The impact of the Fukushima nuclear accident on marine biota: Retrospective assessment of the first year and perspectives

Science of the Total Environment 487 (2014) 143–153

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

The impact of the Fukushima nuclear accident on marine biota: Retrospective assessment of the first year and perspectives Jordi Vives i Batlle a,⁎, Tatsuo Aono b, Justin E. Brown c,g, Ali Hosseini c,g, Jacqueline Garnier-Laplace d, Tatiana Sazykina e, Frits Steenhuisen f, Per Strand c,g a

Biosphere Impact Studies Unit, Belgian Nuclear Research Centre SCK•CEN, Boeretang 200, 2400 Mol, Belgium National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage-ku, Chiba 263-8555, Japan Norwegian Radiation Protection Authority, Grini næringspark 13, 1332 Østerås, Norway d Institute for Radioprotection and Nuclear Safety, Department for research and expertise in environmental risks, PRP-ENV/SERIS, Cadarache, Building 159, 13115 Saint-Paul-Lez-Durance Cedex, France e State Institution Research and Production Association Typhoon, 4 Pobedy Str., Obninsk, Kaluga Region 249038, Russian Federation f Arctic Centre, University of Groningen, Groningen, Netherlands g CERAD Centre of Excellence, Grini næringspark 13, 1332 Østerås, Norway b c

H I G H L I G H T S • • • • • •

UNSCEAR assessment of the Fukushima accident impact on the marine environment. The study covers the period from March 2011 to August 2012. Doses to marine organisms are generally below levels for effects on populations. The only exception is 131I in macroalgae near the plant early after the accident. Exposures to biota in the late phase are below the thresholds for population effects. Further away from the plant, potential effects on biota will be significantly lower.

a r t i c l e

i n f o

Article history: Received 23 January 2014 Received in revised form 31 March 2014 Accepted 31 March 2014 Available online 27 April 2014 Editor: Eddy Y. Zeng Keywords: Fukushima Non-human biota Radiological assessment UNSCEAR

a b s t r a c t An international study under the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) was performed to assess radiological impact of the nuclear accident at the Fukushima-Daiichi Nuclear Power Station (FDNPS) on the marine environment. This work constitutes the first international assessment of this type, drawing upon methodologies that incorporate the most up-to-date radioecological models and knowledge. To quantify the radiological impact on marine wildlife, a suite of state-of-the-art approaches to assess exposures to Fukushima derived radionuclides of marine biota, including predictive dynamic transfer modelling, was applied to a comprehensive dataset consisting of over 500 sediment, 6000 seawater and 5000 biota data points representative of the geographically relevant area during the first year after the accident. The dataset covers the period from May 2011 to August 2012. The method used to evaluate the ecological impact consists of comparing dose (rates) to which living species of interest are exposed during a defined period to critical effects values arising from the literature. The assessed doses follow a highly variable pattern and generally do not seem to indicate the potential for effects. A possible exception of a transient nature is the relatively contaminated area in the vicinity of the discharge point, where effects on sensitive endpoints in individual plants and animals might have occurred in the weeks directly following the accident. However, impacts on population integrity would have been unlikely due to the short duration and the limited space area of the initially high exposures. Our understanding of the biological impact of radiation on chronically exposed plants and animals continues to evolve, and still needs to be improved through future studies in the FDNPS marine environment. © 2014 Elsevier B.V. All rights reserved.

⁎ Corresponding author. Tel.: +32 14 33 88 05; fax: +32 14 32 10 56. E-mail address: [email protected] (J. Vives i Batlle).

http://dx.doi.org/10.1016/j.scitotenv.2014.03.137 0048-9697/© 2014 Elsevier B.V. All rights reserved.

144

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

1. Introduction In the wake of the catastrophic earthquake and tsunami on 11 March 2011, reactor failures at the FDNPS resulted in a significant input of radionuclides to the marine environment. This input occurred both as direct releases into the sea and as deposition of atmospheric releases, given that the direction of the prevailing winds at the time of the accident was towards the sea. This situation presented the urgent need to investigate the radiological impact of accidental releases from Fukushima to marine biota, focussing not only on the first months of the accident when the situation was highly dynamic, but also on the intervening 1-year period, in order to follow-up the impact of radionuclides still persisting in that local environment. Additionally, a longterm prospective assessment would help to manage the situation. A fuller description of the accident is given elsewhere (IAEA, 2011; IRSN, 2012; PMJHC, 2011; Povinec et al., 2013). The scope and magnitude of the marine radioactive releases were relatively well-known from shortly after the accident (Bailly du Bois et al., 2012; Buesseler et al., 2011; Garnier-Laplace et al., 2011; IRSN, 2011; Tsumune et al., 2012; Vives i Batlle, 2011). Several radionuclides such as 131,132I, 134,136,137 Cs and 129,129m,132Te were accidentally released into the marine environment, due to the combined effect of radioactive liquid effluent releases and the settling of the airborne radioactive particles. Some 1016 Bq of 137Cs found their way into the Pacific Ocean, 80% of this input occurring between 11 March and 8 April 2011. Radionuclide levels in the coastal zone were seen to decrease with distance by a factor of roughly 103 over the first 30-km from the source, and with time by factors of about 30 (137Cs) and 200 (131I) over the few weeks postaccident (Buesseler et al., 2011; Garnier-Laplace et al., 2011). By the end of May 2011, the short-lived radioisotopes had largely disappeared and 134,137Cs were the dominant radionuclides. Continued release of various effluents from land resulted in sustained contamination levels in the area during July 2011. A significant fraction of this contamination reached the seabed due to scavenging by suspended sediment or biogenic particles as well as sorption processes (Alekseev et al., 2006; Shiomoto et al., 1998; Vives i Batlle, 2011). An initial screening study suggested that maximum dose rates for 131I, 134Cs, and 137Cs to marine biota in the immediate aftermath of the accident could have ranged from 9 mGy h−1 for marine birds and 110 mGy h− 1 for benthic biota to 190 mGy h− 1 for macroalgae (Garnier-Laplace et al., 2011). This study assumed equilibrium of the biota with the highest seawater concentrations measured. Another early assessment carried out over a slightly longer period (March– May 2011) indicated that dose rates to fish and molluscs from the local coast did not exceed 420 μGy h− 1 and were generally in the order of 80 μGy h−1 (Kryshev and Sazykina, 2011). Monitoring studies indicated that, due to their high concentration capacity (especially for iodine), macroalgae were the marine organisms with the highest activity concentrations, followed by molluscs and fish (Greenpeace, 2012). The dose rates to marine biota calculated in the first screening study (Garnier-Laplace et al., 2011) are generally higher than the ERICA1 screening value of 10 μGy h−1 below which 95% of the species of an ecosystem are exposed to doses less than the ones giving 10% effects on their survival, reproduction or growth (Beresford et al., 2007; Brown et al., 2008) and the UNSCEAR2 level of 400 μGy h−1 which was defined as the “maximum dose rate to a small proportion of the individuals in aquatic populations of organisms that would not have any detrimental effect at the population level”. Hence, this study alerted on potential effects of ionising radiation at various levels of intensity but the exposure dose rates were assessed for a limited space area and a limited time period. Discrepancies were highlighted: for GarnierLaplace et al. since the dose rates reported only for the first 3-week 1 ERICA = “Environmental Risks from Ionising Contaminants: Assessment and management” an EC EURATOM Framework 6 funded project. 2 UNSCEAR = “United Nations Scientific Committee on the Effects of Atomic Radiation.”

