With global warming standing at +1.4  in 2023, we are just a step away from the +1.5  aim set in the Paris Agreement. To buy some time while reaching net-zero carbon emissions, Marine Cloud Brightening (MCB) has been proposed and practiced in Australia to protect the Great Barrier Reef¹. This involves deliberately injecting aerosols into clouds to increase cloud reflectivity, thereby cooling the Earth’s surface. However, the longstanding gap in our understanding of aerosol-cloud interactions (the underlying principles of MCB) induces large uncertainties in aerosol cooling assessments. These large uncertainties have hampered accurate climate projections and mitigations, also have hindered accurate evaluation of effectiveness and side effects of the MCB strategy. Our study points toward a direction for the community to improve our fundamental understanding and provides robust observation-based ground truth to constrain models and achieve this goal. 

Thanks to the availability of long-term satellite observations and recent breakthroughs in our novel machine learning approach², we have unambiguously quantified aerosol’s impacts on clouds over the North Atlantic Ocean. The derived aerosol fingerprint on clouds is robust as it overcomes critical limitations in traditional methods, such as scale-limitation³ or the confounding of meteorological co-variability^4,5. Now, we use this novel method to take a step further and evaluate the efficacy of MCB over the tropical Pacific, where sunlight is strong. In line with the MCB strategy focusing on tropical regions, we selected degassing volcanic eruptions of Kilauea volcano in Hawaii as natural experiments. As Hawaii is located in the trade-wind region, the volcanic plumes were continuously blown to the ocean west of Hawaii, impacting the prevalent trade-wind cumulus clouds (Fig 1a). The machine learning method was trained by cloud properties and meteorology in non-eruption years to predict the cloud properties with eruption-year meteorology but without volcano plumes. Comparing this counterfactual result with satellite observations in the eruption year provides a unique volcanic aerosol fingerprint on trade-wind cumulus in tropical regions.

We found a significant increase in cloud cover by up to 50% relatively in humid and stable meteorological conditions as a response to a 30% increase in cloud drop number concentration, which is a result of volcanic aerosols (Fig 1b). Current global climate models with simplified cloud cover schemes are not able to capture such a strong increase in cloud cover. In other words, the cooling effect associated with aerosol-cloud interactions could be much stronger than previously suggested by the state-of-the-art climate models. This implies a dichotomy in climate science. On the one hand, this implies that the Earth climate system is very sensitive to external climate forcings, as aerosol cooling from anthropogenic emissions likely offsets a considerable amount of the radiative forcing from greenhouse gases. This implies that aerosol emission reductions in the future may unmask significant warming. On the other hand, MCB could be surprisingly effective in mitigating future global warming than current global model estimates, with tropical cloud cover enhanced by up to 50% relatively. However, we are also duty bound to caution that any practical MCB deployment strategy currently has very large uncertainties and therefore potentially unforeseen risks, stemming from the large gaps in our current understanding of aerosol-cloud interactions, particularly through the response of cloud cover.

Fig.1 | Cloud cover responses to aerosol perturbation. (a) Conceptual picture of volcanic aerosol plume interacts with shallow convective marine clouds, leading to an increase of cloud cover, precipitation, and more reflected solar radiation back to space. (b) The responses of cloud cover to 30% increase in cloud droplet number concentration (global average of change from pre-industrial to present-day) depend on meteorology conditions and cloud regimes. The black cross of Chen et al. (2022)²  represents a more generalized analogue for the global cloud regime spectrum, blue and red triangles represent responses of tropical marine clouds with implications for marine cloud brightening. (The figure for this blog is composed of 2 panels. Panel a = Fig 1b from the original paper. We would like to thank Chantal Jackson (University of Birmingham) for her contribution to the conceptual figure. Panel b = Fig 4 from the original paper).

References:

1. Tollefson, J. Can artificially altered clouds save the Great Barrier Reef? Nature. 596, 476-478 (2021). https://doi.org:10.1038/d41586-021-02290-3
2. Chen, Y. et al. Machine learning reveals climate forcing from aerosols is dominated by increased cloud cover. Nature Geoscience. 15, 609–614 (2022). https://doi.org/10.1038/s41561-022-00991-6
3. Glassmeier, F. et al. Aerosol-cloud-climate cooling overestimated by ship-track data. Science 371, 485-489 (2021). https://doi.org:10.1126/science.abd3980
4. Malavelle, F. F. et al. Strong constraints on aerosol–cloud interactions from volcanic eruptions. Nature. 546, 485-491 (2017). https://doi.org:10.1038/nature22974
5. Rosenfeld, D. et al. Aerosol-driven droplet concentrations dominate coverage and water of oceanic low-level clouds. Science. 363, eaav0566 (2019). https://doi.org:10.1126/science.aav0566