lable at ScienceDirect

Journal of Environmental Radioactivity 102 (2011) 26e34

Contents lists avai

Journal of Environmental Radioactivity

journal homepage: www.elsevier .com/locate / jenvrad

Measured and modelled tritium concentrations in freshwater Barnes mussels(Elliptio complanata) exposed to an abrupt increase in ambient tritium levels

T.L. Yankovich a,*, S.B. Kim b, F. Baumgärtner c, D. Galeriu d, A. Melintescu d, K. Miyamoto e, M. Saito f,F. Siclet g, P. Davis b

aAREVA Resources Canada, Safety, Health, Environment and Quality, P.O. Box 9204, 817-45th Street West, Saskatoon, Saskatchewan S7K 3X5, CanadabAtomic Energy of Canada Limited (AECL), Canadac Technical University Munich, Institute of Radiochemistry, Germanyd “Horia Hulubei” National Institute for Physics and Nuclear Engineering (NIPNE), RomaniaeNational Institute of Radiological Sciences (NIRS), JapanfKyoto University Safety Reassurance Academy (SRA), Japang Electricité de France (EDF), France

a r t i c l e i n f o

Article history:Received 12 October 2009Received in revised form3 May 2010Accepted 21 June 2010Available online 12 October 2010

Keywords:Organically-bound tritium, OBTHTOFreshwater musselsModel comparisonDynamic

* Corresponding author. Tel.: þ1 306 343 4050; faxE-mail address: tamara.yankovich@areva.ca (T.L. Y

0265-931X/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.jenvrad.2010.06.014

a b s t r a c t

To improve understanding of environmental tritium behaviour, the International Atomic Energy Agency(IAEA) included a Tritium and C-14 Working Group (WG) in its EMRAS (Environmental Modelling forRadiation Safety) program. One scenario considered by the WG involved the prediction of time-dependent tritium concentrations in freshwater mussels that were subjected to an abrupt increase inambient tritium levels. The experimental data used in the scenario were obtained from a study in whichfreshwater Barnes mussels (Elliptio complanata) were transplanted from an area with background tritiumconcentrations to a small Canadian Shield lake that contains elevated tritium. The mussels were thensampled over 88 days, and concentrations of free-water tritium (HTO) and organically-bound tritium(OBT) were measured in the soft tissues to follow the build-up of tritium in the mussels over time.

The HTO concentration in the mussels reached steady state with the concentration in lake waterwithin one or two hours. Most models predicted a longer time (up to a few days) to equilibrium. Allmodels under-predicted the OBT concentration in the mussels one hour after transplantation, but over-predicted the rate of OBT formation over the next 24 h. Subsequent dynamics were not well modelled,although all participants predicted OBT concentrations that were within a factor of three of the obser-vation at the end of the study period. The concentration at the final time point was over-predicted by allbut one of the models. The relatively low observed concentration at this time was likely due to the loss ofOBT by mussels during reproduction.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Aquatic organisms are occasionally exposed to short-term,elevated tritium concentrations in watersheds that have fluctuatingtritium levels. In general, the rates of tritiated water (HTO) uptakeand organically-bound tritium (OBT) formation in aquatic organismsare not well known under dynamic exposure conditions, but can bestudied by transplanting biomonitoring species from areas with lowcontaminant concentrations to areas of elevated concentrations.Althoughmany such transplantation experiments have been done toassess contaminant dynamics in mussels (e.g., Bayne et al., 1985;

: þ1 306 343 4687.ankovich).

All rights reserved.

Clarke, 1981; Curry, 1977/78; Dechno and Luoma, 1992; Elder andCollins, 1991; Jackim et al., 1977; Greig et al., 1975; Hinch andStephenson, 1987; Kauss et al., 1981; Kauss and Hamdy, 1985;Lobel et al., 1982; Lobel and Wright, 1982; Luten et al., 1986;Matteson, 1948; Phillips, 1979; Rubenstein et al., 1983; Servoset al., 1987; Tatum, 1986; Tessier et al., 1984, 1993), similar studieshave not been done for tritium. That said, such an approach providesa useful way to study tritium dynamics using mussels as bio-monitors, such that changes in HTO and OBT concentrations can bemonitored to quantify responses to dynamic exposure conditions.

Tritium dynamics in aquatic organisms werestudied herethrough transplantation of freshwater Barnes mussels (Elliptiocomplanata) from areas with low tritium concentrations to PerchLake, a small Canadian Shield lake located on Atomic Energy ofCanada Limited (AECL)’s Chalk River Laboratories (CRL) site.

T.L. Yankovich et al. / Journal of Environmental Radioactivity 102 (2011) 26e34 27

The tritium concentration in the lake is approximately 4500 Bq/L(Kim et al., 2004; Yankovich and Kim, 2005). The mussels wereexposed to tritium through the water pathway alone, or via bothwater and sediments. The observations were used to test modelsthat predict temporal changes in HTO and OBT concentrations inaquatic organisms following abrupt increases in their tritiumexposure levels. This model validation exercise involved fivemodelling teams within the Tritium/Carbon-14 Working Group ofthe International Atomic Energy Agency (IAEA)’s EMRAS (Environ-mental Modelling for Radiation Safety) program (Yankovich andKim, 2005).

2. Experimental methodologies

2.1. Site description

Perch Lake is situated downstream of two historic CRL WasteManagement Areas. Tritium from these areas enters the lake in theform of HTO, primarily via groundwater discharge through theunderlying lake sediments and through one of five small inflowingstreams. Tritium measurements are available for Perch Creek,which drains Perch Lake, from the early 1960s until the present.Examination of temporal trends over this period suggests thatcurrent HTO concentrations in Perch Lake are similar to or may beslightly higher than values observed in the 1960s, with lower levelsoccurring in the lake in the late 1970s (Yankovich et al., 2000).During the 1990s, it appears that there may have been a slightincrease in Perch Lake HTO levels; however, this trend may be anartifact, since the sampling site for the Perch Lake outflow waschanged from Perch Lake outlet to a location further downstream atPerch Creek weir in 1988. Although a correction was made toaccount for this change in sampling location, it may have contrib-uted to the observed increase in the 1990s. It is expected thatmussel HTO concentrations will quickly track those measured inPerch Lake surface waters; however, changes in OBT in musseltissues as HTO is converted to OBT will occur over a longer time-frame and will generally integrate fluctuations in HTO over time(Yankovich et al., 2008).

