase from abrupt4xCO2, using data from the Geoengineering Model Intercomparison Project G1 experiment84. Shown are: maximum temperature (part a), minimum temperature (part b), maximum precipitation (part c) and minimum precipitation (part d), all computed for each grid cell. SAG, stratospheric aerosol geoengineering.


Solar geoengineering has maintained interest due to its potential to alleviate many of the consequences of climate change4. While direct model outputs allow quantification of SAG-induced changes in climate-related variables such as temperature, precipitation and sea ice133, synthesizing these into corresponding impacts on food and water security, health, ecosystem services and sea-level rise is more difficult and uncertain134. Gaining consensus on how SAG influences these aspects from impact modelling is difficult due to the diversity of model representations of scale and processes, even in the same sector135. In addition to assessing impacts associated with novel SAG-induced climate regimes, it will be important to assess the direct impact of the aerosols on health (from particulates95, acid rain26 or ultraviolet radiation due to ozone loss24,136,137) and on ecosystems and agriculture (owing to the small reduction in overall sunlight and the increase in diffuse light)138,139.

The incomplete and indirect relevance of available datasets may limit efforts to constrain the impacts of SAG (for example, on agriculture) using volcanic analogues30. Moreover, the impacts of SAG are highly dependent on societal decisions; for example, the potential detrimental effects of SAG on agriculture could be compensated for by changes in fertilizer use29. Coordination between the SAG research community and impact-assessment modellers is essential to reduce uncertainty and improve confidence in our understanding of whether SAG is effective at reducing the impacts of climate change and, if so, where140.

The impacts of climate change typically increase monotonically with increased GHG concentrations. However, solar geoengineering would not decrease all impacts proportionally; known correlations between variables for climate change can be different for solar geoengineering141. As such, single-objective targets may not be appropriate for managing trade-offs in hypothetical deployments. An important component of investigations to create a holistic understanding of the potentials and limitations of solar geoengineering is effective scenario generation142 to capture relatively unexplored issues in climate models, such as distributional justice and multiple stakeholder perspectives143,144.

Conclusions and next steps

Ultimately, the goal of quantifying uncertainty in SAG is to decide what to do about it, of which there are many courses of action. Reducing uncertainty requires research and, implicitly, a prioritization of that research. Managing uncertainty can be accomplished through adaptive methods (such as feedback)21,49,119,145, where geoengineering is adjusted regularly to ensure that the objectives are being met38. Avoiding uncertainty can be accomplished by pursuing methods where SAG is needed less or not at all, such as more aggressive mitigation122, or by choosing strategies for deployment that are less sensitive to particular uncertainties (for example, off-equatorial injection to minimize effects on the QBO)58. Finally, some uncertainties may be irreducible but they can still be quantified and prioritized so that their risks are understood.

The focus here has been to define a path towards quantifying the effects of uncertainty in SAG, following which there may be a diversity of responses in research. In one extreme, there may be a non-zero probability that climate change is catastrophic, providing strong motivation for solar geoengineering research146. Conversely, the probability of showstoppers (that is, catastrophic risks from deployment) may be non-zero, meaning that no further effort should be spent on solar geoengineering research147. However, justifying these two extreme responses, and all responses in between, does not necessitate the complete quantification of all uncertainties; it can be argued that there is insufficient knowledge to support decision making, but decisions can be made in light of uncertainty148. Any decision regarding SAG deployment will involve a risk–risk trade-off149: how do the risks of deploying, including risks introduced by uncertainty, weigh against the risks of not deploying solar geoengineering? Ultimately, this is a question of governance150, and determining what objectives solar geoengineering can and cannot achieve is crucial for understanding what governance mechanisms require further development and expansion151, as well as what observations are needed to inform those governance mechanisms152.

Research recommendations

Based on our current assessment of the state of uncertainty in solar geoengineering, we provide recommendations on several high-priority research directions in an effort to gain greater confidence in present and future conclusions regarding SAG. In doing so, we offer a pathway whereby similar explorations may be applied to other solar geoengineering methods (such as marine cloud brightening).

Recommendations on prioritization

We have described a risk register for prioritizing uncertainties and, thereby, areas of future research. In Table 1, we have identified several research questions that we deem high priority (such as aerosol microphysical growth, sub-grid-scale mixing and ecosystem response), but these are based on our opinions. A more comprehensive effort to thoroughly explore a wider range of uncertainties (including those that fall outside of the realm of natural science), combined with a more objective way of assessing their risks, would be highly informative for building a coordinated, large-scale research agenda.

Recommendations on model intercomparisons

Although model intercomparisons can reveal substantial knowledge about uncertainty through analysing model similarities and differences, most multi-model studies have focused on where models agree on SAG than where they disagree134. Pitari et al.24, by contrast, evaluated differences in ozone changes in different models with different processes, revealing the importance of including specific chemical processes in models24. Studies like this are excellent examples of how model intercomparisons can inform prioritization of uncertainties, and we argue that more studies exploring model differences (for example, simulations with different representations of aerosol microphysics) or looking at a diversity of model responses across high-priority areas (for example, ecosystems) are needed. These may include further analysis of existing model experiments, such as those produced under GeoMIP, as well as carefully constructed simulations to isolate different physical mechanisms, for example, constraining the role of stratospheric heating by introducing specified stratospheric heating rates without the aerosols47 or conducting simulations with idealized aerosols without any heating. In addition, model comparisons can increase robustness in conclusions, for example, repeating GLENS in other, independently developed models that include the relevant processes.

