The price of uncertainty: geoengineering climate change through stratospheric sulfate
Stratospheric sulfate seems to be one of the most promising geoengineering methods to combat climate change. It involves the injection of hydrogen sulfide (H2S), sulfur dioxide (SO2) or other sulfates, into the stratosphere. Similar to what happens after major volcanic eruptions, this would reflect off part of the sun’s energy and cool the Earth, counterbalancing the effect of greenhouse gases (see for instance the “Year without a Summer” that followed the 1815 eruption of Mount Tambora).
It is probably the best geoengineering solution to climate change, in that it’s likely to work, should be technically feasible, can be done by a single nation if need be (no need for global consensus), and is likely to be very cheap – especially in comparison with cutting emissions. But it has a few drawbacks:
- It will have unpredictable effects on the weather across the globe.
- We can’t really test it – the test would be doing it, on a global scale.
- We wouldn’t know if it worked until we’d had about a decade of temperature measurements.
- Once started, it’s extremely dangerous to stop it – especially if carbon emissions keep rising.
So, should we do it? Narrow cost-benefit analysis suggests yes, but that doesn’t take into account the uncertainty, the unknown unknowns – the very likely probability that things will not go as expected, and that we’ll have difficulty dealing with the side effects. This includes the political side effects when some areas of the globe suffer more than others from this process.
How bad does global warming have to get before we consider this type of nearly irreversible geoengineering? If we had to choose between this and cutting emissions, how high would the cost of cutting have to go before we sprang for this instead? In short, what price do we put on avoiding uncertainty on the global scale? Can we estimate a dollar amount, or some alternative measure of the cost – quality-adjusted life years, or some other human-scale estimate? Or is this an illusionary precision, and do our intuitions and qualitative arguments (precautionary principle?) give us a better estimate of whether we should go ahead with this?