Science of the Total Environment 493 (2014) 701–707

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Modeling of leachable 137Cs in throughfall and stemflow for Japaneseforest canopies after Fukushima Daiichi Nuclear Power Plant accident

Nicolas Loffredo a,⁎, Yuichi Onda a, Ayumi Kawamori b, Hiroaki Kato a

a Center for Research in Isotopes and Environmental Dynamics, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8572, Japanb Graduate School of Life and Environmental Sciences, University of Tsukuba

H I G H L I G H T S

• A double exponential model was used to model leachable cesium loss from canopies.• The model could not reproduce variation observed.• Rainfall was identified as the dominant factor controlling the variation.• A rainfall parameter was used to develop an improved double exponential model.• The improved model gives a better estimation of leaf cesium leaching.

⁎ Corresponding author at: CRIED (Center for ResearchDynamics), Faculty of Life and Environmental Sciences, U1-1-1 Sougou Kenkyu A416, Tsukuba, 305-8577, Japan. Te

E-mail address: wataiso@free.fr (N. Loffredo).

http://dx.doi.org/10.1016/j.scitotenv.2014.06.0590048-9697/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 25 March 2014Received in revised form 13 June 2014Accepted 14 June 2014Available online xxxx

Editor: Mae Mae Sexauer Gustin

Keywords:RadionuclideDynamicEnvironmentInitial leachable stockKinetics

The Fukushima accident dispersed significant amounts of radioactive cesium (Cs) in the landscape. Our researchinvestigated, from June 2011 to November 2013, the mobility of leachable Cs in forests canopies. In particular,137Cs and 134Cs activity concentrations were measured in rainfall, throughfall, and stemflow in broad-leaf andcedar forests in an area located 40 km from the power plant.Leachable 137Cs loss was modeled by a double exponential (DE) model. This model could not reproduce the var-iation in activity concentration observed. In order to refine the DE model, the main physical measurable param-eters (rainfall intensity, wind velocity, and snowfall occurrence)were assessed, and rainfall was identified as thedominant factor controlling observed variation. A corrective factor was then developed to incorporate rainfall in-tensity in an improved DE model.With the original DEmodel, we estimated total 137Cs loss by leaching from canopies to be 72± 4%, 67± 4%, and48 ± 2% of the total plume deposition under mature cedar, young cedar, and broad-leaf forests, respectively. Incontrast, with the improved DE model, the total 137Cs loss by leaching was estimated to be 34 ± 2%, 34 ± 2%,and 16± 1% of the total plume deposition undermature cedar, young cedar, and broad-leaf forests, respectively.The improved DE model corresponds better to observed data in literature.Understanding 137Cs and 134Cs forest dynamics is important for forecasting future contamination of forest soilsaround the FDNPP. It also provides a basis for understanding forest transfers in future potential nuclear disasters.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

After the Fukushima Daiichi Nuclear Power Plant (FDNPP) accidenton 11March 2011, about 30 000 km2 of the Japanese territory was con-taminated by the nuclear fallout. Experts estimated emission of morethan 73 radioisotopes (IRSN, 2012). The total amount of atmospheric-released contaminants was 520 PBq of which less than 20% was depos-ited on Japanese land (Steinhauser et al., 2014). This release included

in Isotopes and Environmentalniversity of Tsukuba, Tennodail.: +81 80 6674 3236.

less than 1% cesium (Cs) (Korsakissok et al., 2013). However, due tothe relative long half-life of 134Cs and 137Cs isotopes (2.06 years and30.17 years respectively), these are still a significant source of radioac-tive contamination in the environment, particularly in forests thatcover 75% of the contaminated territory (Yoshihara et al., 2013).

