Journal of Ornithology

, Volume 156, Supplement 1, pp 297–305

Cumulative effects of radioactivity from Fukushima on the abundance and biodiversity of birds


    • Laboratoire d’Ecologie, Systématique et Evolution, CNRS UMR 8079Université Paris-Sud
  • I. Nishiumi
    • Department of ZoologyNational Museum of Nature and Science
  • T. A. Mousseau
    • Department of Biological SciencesUniversity of South Carolina
    • Department of Environmental Biology, Biosciences, and BiotechnologyChubu University

DOI: 10.1007/s10336-015-1197-2


Species differ in their susceptibility to radiation because of differences in their ability to sustain toxic and genetic effects caused by radiation. We censused breeding birds in Fukushima Prefecture, Japan, during 2011-2014 to test whether the abundance and diversity of birds became increasingly negatively affected by radiation over time. The abundance of birds decreased with increasing levels of background radiation, with significant interspecific variation. Even though levels of background radiation decreased over time, the relationship between abundance and radiation became more negative over time. The relationship between abundance and radiation became less negative with increasing trophic levels. These findings are consistent with the hypothesis that the negative effects of radiation on abundance and species richness accumulate over time.


Birds Chernobyl Fukushima Radiation resistance


Radiation has been demonstrated to have diverse negative effects on animals and plants under natural conditions (Møller and Mousseau 2011b). There are highly radiation-resistant bacteria and fungi that are able to cope with extreme natural levels of radiation (Dadachova et al. 2004; Daly 2009), but such organisms seem to be exceptions rather than the rule. While natural levels of radiation vary worldwide with significant effects on the incidence of disease in humans (e.g., Lubin and Boice 1997; Hendry et al. 2009), and by implication also on other organisms (Møller and Mousseau 2011b), there are significant radiation effects involving DNA damage and mutations (e.g., Ghiassi-Nejad et al. 2004), following the original observation that radiation is a powerful mutagen (Nadson and Philippov 1925). Recently, interest in the effects of low-dose radiation on public health but also on free-living organisms has been increasing following the nuclear accidents at Chernobyl and Fukushima.

Acute exposure to radiation has negative effects on a number of physiological processes such as oxidative stress and immune function, while chronic exposure across extended periods of time can result in severe accumulations of effects of mutations during the lifespan of individuals, but also across generations. Current studies, generally conducted under laboratory conditions, typically rely on acute exposure, while the chronic effects of extended exposure are rarely considered. We previously showed that the abundance and diversity of birds and other organisms at Chernobyl were more strongly negatively impacted by a given level of radiation than those affected by similar levels of exposure at Fukushima (Møller et al. 2012). Thus, the difference in the actual levels of radiation in Chernobyl and Fukushima is not the cause since the effects were quantified as effects per unit of background radiation. Some studies suggested that the negative impact of radiation may become ameliorated over time because of adaptation in terms of improved DNA repair (Boubriak et al. 2008). The stronger negative effects at Chernobyl than at Fukushima that we previously documented (Møller et al. 2012) may be associated with a longer history of exposure, although alternative explanations such as differences in radionuclides and their toxicity may also play a role.

Here we report the results of analyses of unique data on the abundance of breeding birds in Fukushima, Japan, in relation to the background level of radiation during the 2011–2014 period. This period covers the radiation effects just after the Fukushima accident, but also the subsequent chronic effects accumulated during the following 3 years. This is the first study quantifying such cumulative temporal effects of radiation over time.

The objectives of this study were to test (1) whether the effect of radiation reduced the abundance and species richness of birds, (2) whether this effect differed among species, and (3) whether such negative effects of radiation accumulated over time. Ionizing radiation at Chernobyl and Fukushima had negative effects on the abundance and species richness of birds and other organisms (Møller and Mousseau 2007a, b; Møller et al. 2011a, b, 2012). Such effects vary considerably among species because of differences in the physiology and ecology (Møller and Mousseau 2007b; Galván et al. 2011, 2014). Unfortunately, there are no population studies or studies of a broad range of species dating back to the period prior to the accidents at Chernobyl in 1986 or Fukushima in 2011. We conducted the first standardized counts of breeding birds in 2006 in Chernobyl, 20 years after the accident. We started bird counts in Fukushima already in 2011, allowing for tests of effects of ionizing radiation directly from the start of radiation exposure. Here we present analyses of these unique data on the effects of radiation since 2011 at Fukushima in an attempt to determine whether such negative effects of radiation on the abundance and diversity of birds accumulate over time.


