Jeffrey LewisRadiation Doses

I’ve been struggling with how to express the risk from radiation doses without sounding like either a fear-mongering lunatic (“You will die from a slow and painful cancer!”) or some nuclear-industry hack (“Really, it’s no worse than smoking a cigarette in Denver while getting a chest CT scan!”).  The key concepts have to do with the concept of cumulative doses and probabilistic risk assessments, or something, but I can’t quite articulate them.

Anyway, XKCD just took a break from being awesomely funny to just be awesome.  What a chart!

Thanks to reader CT for linking to it on Facebook.



  1. bks (History)

    10 microsieverts/hour for one year, 8760 hours, would be a dose of 87600 microsieverts.

    87600 microsieverts is 87.6 millisieverts.

    75 millisieverts is enough to get you into the xkcd lower right “Quadrant of Doom”.

    So a persistent reading over 10 microsieverts/hour is cause for alarm.

    What is the current rate in Sendai, a city of one million, 60 km north of the reactors?


  2. T Nishi (History)

    Reading in Sendai at 9:21 – 9:31 JST on March 21 was 0.16 microsiverts/hour. Readings over 10 microsieverts/hour were observed in some places within 40 km of the plant, and would be cause of alarm if persistent over months. Source: Ministry of Education, Culture, Sports, Science and Technology’s monitoring vehicles.

  3. Seb Tallents (History)

    The “quadrant of doom” area is a little bit strange in that the empirical data going into it tends to be focused into people who have received a large dose over a short time.

    I’m not sure what the verdict is on the elevated levels for a long period of time.

    I think there is another analogy that can be made: a lot of people get scared when they hear that “the radiation is so high that people periodically have to stop work”. This gets translated in peoples head like the workers are next to a fire and have to down tools because the fire is now too hot to be near and will burn them.

    That’s not quite right. People have to stop work because the annual exposure limit is 250mSv, whether you get that in one hour, or over 356 days. Each mSv/Hr translates into less time that a worker can be on the site before he has to go home. The level is around the point where it starts to give you an enhanced chance of cancer, so from the point of view of labor law, this is reasonable. We wouldn’t tolerate an employer requiring you to smoke a pack of cigarettes a day. On the other hand, there are plenty of industries that will increase your risk of cancer if you work there.

    This is less like “to high a temperature to be here, you will get burnt” and more like “I’m afraid you have had enough beer for tonight, would you like me to call you a cab”.

    The question is, how long does it take to sleep off the hangover with radiation? As I understand it, this isn’t known well.

    • Eve (History)

      There is lots of literature out there if you want to go search for it.. but there is little talk about deterministic and stoichastic effects of radiation.

      Often there is controversy because some scientists think that some radiation can be good because it stimulates your repair mechanisms into action and that a no stimulation allows mutations to slip through. Chronic low radiation is going to cause stoichastic effects and will increase the mutation load in the offspring. This is clearly documented from non-human case studies in Chernobyl.

      My opinion is that their work increase to 250 mSv is just too great. Surely they would have to broaden this exposure with time.

      Humans already carry a huge mutation load (around 10 more heterozygous lethals). Perhaps there could be follow up studies on workers recieving such increased levels, to define theories such as the linear no-threshold etc., especially when DNA sequencing is set to become inexpensive.

  4. Taylor Wray (History)

    A LOT of the radiation info in the mass media here and abroad seems to be confused and uncertain, plus there are corporate and political interests trying to control the information going out to the press from Japan’s nuclear sites themselves.

    Is it fair to say that “experts” don’t have much to offer in general about radiation dangers, and that their estimations and advice are really dependent on the particular circumstances of each individual exposure?

  5. JamesL (History)

    I’ve seen no source or info on reactor seawater injection rate. That would equal the liquid loss rate or leakage, an important piece of info.

    • thermopile (History)

      Re: seawater injection rate. This scanned paper (dated March 23):

      has “300 l / min” on it in the upper left hand corner. Could that be an injection rate? Does anyone here read Japanese?

      Other rows in that table, simply guesses based on the numbers and the units:
      Row 2: Water height above/below fuel in core
      Row 3: Reactor vessel pressure?
      Row 4: Coolant temp?
      Row 5: Coolant temp?
      Row 6 (D/W S/C): Pressure, somewhere.
      Row 7 (CAMS): Dose rate, somewhere.
      Rows 8 and 9: Pressure, somewhere else.
      Row 10: Spent fuel pool water temperature.
      Rows 11 and 12: ???

  6. Eve (History)

    There is often too little with the way radiation is, or is not, catergorized and I find this troubling. For instance neutron kinetic energy and neutron activation – would you want to be hit by fast or slow neutrons coming from core materials? Fast neutrons are most likely to cause DNA deletions and chromosome abnormalities. Or what about alpha emitters, how are they harmful? There are some unfortunate examples there.

    Fortunately I would not like to smoke, and I don’t think it’s a fair comparison. Apart from nicotine, which is highly carcinogenic, maleic hydrazide is used to regulate nicotine levels before harvest. Scientists in fact knew about the mutagenic capabilities of this compound in 1967, yet tobacco companies and growers have continually used it, unsparingly – they drench plants 7 days before harvest. Maleic hydrazide is a very efficient generator of double stranded DNA breaks, and people still unknowingly smoke it! A nasty thing about tobacco.

    Perhaps they should break apart the radiation into the biological effects from specific radionucleides, types etc. etc., together with the levels observed in sieverts. Two charts are needed, even though I like the sliding table shown above.