The spatial distribution of HTO in the lake is not well knownquantitatively, although it is believed that the lake is well mixed,with a mean tritium concentration of approximately 4500 Bq/L inthe vicinity of the mussel transplantation cages, which weredeployed in a shallow, sandy substrate where many mussels arefound naturally (Kim et al., 2004; Yankovich and Kim, 2005).

2.2. Mussel cages

Two pairs of mussel transplantation cages were built anddeployed in Perch Lake in early July 2004. The cages had dimen-sions of 96 cm (length)� 96 cm (width)� 12 cm (height). Eachcage was built with an 8� 8 design, resulting in a total of 64compartments per cage (Yankovich and Kim, 2005). Individual cagecompartments had surface area dimensions of 12 cm� 12 cm withone animal per compartment to provide mussels with adequatespace without overcrowding.

2.3. Mussel collection and transplantation

Test mussels were collected on 5e7 July 2004 from a nearbyreference site in the Ottawa River (Fig. 1) with low tritiumconcentrations (less than 10 Bq/L HTO in water). During collection,each mussel was carefully examined to assess its suitability for thestudy. Mussels with total shell lengths of 90e111 mmwere selectedto standardize mussel size and their corresponding filtration rates,and to ensure adequate tissue biomass for tritium analysis.

Damaged or unhealthy mussels (e.g., those incapable of closingtheir shells) were not selected. HTO concentrations in soft tissues ofreference mussels were less than 10 Bq/L. OBT levels were less than15 Bq/L for mussels that were frozen immediately after collection,and 45 Bq/L for mussels that were stored in lidded buckets on theCRL site over a period of three days.

Upon collection, mussels were placed into lidded, plasticbuckets containing water from the reference site to minimizeuptake of tritium prior to initiation of the study. Mussels were thentransported to the CRL site and individuals were quickly measured,weighed and numbered with cage and compartment numbers.Labelledmussels were placed into labelled nylon bags and returnedto the lidded buckets of water from the reference site until initia-tion of the transplantation into Perch Lake, which was carried outon the same day as mussel collection.

Two sets of exposure conditions were established in Perch Lake.These included exposure to tritium via the surface water pathwayonly (Cages 1 and 2), and exposure via both sediments and surfacewater (Cages 3 and 4). Cages 1 and 2 were positioned on cementblocks at a depth of approximately 0.75 m, whereas Cages 3 and 4were placed at the sediment-to-water interface at a depth ofapproximately 0.4 m. Each compartment in Cages 3 and 4 was filledwith 5 to 10 cm of sandy surface sediments originating from thearea surrounding the cages, a depth that enabled mussels to orientthemselves in an upright position with their siphons pointedupwards, as they do in natural systems. The sediments were addedto the cages several hours prior to transplantation of the mussels toallow settling of any suspended particulates.

Cages 1 and 2 were deployed in Perch Lake on 5 July 2004 at14:00 h, whereas Cages 3 and 4 were deployed on 7 July 2004 at14:00 h. Upon initiation of mussel transplantation (Time 0), indi-viduals were first transferred from the lidded buckets containingwater from the reference site to buckets containing water fromPerch Lake, such that all mussels received initial tritium exposure atapproximately the same time, despite the 10 to 15 min time periodrequired to transfer all the mussels from buckets to the numberedcage compartments.

Following transplantation, mussels were visually monitored. Ingeneral, in Cages 3 and 4 (which contained sediments), musselsbegan orienting themselves in an upright position within fiveminutes of transplantation. The mussels in Cages 1 and 2 (withoutsediments) began filtering within less than five minutes. Nomusselmortality occurred in any of the four cages over the course of the86- to 88-day study.

Whole-mussel fresh weights were measured just prior tomussel transplantation into Perch Lake, as well as following musselharvest after mussels had been frozen (Yankovich and Kim, 2005).It was not possible to measure exposed mussels prior to freezingdue to the volatility of HTO. Instead, mussels were immediatelyplaced into an airtight jar and frozen to avoid loss of tritium fromthe samples and mussels were weighed and dissected while stillfrozen.

In general, mussels did not show weight gain over the course ofthe study, as indicated by an arithmetic mean post-harvest-to-initial mussel freshweight ratio of 0.981. This lack of mussel growthwas not surprising, since the mussels used in this study were likelyto be 14e15 years old (e.g. Negus, 1966). The small weight lossesthat were noted for some individual mussels may have been due tothe fact that the mussels had been processed while they were stillfrozen (to prevent exchange of mussel free-water tritium with theatmosphere) and somewater loss may have occurred as icewas lostfrommussel tissues. This icewasmeasured for tritium as part of thesample, but was not included in the tissue weight. In addition, it ispossible that someweight loss occurred as female mussels releasedtheir eggs during reproduction.

Fig. 1. Map depicting the location of the reference site in the Ottawa River where freshwater mussels (Elliptio complanata) were collected, relative to the site of mussel trans-plantation in Perch Lake.

T.L. Yankovich et al. / Journal of Environmental Radioactivity 102 (2011) 26e3428

Perch Lake surface water temperatures were recorded contin-uously from July to October 2004 using a Model 107b CampbellScientific Inc. Temperature Probe and data logger set to integratevalues over 5 min time intervals (Yankovich and Kim, 2005). Theprobe was positioned in the vicinity of the mussel cages a fewcentimetres above the sedimentewater interface.

2.4. Sample collection and processing following transplantation

Surface water, sediment and mussel samples were collected onan expanding time step from Time 0 to the end of the study period.Three mussels were collected from each cage and composited ateach time point to generate two measurements per time point fora given set of exposure conditions (i.e., water only versus water plussediment exposure). Duplicate samples, each consisting of threemussels, were collected for the 19 h, 96 h, 14 day, 25 day/27 day, 42day/41 day and 86 day/84 day time points. Mean� Standard Errortritium activity concentrations were then calculated for a given setof exposure conditions at a given time point. Upon collection,mussels were immediately placed into airtight Mason� jars toavoid tritium exchange with the atmosphere, and the jars werefrozen until tritium analysis could be carried out. In general, it wasnecessary to composite soft tissues from 3 to 4 individual musselsto reach the biomass required for HTO and OBT analysis. The meanwater content of mussel tissue was 89% (by weight), with littlevariability among individual animals.

