Radioactive Mutant Butterflies – Really?

[This post is by Susan Voss. It’s from August 15, 2012. I’ve transferred it over here because I’ve been told it’s not showing up at our other site. Hopefully we will have things straightened out one of these days..]

(updated again – 9:35 am; US MST; added a comparison chart of second generation abnormalities in adults; added link to Scientific Reports; correction last table. Updated Fukushima results Table 2 with correct total for adult butterflies.)

Researchers in Japan published a new study, “The biological impacts of the Fukushima nuclear accident on the pale grass blue butterly”  by Hiyama, A. et al, published by the Nature Publishing Group, a division of Macmillan Publishers Ltd (Scientific Reports – “Hosted on — the home of over 80 journals published by Nature Publishing Group and the destination for millions of scientists globally every month”). The study focuses on the impact of radiation from the Fukushima nuclear power plant accident to the pale grass blue butterfly. In reviewing their data it appears that their results do not fully support their claims.

Figure 1 (Nuclear Diner): From the report – the graph on the left shows % abnormalities vs distance from the reactor for the so-called F1 off-spring from the first generation butterflies taken from May 2011 samples and the figure on the right – is my plot using their data showing abnormalities for F1 first generation offspring as a function of radiation dose levels. The implication from their May 2011 samples for F1 first generation offspring is that the number of abnormalities do not increase with higher radiation dose rate. There does not appear to be any clear correlation between first generation off-spring and radiation dose rates.

Figure 2: Taken from the Japanese report showing % abnormalities vs dose rate for butterflies collected in September 2011. To the right is my plot showing their results for % abnormalities for butterflies collected in May 2011. While the graph from Sept 2011 shows a clear correlation between increasing dose rate and abnormalities, the data from May 2011 does not. Unfortunately, the Japanese authors neglected to include the data from May in the report.

As a lover of butterflies and a believer in nuclear power as an important source of very low carbon output, I was concerned when I read about a new study in Japan linking radiation from the Fukushima nuclear power plant accident and abnormalities in butterflies. I had to take another look –the article seems to imply that the rate of abnormalities is linked to radiation levels but their data clearly shows this is not the case.

As an engineer, I understand data and statistics, therefore I opted to move into the supplemental data provided at the end of the report. The study is based upon a collection of butterflies at several sites around Fukushima and in cities further away. In some cities, they collected butterflies from a single site, while at other cities butterflies were collected from multiple sites. The results are then averaged for each city. For each city, key information is provided including distance from the Fukushima nuclear power plant, ground radiation dose, and the number of male and female butterflies collected for the study. Butterflies were collected from multiple sites on May 13-18, 2011 and September 18-21, 2011, and then from Kobe on October 3-4, 2011.

In reviewing the data I found a number of issues:

  1. Natural background – every location has some level of natural background radiation. The report fails to provide a baseline for any of the locations.
  2. Sample size – looking at the results from the May collection in Table 2 it is surprising how high is the abnormality rate for adult butterflies. The report states that at Fukushima 20% of the wings from female butterflies have abnormalities. That sounds significant – yet what they neglect to say is that the sample size is only 5 female butterflies! In other words – a statistically insignificant sample! The radiation levels at Fukushima were 1/3 to 1/2 of those from two other sites in the study where the abnormalities in male and female wings were zero. This is poor data analysis and representation thereof, since the results were smeared into the overall abnormality results without providing qualifying statements. That is, the radiation level and damage to the wings did not correlate based on a dose rate. In other words, the report denotes “one” butterfly wing abnormality collected at Fukushima, yet there were no abnormalities for those butterflies collected at cities and locations where the radiation levels were significantly higher.
  3. Motomiya, located 58.8 km from Fukushima, had roughly 3 times the radiation dose rate as compared to the Fukushima samples (3.09 and 2.73 μSv/hr vs 1.13 and 1.25 μSv/hr), but had no abnormalities out of the 11 butterflies collected in May – 2 female and 9 males. Yet as the report notes the “mortality rates of larvae, prepupae, and pupae and the abnormality rate of adults were high for Iwaki, Hirono, Motomiya, and Fukushima”. For Motomiya – that implies the offspring from two butterflies – neither of which had abnormalities. Overall the sample sizes for female butterflies are very low yet the authors make far reaching conclusions about their off-spring:
  4. Fukushima – 5 females in May; 8 females in Sept (radiation levels around 1.2 May and 0.7 site 1 and 2.43 μSv/hr site 2)
  5. Iwaki – 6 females in May; 22 females in Sept (radiation levels around 0.4 μSv/hr)
  6. Hirono – 5 females in May; 6 females in May (radiation levels 1.3 μSv/hr in May and .85 μSv/hr in Sept)
  7. Motomiya – 2 females in May; 9 females in September (radiation levels of around 2.9 μSv/hr in May and 1.6 in Sept)
  8. Mixing apples and oranges- The charts look good, but only the ones that correlate to apparently predetermined conclusions are shown. For example:

Figure 1 – that includes my chart on right – shows that for F1 (offspring from the first females) offspring that there is no correlation between radiation dose rates and number of abnormalities. Therefore as the data did not correlate to radiation levels, the author’s chose instead to show the abnormalities as a function of the distance from the nuclear power plant as shown in the graph to the left.

The F1 abnormalities data for the May collection of butterflies have dose rates of:

  • 0.16 μSv/hr with abnormality rates of 39%;
  • dose rates of 0.53 μSv/hr with abnormalities of 65%, and
  • Fukushima with dose rates of 1.19% and abnormalities of 44%.