    • Hairs (History)


      Radiation already IS broken down “…into the biological effects from specific […] types…”. This is done when the energy deposited in your body, which is measured in Grays (= J/kg), is converted into the equivalent harm it does, which is measured in Sieverts.

      I’ve left out your reference to specific radionucleides because there are really only five types of ionising radiation to worry about with regard to protection (alpha, beta, gamma, fission fragments and neutrons) and the only thing that matters is what type of radiation you receive and what energy it deposits in you.

      Thus an absorbed dose of 1 J/kg (i.e. 1 Gy) of gamma radiation gives you an equivalent dose of 1 x 1 = 1 Sv because gamma radiation has a factor 1 by definition. However, an absorbed does of 1 Gy of alpha radiation gives you an equivalent dose of 1 x 20 = 20 Sv because alpha radiation is assessed with a factor of 20. Beta radiation, which is just electrons also gets a factor of 1; fission fragments get a factor of 20 just like alpha; and neutrons are a pain in the proverbial because they get a different factor according to their energy (and that energy changes as they thermalise inside the body, although to a first approximation you can just use the factor for their initial energy).

      In addition to the factors for the radiation type there are also factors for individual parts of the body according to the tissues’ respective sensitivity to ionising radiation. Basically tissues which have a high degree of cell division e.g. the sex cells, bone marrow, epithelial cells in the gut, etc have a higher factor than those which don’t divide much e.g. osteocytes in the bone. By convention the sum of all the tissue factors adds up to 1.0, so that if the radiation is absorbed equally throughout the whole body then the equivalent dose just becomes the absorbed dose multiplied by the radiation-type factor. On the other hand, if there is good reason to believe that only a sub-set of tissues have been irradiated then a factor can be calculated for just those tissues affected. For example, if you put a neutron source close to your eye while in a darkened room so that you can see the scintillation flashes in your vitreous humour (the sort of thing postgrads might do when they’re young, drunk and stupid enough to think it’s a good idea to go to the lab and get the sources out once the pubs close at 11pm) then it’s possible to calculate the factor specific for the eye and brain damage so caused (such a calculation might be the punishment issued by a lab supervisor when said postgrads trigger an alarm by trying to get the Cf-252 out of a cupboard and put it next to the beryllium target to see if it will make “bigger” neutrons)…

      If you Google “equivalent dose you’ll probably find lots of references. And in case you need to do any calculations I can suggest that as a first approximation you assume each human eye is a sphere of water 25mm in diameter…

    • Eve (History)

      You’re right, but what I’m talking is about what happens with discusssion in the media. There is however an additional point I’m trying to make. That is the types of emissions versus their biological effect. Each accident is different in emission spectra and the type of incident radiation. At the same time, it is becoming quite clear from genomic studies that particular types of radiation are exceptionally good at creating mutations while others are not, especially when you consider “equivalent doses”. It is the emission spectra, the type of incident radiation and somewhat the radionuclides type which is least discussed.

  7. krepon (History)

    Almost everything is relative. Every six months I get dosed with a CT scan.

  8. bks (History)

    Is this a lot?

    “A level of 500 millisieverts per hour “days ago” at the No. 2 reactor turbine forced workers to suspend repairs and they have yet to restart, Hidehiko Nishiyama, a spokesman for the Japan Nuclear and Industrial Safety Agency, told reporters today. Tokyo Electric Power Co., operator of the Fukushima Dai- Ichi plant, said it couldn’t confirm the reading, which would mean a worker in the vicinity would receive the maximum recommended lifetime dose in just 30 minutes.”


    • Eve (History)

      yes it is…2 hrs in that and you’ll definitely be sick as per the chart above. You have to think that this is not a static level though and it could go up and down.

  9. bks (History)

    How about this:

    “Edano said at a news conference on Wednesday that a computer forecast system has shown that radiation levels in some areas outside the 30-kilometer zone would exceed 100 millisieverts, which is the level that could affect the human thyroid if a person is exposed to it outdoors for 24 hours.”


    • Eve (History)

      They might be talking in microSieverts? 100 milliSieverts at 30 km!!

  10. MS (History)

    Interesting: NASA dose limits for astronauts are much higher.

    • Eve (History)

      No, the levels they talk about are per mission and fall far below what has been observed at Daiichi (FYI, 1 Sievert = 100 rem)

  11. bks (History)

    This is what a report on radiation levels should look like:


    • Eve (History)

      I don’t know about you, but when I’m writing manuscripts and I’m cold I sit on my hands to keep them warm.

      After looking at the pictures of inside the control and equipment rooms, I don’t think they have the time with putting the error bars in those graphs, let alone get the statistics darn right. It looks like they are cold and a little under stress. When do they get to pee, eat, drink or catch their breath? I think I too would be working with maximum levels, particular since there are other concerns like wiring the place up right in four reactors!

  12. rwendland (History)

    IAEA report that for two of the three workers who stepped in the puddle “the level of local exposure to the workers legs was estimated to be between 2 and 6 sieverts”. (Update 27 March, 03:00 UTC) By the chart above that’s between “severe radiation poisoning” and the mid-point of “survival possible with prompt treatment” and “fatal even with treatment”.

    Nevertheless the IAEA report “While the patients did not require medical treatment, doctors decided to keep them in hospital and monitor their progress over coming days.”

    Obviously(?) legs are more resilient than other parts of the body.(?)