Study sites

We conducted breeding bird censuses at a total of 400 sampling points (in 2011 only 300 points) in forested areas around the Fukushima Daiichi power plants in 2011–2014, totaling 1,500 sampling events (Fig. 1). At least one local ornithologist (Satoe Kasahara, Shin Matsui, Isao Nishiumi or Keisuke Ueda) participated in approximately a quarter of the censuses in Japan to confirm the identity of some difficult bird species. However, all analyses presented here were based on the data recorded by APM. All sampling sites were localized using GPS coordinates, and altitude was estimated to the nearest foot using a GPS.
Fig. 1

Location of census areas around Fukushima, Japan, indicated by lines of dark blue dots in relation to the background radiation level. Circles show distances of 5 and 50 km from the reactors. Radiation level increases from the lowest level for light blue to the highest level for the darkest shade of red. Adapted from, generated by the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT) and local government (color figure online)

Census methods

The point count census method provides reliable information on the relative abundance of birds (Blondel et al. 1970; Møller 1983; Bibby et al. 2005; Voříšek et al. 2010). This method has provided highly repeatable results for birds and other animals at Chernobyl (Møller and Mousseau 2011a). It consists of counts lasting 5 min during which the number of birds seen or heard is recorded. Each census point was separated from the previous point by a minimum distance of 100 m. In Fukushima APM with the help of IN conducted these standard point counts on 11–15 July 2011, 14–19 July 2012, 14–19 July 2013 and 11–16 July 2014. The fact that one person made all counts analyzed here eliminates any variance in results due to interobserver variability.

We directly tested the reliability of our counts by letting two persons independently perform counts, and the degree of consistency was high for both species richness and abundance (details reported by Møller and Mousseau 2007a). The Pearson product-moment correlation between species richness in two series of counts conducted by two different persons was r = 0.99, t = 42.06, df = 8, P < 0.0001, and for abundance it was equally high at r = 0.99, t = 12.47, df = 8, P < 0.0001.

Abundance estimates can be affected by numerous confounding variables (Voříšek et al. 2010); therefore, it is important to control such variables statistically to assess the underlying relationship between radiation and species richness and abundance. We classified habitats in the field directly at the census points immediately following the 5-min count [agricultural habitats with grassland or shrubs (either currently or previously cultivated), deciduous forest or coniferous forest] and estimated to the nearest 10 % ground coverage by herbs, shrubs, trees, agricultural habitat, deciduous forest and coniferous forest within a distance of 50 m from the census points. Weather conditions can affect animal activity and hence census results (Voříšek et al. 2010), and we recorded cloud cover at the start of each point count (to the nearest eighth), temperature (degrees Celsius) and wind force (Beaufort). For each census point, we recorded the time of day when the count was started (to the nearest minute). Our bird counts were concentrated in the morning with counts extending across the entire day depending on other research activities. Because bird activity may show a curvilinear relationship with time of day, for example, with high levels of activity in the morning and to a lesser extent in the evening for birds (Voříšek et al. 2010), we also included time squared as an explanatory variable in the statistical analyses.

Background radiation

Radiation measurements at Fukushima were obtained using the same dosimeters (model: Inspector, SE International, Inc., Summertown, TN, USA) cross-validated with readings from a dosimeter that had been calibrated and certified to be accurate by the factory during the weeks preceding the study (International Medcom, Sebastopol, CA, USA). All radiation measurements were made at the census points immediately after each bird count. We also made a cross-validation test by comparing our own measurements using the Inspector dosimeter with measurements obtained at the same locations with a TCS 171-ALOKA used by Japanese authorities. There was a very strong positive relationship [linear regression on log-log transformed data: F = 2427.97, df = 1, 20, r 2 = 0.99, P < 0.0001, slope (SE) = 1.120 (0.023)]. All data are reported in Electronic Supplementary Material Table S1.


We scored the species as herbivores if they mainly fed on foliage or seeds, primary consumers if they mainly fed on insects, spiders and other invertebrates, and as top consumers if they mainly fed on vertebrates, relying on information presented in del Hoyo et al. (1992–2011).

Statistical analyses

Radiation levels were log10 transformed, and coverage with agricultural land, herbs, shrubs and trees, deciduous forest, coniferous forest and cloud cover was square-root arcsine transformed before analyses.

We quantified the relationship between the abundance of different bird species and level of radiation by estimating the slope of the relationship between abundance and log10-transformed radiation while including potentially confounding variables in the statistical models (coverage by herbs, shrubs, trees, agricultural habitat, deciduous forest and coniferous forest, altitude, cloud cover, temperature, wind force, time of day and time of day squared). The resulting species-specific partial slopes for bird abundance with radiation were used for subsequent analyses. This approach is extremely conservative because it reduces the total counts of the entire study to a single estimate per species or in some analyses a single estimate per species and year. We quantified species richness as the total number of species recorded at a given observation point under the assumption that such estimates would be distributed randomly across radiation levels if there were no negative effects of radiation on species richness.