The data for the September collection is similar. There appears to be no correlation between increased dose rate and higher abnormalities in the F1 samples.The data appears to be random.

From the paper: “Somewhat unexpectedly, the half-eclosion time and half-pupation time were correlated with the distance from the NPP but not with the ground radiation dose (data not shown). The reason for this outcome is not apparent, but it may be that distance values are more resistant to measurement deviations. Dose measurements in a given collection site could sometimes vary more than one order of magnitude within a several-meter range due to uneven distribution of radionuclides in the field.”

In explaining the fact that their data does not correlate to radiation levels this would actually negate their other results saying there is a correlation. The authors note that the results vary as distance from the reactor rather than by dose rate but do not consider other possibilities related to the tsunami that happened concurrently. Could there have been damage to the butterfly population related to the storm or chemicals released in the storm that have nothing to do with radiation?

In reviewing the results for the F1 generation and observed abnormalities, it is important to note how small is the sample size. Beginning with Shiroishi – they have one female butterfly. One.

May F1 first generation abnormalities

# of female butterflies Ground radiation dose μSv/hr Ground radiation dose cal’d averaged μSv/hr Total abnormality rate (%)
Shiroishi 1 0.32 0.32 24
Fukushima 5 1.13, 1.25 1.19 44
Motomiya 2 3.0, 2.7 2.85 49
Hirono 5 1.3 1.3 57
Iwaki 6 0.46, 0.6 0.53 65
Takahagi 5 0.3, 0.4 0.35 35
Mito 2 0.18, 0.14 0.16 37
Tsukuba 5 0.16, 0.15, 0.17, 0.16 0.16 39
  1. Figure 2 – to the left is actually Figure 4b from the report, which shows a nice correlation between ground radiation level and abnormality rate – but they don’t include the equivalent curve for the May data. Why? Possibly because it shows that the abnormalities are just as high, if not higher, for low radiation levels as for high radiation levels. I plotted their data and show the two curves side-by-side. While the September collection data shows a nice curve, the May collection data does not.
  2. Also – the authors change one of the two collection sites in Fukushima in September to one with much higher radiation levels. The dose rates are averaged for the two sites. Most of the butterflies are from Site #1, where the radiation levels are 0.71 Sv/hr versus Site #2, which is 2.43 Sv/hr – yet they AVERAGE the two together to get their nice curve. If the authors were consistent, they would show the number of abnormalities at each site because the radiation levels are significantly different. But they include no qualifications or comments.
    1. The authors claim that the radiation is impacting the second generation of butterflies, I put together a comparison of the ratio abnormal adults to total adults for several different cities with the radiation levels at the site. As you can see, the percentage of abnormalities for second generation were almost the same at Tsukuba as Fukushima yet the radiation levels at Tsukuba were close to background and varied slightly between May and Sept. 2011.

    The background case, Kobe, had 19%** of their second generation adult population with abnormalities even they had what appears to be no radiation from the reactor.

    Per Table 3 supplement: Comparison of 2nd generation F1; number of abnormal adults to total adults Abnormal/Total May 2011 Collection % May 2011 Collection Radiation Levels May 2011 microSv/hr Abnormal/Total Sept 2011 Collection % Sept 2011 Collection Radiation Levels Sept 2011 microSv/hr
    Fukushima 46/239 19% 1.13, 1.25 149/255 58% 0.71, 2.43*
    Tsukuba 28/142 18% 0.15, 0.16, 0.17 0.18
    Takahagi 18/166 11% 0.3, 0.42 131/240 55% 0.24
    Iwaki 57/233 25% 0.46, 0.63 69/109 63% 0.40, 0.42
    Kobe (596 km from the nuclear power plant; background radiation Oct 2011) 43/223** 19%** 0.08
    *different site for Sept collection

    **error in original calculation corrected 8/16/12.

    In summary –

    • the increase in abnormalities for the butterflies that were captured in May 2011 do not correlate to increased radiation.
    • there is no relationship between increased abnormalities between F1 off-spring and increased radiation levels.
    • The number of abnormalities decreases with distance from the reactor, also the site of a tsunami, and yet no other possibilities were considered.

    Cheers Susan

One comment

  1. Peter Walstrom · September 24, 2018

    I am also retired from LANL (in my case in 2015) and since then maintained an office as a Guest Scientist. Recently I have been working part time in AOT Division. I am a physicist, not a radiation biologist, but have tried to educate myself about the subject, especially in relation to nuclear energy. I have been following the Fukushima butterfly story and Otaki et al’s work since it came out in about 2012 and am appalled by how little criticism it has received from the scientific press. I think his methodology is highly questionable and his conclusions baseless.
    I periodically look for the rare articles criticizing Otaki et al. and found your blog today. My understanding is that Otaki et al’s work is inconsistent with years of work on radiation effects on insects, and radiation levels except perhaps right by the reactors (and probably not even there) were not nearly high enough (by many orders of magnitude) to cause statistically significant rates of mutations. This also applies to ingestion of radionuclides, which Otaki et al seem to believe is the main cause of the deformities and so-called mutations.
    In fact, Otaki et al (and many others who quote his work) seem to be confusing congenital deformities with mutations.
    One other person who has criticized Otaki is Timothy Jorgensen of Georgetown University. I exchanged a few E mails with him about Otaki and about another in roughly the same camp, i.e., Mousseau of Chernobyl fame.
    One thing I mentioned to Jorgensen is that is that it seems to be hard to get funding when one’s research does not find large radiation effects, and much easier when one claims to see dramatic bad effects.


Comments are closed.