Both water and sediment samples were collected in triplicate ateach sampling time in the vicinity of each of the mussel cages. Indoing so, water sample bottles were opened at the depth where themussels were filtering and the samples were subsequently leftstanding for at least 4 h to allow suspended sediments to settle.Sediment samples were collected by hand at a depth of 5e10 cmand were placed in Ziplock� bags that were sealed at depth.

On 20 September 2004, three plankton samples consisting ofa mixture of zooplankton and phytoplankton were collected fromthe surface of the Perch Lake water column just offshore of themussel cages using a 0.5-m plankton tow net (Diameter:-Length¼ 1:3) with a 63-mm pore size mesh to quantify tritiumlevels in mussel dietary items. The net was towed behind a boatalong until adequate biomass had been amassed. It was not possibletomeasure OBT in individual plankton samples due to the relativelylarge biomass required for OBT analysis and the relatively largewater content present in the samples. As a result, OBT levels weremeasured in a single composite plankton sample.

2.5. Sample tritium analyses

HTO: HTO concentrations were analyzed in water, sediment,mussel and plankton samples. Surface water samples wereanalyzed for tritium in accordancewith standard liquid scintillationcounting (LSC) procedures (Workman, 2000;Workman and Brown,1992). For OBT measurements, approximately 10 g of dried sample

T.L. Yankovich et al. / Journal of Environmental Radioactivity 102 (2011) 26e34 29

was loaded into a Parr bomb (Model 1121, USA) with oxygen gaspressurized to 20 atmospheres. Following combustion, thecombustion water inside the bomb was collected directly usinga pipette after opening the bomb head in a fume hood. Toluenedistillation of the combustion water removed impurities andeliminated LSC quenching effects. The distilled water (<10 mL) wastransferred into a polyethylene vial (Packard, USA). The volume ofwater was made up to 10 mL using tritium-depleted water and wasmixed with 10 mL Ultima Gold XR (PerkineElmer, USA). The vialswere placed in a Quantulus 1220 LSC and counted for 360 minutes(with a typical background of 1 cpm) to determine the OBT activityof the sample. The minimum detectable activity was approximately10 Bq/L.

The free-water (HTO) of the mussels, plankton and sedimentsamples was extracted using a freeze-drying vacuum system. Thesamples were loaded into vacuum flasks and exposed to dry icetraps at pressures of 10�5e10�4 Torr (or 0.0013e0.013 Pa) andtemperatures of 0 to �4 �C for 24 h. Ten millilitres of the waterobtained was mixed with 10 mL of Ultima Gold LLT (PerkineElmer,USA) and HTO activity concentrations were determined directlyusing a Liquid Scintillation Counter (LSC) (Quantulus 1220) withpolyethylene vials (Packard, USA). The uncertainty in the HTOconcentrations was estimated to be 10%. The minimum detectableactivity for 10 mL of tissue-free water was approximately 2.0 Bq/L.The counting time was 360 min and the typical background was1.0 cpm for these measurements.

OBT: Mussel, sediment and plankton dry matter remaining afterthe HTO analysis was chopped and homogenized using scissors andmixed with 30e50 mL of tritium-free water to remove theexchangeable OBT. The samples were then refrozen and subjectedto a second round of cryogenic distillation under vacuum. Thisprocess was repeated at least twice, until the tritium concentrationof the rinse-water was less than 4.0 Bq/L. The rinsed samples werethen combusted using a Parr bomb (U.S.A.) with pressurizedoxygen. The sediment samples were combusted using a furnacetype combustion tube with oxygen flow. The combustion waterfrom these samples was analyzed by LSC to determine OBTconcentrations.

Counting errors for OBT concentrations were generally less than5%, but additional uncertainty arose due to difficulties in removingexchangeable OBT from the samples, as well as during thecombustion process. The total uncertainty in the OBT measure-ments was estimated to be approximately 25%.

2.6. Observed tritium concentrations

The HTO concentration in Perch Lake water decreased graduallythroughout the study period from about 4800 Bq/L at Time 0 toabout 4000 Bq/L in the last sample. The concentrations measurednear Cages 1 and 2 were consistently a few percent higher thanthose measured near Cages 3 and 4. The sediment HTO concen-trations also decreased, from about 4000 Bq/L to 3600 Bq/L over thestudy period. Again, concentrations were slightly higher at Cages 1and 2. The sediment OBT concentrations fluctuated randomly overtime from a minimum of about 570 Bq/L combustion water toa maximum of 1900 Bq/L. In this case, the concentrations wereusually higher near Cages 3 and 4. HTO levels of 4200, 4100 and4100 Bq/L were measured in the three plankton samples collected.The corresponding HTO concentrations in Perch Lake surfacewaters at the time of plankton sampling were 4100, 4100 and4000 Bq/L, respectively. The OBT level in the single compositeplankton sample was 2900 Bq/L.

The measured HTO and OBT concentrations in the mussels areshown in Table 1. The HTO concentrations increased quickly tolevels similar to those in water after just 1 or 2 h. The OBT

concentrations also increased in the first half of the study, but moreslowly, reaching about 1900 Bq/L at day 41 before dropping off toabout 1100 Bq/L of the last sample time point.

3. Scenario description

A detailed scenario was developed based on the experimentalwork described above andwas provided to the EMRAS Tritium/C-14Working Group (Yankovich and Kim, 2005 and the EMRAS web-page: http://www-ns.iaea.org/projects/emras/). Input datasummarizing initial tritium concentrations in mussel tissues andenvironmental media, temporal changes in Perch Lake watertemperatures and mussel sizes from the time of transplantation tothe time of harvest were provided to the modellers. A total of fivemodelling teams, two from Japan and one each from Romania(Galeriu et al., 2005), France and Germany (Baumgärtner, 2000,2005; Baumgärtner and Donhärl, 2004; Baumgärtner and Kim,2000), participated in the scenario (Table 2). The teams wereasked to predict temporal changes in mussel HTO and OBTconcentrations, along with the 95% confidence intervals on eachprediction, using the input data provided.

4. Description of models

Very brief descriptions of the models used in this scenario aregiven belowand in Table 2. More detailed descriptions can be foundin IAEA (2008).

