> It is partially self-inflicted. Fallout from nuclear incidents, particularly in the US (testing in Nevada) and northern Europe (Chernobyl), is still a measurable contributor to cancer rates.
Other industrial/chemical exposures yes, but this almost certainly isn't it. Outside of specific significant exposures, estimating cancer rates from radiation exposure is just statistical garbage. Anything at the low exposure end relies on the bottom of the linear no-threshold (LNT) model where the model is known to be wrong. (LNT is useful for public policy - you should seek to minimize the exposure from any industrial processes and materials to zero - but it is bad for public health in telling people that any exposure increases their cancer risk.)
Sounds like you might know this but I'll add it for the public dialogue.
LNT is useful because you work with an abundance of caution when it comes to radiation. It's difficult to know what type of radiation someone received and where. Both of these can dramatically change the risk of exposure. It's not hard to measure in a lab, but an accident isn't a lab and you can't just go placing sensors all over every radiation worker's body (at least yet. Small sensors embedded in clothing would change this).
So what do you do? You purposefully over estimate. Because if your estimate is wrong, the human is much more likely to survive if you incorrectly assumed they received more than they actually did than if you error by assuming they got less than they actually did. Failure analysis is a critical part to any engineering or safety plan.
Why not over estimate as much when higher dosages are received? Well that's because it matters a lot less. As dosage increases all those nuances of where and what type matter less (they still matter).
It's still all highly complex and what I'll say is that if you haven't spent at least a year studying this stuff you're more under water than you think. It's great that there's a lot of educational material out there but unfortunately when it comes to complex topics like nuclear many of them do more harm than good. Pro nuclear armchair experts tend to be as uninformed as anti nuclear armchair experts. So like the LNT, it is always good to work with an abundance of caution. Especially when talking about complex subjects on the internet
I wonder how often people don't understand how radiation affects health- because nominal levels don't hit the news. but, oh boy, when a single Fukushima isotope decay is detected on the coast of California- its national news.
We're really good at detecting radiation. Like REALLY good. It's because we spent a lot of money during the Cold War trying to detect nuclear materials. This includes underground weapons testing, being able to detect underground nuclear signatures via satellites, and even very trace amounts on people's clothing because it can help detect spies. It then was found that these things could be used for tons of stuff, such as tracking not spies lol.
But seriously, we can detect levels thousands of times lower than what's dangerous. You can even get pretty good dosimiters for like $100 these days
Other industrial/chemical exposures yes, but this almost certainly isn't it. Outside of specific significant exposures, estimating cancer rates from radiation exposure is just statistical garbage. Anything at the low exposure end relies on the bottom of the linear no-threshold (LNT) model where the model is known to be wrong. (LNT is useful for public policy - you should seek to minimize the exposure from any industrial processes and materials to zero - but it is bad for public health in telling people that any exposure increases their cancer risk.)