A common underlying assumption of most statistical analyses is that each data point provides equally precise information about the deterministic part of total process variation, i.e., the standard deviation of the error term is constant over all values of the predictor variable(s) (Sokal and Rohlf 1995). Because estimates of slopes depend on sample sizes, and because sample sizes vary considerably among species, this can have serious consequences for conclusions (Garamszegi and Møller 2010, 2011). The standard solution to violations of this assumption is to weigh each observation by sampling effort in order to use all data, by giving each datum a weight that reflects its degree of precision due to sampling effort (Draper and Smith 1981; Neter et al. 1996; Garamszegi and Møller 2010). Therefore, we weighted statistical models by sample size in order to use all data in relation to the precision of the estimates. Even a single observation of a species was included in the analyses because such an observation could theoretically be recorded at any of the observation points with a single point having an observation of 1 and all other points an observation of 0. The null hypothesis is that the slope in this case will be zero, while non-random locations of observations will be concentrated at low radiation levels if radiation has a negative effect on the presence and the abundance of birds.

We tested whether the abundance of species was related to the level of radiation, species and the interaction between radiation and species. Next we tested whether there was a temporal trend in slope of the relationship between abundance and background radiation under the prediction that the extent of negative effects would accumulate across years. Finally, we tested whether the slope of the relationship between abundance and radiation differed among categories of main diet: herbivory or carnivory, or herbivores, primary consumers or top consumers. All standard least squares analyses or in case of data that were not normally distributed non-parametric tests were made with JMP (SAS 2012).


Tests for interspecific differences in the effect of radiation on abundance

The number of individual birds of the 57 different species recorded at Fukushima ranged from 1 to 1,715, mean (SE) = 166.7 (48.6), median = 22 individuals.

There was a significant decline in the level of background radiation across years (Fig. 2; F = 1736.70, df = 3, 1097, P < 0.0001) with additional variation among census points (F = 43.53, df = 399, 1097, P < 0.0001).
Fig. 2

Box plots of background radiation (μSv/h) at census points during the years 2011–2014. The box plots show median, quartiles, 5th and 95th percentiles and extreme observations

The abundance of birds at Fukushima differed significantly among species, decreasing with increasing levels of background radiation, and this decrease differed among species as reflected by the significant species by radiation interaction (Table 1). There was a strong negative relationship between species richness and radiation level across the census points (Fig. 3A; F = 6.73, df = 1, 1495, P < 0.0001), and there was an equally strong negative relationship between the total number of individuals and radiation level across the census points (Fig. 3b; F = 18.11, df = 1, 1495, P < 0.0001).
Table 1

Abundance of different species of birds in relation to radiation level in Fukushima


Sum of squares




Estimate (SE)

Radiation (R)





−0.0063 (0.0018)

Species (S)






R × S










The overall model had the statistics F 89,13410 = 55.47, r 2 = 0.27, P < 0.0001
Fig. 3

a Number of bird species and b number of bird individuals at census points in relation to background radiation (μSv/h). The lines are the linear regression lines

The slope of the relationship between abundance and background radiation for 57 different species of birds ranged from −0.525 to +0.107, mean (SE) = −0.162 (0.023) (Table 2). The mean slope differed significantly from zero in an analysis weighted by sample size (one-sample t-test, t = −7.16, df = 56, P < 0.0001). This implies that species were on average less abundant at high levels of radiation.
Table 2

Slope of the relationship between abundance and level of radiation after controlling statistically for potentially confounding variables (see "Methods") based on observations from 1,500 census points, level of significance of this relationship and the number of individuals that were censused for different species of birds




