- Uncertainty
- High
- Decision relevance
- Medium
- Resolvability scale
- Large-scale testing
Observations of SAI’s effects would be essential during any deployment for two main purposes. First, sufficient observations of the aerosols in the stratosphere (their spatial distribution, size distribution, composition, lifetime, and radiative impacts), as well as the concentrations of the precursor gas (e.g. SO2) if released in gaseous form, would be required to enable adjustment of the deployment strategy in response to the real-world response. For example, if aerosol size is larger than expected and lifetime shorter, then the total injection magnitude would need to be increased to achieve a target temperature pathway. Second, a broader set of observational capabilities would be required to test the accuracy of our predictions of the earth system and surface climate response to SAI, and update these predictions in light of the evidence. This uncertainty entry deals with the first set of observations, while the second are discussed in the climate response monitoring uncertainty. The zonal mean, monthly mean aerosol optical depth or top-of-atmosphere shortwave fluxes will be necessary for a comparison against the expected magnitude and north-south pattern of SAI forcing. While in climate models this information (or even simply the pattern of surface temperatures) might be sufficient to update injection strategy controllers, in any real-world deployment, we would want to understand _why_ the pattern of aerosol optical depth or forcing differs from that which we expected, not just _that_ it does so. Understanding this would require more detailed monitoring of features of the aerosol distribution, such as composition and size distribution. It is also possible that better spatial or temporal resolution for aerosol optical depth will be necessary to resolve the mechanisms for the difference.
Metric
Time to define, develop, build, and deploy sufficient stratospheric process observational infrastructure is more than 10 years from first large-scale funding
Uncertainty
There are currently significant uncertainties in exactly what must be observed, and at what resolution and sampling frequency, so it is unclear what infrastructure is needed or how difficult this will be to put in place, including which satellites will carry instruments. This uncertainty is added to by upcoming instrument retirements. The fundamental reason for uncertainty here is not in the physics – we know how to measure the relevant quantities. It is because the work to define observational requirements has not yet been completed, and because it may take a long time to get sufficient assets in place once these requirements are defined. Typical timelines for moving from initial concept to delivering data are over a decade for NASA and ESA satellites. For example, the EarthCARE satellite took over 20 years to move from proposal to its eventual launch in 2024.
Decision relevance
If observational infrastructure for monitoring the aerosols is insufficient, this could either delay deployment until sufficient baselining is done, or could result in poorly monitored deployments. The latter would add significantly to uncertainty in the assessments of SAI's impacts and efficacy.
Further Information
This uncertainty covers aerosol observations by both in-situ (balloon, and aircraft) and remote (satellite) infrastructure. We assume that the required observed characteristics include sulfate aerosol composition and size distribution, as well as the horizontal spatial distribution of the aerosols, their vertically resolved concentrations, and the total column aerosol optical depth. Part of the reason for uncertainty here is that the precise observational requirements (e.g. what spatial resolution and sampling frequency is required) are not yet clear and new research – particularly using Observing System Simulation Experiments (OSSEs; Errico et al., 2012; 2018) – will be required to define these.
Monitoring of background stratospheric conditions and variability would be required at least a few years before any large-scale releases or deployment, in order to define the pre-SAI baseline. Exactly how long a baseline is required for each feature of the stratospheric background conditions depends on its variability, and the presence of the stratospheric quasi-biennial oscillation in winds means that more than two years will be necessary at minimum.
Development time for satellite-borne remote sensing instruments can be significant, and some existing infrastructure is retiring without replacement (e.g. Salawitch et al., 2025). However, time-to-launch can be short for instruments which already have a high Technology Readiness Level (TRL-6 or higher; e.g., the continuity-MLS (NASA, 2022) instrument could be operational within 3.5 years from funding), and various relevant instruments are in development or available, including the solar occultation instrument GLOTemp (Laboratory Gas Filter Correlation Radiometer for Limb Occultation demo for upper atmosphere Temperature). The recently launched EarthCARE mission has a scanning LiDAR that will provide useful aerosol measurements (Wehr et al., 2023). Upcoming selected missions should provide useful data as well, including ALTIUS (ESA’s Atmospheric Limb Tracker for the Investigation of the Upcoming Stratosphere; ESA, 2022) and STRIVE (Stratosphere Troposphere Response using Infrared Vertically-resolved light Explorer; NASA 2026).
In situ observational capabilities are well established, with recent stratospheric airborne missions such as SABRE (Lyu et al., 2026) and DCOTSS (Bowman et al., 2026) having obtained excellent aerosol records. Regular balloon measurements of aerosol number and size distribution are currently occurring at several locations around the world through programs like B2SAP (NOAA; Todt et al., 2023) and BalNeO (NASA). The timeline to develop and deploy these in a more widespread capacity is therefore relatively short, once it becomes clear which of these observations will be necessary on an ongoing basis.
We indicate that this is resolvable "In silico" because it does not require SO2 injection, even though it would not in reality be resolved exclusively via modeling.
References
10Have feedback on this entry?
Spot something off, missing, or worth adding? We’d love to hear from you.
Share feedback