Figure 1. Taipei high dose rate profiles versus Okuma high. First 5 years only.
In our discussion of the Taipei apartment exposure, I made the comment that the dose rate profiles for the people who lived in the Co-60 contaminated flats for at least nine years were similar to the dose rate profiles which would have been incurred at Fukushima had there been no evacuation. That statement turned out not to be very descriptive, to put it most politely.
Figure 1 compares the worse case Taipei dose rate profiles according to Chen\cite{chen-2004} and Hwang\cite{hwang-2006} with the GKG's estimate of the worst case Fukushima profiles to the public assuming no evacuation. This was our Okuma high group (top 10%) which we think started out with an ambient dose rate of about 310 uGy/h. In concocting this profile, we assumed a location factor of 0.2.1 This number is consistent with both Golikov's analysis of Chernobyl and Miyazaki's actual body dosimeter measurements at Date-shi, a town just outside the Fukushima Evacuation Zone. In the latter case, the dose rates measured by the dosimeters that the people carried were on average 15% of the outdoors ambient dose rates with a standard deviation of +/-0.03.\cite{miyazaki-2017}[p 7]
We can be confident that Hwang's worse case was a lower bound on the actual worse case at Taipei. But let's focus on that profile. The Okuma high group starts out with a higher initial dose rate than Hwang's max, but thanks to the rapid decay of Te-132/I-132 and I-131 by day 13 the Okuma profile drops below the Hwang. The Okuma profile then quickly drops to less than half the Hwang before flattening out. Eventually, the Taipei profiles will drop below the Okuma profile. The LNT cancer mortality for Hwang's worst case is 0.1182. The SNT cancer mortality is 0.00029. The same numbers for Okuma high are 0.0206 and 0.00002. Given our inability to reliably detect any increase in cancer in the apartment dwellers, the Fukushima high group almost certainly would not have any detectable risk had they not evacuated.
Figure 2. Fukushima release profile
There are a couple of important caveats.
1) The plant must have a buffer zone. The buffer zone at Fukushima was a little more than a kilometer. The NRC does not worry about buffer zones, but the GKG does. By GKG standards, Fukushima had both too small a buffer zone, and two small a mini-buffer around each unit. For a multiple GW unit plant, the minimum distance from any reactor to a residence should be at least two kilometers, and the units should be at least 200 m apart. But in the Fukushima case, we got away with it.
2. GKG's estimate of the Okuma high profiles does not include inhaled dose during actual plume passage. At Fukushima, the release was mainly in the form of a series of puffs, Figure 2, each lasting about an hour or two or less. The GKG profile implicitly assumes that everybody masks up during plume passage. The NRC does not even mention masking up; and, as far as I know, there was no systematic masking up outside the plant at Fukushima, although the photos indicate many people did.
The cost of masking up is trivial. But despite all the mouthing off about "safety is our over-riding priority", the NRC is not in the safety business. It's in the self-preservation business. That's why it would have evacuated every body within 50 miles of the plant causing immeasurable harm. The NRC is not dangerous; it's deadly.
The location factor is the ratio of the actual dose a person incurs to the ambient dose rate measured outside his residence.
"For a multiple GW unit plant, the minimum distance from any reactor to a residence should be at least two kilometers, and the units should be at least 200 m apart."
Maybe very large plants should be avoided for purposes of decentralizing power generation,
but I still think your guidelines are overly restrictive, and should take into account the reactor technology, inventory, and potential (pressure or chemical) to mobilize any significant quantity of harmful isotopes. In an MSR, the danger is essentially nil; some of the offgas that hasn't yet been bottled. A thermal breeder has minimal fissile inventory, and with ongoing FP removal, the waste inventory can also be made almost arbitrarily small. Actually, most of the isotopes won't even be waste, they'll be products in great demand.
I don't want our most critical infrastructure to be out in the middle of nowhere, connected to load centers by long and vulnerable power lines, where they also can't be used to heat cities. I think we should welcome (suitable) nuclear plants in the center of town in a city park or other green space, and bury the trunk lines. Also, does that 200m apply to 250MWe reactor modules of a GW scale plant? Maybe not initially, but I believe these are a good goals that should not be prohibited on account of older technologies, and we shouldn't make rules that will unnecessarily explode the size of a modular plant. (not talking about so-called SMRs here, which are necessarily huge anyway, at least relative to their power output)
I like that you emphasize that inhaled particles are not in the model. Shame on the NRC and everyone else for not distributing masks in response to the Fukushima release.
Another big difference between Fukushima and Taipei is a a one-time exposure for a few days, vs exposure every day for a long time. I believe the Kramatorsk exposure (20 - 30 mSv/day for years) shows that there is a "saturation" effect overwhelming our DNA repair mechanisms. A single exposure of 20 mSv shows no measurable harm (bomb survivors). The same repeated every day could be fatal.
The Kramatorsk exposures were much higher than Taipei or Fukushima and well beyond anything we might expect from an NPP release. The bomb survivor data is what I would trust in predicting the harm from an NPP release.