Binary categorization is quite appropriate, considering there are IDLH levels. And it is actually highly nonlinear for most contaminants. There's an area of non

Binary categorization is quite appropriate, considering there are IDLH levels. And it is actually highly nonlinear for most contaminants. There's an area of none to low risk, even chronic, then an area of exponentially increasing risk, then area where the levels are known to be harmful chronically, and yet another where they're immediately dangerous. So while not binary, the difference between dangerous and not is usually pretty sharp. For PM 2.5, < 12 ug/m3 is considered safe, while > 35 ug/m3 is considered harmful after just few hours. For chronic exposure, mean levels of < 15 ug/m3 over 3 years is considered safe. (There is probably some level that is certain to produce silicosis in about everyone. I would have to dig on that.) That is how thin the margin is. And these are the laxer NIOSH values. Even AQI which is weighted to be more linear and granular, easily gets pegged at moderate (risky to sensitive people) in clean seeming cities unless there's wind... (That is a level above 15 ug/m3 already by some.)