period after the accident were based on equilibrium with maximum water concentrations for all radionuclides reported from water measurements and all irradiation pathways, they may have been overestimated (Buesseler et al., 2011; Vives i Batlle, 2011) — by at least one order of magnitude, according to other researchers (Kryshev and Sazykina, 2011; Kryshev et al., 2012). However, the latter studies, dealing with the first 2 months after the accident, did not take into account the external irradiation from sediment in the coastal zone. The uncertainties associated with dose estimates to non-human species in the early studies, along with the need to establish the significance of these doses in terms of effects to the marine biota, prompted the UNSCEAR to approach this issue shortly after the accident (Weiss, 2012). There are very few reports of radiation effects in sensitive endpoints of marine biota for both acute high doses and chronic low dose rates. By necessity, one must rely on a combination of data for marine and freshwater species and the reasonable assumption that there will not be a significant difference in the radiosensitivity of freshwater versus marine species. Different endpoints need to be considered including mortality, morbidity and reproductive effects. The UNSCEAR had previously analysed extensively the relevant data in its scientific annexes of the 1996 (UNSCEAR, 1996) and 2008 (UNSCEAR, 2008) reports, concluding that maximum dose rates of less than 400 μGy h−1 to any individual in aquatic populations of organisms would be unlikely to have any detrimental effects at the population level (UNSCEAR, 2011). This is based on the knowledge that there is little consistent and significant evidence for any effects on reproductive capacity at dose rates b200 μGy h− 1 (Copplestone et al., 2008; FREDERICA, 2006; Garnier-Laplace et al., 2008). Other benchmarks for contrasting purposes have been proposed. The ERICA and PROTECT 3 projects (Andersson et al., 2009; Beresford et al., 2007; Garnier-Laplace et al., 2008) suggested a generic dose rate of 10 μGy h− 1 for use in screening out environmental exposure situations of negligible concern. The ICRP4 also published derived consideration reference levels (DCRLs) that can be used to identify where there is likely to be some chance of deleterious effects of exposure to ionizing radiation on individual reference animals and plants (ICRP, 2008). The DCRLs published are broadly consistent with the benchmarks presented above, as UNSCEAR information shows (UNSCEAR, 2011). The present study is the first comprehensive assessment for the marine environment, as part of an overall assessment for the terrestrial and aquatic ecosystems of Fukushima (Strand et al., 2014; UNSCEAR, 2014). It is based on a comprehensive set of monitoring data representative of the first year after the accident (500 sediment, 6000 seawater and 5000 biota data points) and compiled by the UNSCEAR along with other relevant reports and published scientific papers. It is complemented by additional predictive dynamic modelling of radionuclide transfer to biota for the earliest phase of the accident. As added value, the study demonstrates the advantages of using such dynamic transfer modelling in preference to equilibrium-based transfer models in accidental situations, in accordance with what had been observed earlier in the marine environment of Fukushima (Kryshev et al., 2012; Vives i Batlle, 2011; Vives i Batlle and Vandenhove, 2014) and other areas (Vives i Batlle et al., 2007b).

2. Materials and methods To quantify the radiological impact on wildlife, a suite of recently developed approaches (Avila et al., 2004; Brown et al., 2008; ICRP, 2008; ICRP, 2009; Larsson, 2008; Sazykina, 2000; UNSCEAR, 2008; Vives i Batlle et al., 2008b) was applied to calculate exposure and thereafter effects were predicted through comparison with critical effects (or no-effect) values arising from compiled dose/response relationships. 3 PROTECT = “Protection of the Environment from Ionising Radiation in a Regulatory Context” EC EURATOM Framework 6 funded project. 4 ICRP = “International Commission on Radiological Protection.”

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

145

Fig. 1. Location of the seawater, sediment and biota monitoring stations used in the assessment.

2.1. Description of the dataset The primary source of data was the information provided to the UNSCEAR by the government of Japan, and subsequently qualityassured by the UNSCEAR. Some 4950 measurements of 131I, 134Cs and 137Cs were available for 210 different types of marine organisms (fish, crustaceans, molluscs, echinoderms, ascidians, holothurians, plankton and macroalgae) spanning the period 10 May 2011–12 August 2012. Additional activity concentrations in fish, algae and molluscs sampled in the periods 3–9 May and 23–24 June 2011 at various local ports (as well as offshore the Fukushima NPP) (Greenpeace, 2012)

were evaluated in a separate study (Vives i Batlle and Vandenhove, 2014), cited here for comparison purposes. The UNSCEAR also supplied measurements of activity concentration in unfiltered surface seawater (6340 data points between 22 March 2011 and 30 May 2012) and sediment (488 data points between 29 April 2011 and 28 March 2012), scattered over a grid spanning between 35°– 40° N and 140°–145° E with limited overlap between the seawater, sediment and biota sampling in space or in time (Fig. 1). For the period before 22 March 2011, with no available biota monitoring information, we assumed that seawater radioactivity concentrations were constant, so as to allow for a ‘warm-up’ period for the dynamic transfer model.

Table 1 Masses and dimensions for the marine reference organisms used in this assessment. Species code

Reference organisms

Mass (kg f.m.)

Length (m)

Width (m)

Height (m)

Phytoplankton Macroalgae Vascular plant Zooplankton Polychaete Benthic mollusc Crustacean Benthic fish Pelagic fish Small crustacean Octopus1 Squid2 Sea urchin3 Ascidian4 Holothurian5

FASSET phytoplankton ICRP brown seaweed FASSET vascular plant FASSET zooplankton FASSET benthic worm FASSET benthic mollusc ICRP crab ICRP flat fish FASSET pelagic fish Shrimps, prawns Octopus (benthic) Squid (pelagic/benthic) Sea urchin (benthic invertebrate) Sea pineapple Sea cucumber, sea sausage

6.54E−11 6.54E−03 2.62E−02 6.14E−05 1.73E−02 1.64E−02 7.54E−01 1.31E+00 5.65E−01 1.57E−05 1.00E+01 6.54E−01 1.44E−01 2.57E−01 2.92E−01

5.00E−05 5.00E−01 9.29E−02 6.20E−03 2.30E−01 5.00E−02 2.00E−01 3.99E−01 3.00E−01 1.00E−03 1.30E+00 5.00E−01 6.50E−02 1.00E−01 2.75E−01

5.00E−05 5.00E−03 2.32E−02 6.10E−03 1.20E−02 2.50E−02 1.20E−01 2.49E−01 6.00E−02 3.00E−03 1.21E−01 5.00E−02 6.50E−02 7.00E−02 4.50E−02

5.00E−05 5.00E−03 2.32E−02 3.10E−03 1.20E−02 2.50E−02 6.00E−02 2.51E−02 6.00E−02 1.00E−02 1.21E−01 5.00E−02 6.50E−02 7.00E−02 4.50E−02

1 2 3 4 5

http://www.sealifebase.org/summary/octopus-vulgaris.html. http://www.sealifebase.org/summary/todarodes-pacificus.html, assuming 1:10 width to length ratio and 50% pelagic, 50% benthic. Barnes (1982) pp. 961–81, assumed spherical shape (mid-size range, shell only). Barnes (1982) pp. 1028–42 and Kim (1980), assuming a width to length ratio of ~0.7. Barnes (1982) pp. 981–997.

146

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

2.2. Monitoring-based assessment approach The calculation of dose rates using measured radionuclide concentrations in the organisms is the first step towards a robust assessment, as it obviates the use of transfer parameters which is where most of the uncertainty in estimates of exposure concentrates (Avila et al., 2004). The selection of radionuclides was restricted to 131I, 134Cs and 137Cs, due to the fact that the UNSCEAR dataset for marine biota only contained data for these radionuclides. The dose assessment methodology was based on the ERICA integrated approach (Larsson, 2008), making use of the parameter set included in ERICA Tool (Brown et al., 2008), a software with supporting databases developed to calculate radiation doses to non-human biota, to assist with effect analysis and finally to conclude on the ecological risk. Four steps were included in a bespoke quality-assured computer algorithm, as follows: (a) Assigning to each of the 202 biota species an ERICA reference organism [13] chosen from the following: benthic and pelagic fish, mollusc, crustacean, macroalgae, phytoplankton, zooplankton, polychaete worm and vascular plant, or one of six newly-defined reference organisms: small crustacean, octopus, squid, sea urchin, ascidian and holothurian (Table 1). (b) Calculation of internal dose rates for 131I, 134Cs and 137Cs using dose per unit concentration values extracted from the ERICA assessment Tool. (c) Selection of nearby water and sediment data points for each biota measurement by averaging seawater/sediment data within a set radius (10 km) and time interval (2 days for seawater, 10 days for sediment). (d) Calculation of external dose rates using the selected nearby sediment and water concentrations. For the calculation of the dose rate, three separate ‘template’ ERICA model runs were first made by using unit activity concentration in biota, seawater and sediment as input. The resulting “dose per unit concentration” (DPUC) factors were used to multiply the activity concentrations in biota and the medium, giving internal and external dose rates. The DPUCs include factors for organism occupancy in