Four of the five participants used dynamic compartment modelsfor their calculations, with transfers between compartments basedon published rates of tritium uptake and loss by aquatic animals. Ineach of these models, the mussels were assumed to assimilate HTOfrom the water column, plankton and sediments (for those musselsexposed to sediments); plankton and sediments were the sourcesof OBT. Food intake rates depended upon the rate of water filtrationby the mussels and the concentration of plankton in the water.These models differed with respect to the number of compart-ments, the values adopted for the rate constants, the growth ratesof the mussels and the effect of water temperature on the growth,and the characteristics of the mussels themselves.

The NIRS model assumed a filtration rate of 80 kg lake water permussel per day (Miyamoto et al., 2008). The mussels exposed toboth sediment and water ingested three times as much sediment asthe mussels exposed to water alone, with a daily sediment inges-tion rate of 2.4 g per mussel (Inui, 1998; Nakamura, 2001). Based onthe data reported by Kim et al. (2004), the concentration of free-water tritium (HTO) in the mussels was assumed to increase frombackground levels to 90% of the lake water HTO concentration;similarly, the OBT concentration was assumed to increase to 70% ofthe lake water HTO concentration. No direct consideration wasgiven to the contribution of plankton to mussel tritium.

The SRAmodel had two compartments, one for HTO and one forOBT. The rate of tritium uptake into each compartment wasdetermined using transfer coefficients that accounted for thedifference in specific activity between the environment and thebody of the mussel. Coefficients of 0.01 h�1 and 0.001 h�1 for HTOand OBT, respectively, were used in the calculations (IAEA, 2008).The rate of organic hydrogen intake from plankton was assumed tobe twice that from sediment.

The AQUATRIT model (Galeriu et al., 2005) made use of litera-ture data for the edible marine mussel Mytilus edulis to parame-terize the biological attributes of E. complanata as input into themodel (e.g., Ren et al., 2006; Sukhotin and Portner, 2001;Wong andLevinton, 2004; Zotin and Ozernyuk, 2004). The mussel OBTconcentration was estimated using a rate equation that took intoaccount the OBT concentration in food, the transfer coefficients

Table 1HTO and OBT concentrations in mussel soft tissues following transplantation.

Time After Musseltransplantation

Time (days) Cage numbersa Measured HTO(Bq/L)� standard error

Measured OBT(Bq/L)� standard error

0 h 0 Not applicable <10 451 h 0.04 Cages 1 and 2 4400� 5.8 170� 1

Cages 3 and 4 3000� 22 130� 7.52 h 0.08 Cages 1 and 2 4600� 23 150� 7

Cages 3 and 4 4400� 13 240� 124 h 0.17 Cages 1 and 2 4500� 27 180� 19

Cages 3 and 4 4500� 25 230� 307 h 0.29 Cages 1 and 2 4600� 22 160� 1

Cages 3 and 4 4400� 14 230� 1419 h 0.79 Cages 1 and 2 4500� 9.8 210� 16

Cages 3 and 4 4200� 9.2 220� 1124 h 1 Cages 1 and 2 4100� 44 230� 7

Cages 3 and 4 4200� 27 200� 9.548 h 2 Cages 1 and 2 4500� 15 880� 29

Cages 3 and 4 4200� 9.8 930� 13896 h 4 Cages 1 and 2 4100� 19 800� 89

Cages 3 and 4 4400� 17 1100� 378 days 8 Cages 1 and 2 4100� 8.1 1000� 55

Cages 3 and 4 3800� 12 1200� 27014 days 14 Cages 1 and 2 4100� 11 1200� 42

Cages 3 and 4 4000� 17 1000� 11018 days 18 Cages 1 and 2 4100� 20 1200� 23019 days 19 Cages 3 and 4 4200� 26 1300� 1225 days 25 Cages 1 and 2 4200� 31 1200� 10027 days 27 Cages 3 and 4 3900� 18 1500� 8036 days 36 Cages 1 and 2 4200� 13 1700� 19035 days 35 Cages 3 and 4 4000� 13 1800� 5442 days 42 Cages 1 and 2 3900� 19 1800� 13041 days 41 Cages 3 and 4 3900� 13 1900� 9186 days 86 Cages 1 and 2 3800� 21 1200� 13084 days 84 Cages 3 and 4 3700� 16 1000� 69

a Tissues frommussels from Cages 1 and 2, which were only exposed to tritium via the water pathway, were pooled, as were those frommussels from Cages 3 and 4, whichwere exposed via both water and sediments. Three mussels were collected from each cage and composited at each time point to generate two measurements per time pointfor a given set of exposure conditions (i.e., water only versus water plus sediment exposure). In addition, duplicate samples, each consisting of threemussels, were collected forthe 19 h, 96 h, 14 day, 25 day/27 day, 42 day/41 day and 86 day/84 day time points. Tritium activity concentrations represent the mean� standard error for a given set ofexposure conditions at a given time point.

T.L. Yankovich et al. / Journal of Environmental Radioactivity 102 (2011) 26e3430

from HTO in water and OBT in food to OBT in the animal, and theloss rate from the animal. The mussels were assumed to assimilatematerial from bacteria, plankton, detritus and dissolved organicmatter, depending on availability. Calculations were madeassuming that all food items were (i) uncontaminated and (ii) hadtritium levels that were at steady state with Perch Lake HTOconcentrations. The transfer coefficients reflected the metabolicregulation of tritium, and had quite low values due to the age of themussels. The loss rate was assumed to correlate with food avail-ability, as well as with the tritium concentration in the food.

Since fish is the only freshwater animal considered in OURSON,the fish compartment in this model was adapted to estimatetritium concentrations in mussels. The mussel HTO compartment

Table 2Model numbers and modellers participating in the mussel uptake scenario.

