My rationale: the graphs have similar shape. It's hard to imagine anyone denying a connection between BAC and the relative likelihood of a vehicle crash beyond a certain point in BAC. People might argue that they might be better drivers after one drink, especially if the drink makes them more relaxed and less nervous. Beyond the first drink one has no question: the higher the BAC, the worse one's driving gets.
Nobody responsibly denies the connection between the BAC and the impairment of driving beyond a certain point. At any level of atmospheric carbon dioxide there will be some statistical noise in the correlation of temperature to the atmospheric concentration of carbon dioxide. There are volcanic eruptions and sunspot cycles. The big numbers that relate to commonplace vehicle collisions, very frequent events, make allowances for weather, time of day, and the quality of highways. We don't have as big numbers for world temperatures.
The idea is less crazy than it seems. The common thread, of course, being that global climate systems and the human body have a certain amount of resilience/"buffer" that can absorb small to moderate amounts of shock in terms of GHG or booze or whatever without ill effects, but at a certain point you hit an inflection where the natural feedback systems get overwhelmed and any further stress will lead to genuine impairment.
This isn't necessarily meant as a defense of that Wiki-sourced drunk driving chart, of course (which I can't vouch for); just an explanation/defense of the underlying idea.
Bingo!