water and sediment defined as defaults in the ERICA Tool (Beresford et al., 2007; Brown et al., 2008) and are underpinned by default dose conversion coefficients (DCCs) with separate components for lowenergy (b10 keV) β-radiation, high-energy (≥10 keV) β-/γ-radiation and α-radiation. Radiation weighting factors of 3, 1 and 10 respectively were used to represent relative biological effectiveness for deterministic effects in wildlife, as previously suggested (Vives i Batlle et al., 2004). The DCCs assume homogeneous distributions of the radionuclides present both in the organism and the surrounding medium (seawater or sediment) (Ulanovsky and Prohl, 2006; Ulanovsky and Pröhl, 2008; Ulanovsky et al., 2008). The uncertainty introduced by this assumption is well documented (Gómez-Ros et al., 2008) and, in view of the assessment goals, it was considered to be of minor significance. 2.3. Dynamic modelling approach The majority of radiological assessments to non-human biota are carried out by assuming that the activity concentration in an organism of mass M (AO, in Bq kg−1 expressed on a fresh mass basis — f.m.) is proportional to the activity concentration in an adjacent volume V of water (AW, in Bq m−3) via a concentration factor (CF, in units of m3 kg−1 f.m.). This instantaneous equilibrium did not really occur in the case of the early phase of the Fukushima accident because the biota reacted with a time delay to large variations of activity concentration in seawater. The dynamics of the process is determined by a balance between the mean time of availability of the radionuclide in the water in the presence of efficient hydraulic dilution, and the biological half-life of clearance (TB1/2) following intake of the radionuclide. For a single component release, this can be represented by a simple model with two rate constants; kW for uptake and kO for elimination: dAO V dA M ¼ kW AW −ðkO þ λÞAO ; W ¼ −ðkW þ λÞAW þ kO AO M V dt dt where kO ¼ Tln 2 , kW ¼ ðkO þ λÞ M V CF and λ is the radionuclide decay conB1=2 stant and assuming no growth of the organism (no biological dilution) (Vives i Batlle, 2012).

Table 2 Biokinetic parameters used in the dynamic modelling simulations (mass is expressed on a fresh basis). Parameter

Fish

Crustacean

Macroalgae

Mollusc

Source

Mass (kg)

1.31E+00

7.54E−01

6.54E−03

1.64E−02

ERICAa

Modelling parameters for Cs-134 TB1/2 (d) CF (m3 kg−1) DPUCextsed (μGy h−1/Bq kg−1) DPUCextwater (μGy h−1/Bq kg−1) DPUCint (μGy h−1/Bq kg−1) Decay constant (d−1)

6.47E+01 8.60E−02 4.10E−04 4.10E−04 1.70E−04 9.19E−04

5.75E+01 4.10E−02 3.95E−04 3.95E−04 2.00E−04 9.19E−04

5.40E+01 1.20E−01 4.50E−04 4.50E−04 9.50E−05 9.19E−04

1.80E+01 6.60E−02 4.35E−04 4.35E−04 1.20E−04 9.19E−04

Vives Batlle et al. (2008); Vives i Batlle et al. (2007b) ERICAa ERICAa ERICAa ERICAa (ICRP, 1983)

Modelling parameters for Cs-137 TB1/2 (d) CF (m3 kg−1) DPUCextsed (μGy h−1/Bq kg−1) DPUCextwater (μGy h−1/Bq kg−1) DPUCint (μGy h−1/Bq kg−1) Decay constant (d−1)

6.47E+01 8.60E−02 1.50E−04 1.50E−04 1.70E−04 6.32E−05

5.75E+01 4.10E−02 1.45E−04 1.45E−04 1.80E−04 6.32E−05

5.40E+01 1.20E−01 1.70E−04 1.70E−04 1.30E−04 6.32E−05

1.80E+01 6.60E−02 1.60E−04 1.60E−04 1.50E−04 6.32E−05

Vives Batlle et al. (2008); Vives i Batlle et al. (2007b) ERICAa ERICAa ERICAa ERICAa (ICRP, 1983)

Modelling parameters for I-131 TB1/2 (d) CF (m3 kg−1) DPUCextsed (μGy h−1/Bq kg−1) DPUCextwater (μGy h−1/Bq kg−1) DPUCint (μGy h−1/Bq kg−1) Decay constant (d−1)

3.08E+01 3.60E−03 1.00E−04 1.00E−04 1.30E−04 8.64E−02

1.57E+00 3.60E−03 9.50E−05 9.50E−05 1.40E−04 8.64E−02

5.00E+00 4.10E+00 1.15E−04 1.15E−04 1.00E−04 8.64E−02

5.60E+01 1.40E−02 1.05E−04 1.05E−04 1.20E−04 8.64E−02

Vives Batlle et al. (2008); Vives i Batlle et al. (2007b)b ERICAa ERICAa ERICAa ERICAa (ICRP, 1983)

a b

ERICA software Tool version June 2011. Using the value for zooplankton to represent the crustacean.

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

In the present study we used the D-DAT assessment model (Vives i Batlle et al., 2008; Watts et al., 2008), calibrated with generic singlecomponent release data from a detailed review study (Vives i Batlle et al., 2007b, 2009) and dosimetry information from the ERICA Tool (Table 2). A number of models (including D-DAT) can simulate multicomponent release (i.e. radionuclide elimination with more than one biological half-life) (Olsen and Vives i Batlle, 2003; Swift, 1992; Vives i Batlle et al., 2006; Wilson et al., 2007; Wilson et al., 2005) but this approach was avoided due to large uncertainties in the shortterm TB1/2. The D-DAT model calculates internal and external dose rates, using from the latter inputs of activity concentration in seawater as supplied by the UNSCEAR, but external dose from sediment exposure is not included in the model. Dynamic transfer modelling was also performed using the “ECOMOD” approach, which describes the dynamic processes of radionuclide accumulation in aquatic biota as radioactive tracers of stable analogous elements involved in the growth and metabolism of organisms (Kryshev and Ryabov, 2000; Kryshev and Sazykina, 1986; Sazykina, 2000). This model does not use the T B1/2 approach, but assumes that the rate of radionuclide loss is proportional to metabolism and depends on fish mass. The differential equations used by ECOMOD to describe radionuclide uptake and turnover dynamics are fully described elsewhere (Sazykina, 2000). The availability of both ECOMOD and D-DAT for this study presented us with the opportunity to perform an intercomparison between two very different modelling approaches and thus assess the degree of confidence of our modelling results. 2.4. Interpretation of effects and risk characterisation For non-human species, the biological effects of interest are mainly those relevant to species populations and demography i.e. survival, growth and reproduction. They are deterministic, generally dose- or dose rate-triggered, according to a relationship (frequently sigmoidal) between dose (or dose rate) and the intensity of the effects. Moreover, acute (high-dose, short term) and chronic (low dose, long term) exposure situations lead to different biological consequences. Therefore, in the context of an accident, it is essential to distinguish at least two phases contrasted in terms of the levels and types of exposure of living organisms to understand the effects on flora and fauna properly: an acute exposure phase at a high dose lasting a few weeks immediately after the accident and a chronic exposure phase at a low dose lasting many years. The acute phase extends throughout the initial weeks (up to 2–3 months) following the accident. This phase is characterised by the presence of a large quantity of short lived radionuclides likely to generate high dose rates for living organisms, mainly by external irradiation with a significant proportion of the dose delivered by beta-emitting radionuclides. For aquatic organisms, the exposure pathway is direct exposure to water, which is the receiving medium of the release. During this phase, the so-called ‘acute’ effects are likely to be observed. ‘Acute’ effects include any notable biological modification occurring within a few days or weeks of absorption of a significant radiological dose, leading to irreversible damage and, eventually, death. The late phase or chronic phase, during which the contamination levels of the environment change much more slowly, takes place on a scale of several months and years. The effects are many in kind but with large uncertainties inherent in the extrapolations required to interpret the ecological significance of the effects observed at different levels of biological organization. 3. Results and discussion 3.1. Late phase using measured data from biological samples The activity concentrations of radiocaesium in marine biota for the period from May 2011 to June 2012, were too scattered to conclude a