Based on the Chernobyl Nuclear Power Plant (CNPP) accident, Shaw(2007) explained that there are three main phases after deposition ofnuclear released contamination in forests. The ‘early phase’, consists ofrapid mechanical weathering of radionuclides from the canopy to theunderlying soil (1–3 months). The ‘medium term phase’, includes the‘biological decontamination’ of the canopy (growth dilution and litterfall) and the root uptake of radionuclides (2–3 years). The ‘long termphase’, is the steady state that occurs over a period of 3–10+ years

Fig. 1. Location of experimental study relative to the Fukushima Daiichi Nuclear PowerPlant (FDNPP). Source of deposited amounts: the Ministry of Education, Culture, Sports,Science & Technology in Japan (MEXT) from campaign of 31st of May 2012 by airbornemeasurements.

702 N. Loffredo et al. / Science of the Total Environment 493 (2014) 701–707

and includes root uptake. Several surveys are available on the thirdphase after deposition (Steinhauser et al., 2014). Literature demon-strates that 137Cs dynamics in forest ecosystems tends to be slow withbehavior often controlled by slow soil biological mechanisms such asroot uptake and organic matter decomposition and leaf life cycle(Tobler et al., 1988; Tikhomirov and Shcheglov, 1994; Goor and Thiry,2004; Shaw, 2007; Thiry et al., 2009).

Mechanisms of litter fall, stemflow, and throughfall play an impor-tant role in 137Cs and 134Cs dynamics in forests, particularly during thefirst two phases after deposition. Depending on the tree species, thesemechanisms can contribute between 31 and 37% of deposition ontothe ground even several months after the accident (Kato et al., 2012).

Due to significant differences between CNPP and FDNPP accidents(Steinhauser et al., 2014), including deposition chemistry, forest spe-cies, and weather, it is useful to understand the 137Cs and 134Cs dynam-ics in Japanese forests after the FDNPP accident.

Accordingly, the goal of this paper was to model the dynamics ofleachable 137Cs in forest canopies. In particularwemeasured throughfalland stemflow in mature cedar (35 years old), young cedar (15 yearsold), and broad-leaf forests located in the plume deposition path 40km from the Fukushima Daiishi nuclear power plant. Our hypothesiswas that the leachable 137Cs dynamic in forested areas that received fall-out from the FDNPP accident could be estimated by a double exponen-tial (DE) model originally developed for in vitro investigations (Madoz-Escande et al., 2005). In addition,we assumed that althoughweather af-fects this dynamic, the fundamental questionwas: howdoes theweath-er impact DE Model predictions? The objectives were therefore to: (i)measure leachable cesium in canopies after the accident, (ii) under-stand leachable cesium variation observed in throughfall, and (iii)model the 137Cs leaching by taking the rainfall into account and com-pare the predictions with those of the simple DE model.

2. Materials and methods

2.1. Forest sites

Two sites located in the Fukushima prefecture (Fig. 1) were investi-gated: (i) a Japanese cedar forest (Japonica cryptomeria), with youngand mature trees (15 and 35 years), with a population density of2,407 trees/ha and 791 trees/ha, respectively, and (ii) a mixed forest(a broad-leaf forest) composed of oaks (Quercus aliena), beech (Faguscrenata), and pines (Pinus densiflora) with a population density of2,500 trees/ha.

Biomasswas calculated according to the allometric biomass equationsgiven by Lim et al. (2013) for cedars. We calculated the biomass of eachcompartment (root, stem, branch, and foliage) by using themean diame-ter at breast height (DBH). For mature cedar the mean distribution of thetotal biomass (drymass)was: 22±2% for root, 9±1% for foliage, 7±0%for branch, and 62 ± 4% for stem; and for young cedar: 24 ± 1% for root,11 ± 1% for foliage, 8 ± 0% for branch, and 58 ± 2% for stem.

2.2. Monitoring throughfall, rainfall, and stemflow

At each site, stemflow, throughfall, and rainfall were collected. Sevencollectors of throughfall were installed in each cedar forest (mature andyoung), and 6 in the broad-leaf forests. In both forest types, 3 collectorsof stemflowwere used. These collectors were placed under a surface of200 m2 for mature cedar, 70 m2 for young cedar, and 125 m2 for broad-leaf canopies. Three rainfall collectors were placed outside of these for-ests to provide a reference.