Recommendations on field experiments

In the future, there may be a role for SAG field experiments to reduce critical uncertainties. The planned Stratospheric Controlled Perturbation Experiment (SCoPEx)78, for example, is aimed at understanding aerosol-nucleation processes and stratospheric chemical perturbations associated with calcite injections. As another example, in situ observations of aerosol–cloud interactions in cirrus could address critical uncertainties in solar geoengineering (Table 1) and climate science more generally. The societal acceptability and potential risk of harm are greater for field experiments than for modelling or laboratory experiments and, as such, determining the roles of field experiments and whether they should go forward is a matter of governance150. We suggest two potential components of a cost–benefit analysis that could judge whether field experiments proceed: ensuring that the information cannot be obtained by another, less intrusive way and specifying that the field experiments are truly mission driven, aimed at addressing high-priority uncertainties.

Recommendations on natural analogues and observations

The role of natural analogues in bounding risk needs to be understood, especially about unknown unknowns. As an example in SAG, there is not yet a way to compare confidence in the deployment of sulfate (which multiple sources of evidence say are likely to have detrimental side effects) with the deployment of calcite (which may have substantially fewer side effects), and any such conclusions are based on a limited set of modelling studies. Although the usefulness of volcanoes as analogues for SAG is limited, observing future large volcanic eruptions could provide key information on aerosol microphysical growth or stratospheric transport that would help validate model representations of SAG76.

Recommendations on research-capacity building

Research on solar geoengineering has been dominated by developed countries with climate-modelling capacity, which insufficiently captures the breadth of populations who would be affected by any potential deployment149. The lack of a broader geographical diversity of researchers and cultural perspectives may be detrimental to gaining a full appreciation of all sources of uncertainty. Ongoing research activities, such as GeoMIP and the Climate Engineering Conference series, have actively encouraged increased participation from developing countries153,154. Recently, through small research grants, the Developing Country Impacts Modelling Analysis for SRM (DECIMALS) fund has catalysed research into modelled SAG effects on developing countries155, which is an important step towards increasing the diversity of well-informed perspectives in solar geoengineering discussions. Increased diversity in solar geoengineering research will improve confidence that as many uncertainties in solar geoengineering are identified as possible and that the prioritization process will be performed more equitably.

Much of the discussion surrounding new results in solar geoengineering raises questions about their accuracy and relevance to reality. Unfortunately, this reflects how little confidence there is in many of the results. This lack of confidence is inadequate as a basis for climate policy. Although nearly all global-scale decisions are made in the presence of some amount of uncertainty, in the case of solar geoengineering, there is presently not enough confidence in any of the conclusions to design a deployment strategy with the expectation that it will behave as intended. Eventually, with enough research, there may be sufficient confidence in the risks of solar geoengineering that enable well-informed discussions about its role (alongside mitigation and adaptation strategies, and carbon-dioxide removal) in addressing climate change to proceed. Solar geoengineering sits at the frontier of climate science research, with all of the discovery and pitfalls therein; addressing uncertainty in a systematic way will move the field forward and improve its relevance in policy discussions.


  1. 1.

    von Neumann, J. Can we survive technology? Fortune 51, 151–152 (1955).

  2. 2.

    Crutzen, P. J. Albedo enhancement by stratospheric sulfur injections: a contribution to resolve a policy dilemma? Clim. Change 77, 211–220 (2006).

  3. 3.

    Long, J. C. S. & Shepherd, J. G. in Global Environmental Change. Handbook of Global Environmental Pollution Vol. 1 (ed. Freedman, B.) 757–770 (Springer, 2014).

  4. 4.

    Shepherd, J. G. & Working Group on Geoengineering the Climate. Geoengineering the climate: science, governance and uncertainty (Royal Society, 2009).

  5. 5.

    National Research Council. Climate Intervention: Reflecting Sunlight to Cool Earth (The National Academies Press, 2015). Summarized the state of knowledge, providing the base from which to build a solar geoengineering research agenda.

  6. 6.

    Robock, A. Volcanic eruptions and climate. Rev. Geophys. 38, 191–219 (2000).

  7. 7.

    Budyko, M. I. Climatic Changes (American Geophysical Union, 1977).

  8. 8.

    Kravitz, B. et al. Climate model response from the Geoengineering Model Intercomparison Project (GeoMIP). J. Geophys. Res. 118, 8320–8332 (2013).

  9. 9.

    Irvine, P. et al. Halving warming with idealized solar geoengineering moderates key climate hazards. Nat. Clim. Change 9, 295–299 (2019).

  10. 10.

    MacMartin, D. G., Caldeira, K. & Keith, D. W. Solar geoengineering to limit the rate of temperature change. Phil. Trans. R. Soc. A 372, 20140134 (2014).

  11. 11.

    Tilmes, S. et al. The hydrological impact of geoengineering in the Geoengineering Model Intercomparison Project (GeoMIP). J. Geophys. Res. 118, 11036–11058 (2013).

  12. 12.

    Kravitz, B. et al. An energetic perspective on hydrological cycle changes in the Geoengineering Model Intercomparison Project. J. Geophys. Res. 118, 13087–13102 (2013).

  13. 13.

    Moore, J. C. et al. Arctic sea ice and atmospheric circulation under the GeoMIP G1 scenario. J. Geophys. Res. 119, 567–583 (2014).

  14. 14.

    Zhao, L., Yang, Y., Cheng, W., Ji, D. & Moore, J. C. Glacier evolution in high-mountain Asia under stratospheric sulfate aerosol injection geoengineering. Atmos. Chem. Phys. 17, 6547–6564 (2017).

  15. 15.

    Curry, C. L. et al. A multimodel examination of climate extremes in an idealized geoengineering experiment. J. Geophys. Res. 119, 3900–3923 (2014).

  16. 16.