147

definite trend. Activity concentrations range from a few Bq kg−1 f.m. to above 103 Bq kg−1 f.m., with a 134Cs/137Cs ratio of approximately 1. The activity concentrations of 131I tend to decrease faster than radiocaesium due to radioactive decay operating in addition to environmental dispersion and turnover by the biota. We can conclude that there is no obvious substantial decrease in marine biota concentrations during one year of monitoring, in accord with previous observations (Buesseler and Aoyama, 2012; Buesseler et al., 2011). On closer inspection, there is a slight hint of a rise to a maximum around mid-May 2011, possibly followed by a slight decrease in levels. In contrast, changes in seawater concentrations with time for the offshore coastal zone (up to 15–20 km from the FDNPS) are quite dramatic, falling from N 105 Bq m−3 to about 103 Bq m−3 by the end of 2011. Radionuclide concentrations in seawater (Bq L−1) are on average 2 orders of magnitude below levels in sediment (Bq kg−1), which have remained relatively constant. Additional biota activity concentration data (fish, algae, molluscs) sampled in 3–9 May and 23–24 June 2011 generally give dose rates similar to the UNSCEAR-supplied dataset for radiocaesium, and somewhat higher levels for radioiodine for the samples collected in early May 2011. This is accounted for by the intervening decay between the sampling of the two datasets and the fact that the macroalgae species collected in the two surveys are not generally the same (predominantly Sargassum horneri in the former dataset). Estimated weighted absorbed dose rates for selected clusters of measured data for the Fukushima marine coastal area are presented in Fig. 2 (UNSCEAR-supplied data only), showing high variability in time and space and directly reflecting the variations in actual radionuclide concentrations. Total dose rates were derived as the sum of all 131I, 134Cs and 137 Cs internal and external exposure contributions from the compiled arithmetic means for all organism categories. The highest values were approximately in the range 0.17–0.25 μGy h−1 (ascidians, macroalgae, sea urchins and holothurians) and 0.10 to 0.17 μGy h−1 (benthic fish, crustaceans and molluscs). They are dominated by exposures from 134Cs and 137 Cs, following the substantial radioactive decay of 131I after June 2011. An absolute maximum total dose rate of 4.4 μGy h−1 was calculated for benthic fish, namely fat greenling (Hexagrammos otakii) sampled in August 2012. We also calculated the mean internal and external dose rates for all locations and organisms combined on a monthly basis, and from this it is concluded that dose rates have remained relatively constant over the period March 2011–August 2012. The standard deviation for the dose rates is typically a factor of 2–3 times higher than the average values, masking any discernible time trend. The ratio between internal and external dose rate fluctuates highly, directly reflecting biological sample variability. Some 55% of the 131I data have an internal/external dose rate ratio N1, indicating that internal exposure dominates over external exposure. For radiocaesium the percentage is lower at 35%, indicating the preponderance of external dose. However, 65% of the radiocaesium data have a ratio greater than 0.2, indicating that internal and external doses are somewhat comparable in magnitude. The differences between 131I and 137Cs are due to several factors. Firstly, the higher sediment/water distribution coefficient (Kd) for Cs (4 m3 kg−1) compared with I (7 × 10− 2 m3 kg− 1) (Beresford et al., 2007) results in a higher accumulation of Cs in sediment (and hence, a larger external dose component). Secondly, the CF for I for macroalgae (the most represented organism in the monitoring data) is an order of magnitude higher than that for Cs (IAEA, 2004), resulting in greater internal doses for 131I. The additional biological samples collected between 3–9 May and 23–24 June 2011 at various local ports confirm that exposure to macroalgae is dominant, followed by molluscs and fish. The details are reported elsewhere (Vives i Batlle and Vandenhove, 2014). Internal dose rates are b 13 μGy h− 1 for 131I and b0.12 μGy h− 1 for 134,137Cs (highest for 134Cs in fish). On average, dose rates calculated from the UNSCEAR and the more limited, separate set of data collected between 3–9 May and 23–24 June 2011 are within a factor of 2 of each other with the exception of 131I in macroalgae, showing an order of

148

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

Fig. 2. Weighted absorbed dose rates calculated from monitoring data on a combined 131I, 134Cs and 137Cs basis for macroalgae, benthic fish and crustaceans.

magnitude difference attributable to the aforesaid decay and differences in the biological species sampled. The radiocaesium dose rates calculated for biological samples are much lower than those calculated in previous studies using an equilibrium approach, and are thought to be low relative to the relevant benchmarks. At b 0.25 μGy h− 1, they are generally below the 10 μGy h−1 ERICA benchmark (highest risk quotient with respect to the ERICA benchmark =0.4 for benthic fish), and commensurate with background dose rates in the marine environment (Hosseini et al., 2010). Even for 131 I, the dose rates are below the levels at which radiation effects could occur (Copplestone et al., 2008; FREDERICA, 2006), decreasing further with time. The calculated dose rates for marine animals are also below the lower bound of the ad hoc ICRP DCRLs or the UNSCEAR benchmark of 400 μGy h−1 (UNSCEAR, 1996) (as appropriate), suggesting that effects at the population level may have been imperceptible. For comparison purposes, limited data available on mortality effects of chronic irradiation in fish indicate that dose rates b4000 μGy h− 1 at any life stage are unlikely to affect survival. However, there is probably no threshold for some end-points such as gonadosomatic indices, the number of gametogenic cells in fish irradiated as embryos (Real et al., 2004). Very limited data suggest that chronic irradiation-induced genetic damage in fish probably occurs at all dose rates and that radiation sensitivity for this damage is similar to that of other vertebrates, even if the survival of populations is not affected. Chronic effects of ionising radiation in marine animal species are scarcely documented in radiobiology for species besides fish. The lowest value of the chronic dose rate giving 10% effect in reproductive endpoints (EDR10) is equal to 47 μGy h− 1 for the marine species Pleuronectes platessa (Garnier-Laplace et al., 2010). For marine invertebrates, the lowest value of EDR10 is found at 36 μGy h− 1 for

the annelid Ophryotrocha diadema (Knowles and Greenwood, 1994). For macroalgae, there is no chronic effect data in the literature, to our knowledge. As of August 2012, several fish specimens were still being found with 137Cs concentrations exceeding the Japanese regulatory limit of 100 Bq kg−1 f.m. for sale and human consumption (Buesseler, 2012). Some of the highest levels of radiocaesium, as high as 5 × 105 Bq kg−1 f.m., were reported in benthic fish from fishing baskets and in gill nets located at the port entrance of the FDNPS (late February–early March 2013). These reports could not be included in the UNSCEAR assessment, which focussed on the data for the first year after the accident, but would have led to higher estimates of the dose rate for this location, namely in the order of 17–44 μGy h− 1 for 134Cs and 32–82 μGy h− 1 for 137Cs. However, most dose rates are clearly below this maximum, since the data are log-normally distributed with 95% of the total radiocaesium activity below 50 μGy h− 1. These dose rates are below the 400 μGy h−1 UNSCEAR benchmark for the most exposed individuals of an aquatic population below which population detriment is not expected. Additionally, it is very unlikely that the elevated activity concentrations in individual specimens sporadically found close to the FDNPS signal prolonged exposures to whole fish populations. However, estimates are of the same order of magnitude as some EDR10 values in some species and endpoints, and they also exceed the ERICA screening dose rate of 10 μGy h−1. This indicates the need for further assessment as the situation continues to evolve with new emissions potentially convoluting marine biota exposures in the environment close to the FDNPS. The variations of concentration level induced by these hypothetical new emissions would depend on the fish species, their habitat and their position in the food web, given that the time duration for reaching peak concentrations in fish is longer for top-level species than for low-level species.

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

149

Fig. 3. Internal 131I and 137Cs dose rates for Fukushima Dai-ichi N channel and SW Dai-ni Iwasawa shore as predicted by dynamic modelling (black) and equilibrium transfer (grey).