In order to avoid contamination of samples from soil due to splashingand wind, throughfall, and rainfall collectors were installed 50 cm abovethe forest floor. Two types of collectors were used: (i) rain collectors(for throughfall and rainfall), and (ii) snow collectors. Rainfall collectorsincluded a funnel (0.016m2) and a tank (2–10 L for throughfall and rain-fall, and 5.7 L for snowfall collectors). The funnelwas coveredwith amesh

to avoid contamination by litter. Moreover, the tank was protected withan aluminum sheet to prevent algae growth. A ping-pong ball was placedin the funnel to limit evaporation (Kato et al., 2013). Stemflow was col-lected with a collar-type collector, as described in Thimonier (1998).Stemflow collectors were placed at 1.5 m above the base of the tree andconnected to a tank with a probe monitoring water level. Stemflow wascollected at maximum, medium, and minimum tree diameters: 35.3 cm,33.7 cm, and 22 cm at the mature cedar site; 19.3 cm, 13.4 cm, and12.4 cm at the young cedar site; 28 cm, 27 cm, and 25 cm at the broad-leaf site. Between July 2011 and April 2013, 34 samples of rainfall,stemflow, and throughfall were periodically collected.

2.3. Storage and analysis of samples

Samples were directly sieved to 100 μm in order to remove thecoarse organic matter and were stored in U-8 PC containers (150 mL).Then, 137Cs and 134Cs were measured for 24 h using high-purity n-type germanium coaxial gamma-ray detector (EGC25-195-R, CanberraEurisys, Meriden, Connecticut, USA) coupled with an amplifier(PSC822, Canberra-Eurisys) and a multichannel analyzer (DSA1000,Canberra). This non-destructive analytical method, measures both137Cs and 134Cs simultaneously. After decay correction, we observed anear ratio 1:1 between 134Cs and 137Cs for all the period similarly toLepage et al. (2014), and accordingly we only report 137Cs values.

2.4. Double exponential model

There are three main mechanisms of Cs loss from canopies:throughfall, stemflow, and litterfall. Litterfall (not measured)can change the stock of 137Cs in canopies, especially during the early

703N. Loffredo et al. / Science of the Total Environment 493 (2014) 701–707

phasewhere the dynamic of loss is known to be quicker than during themedium term phase (Shaw, 2007). For modeling throughfall andstemflow dynamics, a double exponential model (DE model; Eq. (1))was used. This model gives an apparent description of the first rapid

Fig. 2. CsTF and CsSF under mature cedar, young cedar, and broad-leaf forests; arrows correspvalues measured for each collector: seven values for throughfall of cedar forests, six for througthe volume of water collected in rainfall collectors by the surface area of the collectors; snowfasum of the snow heights measured between two sampling events; maximumwind velocity w(40 km from the study sites).

loss (during the early phase), followed by a slow loss (during themedi-um term phase; Madoz-Escande et al. (2005)):

CsDEmod ¼ A1 � 1−e−k1�t� �

þ A2 � 1−e−k2�t� �

ð1Þ

ond to the events discussed in Section 3.2; errors correspond to the standard deviation ofhfall of broad leaf forest, and 3 for stemflow; rainfall intensity was calculated by dividingll was measured outside each forest after a snow event with values corresponding to theas measured at a meteorological station (Japanese Meteorological Agency) at Nihonmatsu

Fig. 3. CsTF under mature cedar, young cedar, and broad-leaf forests. DE Model describes in Eq. (1); DE model was adjusted on the experimental data by using least square method (pa-rameters summarized in Table 1).