    Dagon, K. & Schrag, D. P. Regional climate variability under model simulations of solar geoengineering. J. Geophys. Res. 122, 12106–12121 (2017).

  17. 17.

    Moore, J. C. et al. Atlantic hurricane surge response to geoengineering. Proc. Natl Acad. Sci. USA 112, 13794–13799 (2015).

  18. 18.

    Kravitz, B. et al. A multi-model assessment of regional climate disparities caused by solar geoengineering. Environ. Res. Lett. 9, 074013 (2014).

  19. 19.

    MacMartin, D. G., Keith, D. W., Kravitz, B. & Caldeira, K. Management of trade-offs in geoengineering through optimal choice of non-uniform radiative forcing. Nat. Clim. Change 3, 365–368 (2013).

  20. 20.

    Moreno-Cruz, J. B., Ricke, K. L. & Keith, D. W. A simple model to account for regional inequalities in the effectiveness of solar radiation management. Clim. Change 110, 649–668 (2012).

  21. 21.

    Kravitz, B. et al. First simulations of designing stratospheric sulfate aerosol geoengineering to meet multiple simultaneous climate objectives. J. Geophys. Res. 122, 12616–12634 (2017).

  22. 22.

    Haywood, J. M., Jones, A., Bellouin, N. & Stephenson, D. Asymmetric forcing from stratospheric aerosols impacts Sahelian rainfall. Nat. Clim. Change 3, 660–665 (2013).

  23. 23.

    Tilmes, S., Müller, R. & Salawitch, R. The sensitivity of polar ozone depletion to proposed geoengineering schemes. Science 320, 1201–1204 (2008).

  24. 24.

    Pitari, G. et al. Stratospheric ozone response to sulfate geoengineering: results from the Geoengineering Model Intercomparison Project (GeoMIP). J. Geophys. Res. 119, 2629–2653 (2014). One of the few papers to carry out an in-depth analysis of model differences, instead of focusing on similarities.

  25. 25.

    Visioni, D., Pitari, G., di Genova, G., Tilmes, S. & Cionni, I. Upper tropospheric ice sensitivity to sulfate geoengineering. Atmos. Chem. Phys. 18, 14867–14887 (2018).

  26. 26.

    Visioni, D., Pitari, G., Tuccella, P. & Curci, G. Sulfur deposition changes under sulfate geoengineering conditions: quasi-biennial oscillation effects on the transport and lifetime of stratospheric aerosols. Atmos. Chem. Phys. 18, 2787–2808 (2018).

  27. 27.

    Trisos, C. H. et al. Potentially dangerous consequences for biodiversity of solar geoengineering implementation and termination. Nat. Ecol. Evol. 2, 475–482 (2018).

  28. 28.

    Fasullo, J. T. et al. Persistent polar ocean warming in a strategically geoengineered climate. Nat. Geosci. 11, 910–914 (2018).

  29. 29.

    Xia, L. et al. Solar radiation management impacts on agriculture in China: a case study in the Geoengineering Model Intercomparison Project (GeoMIP). J. Geophys. Res. Atmos. 119, 8695–8711 (2014).

  30. 30.

    Proctor, J., Hsiang, S., Burney, J., Burke, M. & Schlenker, W. Estimating global agricultural effects of geoengineering using volcanic eruptions. Nature 560, 480–483 (2018).

  31. 31.

    Parker, A. & Irvine, P. J. The risk of termination shock from solar geoengineering. Earths Future 6, 456–467 (2018).

  32. 32.

    Jones, A. et al. The impact of abrupt suspension of solar radiation management (termination effect) in experiment G2 of the Geoengineering Model Intercomparison Project (GeoMIP). J. Geophys. Res. 118, 9743–9752 (2013).

  33. 33.

    Keith, D. W. Geoengineering the climate: history and prospect. Annu. Rev. Energy Environ. 25, 245–284 (2000).

  34. 34.

    Rasch, P. J. et al. An overview of geoengineering of climate using stratospheric sulphate aerosols. Phil. Trans. R. Soc. A 366, 4007–4037 (2008).

  35. 35.

    Vaughan, N. E. & Lenton, T. M. A review of climate geoengineering proposals. Clim. Change 109, 745–790 (2011).

  36. 36.

    Caldeira, K., Bala, G. & Cao, L. The science of geoengineering. Annu. Rev. Earth Planet. Sci. 41, 231––256 (2013).

  37. 37.

    Irvine, P. J., Kravitz, B., Muri, H. & Lawrence, M. G. An overview of the Earth system science of solar geoengineering. Wiley Interdiscip. Rev. 7, 815–833 (2016).

  38. 38.

    MacMartin, D. G. & Kravitz, B. The engineering of climate engineering. Annu. Rev. Control Robot. Auton. Syst. 2, 445–467 (2019).

  39. 39.

    MacMartin, D. G., Kravitz, B., Long, J. C. S. & Rasch, P. J. Geoengineering with stratospheric aerosols: What don’t we know after a decade of research? Earths Future 4, 543–548 (2016). Provides a state of the science from a field that was largely curiosity driven, providing motivation for moving to a mission-driven approach.

  40. 40.

    Hawkins, E. & Sutton, R. The potential to narrow uncertainty in regional climate predictions. Bull. Am. Meteorol. Soc. 90, 1095–1108 (2009).

  41. 41.

    Stocker, T. F. et al. in TAR Climate Change 2001: The Scientific Basis (eds Houghton, J. T. et al.) 419–457 (Cambridge Univ. Press, 2001).

  42. 42.

    Kravitz, B. et al. The climate effects of increasing ocean albedo: an idealized representation of solar geoengineering. Atmos. Chem. Phys. 18, 13097–13113 (2018).

  43. 43.