Model no. Model name Lead modeller Affiliation Coun

1 NIRS Kiriko Miyamoto National Institute of RadiologicalSciences (NIRS)

Japa

2 SRA Masahiro Saito Kyoto University SafetyReassurance Academy (SRA)

Japa

3 AQUATRIT Dan Galeriu andAnca Melintescu

”Horia Hulubei” National Institutefor Physics and NuclearEngineering (NIPNE)

Rom

4 OURSON Françoise Siclet Electricité de France (EDF) Fran

5 BIOCHEM Franz Baumgärtner Technical University Munich,nstitute of Radiochemistry

Germ

was assumed to reach equilibrium with the surrounding waterwithin 1 h of transplantation due to the high initial filtration rate,which decreased at later times. The animals gradually incorporatedOBT from plankton at a rate that was proportional to the feedingrate. The OBT concentration in the plankton was assumed to be atequilibrium with HTO levels in the water. A second OBT compart-ment, comprised of the food particles inside the mussel, wasincluded to address an initial under-prediction of OBT concentra-tions shortly after transplantation. 30% of the mussel soft tissueweight was assumed to bemade up of food particles. Turnover ratesof 0.02 day�1 were assumed for the soft tissue OBT compartment(Tremblay et al., 1998) and 0.33 day�1 for the food particle OBTcompartment, assuming a filtration rate of 38 L day�1 and

try Type of model Tritiumcompartments

Does modelaccount forwatertemperature?

Uptake pathways

n Dynamic HTO, 2 OBTcompartments

No Water, sediment

n Dynamic HTO, OBT No Water, sediment,plankton

ania Dynamic HTO, OBT Yes water, sediment,plankton

ce Dynamic HTO, OBTsoft tissue,OBTfood particles

No Water, sediment,plankton

any Steady state HTO, XBT, YBT,CBT (assumed negligible)

Yes Water, plankton

Fig. 2. Modelled-to-measured HTO concentrations in soft tissues of transplantedmussels.

T.L. Yankovich et al. / Journal of Environmental Radioactivity 102 (2011) 26e34 31

a suspended particle concentration in water of 10 mg L�1. Themussels exposed via water only were modelled in the same way asmussels exposed to both water and sediment.

The BIOCHEM model was the only steady-state model used inthe scenario. In BIOCHEM, penetration of HTO from the water intothe mussel cells was assumed to occur by diffusion over a few dayson a linear time scale. OBT was assumed to be made up of threecomponents: carbon bound (non-exchangeable) OBT (CBT);hydrate-bound tritium (YBT); and tritium bound to oxygen,nitrogen or sulphur atoms (XBT). The tritium nuclei of YBT andXBT were assumed to be “buried” in the sense that they did notexchange with hydrogen nuclei in water during the rinsing processof OBT analysis because they were buried inside biopolymers andinaccessible (Baumgärtner and Donhärl, 2004). Factors describingthe distribution of tritium among the three OBT compartmentswere estimated based on theoretical information and mathemat-ical constants reported by Griffiths et al. (1993), Saenger (1987),Klapper (1977) and Baumgärtner (2005). The formation of YBTand XBT in the mussels was assumed to proceed spontaneouslywith the ingress of Perch Lake water into mussel cells. The outputof the model was the concentration of buried tritium (YBTþXBT),assuming that mussel soft tissue consisted of proteins and DNAbut not carbohydrates, and that CBT was negligible.

5. Results: modelled-to-measured comparisons

5.1. Tritium concentrations

HTO: Both the NIRS and the SRAmodels under-predictedmusselHTO concentrations over the initial 8 days of the study period, afterwhich both models over-predicted slightly for all mussel cages(Fig. 2). By comparison, mussel HTO concentrations predicted byAQUATRIT were very close to measured values at all samplingpoints, with modelled-to-measured ratios that ranged from 0.7 to1.2 (for Cages 1 and 2) and from 0.8 to 1.2 (for Cages 3 and 4). Withthe exception of the one-hour time point for Cages 3 and 4 (forwhich predicted values were approximately 1.5-fold higher thanmeasured values), the OURSON and BIOCHEMmodels showed goodpredictive power throughout the study. All models predicted thatthemussel HTO concentrationwas similar formussels that receivedtritium exposure fromwater only (in Cages 1 and 2) and fromwaterand sediments (in Cages 3 and 4).

OBT: The NIRS model under-estimated mussel OBT concentra-tions by up to 6-fold over the first three sampling points, then over-estimatedmeasured values by 2- to 7-fold over the rest of the studyperiod (Fig. 3). The SRA model also initially under-estimated OBTfor the first 18 days for Cages 1 and 2, and for the first 14 days forCages 3 and 4, after which predicted values fell close to measuredvalues.

OBT concentrations in mussels receiving tritium exposure viawater only were under-estimated by AQUATRIT over the entirestudy period by factors of 210 after 1 h to 1.3 after 86 days (Fig. 3).This was because the predictions were initially made assuming themussels contained no background OBT at the start of the experi-ment. When the predictions were corrected to account for the OBTbackground in the mussels at the start of experiment, the predic-tions fell close to the observations. In Fig. 4, these twocases are distinguished as AQUATRIT (uncorrected) and AQUATRIT(corrected). The modelled-to-measured ratio for the correctedpredictions ranged from 0.3 to 1.06 (excluding the 86-day timepoint).

OBT concentrations were under-estimated by OURSON byapproximately 2- to 3-fold during the first seven hours followingtransplantation, after which the predicted values fell within a factorof 2 of the measured values for all time points except the last.

Over the first day, BIOCHEM under-estimated mussel OBTconcentrations by factors of approximately 1.5e2 (Fig. 3), afterwhichmodelled-to-measured OBT ratios of 0.7 to 1.1 were predicted.

Most of the models over-predicted the final OBT concentrationregardless of whether the mussels received tritium exposure fromwater only (in Cages 1 and 2) or fromwater and sediments (in Cages3 and 4) (Fig. 3). The drop in the modelled-to-measured ratio for allmodels except BIOCHEM at 48 h is due to the relatively high OBTconcentration that was measured at this time (Table 1).

5.2. Tritium dynamics

In evaluating the performance of the models, it is important toappraise not only any similarities and differences between pre-dicted and observed concentrations at a given time, but also thedynamics of HTO and OBT build-up in the mussels.

HTO: The experimental data showed that the HTO concentra-tions in the mussels quickly reached steady state with theconcentrations in lake water. All models reproduced this result, butwith the exception of BIOCHEM and OURSON, over-estimated thetime required to achieve equilibrium by intervals ranging from onehour to a few days.

The predicted dynamics of the HTO concentrations were similarfor the NIRS and SRA models. AQUATRIT, OURSON and BIOCHEMalso showed similar dynamics, but with a pattern that differed fromthat of NIRS and SRA (Fig. 2).

