Much of my research work deals either with climate change or air pollution. That I work on both topics is not an accident. Combating pollution and combating climate change are often treated as two sides of the same coin. Reduce emissions in the name of one and you usually get benefits for the other. That is the logic of climate (or pollution) co-benefits.
But it is not always that simple. Aerosols interact with sunlight in ways that can complicate the story. Aerosols can cool the atmosphere and the ocean by reflecting sunlight and by changing clouds. At a global scale this is the logic of solar radiation management. But non-stratospheric aerosols can also have a cooling effect and that cooling can mute warming regionally. It can also change tropical cyclones by altering sea surface temperatures. Recent work focused on South Asia and the western North Pacific shows these interactions clearly.
Missing warming in South Asia
Among climate scientists there has long been a recognition of “missing heat” in South Asia. Observed surface temperatures in parts of the region have risen less than many models would imply given the extent of global forcing. There are several mechanisms that could lead to this divergence between observed temperatures and predicted temperatures. But it is an empirical fact that aerosol loading is high in South Asia. Sulfates, nitrates, and other particles reflect radiation and can seed cloud changes that enhance cooling.
This aerosol loading has long been shown in models to offset some greenhouse warming at regional scales. Given the high levels in South Asia it is very plausible that the high levels of pollution are leading to surface cooling by increasing reflectivity. If you cool the surface, you reduce the near-term expression of greenhouse warming. That creates an apparent gap between predicted and observed temperature trends.
It is worth noting that there are other mechanisms that have been proposed to explain that gap in South Asia, including high levels of irrigation, meteorological changes, and errors in the basic climate models. It is also true that the make-up of air pollution influences the extent to which it reflects or absorbs heat. Some argue that because of high levels of black carbon in the particulate mix, pollution in South Asia could be increasing warming.
But there is a related precedent for the reflectivity hypothesis on the oceans. International shipping burned high-sulfur fuel for decades. Ship tracks seeded clouds and increased reflection over busy sea lanes. The International Maritime Organization tightened sulfur limits in 2020. Early evidence suggests the resulting drop in sulfate aerosols reduced that localized cooling and lead to increases in sea surface temperatures near heavy shipping lanes. That is a parallel to what South Asia could experience as pollution control improves.
Fewer and weaker cyclones in the western North Pacific
A similar story seems be emerging for tropical cyclones. The mechanism is straightforward. Tropical cyclones draw energy from warm surface waters. If aerosols cool the surface, potential intensity falls and cyclone frequency and power can drop. There is some evidence that increases in anthropogenic aerosols over South and East Asia likely cooled the western North Pacific. That cooling coincides with declines in cyclone activity there. Over the same period, Europe and the United States cut aerosol emissions. That reduced cooling in the North Atlantic and helped create conditions for stronger and more frequent Atlantic cyclones.
What should we do with this information?
I do not think the lesson of the evidence for this relationship between aerosols and climate change is that we should avoid reducing pollution. The health, productivity, and welfare costs of PM2.5 are very high. Air quality policy delivers large, immediate gains. That remains true.
But it does imply a potential trade-off. Cleaning up local air pollution comes with large benefits. But in regions where aerosols have been reducing heat, lowering pollution might lead to a snap-back (in the language of geoengineering, a termination shock). All else equal, ambient temperatures will rise toward the greenhouse baseline. That can push temperatures higher and lead to increased heat stress.
There are other practical implications as well. First, we should not treat the recent temperature record in high-aerosol regions as the best guide to the near future, especially under climate change and in the face of policies to improve air quality. The removal of aerosol cooling can change local trends quickly. Second, adaptations to higher heat may be needed along with improvements in air quality. Increasing adaptation to heat could reduce any heat related increases in mortality that might otherwise offset some of the mortality benefits of reducing pollution.
We should also set expectations correctly. In regions where aerosols have been leading to meaningful cooling the aggregate, net benefits of pollution control might be smaller than expected if we ignore effects on heat. Heat mortality could rise as air quality improves. Cyclone risk could increase in basins where aerosol cooling had been suppressing storm power.
I think this also underlines the importance of moving forward on adaptation regardless of what happens on emissions mitigation. The argument that mitigation should be prioritized over adaptation isn’t as common as it used to be, but this should put it to rest. To the extent that aerosols are leading to meaningful cooling, emissions reductions will lead to short-term increases in temperature that can only be addressed with adaptation.
Implications for economists
Beyond the fact that we may want to think about how to couple climate adaptation with air quality improvements, I think the implications of the relationship between aerosols and cooling for economists are less clear than some of the other topics I’ve written about recently. But I also think that this relationship is a somewhat surprising fact about the world that may be important for places where many of us work. So perhaps there are more direct implications than I’ve thought of and this will lead to new research directions for some of you.
Join the Conversation