    MacMartin, D. G., Kravitz, B. & Rasch, P. J. On solar geoengineering and climate uncertainty. Geophys. Res. Lett. 42, 7156–7161 (2015).

  44. 44.

    Hofmann, D. J. & Solomon, S. Ozone destruction through heterogeneous chemistry following the eruption of El Chichón. J. Geophys. Res. 94, 5029–5041 (1989).

  45. 45.

    Tie, X. X. & Brasseur, G. The response of stratospheric ozone to volcanic eruptions: Sensitivity to atmospheric chlorine loading. Geophys. Res. Lett. 22, 3035–3038 (1995).

  46. 46.

    Heckendorn, P. et al. The impact of geoengineering aerosols on stratospheric temperature and ozone. Environ. Res. Lett. 4, 045108 (2009).

  47. 47.

    Simpson, I. R. et al. The regional hydroclimate response to stratospheric sulfate geoengineering and the role of stratospheric heating. J. Geophys. Res. https://doi.org/10.1029/2019JD031093 (2019). One of the first studies to isolate some of the surface-climate effects of stratospheric heating.

  48. 48.

    Kalidindi, S. et al. Modeling of solar radiation management: a comparison of simulations using reduced solar constant and stratospheric sulphate aerosols. Clim. Dyn. 44, 2909–2925 (2015).

  49. 49.

    Kravitz, B., MacMartin, D. G., Wang, H. & Rasch, P. J. Geoengineering as a design problem. Earth Syst. Dyn. 7, 469–497 (2016).

  50. 50.

    Rasch, P. J., Latham, J. & Chen, C.-C. Geoengineering by cloud seeding: influence on sea ice and climate system. Environ. Res. Lett. 4, 045112 (2009).

  51. 51.

    Jones, A., Haywood, J. & Boucher, O. Climate impacts of geoengineering marine stratocumulus clouds. J. Geophys. Res. 114, D10106 (2009).

  52. 52.

    Alterskjær, K. & Kristjánsson, J. E. The sign of the radiative forcing from marine cloud brightening depends on both particle size and injection amount. Geophys. Res. Lett. 40, 210–215 (2013).

  53. 53.

    Visioni, D., Pitari, G. & Aquilam, V. Sulfate geoengineering: a review of the factors controlling the needed injection of sulfur dioxide. Atmos. Chem. Phys. 17, 3879–3889 (2017). Provides a multi-model summary of SAG from the lens of solar geoengineering as a design problem.

  54. 54.

    Garfinkel, C. I. & Hartmann, D. L. The influence of the quasi-biennial oscillation on the troposphere in winter in a hierarchy of models. Part II: Perpetual winter WACCM runs. J. Atmos. Sci. 68, 2026–2041 (2011).

  55. 55.

    Seo, J., Choi, W., Youn, D., Park, D.-S. R. & Kim, J. Y. Relationship between the stratospheric quasi-biennial oscillation and spring rainfall in the western North Pacific. Geophys. Res. Lett. 40, 5949–5953 (2013).

  56. 56.

    Aquila, V., Garfinkel, C. I., Newman, P. A., Oman, L. D. & Waugh, D. W. Modifications of the quasi-biennial oscillation by a geoengineering perturbation of the stratospheric aerosol layer. Geophys. Res. Lett. 41, 1738–1744 (2014).

  57. 57.

    Kleinschmitt, C., Boucher, O. & Platt, U. Sensitivity of the radiative forcing by stratospheric sulfur geoengineering to the amount and strategy of the SO2 injection studied with the LMDZ-S3A model. Atmos. Chem. Phys. 18, 2769–2786 (2018).

  58. 58.

    Kravitz, B. et al. Comparing surface and stratospheric impacts of geoengineering with different SO2 injection strategies. J. Geophys. Res. 124, 7900–7918 (2019). Illustrates the dependence of uncertainty on the SAG scenario and strategy.

  59. 59.

    Dai, Z., Weisenstein, D. K. & Keith, D. W. Tailoring meridional and seasonal radiative forcing by sulfate aerosol solar geoengineering. Geophys. Res. Lett. 45, 1030–1039 (2018).

  60. 60.

    MacMartin, D. G. et al. The climate response to stratospheric aerosol geoengineering can be tailored using multiple injection locations. J. Geophys. Res. Atmos. 122, 12574–12590 (2017).

  61. 61.

    Richter, J. H. et al. Stratospheric response in the first geoengineering simulation meeting multiple surface climate objectives. J. Geophys. Res. 123, 5762–5782 (2018).

  62. 62.

    Irvine, P. J., Sriver, R. L. & Keller, K. Tension between reducing sea-level rise and global warming through solar-radiation management. Nat. Clim. Change 2, 97–100 (2012).

  63. 63.

    Wong, T. E. et al. BRICK v0.2, a simple, accessible, and transparent model framework for climate and regional sea-level projections. Geosci. Model Dev. 10, 2741–2760 (2017).

  64. 64.

    Boucher, O., Kleinschmitt, C. & Myhre, G. Quasi-additivity of the radiative effects of marine cloud brightening and stratospheric sulfate aerosol injection. Geophys. Res. Lett. 44, 11158–11165 (2017).

  65. 65.

    Cao, L., Duan, L., Bala, G. & Caldeira, K. Simultaneous stabilization of global temperature and precipitation through cocktail geoengineering. Geophys. Res. Lett. 44, 7429–7437 (2017).

  66. 66.

    Gao, Y., Gao, X. & Zhang, X. The 2 °C global temperature target and the evolution of the long-term goal of addressing climate change—from the United Nations framework convention on climate change to the Paris agreement. Engineering 3, 272–278 (2017).

  67. 67.

