Comment by atoav

> It's not clear to me if the study is isolating the variable they're measuring properly.

> Surely there's a "control" shipping lane somewhere that was cleaner to begin with or never cleaned up.

As mentioned in the first paragraph of the article they are using the Global Lightning Detection Network, which is well, global. Then you just need a map of SO2 concentration and compare shipping lanes against non-shipping lanes. You don't need an explicit control group if your data includes the whole planet, since you can just compare shipping lanes against similar areas with less/no shipping. Since both lightning and SO2 also varies over time you can also correlate this way with enough data.