What Can We Learn About SRM From Climate Modelling?

Climate models are a tool I’ve been fascinated with since the first time I came across one – thousands of lines of code describing the laws of fluid dynamics and all the many interactions between the various components of the Earth system. They capture processes ranging from the tiniest scales – how photons interact with molecules of gas – to the largest – how colossal streams of oceanic waters flow across the world and disappear into untold depths. Despite blunders and imperfections, the fact that we manage to pull all this together into a coherent whole and produce features that resemble the real world is pretty cool.

Aside from their beauty, climate models are also incredibly useful. We can use them to explore the past and potential futures and to tease out the inner workings of real-world phenomena. For example, we can explore how (and why) El Niño affects rainfall patterns in New York, or how the rising concentrations of carbon dioxide change the frequency and likelihood of heatwaves over Mexico.

Mind you, climate models are not crystal balls predicting an inexorable future with full certainty, nor are they guides to what climate policy is best. They can, however, help us explore the consequences of human choices, providing an essential input alongside other more subjective, but no less important, factors.

This is also how we should think about using climate models to understand and explore SRM. Nobody will ever run a climate model experiment that would tell us definitively whether or not to do SRM. Those kinds of decisions do not rest in the hands of climate scientists or their models. But the understanding we can glean from climate modeling of SRM can help support informed decisions about its future use or non-use.

There are several ways climate models can help to support these decisions. They can be used to evaluate the consequences of SRM deployment scenarios as realistically as possible, simulating both well-intentioned deployments as well as messier geopolitical possibilities in great detail. They can also be used to simulate clear, simple experiments where SRM is boiled down to “less sunlight reaches the Earth”, which can help us to unpick the physical mechanisms driving the more complex responses we see in more realistic scenarios.

Analyzing climate model simulations highlights parts of the climate system we don’t understand too well – or not at all – helping to identify where more work is needed. For example, identifying where model developments are needed to capture missing processes, where new or more consistent observations would provide some ground truth, or where targeted field experiments could test key processes.

Looking back over the tens of SRM climate model simulations and hundreds of modeling papers, a few key high-level findings stand out for me:

• Reducing incoming sunlight would cool the planet. We’ve found a surprisingly simple rule of thumb: a 1% reduction in the amount of incoming sunlight reaching the Earth would cool the surface atmosphere by about 1°C.¹
• A uniform reduction in sunlight would not produce a uniform cooling. As more sunlight reaches the tropics than the high latitudes, a uniform reduction in incoming sunlight offsetting an increase in greenhouse warming would result in a cooler tropical belt and some residual warming at high latitudes.² However, both would see much less change than with just the greenhouse warming.
• A uniform cooling can be produced by carefully distributing a stratospheric aerosol layer. You can reduce this disparity if you, somehow, reduce sunlight more at high latitudes than you do at low latitudes. For instance, if you want to offset 1°C of greenhouse warming evenly, you might need to reduce sunlight by 2% close to the poles, and by 0.5% close to the equator.³ Modeling shows that such reductions could be achieved with SAI by carefully tuning the injection of aerosols at different latitudes in the stratosphere.³
• A uniform cooling could reduce many aspects of climate change but would still not be perfect. Under climate change, along with warming we expect other climate changes. For example, many areas will experience large changes in rainfall and rainfall extremes, leading to some regions getting much wetter and others much drier, as well as most regions seeing both more floods and more droughts.^4,5 SAI, if aimed at curbing global warming through a cooling as uniform as possible, would offset some of these changes to rainfall and rainfall extremes, but would result in a reduction in global rainfall as well as some shifts in where precipitation happens; for many regions, these changes look smaller than under climate change, but this might not be true everywhere.^6–8
• Asymmetric cooling could produce very large shifts in rainfall. If the cooling from SRM is not uniform but contained to only one hemisphere (say, because aerosols are injected only at 45°N), this would lead to dramatic shifts in tropical precipitation. The tropical rainfall belt would shift towards the warmer hemisphere due to the imbalance of the interhemispheric energy fluxes.^9,10
• Stopping SRM suddenly would lead to a rapid warming. If SRM were to be stopped suddenly and permanently, the warming that it had offset would come back in a matter of years.¹¹ If the amount of cooling was significant, then the rapid warming would result in large, rapid climatic shifts that could have large impacts on ecosystems.¹²

These results are not one-offs. They have been derived from multiple climate models and analyzed by multiple groups all over the world. Just like similar results for climate change, these kinds of broad, high-level findings seem likely to stand the test of time, though they are always open to further scrutiny and refinement.

While we can draw some robust, high-level conclusions about SRM in general, we are not yet at the stage where we have much confidence in the details of specific scenarios at regional scales. These kinds of broad conclusions that climate modeling of SRM can provide today will get more complex and uncertain as one looks at different scenarios and more local scales. For some of these results, we have created a web-based simulator that allows anyone to explore the potential impacts of SRM on temperature and precipitation.

Despite the uncertainties, the high-level view from “model land” indicates that a well-managed SRM effort could be a useful tool in a broader portfolio of policies to manage the risks of climate change. Even used well, SRM is not a panacea, having many risks and shortcomings. Understanding whether it would be well-managed – whether it might undermine emissions cuts or how it could interact with geopolitical tensions – well, that lies far beyond what any number of lines of code can tell us.

The views expressed by Perspective writers are their own and are not necessarily endorsed by SRM360. The goal of our Perspectives is to present ideas from diverse viewpoints, further supporting informed discussion of sunlight reflection methods.