3.2. Early phase: dynamic modelling predictions Dynamic modelling calculations using the D-DAT model are shown in Fig. 3. To better represent the difference between equilibrium and dynamic transfer modelling, only the internal dose rates are represented. A summary of the maximum total doses is given separately in Table 3. It was not possible to calculate activities in biota prior to the time of the first seawater measurements. The highest dose rates for the marine environment in the early phase, calculated with the D-DAT kinetic model, were for the station 30 m north of the discharge channel of FDNPS units 5 and 6. There, the maximum total dose rate for fish (sum of the internal and external dose and sum of both radiocaesium and 131I contributions), calculated to be 140 μGy h− 1, would have occurred within the first five weeks

after the accident. Accumulated absorbed doses to fish, crustaceans and molluscs at that location were estimated to be around 0.32, 0.18 and 0.25 Gy respectively over the entire 1-year modelling period. The corresponding values for an integration period of 90 days are 0.12, 0.08 and 0.15 Gy, respectively. In the early days post-release the dose rate in macroalgae was dominated by the presence of 131I. The maximum exposures occurred at 23 days, in early April 2011, and exceeded 20 mGy h−1 at their peak but fell rapidly to levels below 10 mGy h−1 by 32 days, largely reflecting a highly elevated uptake followed by rapid decay of short-lived radioiodine. The accumulated dose for macroalgae was estimated to be approximately 7.1 Gy over the entire modelling period, with most of it (6.8 Gy) concentrating over the first 90 days. Dynamically modelled radiocaesium and radioiodine dose rates in animal species outside the vicinity of the FDNPS North Channel were

Table 3 Maximum total dose rates received by fish (F), crustaceans (C), macroalgae (Ma) and Mollusca (Mo) as predicted by dynamic modelling. Location

Maximum total dose rate (μGy h−1) 131

134

I

Dai-ichi N Ch. Dai-ichi S Ch. SW Dai-ni Iwasawa Dai-ichi 15 km offshore Dai-ni 15 km offshore

137

Cs

Cs

F

C

Ma

Mo

F

C

Ma

Mo

F

C

Ma

Mo

27 34 1 0.09 0.09

46 57 1 0.11 0.11

20,109 24,298 633 70 67

66 77 2 0.27 0.25

67 44 3 0.48 0.56

50 34 2 0.32 0.38

65 44 2 0.44 0.52

88 57 3 0.58 0.71

49 31 2 0.43 0.48

31 20 1 0.25 0.28

60 38 3 0.71 0.86

86 52 4 0.68 0.79

150

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

generally b3 μGy h−1, whereas 131I doses for seaweed exceeded 100 μGy h−1 over the first 1 ½ months following the accident but decayed rapidly afterwards. A maximum of 633 μGy h−1 for macroalgae south-west of Fukushima Dai-ni (Iwasawa shore) is predicted to have occurred 28days post-accident but decreasing by 2 orders of magnitude a month afterwards. The dynamic modelling calculations suggest that the external dose from exposure to seawater is significantly lower than the internal dose, generally by 2–3 orders of magnitude (note that external dose from sediment could not be calculated with the available models). At present, effect benchmark values dedicated to acute exposure are scarce in the literature. What can be affirmed is that the 3-month integrated doses calculated here are far below (for animal species) or in the same order of magnitude (for macroalgae) the ERICA acute threshold value proposed at the ecosystem level for the marine environment (Garnier-Laplace and Gilbin, 2006), also reported by the UNSCEAR (UNSCEAR, 2008) (4.84 Gy). This is an indication that acute effects are unlikely in marine animals. Consistently, no drastic and immediate acute effect was reported from the field for the acute phase of the accident. However, transient exposures for macroalgae at locations very close to the discharge point around 20 mGy h− 1 are within the range of 10–100 mGy d−1 considered to cause potential effects on reproduction and growth rate, with 3-month integrated doses of 6.8 Gy signalling potential acute effects (ICRP, 2008), at least in theory. Therefore, the calculated exposures for macroalgae could have been at a level where acute impact could have occurred, but for the fact that these exposures were transient as most of the 131I quickly decayed. Further confirmation of effects on marine macroalgae in the vicinity of Fukushima is tempered by the present scarcity of available observations of such effects at the site. Moreover, the green algae Acetabularia mediterranea is the only marine algal species in the FREDERICA radiation dose effects database with data allowing the establishment of dose–effect relationships (FREDERICA, 2006). The minimum value of the ED50 (or effect dose giving a 50% change in observed effect) for this organism is equal to 160 Gy. Acute ED50s for marine crustaceans, molluscs or fish are all higher than the calculated 3-month doses. To provide an indication of the time evolution of dose rates for selected biota, results from the ECOMOD and equilibrium models are presented in Fig. 4. ECOMOD estimates peak dose rates for fish to be an order of magnitude below those derived using CFs. After ~50 days, the relationship reverses, a behaviour also observed in prior (Vives i Batlle and Vandenhove, 2014) and current D-DAT simulations for the FDNPS. In general, dose rates for pelagic fish, molluscs and macroalgae fell rapidly from their peaks, which occurred within the first weeks of the accident. The highest absorbed dose rate is calculated for

Fig. 4. Estimated total dose rates (sum of 131I, 134Cs and 137Cs) for fish, macroalgae and molluscs in the Fukushima Dai-Ichi South Channel using the ECOMOD and equilibrium CF models.

Fig. 5. Comparison between the activity concentrations (Bq kg−1) predicted by D-DAT and ECOMOD for the Fukushima Dai-ichi southern drainage channel, along with the equilibrium model prediction using a CF of 8.6 × 10−2 m3 kg−1.

macroalgae. The calculated dose rates for fish using dynamic modelling differ substantially from those presented by Garnier-Laplace et al. (2011), partly reflecting the fact that these authors assumed equilibrium between organisms and seawater and took account for all exposure pathways and a larger series of short-lived radioisotopes than just 131I.

3.3. Dynamic modelling validation The dose rates calculated via the two dynamic models D-DAT and ECOMOD show high consistency in results, imbuing some confidence to both models (Fig. 5). This is true both in terms of the predicted activity concentrations in biota and the time at which these dose rates reach a peak before the post-peak period where the models predict a quasiexponential rate of loss governed by the biological half-life of elimination. The ECOMOD model predictions show a slightly faster rate of loss than D-DAT, explained by the use of a fish biological half-life of 65 days in D-DAT compared with a metabolically-driven rate of radionuclide loss of 50 days in ECOMOD. In comparison with the equilibrium model prediction (using the ERICA default CF for benthic fish of 8.6 × 10− 2 m3 kg−1), the D-DAT model predicts activity concentrations a factor of ~ 100–300 lower at peak times for the period 10–30 days when the discharges were close to their maximum, inverting to a factor ~ 10–60 higher in the subsequent period. Consequently, the equilibrium model gives an overtly conservative prediction when discharges are rising and otherwise under-predicts till both models converge to one another at equilibrium, demonstrating in a practical way the need to use a dynamic model to predict transfer to biota in the acute phase of an accidental situation. It is not really possible to compare the modelled dose rates in biota with those calculated from the monitoring data for the early period where discharges were at their peak. This is because they originate from different locations and generally the monitored samples are further away from the source points (Dai-ichi and Dai-ni drainage channels). In addition, there is little overlap in the UNSCEAR data between the seawater and biota concentrations for the early period. Nevertheless, in a limited exercise, the modelled total dose rates outside the drainage channels but still close to the Dai-ni nuclear plant (SW Daini Iwasawa shore and Dai-ni N channel stations) were compared with dose rates calculated directly from the additional samples collected at the coastal stations less than 50 km from Fukushima Daiichi (Hisanohama, Yotsukura, Ena ports) for macroalgae (n = 5) and molluscs (n = 3) (Vives i Batlle and Vandenhove, 2014). The results are shown in Table 4. Despite the uncertainties associated with statistical dispersion of the monitoring data, it is clear that the dynamic model predicts dose rates much closer to those derived from monitoring data

