[jfengel]

I'll admit, I'm also skeptical about our ability to manage climate by pumping even more into the atmosphere. The CO2 models have been very accurate, but doubling the number of variables more-than-doubles the complexity. SO2, for example, is even more acidic than CO2, so we'd risk helping the temperature while making the oceans worse.

SO2's instability is somewhat reassuring, and I'm willing to consider a plan, especially if atmospheric scientists can come to a similar overwhelming consensus as they have on CO2. But until such a plan exists, I believe we need to focus heavily on decarbonization because that's a plan we are absolutely certain to be effective. To the extent that geoengineering reduces the urgency of that, it risks making things worse.

[labster]

Not sure what you mean by doubling the variables. Climate models have to take as input levels of all sorts of greenhouse gasses: CFCs, methane, water, and CO2. And land use change, which changes albedo and CO2 and water. And aerosol emissions, which are often point sources, which vary from light reflective particles to black carbon on snow. And also feedback mechanisms like deglaciation and sea ice change and three different cloud feedbacks. And change in aviation as a direct creator of clouds.

The amount of geoengineering we are already doing is massive and multivariate. If you don’t believe we can handle one more variable of a well known aerosol’s effect in our models, then you should consider disbelieving all climate models.

[jfengel]

Here's another way to put it: The "zero dimensional" model of CO2 is incredibly straightforward. You add CO2, temperature goes up. It's trivial and undeniable, and all by itself tracks fairly well with the observed warming.

Better modeling adds precision, and has been even more accurate, but is difficult for the non-expert to evaluate. It should be easy to trust its track record, but without that positive gut feeling I get from "We've been burning fossil fuels and it would be bizarre if the temperature didn't go up".

So I'm able to base my confidence in the models on a simpler model that I do understand. It's easy to accept that the existing additions (CFCs, methane, ocean currents, land, ice, etc) are valid (especially since they also track the data so well). But SO2 would be a brand-new variable, so I don't have as good an intuition.

I know it's not brand-new. As with CO2, the basic physics of SO2 are well understood, we've observed natural experiments with SO2 emission several times. It's definitely promising.

But I really want to see a full plan in place and become comfortable with it before we begin to rely on it. Because otherwise, having watched people deny that trivial, obvious stuff for ideological reasons, I expect them to seize on geoengineering as "stage 6 denialism": "It's real, it's our fault, it's not good, reducing CO2 would help, it's not too late... but only because somebody will dump SO2 into the atmosphere and so we should start all the coal factories up again".