OBT: With the exception of the BIOCHEMmodel (for both sets ofcages) and the AQUATRIT (corrected) model, all the models under-

Predicted Mussel HTO

Modelled-to-Measured HTO in Cage 1 and 2 Mussels

(Water Only Exposure)

sl

es

su

M4

dn

a3

eg

aC

ni

OT

Hd

er

us

ae

M-

ot

-d

el

le

do

M

)e

ru

so

px

Er

et

aW

dn

at

ne

mi

de

S(

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

NIRS

SRA

AQUATRIT

OURSON

BIOCHEM

Predicted Mussel OBT

Modelled-to-Measured OBT in Cage 1 and 2 Mussels

(Water Only Exposure)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

0.001 0.01 0.1 1 10 100

sl

es

su

M4

dn

a3

eg

aC

ni

TB

Od

er

us

ae

M-

ot

-d

el

le

do

M

)e

ru

so

px

Er

et

aW

dn

at

ne

mi

de

S(

0.001

0.01

0.1

1

10

100

NIRS

SRA

AQUATRIT (uncorrected)

AQUATRIT (corrected)

OURSON

BIOCHEM

Fig. 4. Comparison of predictions for mussels in Cages 1 and 2 relative to those inCages 3 and 4. Cage 1 and 2 mussels received tritium exposure from the water pathwayonly. Cage 3 and 4 mussels received tritium exposure from both the sediment and thewater pathways. Uncorrected values predicted by AQUATRIT did not account forelevated tritium concentrations in mussel dietary items in Perch Lake, whereas cor-rected values assumed that mussels were feeding on food with tritium levels that wereat steady state with HTO in Perch Lake water.

Fig. 3. Modelled-to-measured OBT concentrations in soft tissues of transplantedmussels. Uncorrected values predicted by AQUATRIT did not account for elevatedtritium concentrations in mussel dietary items in Perch Lake, whereas corrected valuesassumed that mussels were feeding on food with tritium levels that were at steadystate with HTO in Perch Lake water.

T.L. Yankovich et al. / Journal of Environmental Radioactivity 102 (2011) 26e3432

predicted the OBT concentration in the mussels one hour aftertransplantation, often by large amounts, indicating that OBT buildsup much more quickly than expected in the mussels (Fig. 3). Incontrast, most of the models over-predicted the rate of OBTformation over the next 24 h. The OBT dynamics beyond 24 h werenot well modelled, although all of the participants predicted OBTconcentrations that werewithin a factor of three of the observationat the final sampling point.

The OBT dynamics predicted by most models were generallysimilar, with only those of BIOCHEM clearly different. In addition,all models, except BIOCHEM, slightly over-predicted the OBT levelsat the final 88-day experimental time point (Fig. 3), perhaps due tothe loss of OBT by female mussels during egg production andrelease (which was accounted for by the BIOCHEM model)(Baumgärtner et al., 2009).

5.3. Exposure pathways

All models predicted similar HTO concentrations for musselsexposed to water alone and to water plus sediment (Fig. 4). Simi-larly, all models except AQUATRIT predicted similar OBT concen-trations for the two exposure pathways. In the case of AQUATRIT,slightly higher OBT concentrations were predicted for musselsexposed to sediments and water than for mussels exposed to wateralone, indicating that this model assumes some OBT contributionfrom the sediments. The higher OBT concentrations that were

predicted in the AQUATRIT (corrected) model compared to thosepredicted in the AQUATRIT (uncorrected) model demonstrate theimportance of OBT background in mussels in prediction of OBT inmussel soft tissues.

6. Discussion

Themodel predictionswere consistent in some respects, leadingto under-estimates or over-estimates of the measured tritiumconcentrations. Such mis-predictions were evaluated to determinewhether they could be attributed to similarities in model structureor parameter values.

6.1. Under-estimation of initial mussel tritium uptake rates

For all models, the rate of OBT formation and, to a lesser extent,HTO uptake between the 1-h and 2-h time points were predicted tooccur relatively more slowly in mussels that received tritium

T.L. Yankovich et al. / Journal of Environmental Radioactivity 102 (2011) 26e34 33

exposure via both water and sediments (in Cages 3 and 4) than inmussels exposed to water only (in Cages 1 and 2) (Fig. 3). This ispossibly due to differences in assumed mussel filtration rates orfactors that potentially affect mussel filtration rates, such as musselorientation, particle quality or density of suspended particulates.

For HTO, the models can be readily optimised by increasing therate at which HTO is transferred from water to the mussels. In thecase of OBT, however, the observed trends suggest the possibility ofslow and fast OBTcompartments corresponding to differing rates ofOBT dynamics. For example, OBT concentrations were better pre-dicted bymodels that distinguished such compartments within themussels (e.g., NIRS, OURSON, BIOCHEM). It is interesting to notethat recent human dosimetry models for OBT also include fast andslow OBT compartments (e.g., Etnier et al., 1984; Hill and Johnson,1993; McKone et al., 1997). Similarly, fast and slow compartmentshave been reported in other types of mammals (Galeriu et al.,2009).

In addition, it has been suggested by Baumgärtner et al. (2009)that the initial rate of OBT increase in mussels is relatively rapid,with the prompt formation of OBT due to surface effects. Relativelyslower OBT formation mechanisms related to diffusion and diges-tion are expected to become more prominent at later times. Suchinformation may be useful in the development of biokinetic dosemodels for tritium in freshwater mussels and can also be used toestablish the relative contributions of HTO versus OBT to biota dose.

Further exercises with AQUATRIT have shown that improve-ments in model predictions can be obtained by assuming the OBTconcentration of the the new organic matter in the mussel was thesame as for the food consumed (which was assumed equal to theHTO concentration in the lake water reduced by a discriminationfactor) and by doubling the biological half-time.

6.2. Over-estimation of OBT concentrations in mussels at the finaltime point

With the exception of BIOCHEM and Cage 1 and 2 mussels forAQUATRIT, all models over-estimated mussel OBT concentrations atthe final experimental time point (Fig. 3). These over-estimationswere likely to be related to unexpectedly low measured OBT levelsin harvested mussels at this time compared to previous time points(Yankovich et al., 2006), whichmay be attributed tomussel biology.Based on the shell lengths used in this study, it is likely that themussels were more than 14 years old (Negus, 1966). Since unionidmussels, such as E. complanata, typically reach sexual maturitybetween 6 and 12 years of age, it is likely that the transplantedmussels were sexually mature and expended a relatively largeproportion of their energy towards reproduction, as opposed togrowth of somatic tissues.