    MacMartin, D. G. & Kravitz, B. Mission-driven research for stratospheric aerosol geoengineering. Proc. Natl Acad. Sci. USA 116, 1089–1094 (2019).

  68. 68.

    Axelos. Managing Successful Projects with PRINCE2 (The Stationery Office, 2017).

  69. 69.

    Knight, F. H. Risk, Uncertainty and Profit (Univ. Chicago Press, 1921).

  70. 70.

    Cairns, R. C. Climate geoengineering: issues of path-dependence and socio-technical lock-in. Wiley Interdiscip. Rev. Clim. Change 5, 649–661 (2014).

  71. 71.

    Visioni, D. et al. Sulfate geoengineering impact on methane transport and lifetime: results from the Geoengineering Model Intercomparison Project (GeoMIP). Atmos. Chem. Phys. 17, 11209–11226 (2017).

  72. 72.

    Robock, A., MacMartin, D. G., Duren, R. & Christensen, M. W. Studying geoengineering with natural and anthropogenic analogs. Clim. Change 121, 445–458 (2013).

  73. 73.

    English, J. M., Toon, O. B. & Mills, M. J. Microphysical simulations of sulfur burdens from stratospheric sulfur geoengineering. Atmos. Chem. Phys. 12, 4775–4793 (2012).

  74. 74.

    Duan, L., Cao, L., Bala, G. & Caldeira, K. Climate response to pulse versus sustained stratospheric aerosol forcing. Geophys. Res. Lett. 46, 8976–8984 (2019).

  75. 75.

    Keith, D. W., Weisenstein, D. K., Dykema, J. A. & Keutsch, F. N. Stratospheric solar geoengineering without ozone loss. Proc. Natl Acad. Sci. USA 113, 14910–14914 (2016).

  76. 76.

    Robock, A. Blowin’ in the wind: research priorities for climate effects of volcanic eruptions. Eos Trans. AGU 83, 472 (2002).

  77. 77.

    Pope, F. D. et al. Stratospheric aerosol particles and solar-radiation management. Nat. Clim. Change 2, 713–719 (2012).

  78. 78.

    Dykema, J. A., Keith, D. W., Anderson, J. G. & Weisenstein, D. Stratospheric controlled perturbation experiment: a small-scale experiment to improve understanding of the risks of solar geoengineering. Phil. Trans. R. Soc. A Math. Phys. Eng. Sci. 372, 20140059 (2014).

  79. 79.

    Keith, D. W., Duren, R. & MacMartin, D. G. Field experiments on solar geoengineering: report of a workshop exploring a representative research portfolio. Phil. Trans. R. Soc. A Math. Phys. Eng. Sci. 372, 20140175 (2014).

  80. 80.

    Russell, L. M. et al. Eastern pacific emitted aerosol cloud experiment. Bull. Am. Meteorol. Soc. 94, 709–729 (2013).

  81. 81.

    Wiertz, T. Visions of climate control: solar radiation management in climate simulations. Sci. Technol. Hum. Values 41, 438–460 (2016).

  82. 82.

    Ferraro, A. J., Highwood, E. J. & Charlton-Perez, A. J. Stratospheric heating by potential geoengineering aerosols. Geophys. Res. Lett. 38, L24706 (2011).

  83. 83.

    Sukumara-Pillai, K. K.-P., Bala, G., Cao, L., Duan, L. & Caldeira, K. Climate system response to stratospheric sulfate aerosols: sensitivity to altitude of aerosol layer. Earth Syst. Dynam. Discuss. https://doi.org/10.5194/esd-2019-21 (2019).

  84. 84.

    Kravitz, B. et al. The Geoengineering Model Intercomparison Project (GeoMIP). Atmos. Sci. Lett. 12, 162–167 (2011).

  85. 85.

    Kay, J. E. et al. The Community Earth System Model (CESM) large ensemble project: A community resource for studying climate change in the presence of internal climate variability. Bull. Am. Meteorol. Soc. 96, 1333–1349 (2015).

  86. 86.

    Tilmes, S. et al. CESM1(WACCM) stratospheric aerosol Geoengineering Large Ensemble project. Bull. Am. Meteorol. Soc. 99, 2361–2371 (2018).

  87. 87.

    MacMartin, D. G. & Kravitz, B. Dynamic emulators for solar geoengineering. Atmos. Chem. Phys. 16, 15789–15799 (2016).

  88. 88.

    Masson, D. & Knutti, R. Climate model genealogy. Geophys. Res. Lett. 38, L08703 (2011).

  89. 89.

    Knutti, R., Masson, D. & Gettelman, A. Climate model genealogy: Generation CMIP5 and how we got there. Geophys. Res. Lett. 40, 1194–1199 (2013).

  90. 90.

    Mastrandrea, M. D. et al. Guidance Note for Lead Authors of the IPCC Fifth Assessment Report on Consistent Treatment of Uncertainties (IPCC, 2010). Provides general guidance for how to describe and assess confidence and uncertainty.

  91. 91.

    Jiang, J., Zhang, H. & Cao, L. Simulated effect of sunshade solar geoengineering on the global carbon cycle. Sci. China Earth Sci. 61, 1306–1315 (2018).

  92. 92.

    Plazzotta, M., Séférian, R. & Douville, H. Impact of solar radiation modification on allowable CO2 emissions: What can we learn from multimodel simulations? Earths Future 7, 664–676 (2019).

  93. 93.

    Dagon, K. & Schrag, D. P. Exploring the effects of solar radiation management on water cycling in a coupled land–atmosphere model. J. Clim. 29, 2635–2650 (2016).

  94. 94.

    Dagon, K. & Schrag, D. P. Quantifying the effects of solar geoengineering on vegetation. Clim. Change 153, 235–251 (2019).