No. individuals

Accipiter gentilis




Acrocephalus arundinaceus




Aegithalos caudatus




Alauda arvensis




Alcedo atthis




Anas poecilorhyncha




Apus affinis




Ardea cinerea




Bambusicola thoracica




Butastur indicus




Buteo buteo




Carduelis sinica




Cettia diphone




Cinclus pallasii




Cisticola juncidis




Corvus corone




Corvus macrorhynchos




Cuculus canorus




Cuculus poliocephalus




Cuculus saturatus




Cyanopica cyanus




Cyanoptila cyanomelana




Delichon urbica




Dendrocopos kizuki




Emberiza cioides




Eophona personata




Falco peregrinus




Falco tinnunculus




Ficedula narcissina




Garrulax canorus




Garrulus glandarius




Hirundo rustica




Hypsipetes amaurotis




Lanius bucephalus




Milvus migrans




Motacilla alba




Motacilla cinerea




Motacilla grandis




Parus ater




Parus major




Parus montanus




Parus varius




Passer montanus




Pericrocotus divaricatus




Phalacrocorax carbo




Phasianus colchicus




Phasianus soemmeringii




Picus awokera




Regulus regulus




Streptopelia orientalis




Sturnus cineraceus




Tarsiger cyanurus




Terpsiphone atrocaudata




Troglodytes troglodytes




Turdus cardis




Urosphena squameiceps




Zosterops japonica




Differences in slopes among years

Correlation coefficients between abundance and radiation differed among years in an analysis weighted by sample size (Fig. 4; F = 5.25, df = 3, 107, r 2 = 0.05, P = 0.0020) in a model that accounted for species (F = 73.17, df = 56, 107, P < 0.0001).
Fig. 4

Box plots of the correlation coefficients between abundance and background radiation during the years 2011–2014. The box plots show median, quartiles, 5th and 95th percentiles and extreme observations

We compared the slope of the relationship between abundance and radiation for the same species in different years under the assumption that the relationship for slopes of different species in subsequent years should be positive with a slope of one. Any deviation from this null expectation would imply that factors other than statistical dependence were involved. There was a significant change in slope across years with 57 species for which there were data for at least 3 years (Kendall τ = −0.192, SE = 0.074, P = 0.012). While the relationship between abundance and background radiation on average was weakly negative in 2011, it became on average more strongly negative in 2012 (Fig. 5a), and on average it became even more strongly negative in 2013 compared to 2012 (Fig. 5b). Please note that the relationship between abundance and radiation for the different species in 2011 was positively related to the relationship for different species in 2012, and the relationship between abundance and radiation for different species in 2012 was positively related to the relationship for different species in 2013. Thus, species tended to become more negatively affected by radiation over time. In contrast, there was no evidence that the relationship between abundance and background radiation changed with sample size in the species for which there were data for at least 3 years (Kendall τ = −0.060, SE = 0.098, P = 0.542).
Fig. 5

Relationship between correlation coefficients for abundance of different bird species and background radiation (μSv/h) at census points during (a) 2011 and 2012 and (b) 2012 and 2013

There was a significant difference in slope between species with a herbivorous and a carnivorous diet (F = 25.20, df = 1, 55, r 2 = 0.30, P < 0.0001). Surprisingly herbivorous species had a more strongly negative slope (−0.32 (SE = 0.04), N = 10) than carnivorous species (−0.10 (0.02), N = 47). There was also a significant positive association between the slope of the relationship between abundance and radiation and trophic level (F = 12.52, df = 2, 54, r 2 = 0.29, P < 0.0001). The mean slope was the steepest for primary consumers [−0.32 (SE = 0.04, N = 10) over low level predators (−0.10 (0.02), N = 38) to top predators (−0.04 (0.13), N = 9)].


The main results of this study of radiation and species richness and abundance of birds at Fukushima, Japan, during 2011–2014 were that (1) overall abundance and diversity of species on average decreased with increasing levels of background radiation, (2) the relationship differed among species, with most species decreasing, but some species increasing in abundance with increasing levels of radiation, and (3) the relationship became more strongly negative across years, while there was no effect of change in abundance.

The overall negative relationship between abundance and level of background radiation differed among species. This result parallels previous findings from Chernobyl (Møller and Mousseau 2007a). We conducted our censuses under the assumption that there would be no significant difference in abundance because radiation should be distributed randomly across habitats that were censused. Therefore, we should expect the effect of radiation to remain after controlling statistically for the confounding effects of habitat, weather and time of day. We suggest that the difference in the effect of radiation on abundance between Fukushima and Chernobyl could be ascribed to differences in duration of exposure to radiation with (1) mutations accumulating for longer time in Chernobyl and (2) selection due to radiation not having acted for equally long time in the two areas.

We had expected an effect of bioaccumulation of radionuclides in the food web because bioaccumulation is common (Voitovich and Afonin 2002; Yakushev et al. 1999) and animals at higher trophic levels generally have higher levels of radionuclide concentrations than animals at lower levels (e.g., Kryshev and Ryabov 1990; Kryshev et al. 1992; Smith et al. 2002). However, the evidence suggested an opposite effect, with more strongly negative effects of radiation on abundance at low trophic levels. Our findings are therefore possibly more consistent with the hypothesis that oxidative stress in contaminated areas and reduced antioxidant levels (Møller et al. 2005) are a consequence of chronic radiation exposure (Møller and Mousseau 2007b), with negative effects on reproduction and survival and ultimately population trends.