704 N. Loffredo et al. / Science of the Total Environment 493 (2014) 701–707

The initial point of the DEmodel (t0) corresponds to the initial phaseof investigation (June 2011). All parameters of thismodel were estimat-ed by the least square method based on measured data; k1 and k2 rep-resent the kinetic parameters of the DE model. The stock of Csdecreased in the canopy with time due to other mechanisms of loss(e.g. litterfall), but A1 and A2, represent the available leachable stock of137Cs at t0 because theywere fitted on data. Therefore, mechanisms (lit-ter fall, root uptake, etc.), which can affect the leachable stock of 137Cs incanopies, are implicitly taken into account in these parameters.

3. Results and discussion

3.1. Estimation of leachable cesium in canopies after the accident

Fallout densities (Bq/m2) of 137Cs in throughfall (CsTF) and stemflow(CsSF) were calculated for each forest (cedar and broad-leaf; Fig. 2).Slow and fast kinetics can be observed during theperiod of investigationdemonstrating that the DE model is well adapted to describe the 137Cs

Fig. 4. CsTF as a function of rainfall under mature cedar, young cedar and broad-leafforests.

dynamic. For CsTF, cumulative leaching was highest from September2011 to March 2012, and decreased after March. On the contrary, forstemflow, the dynamic was slower from September 2011 to March2012.

During fieldwork, activity concentrations of 137Cs in the rainfallalone were below detection limits (b1 Bq/kg; data not shown). Intakeof 137Cs by rainfall was therefore negligible, and the measured CsTFand CsSF were related to canopy release mechanisms. Throughfall isthe main mechanism of 137Cs leaching and regardless of the forest site,it represented more than 99% of total leaching; hence we used the DEmodel only to describe the CsTF values (Fig. 3).

Estimated 137Cs loss by leaching during the early phase (estimationby extrapolation to the date of the accident with the DE model) was196 kBq/m2 for mature cedar, 173 kBq/m2 for young cedar, and92 kBq/m2 for broad-leaf forests. Initial deposition of 137Cs under ourstudy sites was calculated by using airborne monitoring data in May2011 (NRA, 2011). The airborne data measured in our study sites (4values per site) were converted to fallout density concentrations (kBq/m2) according to the linear relationship shown in the MEXT (2011) re-port and estimated to be 442 ± 30 kBq/m2 for cedar forests and 451 ±17 kBq/m2 for broad-leaf forests. 137Cs loss by leaching percentagestherefore correspond to 44 ± 3%, 39 ± 3%, and 20 ± 1% of thetotal plume deposition for mature cedar, young cedar, and broad-leafforests respectively. The estimation of the total CsTF (137Cs loss inearly phase + A1 + A2) should correspond to 72 ± 4%, 67 ± 4%, and48 ± 2% of the total plume deposition.

After the FDNPP accident, Kato et al. (2012) examined 137Cs loss in acedar forest during the early phase located only 150 km southwest ofour study site. The tree density of the cedar forest (1300 trees/ha) stud-ied was similar to this study (791–2407 trees/ha). Kato et al. (2012) es-timated the total CsTF to be 40% of the total plume deposition, which isabout 30% less than the estimation of the DE model. If we assume thecanopy interception was similar for both sites, the DE model mayover-estimate the total stock of leachable 137Cs (A1 and A2) in cedarcanopies.

Fig. 5. Ratio R, calculated with Eq. (2), in function of rainfall under mature cedar, youngcedar and broad-leaf forests; R represents the variation relative to the DEmodel and cor-responding to CsTF(t)measured (CsTFme(t), kBq/m2) onCsTF(t)modeling (CsTFmo(t), kBq/m2; by using the DE model; Eq. (1)).

