Radiation Harm Models don't know Safe

Correction: the incorrect statement about EPA’s CERCLA limit was inadvertently left in. See second footnote. This version has the offending statement deleted. Mea culpa.

"Safe" is a four letter word that should be banned from polite discourse. Safe is little more than an emotion. Safe is so rubbery that it's a favorite monosyllable of advertisers and politicians and other misleaders. I try to avoid s**e like sin in my writing; but for this piece I had to make an exception.

What's "safe" is a subjective. Most of us regard flying commercial as "safe"; some of us do not. Both are right. What we can agree on is that the fatal casualty rate is about 0.2 per million departures.\cite{boeing-2019} 0.2 per million is just a number. Whether that number is "safe" or not is a separate, subjective judgement.

You will often hear LNT called "the model that claims there is 'no safe dose'". LNT makes no such claim. LNT converts a dose rate profile into a cancer incidence prediction. That's it. LNT like many other possible models generates a positive cancer incidence for any dose rate profile whose cumulative dose is positive. So in calling LNT the "no safe dose" model the caller is making his (not LNT's) judgement that any non-zero dose is "unsafe". If the caller is a pro-nukie, he should rethink his support of nuclear power.

A radiation harm model is a mechanism for converting an individual's dose rate profile into a prediction of cancer incidence. Nothing more, nothing less. The model should be judged on its consistency with established biology and the accuracy of its predictions. Period.

What level of risk is acceptable in return for the benefits of what ever caused that risk is a totally separate issue. Ideally the acceptable level of risk would depend on the benefits. It is axiomatic that we should accept more risk for something that is highly beneficial than we would for something that results in very little benefit. Almost always one of those benefits is the reduction in harm by eliminating what ever the alternative is to the activity under consideration. Someone may drive rather than take an airplane but, unless the trip is very short, she will almost certainly increase her risk of death. Replacing nuclear electricity with some other form of power almost always increases overall risk.

But that's not the way regulators think. They try to come up with direct risk levels that are independent of the benefits, but are "acceptable" or "tolerable" in some sense, Table 1 shows some of these attempts.

The UK and EPA numbers are not directly comparable: one is annual all cause deaths, the other is lifetime cancer incidence. But if we assume an 80 year life and half of cancers result in death, the UK 1.0e-6 "broadly acceptable" annual is roughly equivalent to a 0.4e-4 life time risk of cancer.

Table 2 shows the SNT and LNT constant 80 yr dose rates associated with a range of cancer incidences. According to both the EPA and the UK HSE, the tolerable cancer risk is somewhere between 1.0e-4 and 1.0e-5. For LNT the lifetime constant dose rates associated with such risks are between 0.0125 and 0.00125 mSv/y.¹

LNT does not make any statement about what's "safe"; but, if your definition of "safe" is the EPA's "generally acceptable", then these are the "safe" LNT dose rates.² They are also nonsense. They are a tiny fraction of background dose rates, and far, far below the dose rates at which we have ever detected increased cancer incidence.

The SNT flat lifetime dose rates associated with these risks are between 40 and 14 mSv/y. The most horribly contaminated radiation clean up site in the USA is Hanford. The maximum dose rate at Hanford is 6 mSv/y. Under an SNT guided EPA, all the radiation clean up programs would be shut down.

BTI and others have attempted to interpret Congress's use of one in a million in the Clean Air Act is a statement that such risks are below regulatory concern. That's not what the Act says. The actual wording is opaque legalese. Here's an example.

42 USC 7412 (f)(2) If standards promulgated pursuant to subsection (d) of this section and applicable to a category or subcategory of sources emitting a pollutant (or pollutants) classified as a known, probable or possible human carcinogen do not reduce lifetime excess cancer risks to the individual most exposed to emissions from a source in the category or subcategory to less than one in one million, the Administrator shall promulgate standards under this subsection for such source category.

But reading between the lines a bit, Congress pretty clearly felt that a lifetime risk of 1 in a million need not be regulated. Unfortunately, it did not say so in anything close to an unambiguous fashion. But they could correct that. Let's assume they do,

Under LNT, nothing would change. The LNT dose rates associated with an 80 year exposure that results in this risk are so tiny, that no activity involving radiation would be exempted from regulation. But under SNT, the lifetime dose rate associated with the one in a million risk is about 5 mSv/y. Quite a few activities would below regulatory concern. For example, it would be easy and cheap to come up with Low Level Waste landfills that met this requirement. Moving from LNT to SNT would be transformative.

Neither LNT nor SNT says anything about what "safe". and I am not a fan of regulating by arbitrary "safe" limits. Each activity has a different optimal level of radiation exposure. What we need is a balancing process that reacts to specific benefits and risks. This is precisely what the insurance market does. But if we must go the arbitrary limit route, and we anthropomorphically project our vague, confused, emotive ideas of safety on a few lines of computer code, we do end up with very different policy implications.

But models must not be chosen on their policy implications. That's how we got LNT. Models must be chosen on how well they replicate reality.

1. Quoting radiation dose rates on an annual basis is biological nonsense and dangerous. This substack breaks all my rules. Destroy after reading.

2. In practice, the dose rate limits EPA promulgates are far higher. For example, the EPA clean up level for CERCLA sites is 0.15 mSv/y. Such limits tell me that EPA knows LNT is nonsense.