Mussel embryos are retained inside the females for a shortperiod during their early stages of development, which is a period ofrapid growth. It is possible that the growing embryos incorporatedOBT at a faster rate than other tissues. Moreover, mussel larvae arereleased by females into the water column, with a correspondingloss of OBT, between mid-August (when the second last musselsample was taken) and early October (when the final sample wascollected). This interpretation is supported by the fact that, in somecases, mussel fresh weights declined slightly over the course of thestudy, suggesting that female mussels may have lost weight withrelease of larvae into the water column. The one model (BIOCHEM)that took account of these processes tended to produce the best OBTpredictions, both over time and at the final sampling point. Thereproductive tissue compartment in mussels could be comparableto the fast-fat OBT compartment that has been included in humanbiodosimetry models for tritium (Etnier et al., 1984).

6.3. Other reasons for mis-predictions

A number of factors could account for some of the otherdifferences between modelled and measured HTO and OBTconcentrations: the assumed HTO transfer rate from water intomussel tissue (which could, in turn, be influenced by physicalfactors such as diffusion rates and/or concentration gradients, aswell as biological properties such as mussel filtration rates); theassumed OBT formation rates; the assumed contribution of the twoexposure pathways to the HTO and OBT inventories in musseltissues; and/or assumptions with respect to tritium speciation inkey biological compartments of the body. Such factors were notcaptured by all models, and in cases where similar processes wereassumed to occur, the numerical values of the relevant transferparameters were not always the same among models.

The uptake and formation of OBT in aquatic organisms is rela-tively complex. OBT can be taken in through direct dietary uptakeor formed in the body by various metabolic processes followingHTO uptake. Furthermore, new OBT species (e.g., buried tritium andhydrate-bound tritium) have recently been proposed, and there issome evidence that OBT uptake/formation rates are much higherimmediately after exposure than at later times (Baumgärtner, 2000,2005; Baumgärtner and Donhärl, 2004; Baumgärtner and Kim,2000; Baumgärtner et al., 2009). These factors should be consid-ered in predictions of tritium doses to humans and biota, but withthe exception of BIOCHEM, were not taken into account in themodels participating in this study.

7. Future work

To further improve the understanding and modelling of tritium(especially OBT) in abiotic and biotic environments, more scenariosand accurate datasets are required over longer time periods. Futurework could focus on characterization of key parameter values, suchas the biological half-life of OBT, OBT formation rates over varioustime scales and the influence of exposure pathway on OBT uptakeand formation. Furthermore, although the modelled-to-meanmeasured tritium values was relatively low, uncertainty andsensitivity analyses are required to gain additional insights into thekey parameters that should be included in OBT models.

Such information could lead to the development biokineticmodels to describe tritium dynamics in freshwater mussels. Addi-tional work could be done to assess age-structured tritiumdynamics to assess HTO uptake and OBT formation in musselpopulations, which can form the basis of population models topredict the propensity of different life stages to incorporate tritiumat different rates.

Acknowledgements

The authors would like to thank Dr. Ian Swainson for his thor-ough review of this paper, as well as CANDU Owners Group (COG)and Atomic Energy of Canada Limited (AECL) for their support ofthis work.

References

Baumgärtner, F., 2000. Theoretical foundation and experimental proof of theaccumulating transfer of tritium from water into DNA and other biomoleculesin vitro and in vivo. Radiation Biology and Radioecology 40 (5), 495e499.

Baumgärtner, F., 2005. Accumulative tritium transfer from water into biosystems.Fusion Science and Technology 48 (1), 787e790.

Baumgärtner, F., Donhärl, W., 2004. Non-exchangeable organically-bound tritium(OBT): its real nature. Analytical and Bioanalytical Chemistry 379, 204e209.

Baumgärtner, F., Kim, M.-A., 2000. Tritium/protium fractionation near and insideDNA. Journal of Radioanalytical & Nuclear Chemistry 243, 95e98.

T.L. Yankovich et al. / Journal of Environmental Radioactivity 102 (2011) 26e3434

Baumgärtner, F., Yankovich, T.L., Kim, S.B., 2009. Logistic analysis of OBT dynamics.Health Physics 97 (6), 572e580.

Bayne, B.L., Brown, D.A., Burns, K., Dixon, D.R., Ivanovici, A., Livingstone, D.R.,Lowe, D.M., Moore, M.N., Stebbing, A.R., Widdows, J., 1985. The Effects of Stressand Pollution on Marine Animals. Praeger, New York, 381 pp.

Clarke, A.H., 1981. Elliptio complanata (Lightfoot, 1786) Eastern Elliptio. The Fresh-water Molluscs of Canada. National Museum of Natural Sciences, Ottawa.266e269.

Curry, C.A., 1977/78. The freshwater clam (Elliptio complanata). A practical tool formonitoring water quality. Water Pollution Research Canada 13, 45e52.

Dechno, A.W., Luoma, S.N., 1992. Geopolymers (humic and fulvic acids): sink ortrophic transfer-vehicle for metals to bivalves. In: Proceedings of the 13thAnnual Meeting of the Society of Environmental Toxicology and Chemistry.Cincinnati, Ohio.

Elder, J.F., Collins, J.J., 1991. Freshwater molluscs as indicators of bioavailability andtoxicity of metals in surface-water systems. Reviews of EnvironmentalContamination and Toxicology 122, 37e79.

Etnier, E.L., Travis, C.C., Hetrick, D.M., 1984. Metabolism of organically-boundtritium in man. Radiation Research 100, 487e502.

Galeriu, D., Heling, R., Melintescu, A., 2005. The dynamics of tritium e including OBTe in the aquatic food chain. Fusion Science and Technology 48 (1), 779e782.

Galeriu, D., Melintescu, A., Beresford, N.A., Takeda, H., Crout, N.M.J., 2009. Thedynamic transfer of 3H and 14C in mammals e a proposed generic model.Radiation and Environmental Biophysics 48, 29e45.