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

151

Table 4 Comparison of dynamic and equilibrium modelling approaches (uncertainties are quoted as ± 1σ). Location

Mean total dose rate (μGy h−1)

Model prediction/ measurement

Macroalgae

Mollusc

Macroalgae

Mollusc

I-131 dose rates Mean Iwasawa/ Dai-ni (equilibrium) Mean Iwasawa/ Dai-ni (dynamic) Mean monitoring data

(5.1 ± 0.2)E+04 (9.24 ± 0.01)E+00 (6 ± 5)E+00

(1.76 ± 0.07)E+02 (2.25 ± 0.08)E−01 (1.1 ± 0.6)E−01

(8 ± 7)E+03 (1.4 ± 1.2)E+00 –

(1.6 ± 0.8)E+03 (2.0 ± 1.0)E+00 –

Cs-134 dose rates Mean Iwasawa/ Dai-ni (equilibrium) Mean Iwasawa/ Dai-ni (dynamic) Mean monitoring data

(3.0 ± 0.7)E+03 (1.7 ± 0.2)E+00 (1.3 ± 0.5)E−01

(1.7 ± 0.4)E+03 (1.6 ± 0.1)E+00 (1.1 ± 0.5)E−01

(2.4 ± 1.0)E+04 (1.3 ± 0.5)E+01 –

(1.5 ± 0.7)E+04 (1.4 ± 0.6)E+01 –

Cs-137 dose rates Mean Iwasawa/ Dai-ni (equilibrium) Mean Iwasawa/ Dai-ni (dynamic) Mean monitoring data

(3.8 ± 0.5)E+03 (2.2 ± 0.3)E+00 (1.3 ± 0.5)E−01

(2.1 ± 0.3)E+03 (2.0 ± 0.3)E+00 (9 ± 3)E−02

(2.8 ± 1.1)E+04 (1.7 ± 0.1)E+01 –

(2.3 ± 0.9)E+04 (2.2 ± 0.8)E+01 –

than the equilibrium model, particularly for 131I, reducing the excess conservatism of the equilibrium approach by some 2 orders of magnitude and bringing the modelled dose rate predictions in line with the monitoring-based dose rates within a factor of 1.4–22 of each other. Since the total dose rate is dominated by the internal component (owing to the relative insignificance of the external dose shortly after the accident), this result reflects the advantages of considering dynamic transfer rather than assuming equilibrium. 4. Study limitations and uncertainties The present assessment is subject to limitations. It was not possible to account for some radionuclides present in the initial post-accident period but conspicuously absent from the monitoring data, such as 89 Sr, 90Sr, 129Te, 129mTe, 136Cs or the actinides. The exposure in the late phase is certainly dominated by 134Cs and 137Cs, but given the short biological half-life of radiocaesium compared with bone-seekers such as 90Sr and plutonium, the latter are of some interest for future assessments, even if they are present in much smaller quantities (Zheng et al., 2013). Likewise, it was not possible to include exposures from sediment when modelling external doses. The lack of equilibrium between radionuclide activities in the water–sediment system renders the simple modelling of sediment concentrations using distribution coefficients (Kds) unfeasible. However, with the aid of a previous sediment model (Vives i Batlle et al., 2008a), we calculated that it takes ~ 250 days for 137Cs in surface sediment to attain 10% of its maximum equilibrium concentration in response to a constant activity concentration in seawater. This, combined with a low Kd for radiocaesium of 4 m3 kg−1 (Beresford et al., 2007) and the fact that sedimentation of suspended particles transporting adsorbed radionuclides to the sea bed is a slow process (Vives i Batlle, 2011), raises the hypothesis that the gradual build-up of radionuclides in sediment in the earliest phase of the accident may not have been enough to constitute a significant source of exposure in comparison with seawater. The extent to which marine sediments will act as a longterm source of external exposure (and possibly radionuclide uptake) by biota remains to be determined in future investigations. For the late phase, our dose evaluation is based on limited radiation effect information for fish and other marine biota at low levels of radiation (Copplestone et al., 2008; FREDERICA, 2006; Garnier-Laplace et al., 2008). A reasonable consensus has formed around the relevant chronic dose benchmarks, but their application to average exposed organisms in an accidental situation is novel and, in the case of the UNSCEAR 400 μGy h−1 benchmark, it is potentially open to scientific questioning due to its intended applicability to the most exposed organisms. Moreover, although alterations to population integrity are deemed unlikely, more subtle effects at the individual level cannot be totally ruled out, and effects on reproductive endpoints that may have impacts on longer

term population viability cannot yet be assessed. For the early phase, acute effect benchmarks dedicated to marine ecosystems are scarce and building consensus on values is an on-going challenge at the international level. Overall uncertainties of the monitoring-based assessment, particularly those involving a biological transfer, are normally large as typified by the observation that estimations using different models often differ from one another by more than an order of magnitude (Linkov and Burmistrov, 2003). The main error is incurred when applying a concentration factor to calculate the activity in biota based on activity concentration in the medium, a problem obviated by our direct use of monitoring data. Based on previous intercomparisons of non-human dosimetry models, additional uncertainty relating to the dose conversion factors can be estimated as ±25% for internal and ±120% for external exposure (Beresford et al., 2008; Vives i Batlle et al., 2007a, 2011). To this one must add the effect on external dose rates of estimating the activity in seawater and sediment from an area 10 km around the location of individual biota monitoring collection points. Assuming a 50% dispersion error in data within this radius of influence, and given the fact that the external doses were typically b70% of the total dose, we can estimate the likely effect as not exceeding ±35%. We consider these uncertainties to be acceptable for dose assessments of this kind. The modelling-based assessment has its own limitations due to the necessity to calculate the activity concentration in biota using seawater activity as an input, which was not available for the first days after the accident. The effect of this was mathematically simulated and we found it to be not significant in our 2-month projections (a factor of 2 difference at 10 days, decreasing to b 10% difference at 20 days). Another potential limitation is the aforesaid exclusion of radionuclides such as 129,129m,132Te, 136 Cs and 133I known to have been released, due to lack of data availability. However, their short decay half-lives mean that their contribution to dose must have been limited to the earliest exposures. A detailed parameter variation sensitivity analysis carried out with the D-DAT tool indicates that the concentration factor CF and the biological half-life TB1/2 are, in this order, the most sensitive model parameters (Vives Batlle et al., 2008). To this one must add the aforementioned uncertainties in the dosimetry (± 25% and ± 120% for internal and external doses, respectively). As internal exposure tends to dominate, the uncertainty in the total dose should be closer to 25%. The dynamic models D-DAT and ECOMOD correspond closely within a factor of b2 over the first 60 days of simulations, supporting a factor of ~ 2 for the modelling-related uncertainty in our assessment. Indications of a degree of correspondence between dynamically modelled and monitored radionuclide concentrations from the additional samples gathered at various local ports (Greenpeace, 2012) bring additional confidence to the modelling results. Considerations of the impact of radiation to the overall environment (including the abiotic component) lie beyond the scope of this study as