This study has implications for the assessment of biological effects of ionizing radiation on free-living organisms. We have shown that the negative impact of radiation on abundance and species richness changes over time, and this effect varies among species. If we had not conducted the first breeding bird census in 2011 immediately after the Fukushima accident, we would not have been able to document a temporal change in abundance as we have reported here. Although bird species on average declined in abundance with increasing background radiation, there were several species that clearly increased in abundance. The reason for such changes can be changes in land use or release from competition due to reductions in the abundance of other species. The patterns of change in abundance with radiation level were only weakly species-specific, and even closely related species such as barn swallow Hirundo rustica and house martin Delichon urbica, carrion crow Corvus corone and jungle crow Corvus macrorhynchos, and great tit Parus major and varied tit Parus varius varied significantly in the impact of radiation on abundance (Table 2). This makes it unlikely that competitive release is an important factor.

Although there has been great public interest concerning the ecological, genetic and potential health consequences of the Fukushima radiological disaster, basic research to date has been surprisingly limited with only a handful of studies published since the disaster. A recent study of bull sperm and testis from the Fukushima region found no evidence for significant histological changes in the testes or sperm morphology (Yamashiro et al. 2013), although this study was very preliminary with only two bulls from a relatively uncontaminated part of Fukushima represented for the analysis of sperm. A study of aphids revealed large effects of radiation on morphology, although aberrant forms were only reported for one location in an area of relatively low contamination (Akimoto 2014). Similarly, a recent study of Japanese macaques found evidence for radiation effects on various characteristics of their blood, but individuals used in this study were obtained from areas surrounding Fukushima City where contamination levels, though measurable, are low relative to other parts of Fukushima Prefecture (Ochiai et al 2014). Ishida (2013) reported on surveys of some bird populations living in more heavily contaminated areas of the region and suggested that there was no evidence of significant declines resulting from the disaster. However, the number of sites surveyed was relatively few (56 in May and 38 in June 2012), and the analyses did not control for the many other potentially confounding factors that influence bird abundance and distribution, making this study very preliminary. Recent seminal studies of butterflies exposed to radioactive contaminants associated with the Fukushima disaster found strong evidence for increased mutation rates, developmental abnormalities and population effects as a direct consequence of exposure to radionuclides (Hiyama et al. 2012, 2013). These studies by Hiyama et al. (2012, 2013) were greatly strengthened by laboratory experiments that used both internal and external radiation sources, and these unambiguously supported observations of the elevated mutation rates and phenotypic effects observed in the field (Møller and Mousseau 2013), although, as with other studies, the number of populations studied, and hence the level of replication of observations, was very limited. Murase et al. (2015) made an equally compelling case for radiation having a negative impact on reproductive performance in the decline of Japanese goshawks Accipiter gentilis fujijamae compared to the pre-accident years, the progressive decline over time being directly linked to the air dose rate.

In conclusion, we have shown substantial evidence based on rigorous and highly replicated observations across space and time that is consistent with the hypothesis that the species richness and abundance of different species of birds were suppressed at high levels of background radiation in Fukushima. The relationship between abundance and radiation differed significantly among species, with most species decreasing, but some species increasing in abundance with increasing levels of radiation. Importantly, the relationship between abundance and radiation became more strongly negative over the 4 years studied, while there was no change in the effect of radiation on abundance with change in abundance over years.

Acknowledgments We gratefully acknowledge the logistic support and help in Japan by Prof. A. Hagiwara, Prof. K. Ueda, Prof. K. Kawatsu, S. Matsui, S. Kasahara, T. Kanagawa, K. Kawai, H. Suzuki and K. Koyama. We are especially grateful to the people of Fukushima Prefecture who permitted us to conduct this study. We thank Dr. S. Welch for help in preparing the contamination map of Fukushima Prefecture. We gratefully acknowledge support from the US National Science Foundation, the University of South Carolina School of the Environment, the NATO CLG program, the CRDF, the Fulbright Program, the National Geographic Society and the Samuel Freeman Charitable Trust for Research in Chernobyl, and the University of South Carolina, the Samuel Freeman Charitable Trust, Qiagen GmbH, the Chubu University Science and Technology Center and anonymous gifts from individual citizens for research in Japan.

Supplementary material

10336_2015_1197_MOESM1_ESM.doc (224 kb)
Supplementary material 1 (DOC 369 kb)

Copyright information

© Dt. Ornithologen-Gesellschaft e.V. 2015