Table 2Values of b1 (Eq. (2)) formature cedar, young cedar andbroad-leaf forests;Sb1 correspondsto the standard deviation of b1 (Eq. (3)); p-values were calculated with the tdist functionof Excel used on the calculated t-values (Puth et al., 2014).

b1 (1/mm) Sb1 r2 p-value

Mature cedar 0.0172 0.0031 0.52 0.03Young cedar 0.0178 0.0037 0.42 0.04Broad leaf 0.0126 0.0027 0.29 0.19

705N. Loffredo et al. / Science of the Total Environment 493 (2014) 701–707

3.2. Description of leachable cesium variation in throughfall

Two distinct sampling periods where CsTF increases were observed(arrows in Fig. 2), leading to inaccuracies in the DE model (Fig. 3). Dur-ing the autumn of 2011, CsTF reached more than 50 kBq/m2, 15 kBq/m2, and 5 kBq/m2 under mature cedar, young cedar, and broad-leaf for-ests respectively. During the spring of 2012, values exceeded 20 kBq/m2, 10 kBq/m2, and 5 kBq/m2. For the remainder of the year, the

Table 1Parameters used in the double exponential model (DE) in order to describe CsTF; p-valueswere calculatedwith the tdist function of Excel used on the calculated t-values (Puth et al.,2014).

Mature cedar Young cedar Broad leaf

A1 (Bq/m2) 62 99 57A2 (Bq/m2) 61 25 8k1 (1/d) 5.0E-04 4.2E-04 2.1E-04k2 (1/d) 1.2E-02 1.7E-02 2.0E-02r2 0.986 0.995 0.988p-value b 0.001

mean CsTF was 2 kBq/m2, 2 kBq/m2, and 1 kBq/m2 under maturecedar, young cedar, and broad-leaf forests, respectively.

The same mechanism was observed for potassium that has similarforest dynamics compared to Cs (Myttenaere et al., 1993; Sombréet al., 1994). Sase et al. (2008) concluded that leaf permeability inducedby wax degradation could be one of the key factors governing leachingof potassium in Japanese cedar forests (Japonica cryptomeria). They re-ported that for year 1 leaves, all the crystal structure of the wax isdestroyed in autumn. The wax could also be a barrier for Cs, and degra-dation would allow Cs to be more available for leaching in autumn.

In spring, similarly to potassium, Cs can be largely directed throughphloem from the older to the younger plant tissues (Zörb et al., 2014). Atagging experiment of trees by Auerbach and Olson (1964), revealedthat leaves or needles accumulated more 137Cs in spring and thereforeare likely more leachable at this time. Both mechanisms are difficult toqualify, but the influence of rainfall is likely dominant.

3.3. Variation of 137Cs in throughfall linkedwith meteorological parameters

We evaluated the impact of wind, rainfall intensity, and snowfall on137Cs remobilization from canopies. Regardless of the forest, there wereno evident correlations between CsTF/CsSF and snowfall and wind ve-locity (Fig. 2). Although,Delphis and Levia (2003) found that the contacttime of snow on trees can increase the nutrient release in cedar andbroad-leaf forests, we found that 137Cs in throughfall and stemflowdid not deviate in snowfall episodes. The strongest correlation observedwas with rainfall intensity. However, this correlation is not visible be-tween rainfall intensity and CsTF (Fig. 4) but between rainfall intensityand the variation relative to the DE model.

The ratio (R)which represents the variation relative to theDEmodelwas calculated (Eq. (2)). When R is equal to 1, there is no variation be-cause the CsTF(t) measured (CsTFme(t)) is equal to the CsTF(t) calculat-ed by the DE model (CsTFmo(t)). On the contrary, there is a variationwhen R is different from 1.

In Fig. 5, a linear relationship is observed between rainfall and theratio R. A linearmodel was therefore used to describe R as a linear func-tion of rainfall intensity (Eq. (2)).

R rfð Þ ¼ CsT Fme tð ÞCsT Fmo tð Þ≈b1 � rf ð2Þ

b1 is the slope of the linear Eq. (2) and rf is the rainfall intensity (mm).The values of b1 for each forest are given in Table 2 and were estimatedby a simple linear regression. Standard deviations of b1 Sb1

� �were cal-

culated by using Eq. (3).