152

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153

the applicability of the radiation protection framework to the abiotic environment is still open to discussion (IAEA, 2002). Radiation dose cannot capture impacts to the actual value that mankind assigns to that environment. For the above reasons, we have avoided adding qualifiers such as “negligible” and “insignificant” to the exposures calculated in this paper. 5. Conclusions The doses calculated in this study are generally below the amounts necessary to cause a measurable effect on populations. The only exception is radio-iodine in macroalgae close to the discharge point, limited to the earlier phase after the accident. Exposures for marine biota during the late phase fall below thresholds for which population effects are deemed likely. Further away from the FDNPS, the potential for effects on biota will consequently be lower. It would seem that the possibility of effects on non-human biota is, at most, geographically constrained to a few hotspot locations, possibly linked to the delayed inputs of radioactivity currently under investigation. If this study has reduced some of the initial estimates of exposures from previous studies, it has also revealed that enhanced levels of radioactivity in biota persist in the vicinity of the FDNPS. There is a need to characterise any local hotspots and understand the resilience of radioactivity in the most exposed biota. Furthermore, our present understanding of the biological impacts of radiation on chronically exposed plants and animals is largely based upon limited high-exposure data collated under controlled laboratory conditions. Non-human biota exhibit a wide range of inter-species radiosensitivity and may react according to a complex dynamic of interactions between absorbed doses (or dose rates) and radiotoxic responses. These interactions may be expressed at different levels of biological and ecological organization, signalling the need for future investigations. Acknowledgements This work was conducted under the auspices of the UNSCEAR and a more comprehensive version of the assessment presented in this article is reported within the UNSCEAR report (UNSCEAR, 2014). The role that UNSCEAR played in providing the datasets for analysis and in particular the contribution of Expert Group A of the UNSCEAR assessment team is gratefully acknowledged. References Alekseev AV, Khrapchenkov FF, Baklanov PJ, Blinov YG, Kachur AN, Medvedeva IA, et al. Oyashio current — GIWA regional assessment 31: regional definition. Global International Waters Assessments (GIWA) regional report; 2006. p. 13–21 [Available from http://www.unep.org/dewa/giwa/areas/reports/r31/regional_definition_giwa_ r31.pdf [Accessed 7 January 2014]]. Andersson P, Garnier-Laplace J, Beresford N, Copplestone D, Howard B, Howe P, et al. Protection of the environment from ionising radiation in a regulatory context (PROTECT): proposed numerical benchmark values. J Environ Radioact 2009; 100:1100–8. Avila R, Beresford NA, Aguero A, Broed R, Brown J, Iospje M, et al. Study of the uncertainty in estimation of the exposure of non-human biota to ionizing radiation. J Radiol Prot 2004;24:A105–22. Bailly du Bois P, Laguionie P, Boust D, Korsakissok I, Didier D, Fiévet B. Estimation of marine source-term following Fukushima Dai-ichi accident. J Environ Radioact 2012;114:2–9. Barnes RD. Invertebrate zoology. 4th ed. Philadelphia: Holt-Saunders International; 1982 [1089 pp.]. Beresford NA, Brown J, Copplestone D, Garnier-Laplace J, Howard B, Larsson C-M, et al. D-ERICA: an integrated approach to the assessment and management of environmental risks from ionising radiation. A deliverable of the ERICA project FI6R-CT-2004-508847; 2007 [88 pp. Available from: https://wiki.ceh.ac.uk/download/attachments/115017395/D-Erica.pdf?version=1 [Accessed 7 January 2014]]. Beresford NA, Barnett CL, Brown J, Cheng JJ, Copplestone D, Filistovic V, et al. Intercomparison of models to estimate radionuclide activity concentrations in nonhuman biota. Radiat Environ Biophys 2008;47(4):491–514. Brown JE, Alfonso B, Avila R, Beresford NA, Copplestone D, Pröhl G, et al. The ERICA Tool. J Environ Radioact 2008;99(9):1371–83. Buesseler K. Fishing for answers off Fukushima. Science 2012;338:480–2.

Buesseler K, Aoyama M. Fukushima derived radionuclides in the ocean. Proc. Fukushima ocean impacts symposium. Exploring the impacts of the Fukushima Dai-ichi nuclear power plants on the ocean, University of Tokyo, November 12–13, 2012; 2012. [Available from: http://www.whoi.edu/fileserver.do?id=138584&pt=2&p=141569 [Accessed 7 January 2014]]. Buesseler K, Aoyama M, Fukasawa M. Impacts of the Fukushima nuclear power plants on marine radioactivity. Environ Sci Technol 2011;45:9931–5. Copplestone D, Hingston JL, Real A. The development and purpose of the FREDERICA radiation effects database. J Environ Radioact 2008;99(9):1456–63. FREDERICA. FREDERICA radiation effects database. 2006. [Available from: http://www. frederica-online.org/mainpage.asp [Accessed 7 January 2014]]. Garnier-Laplace J, Gilbin R. Derivation of predicted no effect dose rate values for ecosystems (and their sub-organisational levels) exposed to radioactive substances. Deliverable D5. European Commission, 6th Framework Project ERICA, Contract N°FI6RCT-2003-508847; 2006. Garnier-Laplace J, Copplestone D, Gilbin R, Alonzo F, Ciffroy P, Gilek M, et al. Issues and practices in the use of effects data from FREDERICA in the ERICA integrated approach. J Environ Radioact 2008;99(9):1474–83. Garnier-Laplace J, Della-Vedova C, Andersson P, Copplestone D, Cailes C, Beresford NA, et al. A multi-criteria weight of evidence approach for deriving ecological benchmarks for radioactive substances. J Radiol Prot 2010;30(2):215–33. Garnier-Laplace J, Beaugelin-Seiller K, Hinton TG. Fukushima wildlife dose reconstruction signals ecological consequences. Environ Sci Technol 2011;45:5077–8. Gómez-Ros JM, Prohl A, Ulanovsky A, Lis M. Uncertainties of internal dose assessment for animals and plants due to non-homogeneously distributed radionuclides. J Environ Radioact 2008;99(9):1449–55. Greenpeace. Marine testing lab results 1. published 25 May 2011 http://www.greenpeace. org/international/Global/international/publications/nuclear/2011/Report%20SCK% 20CEN.pdf, 2012. [Accessed 7 January 2014]. Hosseini A, Beresford NA, Brown JE, Jones DG, Phaneuf M, Thørring H, et al. Background dose-rates to reference animals and plants arising from exposure to naturally occurring radionuclides in aquatic environments. J Radiol Prot 2010;30(2): 235–64. IAEA. Ethical considerations in protecting the environment from the effects of ionizing radiation: a report for discussion. International Atomic Energy Agency Technical Document 1270; 2002 [Vienna, 29 pp.]. IAEA. Sediment distribution coefficients and concentration factors for biota in the marine environment. Technical reports series no. 422. Vienna: International Atomic Energy Agency; 2004. IAEA. IAEA International fact finding expert mission of the Fukushima Dai-ichi NPP accident following the Great East Japan earthquake and tsunami. Tokyo, Fukushima Dai-ichi NPP, Fukushima Dai-ni NPP and Tokai Dai-ni NPP, Japan, 24 May–2 June 2011. Vienna: International Atomic Energy Agency; 2011. ICRP. Radionuclide transformations — energy and intensity of transmissions. ICRP publication 38 (Annals of the ICRP 11). Oxford: Pergamon Press; 1983. ICRP. The concept and use of reference animals and plants for the purposes of environmental protection. In: Valentin J, editor. International commission on radiological protection publication 108. Annals of the ICRPOxford: Elsevier; 2008. [76 pp.]. ICRP. Transfer parameters for reference animals and plants. International Commission on Radiological Protection Environmental protection publication 114Annals of the ICRP 39(6). Oxford: Elsevier; 2009. IRSN. Synthèse actualisée des connaissances relatives à l'impact sur le milieu marin des rejets radioactifs du site nucléaire accidenté de Fukushima Dai-ichi. Available from: http://www.irsn.fr/FR/Actualites_presse/Actualites/Documents/IRSN-NIImpact_accident_Fukushima_sur_milieu_marin_26102011.pdf, 2011. [accessed 7 January 2014]. IRSN. Fukushima, one year later — initial analyses of the accident and its consequences. French institute for radiological protection and nuclear safety report IRSN/DG/ 2012-003, 12 March 2012; 2012 [188 pp.]. Kim YG. Ecological study on the trsansplantation of Sea Squirt, Halocynthia roretzi (v. Drasche) to Gogunsan Islands. Bull Korean Fish Soc 1980;13(2):57–64. Knowles JF, Greenwood LN. The effects of chronic irradiation on the reproductive performance of Ophryotrocha diadema (Polychaeta, Dorvilleidae). Mar Environ Res 1994;38:207–24. Kryshev AI, Ryabov IN. A dynamic model of 137Cs accumulation by fish of different age classes. J Environ Radioact 2000;50(3):221–33. Kryshev II, Sazykina TG. Mathematical modeling of radionuclide migration in aquatic ecosystems. Energoatomizdat 1986:1–152. Kryshev A, Sazykina T. Evaluation of the irradiation dose rate for marine biota in the region of the destroyed Fukushima reactor (Japan) in March–May 2011. At Énergiya 2011;111(1):41–5. Kryshev II, Kryshev AI, Sazykina TG. Dynamics of radiation exposure to marine biota in the area of the Fukushima NPP in March–May 2011. J Environ Radioact 2012;114:157–61. Larsson CM. An overview of the ERICA integrated approach to the assessment and management of environmental risks from ionising contaminants. J Environ Radioact 2008;99(9):1364–70. Linkov I, Burmistrov D. Model uncertainty and choices made by modelers: lessons learned from the International Atomic Energy Agency model intercomparisons. Risk Anal 2003;23(6):1297–308. Olsen YS, Vives i Batlle J. A model for the bioaccumulation of 99Tc in lobsters (Homarus gammarus) from the West Cumbrian coast. J Environ Radioact 2003;67(3):219–33. PMJHC. Report of Japanese Government to the IAEA ministerial conference on nuclear safety. The accident at TEPCO's Fukushima nuclear power stations. Prime Minister of Japan and His Cabinet. Nuclear Emergency Response Headquarters, Government of Japan; 2011 [June, 387 pp.].