  95. 95.

    Eastham, S. D., Keith, D. W. & Barrett, S. R. H. Mortality tradeoff between air quality and skin cancer from changes in stratospheric ozone. Environ. Res. Lett. 13, 034035 (2018).

  96. 96.

    Pierce, J. R., Weisenstein, D. K., Heckendorn, P., Peter, T. & Keith, D. W. Efficient formation of stratospheric aerosol for climate engineering by emission of condensible vapor from aircraft. Geophys. Res. Lett. 37, L18805 (2010).

  97. 97.

    Hamill, P., Jensen, E. J., Russell, P. B. & Bauman, J. J. The life cycle of stratospheric aerosol particles. Bull. Am. Meteorol. Soc. 78, 1395–1410 (1997).

  98. 98.

    Kuebbeler, M., Lohmann, U. & Feichter, J. Effects of stratospheric sulfate aerosol geo-engineering on cirrus clouds. Geophys. Res. Lett. 39, L23803 (2012).

  99. 99.

    Cirisan, A. et al. Microphysical and radiative changes in cirrus clouds by geoengineering the stratosphere. J. Geophys. Res. Atmos. 118, 4533–4548 (2013).

  100. 100.

    Joshi, M. M. & Shine, K. P. A GCM study of volcanic eruptions as a cause of increased stratospheric water vapor. J. Clim. 16, 3525–3534 (2003).

  101. 101.

    Mills, M. J. et al. Radiative and chemical response to interactive stratospheric sulfate aerosols in fully coupled CESM1(WACCM). J. Geophys. Res. Atmos. 122, 13061–13078 (2017).

  102. 102.

    Timmreck, C. Modeling the climatic effects of large explosive volcanic eruptions. Wiley Interdiscip. Rev. Clim. Change 3, 545–564 (2012).

  103. 103.

    Arfeuille, F. et al. Modeling the stratospheric warming following the Mt. Pinatubo eruption: uncertainties in aerosol extinctions. Atmos. Chem. Phys. 13, 11221–11234 (2013).

  104. 104.

    Rosenlof, K. H. et al. Stratospheric water vapor increases over the past half-century. Geophys. Res. Lett. 28, 1195–1198 (2001).

  105. 105.

    Solomon, S. et al. Contributions of stratospheric water vapor to decadal changes in the rate of global warming. Science 327, 1219–1223 (2010).

  106. 106.

    Dessler, E., Schoeberl, M. R., Wang, T., Davis, S. M. & Rosenlof, K. H. Stratospheric water vapor feedback. Proc. Natl Acad. Sci. USA 110, 18087–18091 (2013).

  107. 107.

    Stuber, N., Ponater, M. & Sausen, R. Is the climate sensitivity to ozone perturbations enhanced by stratospheric water vapor feedback? Geophys. Res. Lett. 28, 2887–2890 (2001).

  108. 108.

    Tilmes, S. et al. Sensitivity of aerosol distribution and climate response to stratospheric SO2 injection locations. J. Geophys. Res. Atmos. 122, 12591–12615 (2017).

  109. 109.

    Chapman, S. On ozone and atomic oxygen in the upper atmosphere. Philos. Mag. 10, 369–383 (1930).

  110. 110.

    Chapman, S. The photochemistry of atmospheric oxygen. Rep. Prog. Phys. 9, 92–100 (1942).

  111. 111.

    Groves, K. S., Mattingly, S. R. & Tuck, A. F. Increased atmospheric carbon dioxide and stratospheric ozone. Nature 273, 711–715 (1978).

  112. 112.

    Eyring, V. et al. Comprehensive summary on the workshop on “Process-oriented validation of coupled chemistry-climate models” (SPARC, 2010).

  113. 113.

    Butchart, N. The Brewer-Dobson circulation. Rev. Geophys. 52, 157–184 (2014).

  114. 114.

    Diallo, M., Legras, B. & Chédin, A. Age of stratospheric air in the ERA-Interim. Atmos. Chem. Phys. 12, 12133–12154 (2012).

  115. 115.

    Sassen, K. et al. The 5–6 December 1991 FIRE IFO II jet stream cirrus case study: possible influences of volcanic aerosols. J. Atmos. Sci. 52, 97–123 (1995).

  116. 116.

    Luo, Z., Rossow, W. B., Inoue, T. & Stubenrauch, C. J. Did the eruption of the Mt. Pinatubo volcano affect cirrus properties? J. Clim. 15, 2806–2820 (2002).

  117. 117.

    Zanchettin, D. et al. The Model Intercomparison Project on the climatic response to Volcanic forcing (VolMIP): experimental design and forcing input data for CMIP6. Geosci. Model. Dev. 9, 2701–2719 (2016).

  118. 118.

    Timmreck, C. et al. The Interactive Stratospheric Aerosol Model Intercomparison Project (ISA-MIP): motivation and experimental design. Geosci. Model. Dev. 11, 2581–2608 (2018).

  119. 119.

    MacMartin, D. G., Kravitz, B., Keith, D. W. & Jarvis, A. Dynamics of the coupled human–climate system resulting from closed-loop control of solar geoengineering. Clim. Dyn. 43, 243–258 (2014).

  120. 120.

    Kravitz, B., MacMartin, D. G., Rasch, P. J. & Jarvis, A. J. A new method of comparing forcing agents in climate models. J. Clim. 28, 8203–8218 (2015).

  121. 121.

    Yu, X. et al. Impacts, effectiveness and regional inequalities of the GeoMIP G1 to G4 solar radiation management scenarios. Glob. Planet. Change 129, 10–22 (2015).

  122. 122.