Sb1 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiX

yi−yð Þ2

n−2ð Þ �X

xi−xð Þ2

vuut ð3Þ

yi and xi are R(rf) and rf values; y are the recalculated R(rf) valueswith Eq. (2); n is the number of data; x is the mean of rf.

The linear relationship between rainfall and R is more suited tothe cedar forests (p-value b 0.05) than to the broad-leaf forest (p-value N 0.05; Table 2).

706 N. Loffredo et al. / Science of the Total Environment 493 (2014) 701–707

This could be due to the differences of behavior between decidu-ous and evergreen forest. Some authors have postulated that canopydensity is as an important factor influencing nutrient depositionon different forest floors (Potter et al., 1991; Whelan et al., 1998).For the cedar forests, we assume that canopy density is constantand that rainfall drives variations. However, for broad-leaf forest,canopy density varies and may have an additional impact on 137Csleaching.

3.4. Modeling 137Cs leaching

We added rainfall to the DE model (CsTFmo2(t)). By multiplyingEq. (2) by CsTFmo(t), the result nearly equals CsTFme(t):

CsT Fmo2 tð Þ ¼ R� CsT Fmo tð Þ→ CsT Fme tð Þ ð4Þ

CsTFmo2(t) corresponds to the corrected 137Cs flux of canopies. Thecumulative CsTFmo2(t) was calculated by the improved DE modelfound in Eq. (5), with the terms of Eq. (4) replaced by their respectiveterms (Eqs. (1) and (2)).

CsT Fmo2 tð Þ ¼Ztfinal

t0

b1 � rfð Þ � A1 � 1−e−k1�t� �

þ A2 � 1−e−k2�t� �h ih i

ð5Þ

The fit between CsTFme(t) and CsTFmo2(t) (Fig. 6) followed seasonaltrends observed in Fig. 2. The correlation coefficients between observedandmodeled data are 0.98, 0.98, and 0.90 (p-values b 0.001) for maturecedar, young cedar, and broad-leaf forests, respectively. Rainfall intensi-ty was themost important factor describing the seasonal 137Cs variationin throughfall. Although some authors have suggested that leaf perme-ability could be an important factor for nutrient leaching (Sase et al.,2008), this mechanism does not appear to be significant for Cs mobilityin cedar and broad-leaf forests after deposition. Indeed, themechanismof canopydecontamination by leaching ismore likely to bephysical (e.g.

Fig. 6. CsTF under mature cedar, young cedar and broad-leaf forests. Improved DEmodel descrisummarized in Table 2.

erosion by desquamation of cuticles; (Madoz-Escande et al. (2004))than biological.

The total 137Cs loss by leaching during the early phase after deposi-tion was recalculated using the improved DE model and rainfall at theYamakiya site (Ministry of Land, Infrastructure and Transport of Japan(2014)). 137Cs loss during the early phase was modeled to be to31 kBq/m2 for mature cedar, 27 kBq/m2 for young cedar, and 6 kBq/m2 for broad-leaf forests. Thus, 137Cs loss by leaching percentageswould correspond to 7 ± 0%, 6 ± 0%, and 1 ± 0% of the total plume de-position. Estimation of total CsTF (137Cs loss in early phase + A1 + A2)corresponds to 34 ± 2%, 34 ± 2%, and 16 ± 1% of the total plumedeposition.

Compared to the original DE model, the improved DE model pro-vides a lower estimation of the total CsTF (DE model: 72 ± 4%, 67 ±4%, and 48± 2% for mature cedar, young cedar, and broad-leaf forests).These estimations are more consistent with the study of Kato et al.(2012) who reported total CsTF corresponds to 40% of the total plumedeposition for the same period. The improved DEmodel provides a bet-ter estimation of total 137Cs loss by leaching than the DEmodel throughthe incorporation of a rainfall parameter.