J. Vives i Batlle et al. / Science of the Total Environment 487 (2014) 143–153 Povinec PP, Hiroshe K, Aoyama M. Fukushima accident: radioactivity impact on the environment. Elsevier; 2013 [400 pp.]. Real A, Sundell-Bergman S, Knowles JF, Woodhead DS, Zinger I. Effects of ionising radiation exposure on plants, fish and mammals: relevant data for environmental radiation protection. J Radiol Prot 2004;24:A123–37. Sazykina TG. ECOMOD — an ecological approach to radioecological modelling. J Environ Radioact 2000;50(3):207–20. Shiomoto A, Hashimoto S, Murakami T. Primary productivity and solar radiation off Sanriku in May 1997. J Oceanogr 1998;54:539–44. Strand P, Aono T, Brown J, Garnier-Laplace J, Hosseini A, Sazykina T, et al. Assessment of Fukushima-derived radiation doses and effects on wildlife in Japan. Environmental Science & Technology Viewpoint 2014;1:198–203. Swift DJ. The accumulation of plutonium by the European lobster (Homarus gammurus L.). J Environ Radioact 1992;16:1–24. Tsumune D, Tsubono T, Aoyama M, Hirose K. Distribution of oceanic 137Cs from the Fukushima Dai-ichi nuclear power plant simulated numerically by a regional ocean model. J Environ Radioact 2012;111:100–8. Ulanovsky A, Prohl G. A practical method for assessment of dose conversion coefficients for aquatic biota. Radiat Environ Biophys 2006;45(3):203–14. Ulanovsky A, Pröhl G. Tables of dose conversion coefficients for estimating internal and external radiation exposures to terrestrial and aquatic biota. Radiat Environ Biophys 2008;47(2):195–203. Ulanovsky A, Prohl G, Gomez-Ros JM. Methods for calculating dose conversion coefficients for terrestrial and aquatic biota. J Environ Radioact 2008;99(9): 1440–8. UNSCEAR. Sources and effects of ionizing radiation. United Nations Scientific Committee on the Effects of Atomic Radiation report to the general assembly, with scientific annex: effects of radiation on the environment. United Nations sales publication E. 96.IX.3; 1996 [New York, 86 pp.]. UNSCEAR. Sources and effects of ionizing radiation. Report to the general assembly, with scientific annexes; annex e: effects of ionizing radiation on non-human biota. New-York: United Nations Scientific Committee on the Effects of Atomic Radiation, United Nations; 2008 [313 pp.]. UNSCEAR. Sources and effects of ionizing radiation. Volume II: effects, scientific annexes C, D and E. United Nations Scientific Committee on the Effects of Atomic Radiation. sales publication E.11.IX.3. New York: United Nations; 2011. UNSCEAR. Sources, Effects and Risk of Ionizing Radiation. Volume I: Report to the General Assembly, Scientific Annex A: Levels and effects of radiation exposure to the nuclear accident after the 2011 great east-Japan earthquake and tsunami. Vienna: United Nations Scientific Committee on the Effects of Atomic Radiation, Report to the 68th session of the United Nations General Assembly A/68/46; 2014 [311 pp.]. Vives Batlle J, Wilson RC, Watts SJ, Jones SR, McDonald P, Vives-Lynch S. Dynamic model for the assessment of radiological exposure to marine biota. J Environ Radioact 2008; 99:1711–30. Vives i Batlle J. Impact of nuclear accidents on marine biota. Integr Environ Assess Manag 2011;7(3):365–7.

153

Vives i Batlle J. Radioactivity in the marine environment. In: Meyers RA, editor. Encyclopaedia of sustainability science and technology. Springer Science & Business Media, LLC; 2012. p. 8387–425. Vives i Batlle J, Vandenhove H. Dynamic modelling of the radiological impact of the Fukushima accident on marine biota. Ann Belg Ver Stralingsbescherming 2014; 38(3):299–312. Vives i Batlle J, Jones SR, Gomez-Ros JM. A method for calculation of dose per unit concentration values for aquatic biota. J Radiol Prot 2004;24(4A):A13–34. Vives i Batlle J, Wilson RC, McDonald P, Parker TG. A biokinetic model for the uptake and release of radioiodine by the edible periwinkle Littorina littorea. In: Povinec PP, Sanchez-Cabeza JA, editors. Radionuclides in the environment, Volume 8. Elsevier; 2006. p. 449–62. Vives i Batlle J, Balonov M, Beaugelin-Seiller K, Beresford NA, Brown J, Cheng J-J, et al. Inter-comparison of unweighted absorbed dose rates for non-human biota. Radiat Environ Biophys 2007a;46(4):349–73. Vives i Batlle J, Wilson RC, McDonald P. Allometric methodology for the calculation of biokinetic parameters for marine biota. Sci Total Environ 2007b;388(1-3):256–69. Vives i Batlle J, Bryan S, McDonald P. A process-based model for the partitioning of soluble, particulate and sediment fractions of plutonium and radiocaesium in the Eastern Irish Sea near Sellafield. J Environ Radioact 2008a;99(1):1464–73. Vives i Batlle J, Wilson RC, Watts SJ, Jones SR, McDonald P, Vives-Lynch S. Dynamic model for the assessment of radiological exposure to marine biota. J Environ Radioact 2008b;99:1711–30. Vives i Batlle J, Wilson RC, Watts SJ, McDonald P, Craze A. Derivation of allometric relationships for radionuclides in marine phyla. Radioprotection 2009;44(5):47–52. Vives i Batlle J, Barnett CL, Beaugelin-Seiller K, Beresford NA, Copplestone D, Horyna J, et al. Absorbed dose conversion coefficients for non-human biota: an extended inter-comparison of data. Radiat Environ Biophys 2011;50(2):231–51. Watts SJ, Vives i Batlle J, Wilson RC, Jones SR, McDonald P. Dynamic modelling for the assessment of radiation impacts on biota and man using Excel VBA, and related extensions of the ERICA assessment methodology. In: Strand P, Brown J, Jolle T, editors. Proc. international conference on radioecology and environmental radioactivity, Bergen, Norway, 15–20 June 2008Publ. Norwegian Radiation Protection Authority, Part 1 978-82-90362-25-1; 2008. p. 289–93. Weiss W. Preparing a scientific report to the general assembly on ‘exposures due to the nuclear accident following the Great East-Japan earthquake and tsunami’. J Radiol Prot 2012;32:113–8. Wilson RC, Vives i Batlle J, McDonald P, Parker TG. Uptake and depuration of I-131 by the edible periwinkle Littorina littorea: uptake from labelled seaweed (Chondrus crispus). J Environ Radioact 2005;80(3):259–71. Wilson RC, Vives Batlle J, Watts SJ, McDonald P, Parker TG. Uptake and depuration of 131I from labelled diatoms (Skeletonema costatum) to the edible periwinkle (Littorina littorea). J Environ Radioact 2007;96:75–84. Zheng J, Tagami K, Uchida S. Release of plutonium isotopes from the Fukushima Daiichi nuclear power plant accident: what is known and what needs to be known. Environ Sci Technol 2013;47:9584–95.