    MacMartin, D. G., Ricke, K. L. & Keith, D. W. Solar geoengineering as part of an overall strategy for meeting the 1.5 °C Paris target. Phil. Trans. R. Soc. A. Math. Phys. Eng. Sci. 376, 20160454 (2018).

  123. 123.

    Govindasamy, B. & Caldeira, K. Geoengineering Earth’s radiation balance to mitigate CO2-induced climate change. Geophys. Res. Lett. 27, 2141–2144 (2000).

  124. 124.

    Henry, M. & Merlis, T. M. Forcing dependence of atmospheric lapse rate changes dominates residual polar warming in solar radiation management scenarios. Nat. Commun. (under review).

  125. 125.

    Bala, G., Duffy, P. B. & Taylor, K. E. Impact of geoengineering schemes on the global hydrological cycle. Proc. Natl Acad. Sci. USA 105, 7664–7669 (2008).

  126. 126.

    Richter, J. H. et al. Stratospheric dynamical response and ozone feedbacks in the presence of SO2 injections. J. Geophys. Res. Atmos. 122, 12557–12573 (2017).

  127. 127.

    Driscoll, S., Bozzo, A., Gray, L. J., Robock, A. & Stenchikov, G. Coupled Model Intercomparison Project 5 (CMIP5) simulations of climate following volcanic eruptions. J. Geophys. Res. 117, D17105 (2012).

  128. 128.

    Glienke, S., Irvine, P. J. & Lawrence, M. G. The impact of geoengineering on vegetation in experiment G1 of the GeoMIP. J. Geophys. Res. 120, 10196–10213 (2015).

  129. 129.

    Hall, A., Cox, P., Huntingford, C. & Klein, S. Progressing emergent constraints on future climate change. Nat. Clim. Change 9, 269–278 (2019).

  130. 130.

    Robock, A., Bunzl, M., Kravitz, B. & Stenchikov, G. L. A test for geoengineering? Science 327, 530–531 (2010).

  131. 131.

    MacMynowski, D. G., Keith, D. W., Caldeira, K. & Shin, H.-J. Can we test geoengineering? Energy Environ. Sci. 4, 5044–5052 (2011).

  132. 132.

    MacMartin, D. G. et al. Timescale for detecting the climate response to stratospheric aerosol geoengineering. J. Geophys. Res. 124, 1233–1347 (2019).

  133. 133.

    IPCC in Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds Field, C. B. et al.) 1–32 (Cambridge University Press, 2014).

  134. 134.

    Kravitz, B. et al. The Geoengineering Model Intercomparison Project – introduction to the second special issue. Atmos. Chem. Phys. https://doi.org/10.5194/acp-special_issue376-preface (2018).

  135. 135.

    Warszawski, L. et al. The inter-sectoral impact model intercomparison project (ISI–MIP): project framework. Proc. Natl Acad. Sci. USA 111, 3228–3232 (2014).

  136. 136.

    Madronich, S., Tilmes, S., Kravitz, B., MacMartin, D. G. & Richter, J. H. Response of surface ultraviolet and visible radiation to stratospheric SO2 injections. Atmosphere 9, 432 (2018).

  137. 137.

    Nowack, P. J., Abraham, N. L., Braesicke, P. & Pyle, J. A. Stratospheric ozone changes under solar geoengineering: implications for UV exposure and air quality. Atmos. Chem. Phys. 16, 4191–4203 (2016).

  138. 138.

    Mercado, L. M. et al. Impact of changes in diffuse radiation on the global land carbon sink. Nature 458, 1014–1017 (2010).

  139. 139.

    Xia, L., Robock, A., Tilmes, S. & Neely, R. R. III. Stratospheric sulfate geoengineering could enhance the terrestrial photosynthesis rate. Atmos. Chem. Phys. 16, 1479–1489 (2016).

  140. 140.

    Irvine, P. J. et al. Towards a comprehensive climate impacts assessment of solar geoengineering. Earths Future 5, 93–106 (2017).

  141. 141.

    Mengis, N., Keller, D. P. & Oschlies, A. Systematic Correlation Matrix Evaluation (SCoMaE) – a bottom–up, science-led approach to identifying indicators. Earth Syst. Dyn. 9, 15–31 (2018).

  142. 142.

    Sugiyama, M., Arino, Y., Kosugi, T., Kurosawa, A. & Watanabe, S. Next steps in geoengineering scenario research: limited deployment scenarios and beyond. Clim. Policy 18, 681–689 (2018). Describes the importance of and process towards building inclusive scenarios.

  143. 143.

    Talberg, A. et al. How geoengineering scenarios frame assumptions and create expectations. Sustain. Sci. 13, 1093–1104 (2018).

  144. 144.

    McLaren, D. P. Whose climate and whose ethics? Conceptions of justice in solar geoengineering modelling. Energy Res. Soc. Sci. 44, 209–221 (2018).

  145. 145.

    Kravitz, B., MacMartin, D. G., Leedal, D. T., Rasch, P. J. & Jarvis, A. J. Explicit feedback and the management of uncertainty in meeting climate objectives with solar geoengineering. Environ. Res. Lett. 9, 044006 (2014).

  146. 146.

    Smith, S. J. & Rasch, P. J. The long-term policy context for solar radiation management. Clim. Change 121, 487–497 (2013).

  147. 147.

    Ribeiro, S. Against geoengineering. Geoengineering Monitor http://www.geoengineeringmonitor.org/2018/11/against-geoengineering/ (2018).

  148. 148.

    Lempert, R. J., Popper, S. W. & Bankes, S. C. Shaping the Next One Hundred Years (RAND Corporation, 2003). A fundamental reference for adaptive management strategies in climate science.

  149. 149.