4. Conclusions

137Cs leaching from the canopies of cedar and broad-leaf forests afterthe FDNPP accident was modeled using a DEmodel. This model did notaccurately describe the variation observed in throughfall that represent-ed 99% of 137Cs leaching from canopies. Rainfall intensity explained theobserved variation and accordingly a corrective factor was determinedand applied to modify the DE model. The total loss of 137Cs throughthroughfall during the early phase was recalculated with an improvedDE model (including rainfall parameters). Estimation of total CsTF was34±2%, 34± 2%, and 16± 1% of the total plume deposition formaturecedar, young cedar, and broad-leaf forests respectively.

To further document the uncertainty of the improved DE model itwould be worthwhile to conduct a sensitivity analysis and a thoroughcomparison of model results to observations for different forests in the

bed in Eq. (4). Kinetics parameters are summarized in Table 1, and rainfall parameters are

707N. Loffredo et al. / Science of the Total Environment 493 (2014) 701–707

region. Future research should focus on determining the relationshipsbetween the DE model parameters (A1, A2, k1 and k2) and physical pa-rameters in forests in order to improve our understanding Cs storageand deposition in different components of forest landscapes.

Acknowledgements

This workwas commissioned by the JAEA as a part of the distributionmapping project financially supported by theMinistry of Education, Cul-ture, Sports, Science, and Technology (Japan). We would like to thankour colleagues for their valuable help: F. Coppin from IRSN (Institute/of Radioprotection and Nuclear Safety; France) and F. Koiwa, A. Okasakiand N. Oshima from Tsukuba University (Japan). We would like to ac-knowledge the French Laboratoire des Sciences du Climat et del’Environnement (LSCE) for their timely and useful insights on thismanuscript.

References

Auerbach SI, Olson JS. Landscape investigations using caesium-137. Nature 1964;201.Delphis F, Levia J.Winter stemflow leaching of nutrient-ions from deciduous canopy trees

in relation to meteorological conditions. Agric For Meteorol 2003;117:39–51.Goor F, Thiry Y. Processes, dynamics and modelling of radiocaesium cycling in a

chronosequence of Chernobyl-contaminated Scots pine (Pinus sylvestris L.) planta-tions. Sci Total Environ 2004;325:163–80.

IRSN. Fukushima, un an après. Premières analyses de l’accident et de ses conséquences.Report IRSN/DG/2012-001; 2012.

Kato H, Onda Y, Takashi G. Interception of the Fukushima reactor accident-derived 137 Cs,134 Cs and 131 I by coniferous forest canopies. Geophys Res Lett 2012;39.

Kato H, Onda Y, Nanko K, Gomi T, Yamanaka T, Kawaguchi S. Effect of canopy interceptionon spatial variability and isotopic composition of throughfall in Japanese cypressplantations. J Hydrol 2013;504:1–11. [Nov 11].

Korsakissok I, Mathieu A, Didier D. Atmospheric dispersion and ground deposition in-duced by the Fukushima Nuclear Power Plant accident: a local-scale simulation andsensitivity study. Atmos Environ 2013;70:267–79.

Lepage H, Evrard O, Onda Y, Patin J, Chartin C, Lefèvre I, et al. Environmental mobility of110mAg: lessons learnt from Fukushima accident (Japan) and potential use for track-ing the dispersion of contamination within coastal catchments. J Environ Radioact2014;130:44–55. [Apr].

Lim H, Lee K-H, Lee KH, Park IH. Biomass expansion factors and allometric equations in anage sequence for Japanese cedar (Cryptomeria japonica) in southern Korea. J For Res2013;18(4):316–22. [Aug 1].

Madoz-Escande C, Henner P, Bonhomme T. Foliar contamination of Phaseolus vulgariswith aerosols of 137Cs, 85Sr, 133Ba and 123mTe: influence of plant developmentstage upon contamination and rain. J Environ Radioact 2004;73:49–71.