    Flegal, J. A. & Gupta, A. Evoking equity as a rationale for solar geoengineering research? Scrutinizing emerging expert visions of equity. Int. Environ. Agreem. 18, 45–61 (2018).

  150. 150.

    Reynolds, J. L. Solar geoengineering to reduce climate change: A review of governance proposals. Proc. R. Soc. A 475, 20190255 (2019).

  151. 151.

    MacMartin, D. G., Irvine, P. J., Kravitz, B. & Horton, J. B. Technical characteristics of a solar geoengineering deployment and implications for governance. Clim. Policy 19, 1325–1339 (2019).

  152. 152.

    Lo, Y. T. E., Charlton-Perez, A. J., Lott, F. C. & Highwood, E. J. Detecting sulphate aerosol geoengineering with different methods. Sci. Rep. 6, 39169 (2016).

  153. 153.

    Boettcher, M., Schäfer, S., Low, S. & Parker, A. Climate Engineering Conference 2017. Conference report (CEC, 2017).

  154. 154.

    Rahman, A. A., Artaxo, P., Asrat, A. & Parker, A. Developing countries must lead on solar geoengineering research. Nature 556, 22–24 (2018). Describes the importance of building solar geoengineering research capacity in developing countries.

  155. 155.

    Solar Radiation Management Governance Initiative. DECIMALS Fund. http://www.srmgi.org/decimals-fund/, last accessed 15 September 2019.

  156. 156.

    Robock, A., Marquardt, A., Kravitz, B. & Stenchikov, G. Benefits, risks, and costs of stratospheric geoengineering. Geophys. Res. Lett. 36, L19703 (2009).

  157. 157.

    Smith, W. & Wagner, G. Stratospheric aerosol injection tactics and costs in the first 15 years of deployment. Environ. Res. Lett. 13, 124001 (2018).

  158. 158.

    Marshall, L. et al. Multi-model comparison of the volcanic sulfate deposition from the 1815 eruption of Mt. Tambora. Atmos. Chem. Phys. 18, 2307–2328 (2018).

  159. 159.

    Hardiman, S. C. et al. Processes controlling tropical tropopause temperature and stratospheric water vapor in climate models. J. Clim. 28, 6516–6535 (2015).

  160. 160.

    Oman, L., Waugh, D. W., Pawson, S., Stolarski, R. S. & Nielsen, J. E. Understanding the changes of stratospheric water vapor in coupled chemistry–climate model simulations. J. Atmos. Sci. 65, 3278–3291 (2008).

  161. 161.

    Hansen, J. E. et al. Efficacy of climate forcings. J. Geophys. Res. 110, D18104 (2005).

  162. 162.

    Waliser, D. et al. Cloud ice: a climate model challenge with signs and expectations of progress. J. Geophys. Res. 114, D00A21 (2009).

  163. 163.

    Seinfeld, J. H. et al. Improving our fundamental understanding of the role of aerosol–cloud interactions in the climate system. Proc. Natl Acad. Sci. USA 113, 5781–5790 (2016).

  164. 164.

    Wennberg, P. O. et al. Removal of stratospheric O3 by radicals: in situ measurements of OH, HO2, NO, NO2, ClO, and BrO. Science 266, 398–404 (1994).

  165. 165.

    Pickett, H. M. & Peterson, D. B. Comparison of measured stratospheric OH with prediction. J. Geophys. Res. 101, 16789–16796 (1996).

  166. 166.

    Thompson, D. W. J. et al. Signatures of the Antarctic ozone hole in Southern Hemisphere surface climate change. Nat. Geosci. 4, 741–749 (2011).

  167. 167.

    Hartmann, D. L., Wallace, J. M., Limpasuvan, V., Thompson, D. W. J. & Holton, J. R. Can ozone depletion and global warming interact to produce rapid climate change? Proc. Natl Acad. Sci. USA 97, 1412–1417 (2000).

  168. 168.

    Polvani, L. M., Banerjee, A. & Schmidt, A. Northern Hemisphere continental winter warming following the 1991 Mt. Pinatubo eruption: reconciling models and observations. Atmos. Chem. Phys. 19, 6351–6366 (2019).

  169. 169.

    McCusker, K. E., Armour, K. C., Bitz, C. M. & Battisti, D. S. Rapid and extensive warming following cessation of solar radiation management. Environ. Res. Lett. 9, 024005 (2014).

  170. 170.

    Russell, L. M. et al. Ecosystem impacts of geoengineering: a review for developing a science plan. Ambio 41, 350–369 (2012).

  171. 171.

    Daily, G. C. Nature’s Services: Societal Dependence on Natural Ecosystems (Island Press, 1997).

  172. 172.

    McClellan, J., Keith, D. W. & Apt, J. Cost analysis of stratospheric albedo modification delivery systems. Environ. Res. Lett. 7, 034019 (2012).

  173. 173.

    Moriyama, R. et al. The cost of stratospheric climate engineering revisited. Mitig. Adapt. Strateg. Glob. Change 22, 1207–1228 (2017).

Download references


The authors thank D. Visioni for helpful comments on the manuscript. Support for B.K. 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. The Pacific Northwest National Laboratory is operated for the US Department of Energy by Battelle Memorial Institute under contract DE-AC05-76RL01830. Support for D.G.M. was provided by the Atkinson Center for a Sustainable Future at Cornell University and by the National Science Foundation through agreement CBET-1818759.

Author information

The authors contributed equally to all aspects of the article.

Correspondence to Ben Kravitz.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kravitz, B., MacMartin, D.G. Uncertainty and the basis for confidence in solar geoengineering research. Nat Rev Earth Environ 1, 64–75 (2020). https://doi.org/10.1038/s43017-019-0004-7

Download citation