Madoz-Escande C, Garcia-Sanchez L, Bonhomme T, Morello M. Influence of rainfall char-acteristics on elimination of aerosols of cesium, strontium, barium and tellurium de-posited on grassland. J Environ Radioact 2005;84:1–20.

Ministry of Land, Infrastructure and Transport, Japan. Water information system. web site:http://www1.river.go.jp/cgi-bin/SrchRainData.exe?ID=102011282218120&KIND=2&PAGE=0, 2014.

Myttenaere C, Schell WR, Thiry Y, Sombre L, Ronneau C, van der Stegen de Schrieck J.Modelling of Cs-137 cycling in forests: recent developments and research needed.Sci Total Environ 1993;136:77–91.

NRA. Evironment Monitoring Database for the Distribution of Radioactive Substances Re-leased by the TEPCO Fukushima Daishi NPP Accident. Database of Fukushima prefec-ture: https://mapdb.jaea.go.jp/mapdb/en/portals/73/?prefectures=07&category_b=b202&year=2011, 2011.

Potter CS, Ragsdale HL, Swank WT. Atmospheric deposition and foliar leaching in aregenerating southern Appalachian forest canopy. J Ecol 1991;79:97–115.

Puth M-T, Neuhäuser M, Ruxton GD. Effective use of Pearson’s product–moment correla-tion coefficient. Anim Behav 2014;93:183–9. [Jul].

Sase H, Takahashi A, Sato M, Kobayashi H, Nakata M, Totsuka T. Seasonal variation in theatmospheric deposition of inorganic constituents and canopy interactions in aJapanese cedar forest. Environ Pollut 2008;152:1–10.

Shaw G. Radionuclides in forest ecosystems. In: George S, editor. Radioact Environ [Inter-net]. Elsevier; 2007. p. 127–55. [Available from: http://www.sciencedirect.com/sci-ence/article/pii/S1569486006100066].

Sombré L, Vanhouche M, de Brouwer S, Ronneau C, Lambotte JM, Myttenaere C. Long-term radiocesium behaviour in spruce and oak forests. Sci Total Environ 1994;157:59–71.

Steinhauser G, Brandl A, Johnson TE. Comparison of the Chernobyl and Fukushima nucle-ar accidents: a review of the environmental impacts. Sci Total Environ 2014;470–471:800–17. [Feb 1].

Thimonier A. Measurement of atmospheric deposition under forest canopies: some rec-ommendations for equipment and sampling design. Environ Monit Assess 1998;52:353–87. [Sep 1].

Thiry Y, Colle C, Yoschenko V, Levchuk S, Van Hees M, Hurtevent P, et al. Impact of Scotspine (Pinus sylvestris L.) plantings on long term 137Cs and 90Sr recycling from awaste burial site in the Chernobyl Red Forest. J Environ Radioact 2009;100:1062–8.

Tikhomirov FA, Shcheglov AI. Main investigation results on the forest radioecology in theKyshtym and Chernobyl accident zones. Sci Total Environ 1994;157:45–57.

Tobler L, Bajo S, Wyttenbach A. Deposition of 134,137Cs from Chernobyl fallout onNorway spruce and forest soil and its incorporation into spruce twigs. J EnvironRadioact 1988;6:225–45.

Whelan MJ, Sanger LJ, Baker M, Anderson JM. Spatial patterns of throughfall and mineralion deposition in a lowland Norway spruce (Picea abies) plantation at the plot scale.Atmos Environ 1998;32:3493–501.

Yoshihara T, Matsumura H, Hashida S, Nagaoka T. Radiocesium contaminations of 20wood species and the corresponding gamma-ray dose rates around the canopies at5 months after the Fukushima nuclear power plant accident. J Environ Radioact2013;115:60–8.

Zörb C, Senbayram M, Peiter E. Potassium in agriculture – Status and perspectives. J PlantPhysiol 2014 May 15;171(9):656–69.