Greig, R.A., Nelson, B.A., Nelson, D.A., 1975. Trace metal content in the Americanoyster. Marine Pollution Bulletin 6 (5), 72e73.

Griffiths, A.J.F., Miller, J.H., Suzuki, D.T., Lewontin, R.C., Gelbart, W.M., 1993. AnIntroduction to Genetic Analysis, fifth ed. W.H. Freeman and Comp, New York,pp. 308.

Hill, R.L., Johnson, J.R., 1993. Metabolism and dosimetry of tritium. Health Physics65, 628e647.

Hinch, S.G., Stephenson, L.A., 1987. Size- and age-specific patterns of trace metalconcentrations in freshwater clams from an acid-sensitive and a circumneutrallake. Canadian Journal of Zoology 65, 2436e2442.

IAEA, 2008. Inter-model comparison of tritium concentrations in freshwater Barnesmussels (Elliptio complanata) following an abrupt increase in ambient tritiumexposure conditions. Final report. August 2008.

Inui, M., 1998. Fisheries and environmentdenvironmental conservation for fishery,no.2. Elimination of suspended matters by filtering feeding animals. FisheriesAdvancement (in Japanese) 32 (9), 24e26.

Jackim, E., Morrison, G., Steele, R., 1977. Effects of environmental factors on radio-cadmiumuptake by four species of marine bivalves. Marine Biology 40, 303e308.

Kauss, P.B., Griffiths, M., Melkic, A., December 1981. Use of freshwater clams inmonitoring trace contaminant source areas. In: Proceedings of the 2ndOntario Ministry of Environment, Technology Transfer Conference, Toronto,Ontario, pp. 371e388.

Kauss, P.B., Hamdy, Y.S., 1985. Biological monitoring of organochlorine contami-nants in the St. Clair and Detroit Rivers using introduced clams, Elliptio com-planata. Journal of Great Lakes Research 11 (3), 247e263.

Kim, S.B., Workman, W.J.G., Davis, P.A., Yankovich, T.L., July 2004. Tritiumconcentrations in the Perch Lake aquatic ecosystem. AECL Technical Report,CTD-03700-ENA-003 (Rev. 0).

Klapper, M.H., 1977. The independent distribution of amino acid near neighbor pairsinto polypeptides. Biochemical and Biophysical Research Communications 78,1018e1024.

Lobel, P.B., Mogie, P., Wright, D.A., Wu, B.L., 1982. Metal accumulation in fourmolluscs. Marine Pollution Bulletin 13 (5), 170e174.

Lobel, P.B., Wright, D.A., 1982. Total body zinc concentration and allometric growthratios in Mytilus edulis collected from different shore levels. Marine Biology 66,231e236.

Luten, J., Bouquet, W., Burggraaf, M., Rus, J., 1986. Accumulation, elimination andspeciation of Cd and Zn in mussels: Mytilus edulis in the natural environment.Bulletin of Environmental Contamination and Toxicology 37, 579e586.

Matteson, M.R., 1948. Life history of Elliptio complanatus (Dillwyn, 1817). AmericanMidland Naturalist 40 (3), 690e723.

McKone, T.E., Brand, K.P., Shan, C., April 1997. Environmental health-risk assessmentfor tritium releases at the National Tritium Labeling Facility at Lawrence Ber-keley National Laboratory. LBL-37760/UC-2000.

Miyamoto, K., Inoue, Y., Takeda, H., Yanagisawa, K., Fuma, S., Ishii, N., Kuroda, N.,Yankovich, T.L., Kim, S.B., Davis, P., 2008. Development and validation ofa model for tritium accumulation by a freshwater bivalve using the IAEARMRAS scenarios. Fusion Science and Technology 54, 265e268.

Nakamura, M., 2001. Environment and fisheries in the ocean, no.3. Principle ofbiological activity. Fisheries Advancement (in Japanese) 35 (3), 52e55.

Negus, C., 1966. A quantitative study of growth and production of unionid musselsin the River Thames at Reading. Journal of Animal Science 35, 513e532.

Phillips, D.J.H., 1979. Trace metals in the common mussel Mytilus edulis and in thealga Fucus vesiculosus from the region of the sound (Oresund). EnvironmentalPollution (Series A) 18, 31e43.

Ren, J.S., Ross, A.H., Hayden, B.J., 2006. Comparison of assimilation efficiency ondiets of nine phytoplankton species of the greenshell mussel, Perna canaliculus.Journal of Shellfish Research 25 (3), 887e892.

Rubenstein, N.I., Lores, E., Gregory, N.R., 1983. Accumulation of PCBs, mercury andcadmium by Nereis virens, Mercenaria mercenaria and Palaemonetes pugio fromcontaminated harbor sediments. Aquatic Toxicology 3 (3), 249e260.

Saenger, W., 1987. Structure and dynamics of water surrounding biomolecules.Annual Review of Biophysics and Biophysical Chemistry 16, 93e114.

Servos, M.R., Malley, D.F., Mackie, G.L., LaZerte, B.D., 1987. Lack of bioaccumulationof metals by Elliptio complanata (Bivalvia) during snowmelt in three South-Central Ontario streams. Bulletin of Environmental Contamination and Toxi-cology 38, 762e768.

Sukhotin, A.A., Portner, H.-O., 2001. Age-dependence of metabolism in musselsMytilus edulis (L.) from the White Sea. Journal of Experimental Marine Biologyand Ecology 257 (1), 53e72.

Tatum, H.E., 1986. Bioaccumulation of polychlorinated biphenyls and metals fromcontaminated sediment by freshwater prawns, Macrobrachium rosenbergii andclams, Corbicula fluminea. Archives of Environmental Contamination and Toxi-cology 15 (2), 171e183.

Tessier, A., Campbell, P.G.C., Auclair, J.C., Bisson, M., 1984. Relationships between thepartitioning of trace metals in sediments and their accumulation in the tissuesof the freshwater mollusc Elliptio complanata in a mining area. Canadian Journalof Fisheries and Aquatic Sciences 41, 1463e1472.

Tessier, A., Couillard, Y., Campbell, P.G.C., Auclair, J.C., 1993. Modeling Cd partitioningin oxic lake sediments and Cd concentrations in the freshwater bivalve Ano-donta grandis. Limnology and Oceanography 38 (1), 1e17.

Tremblay, R., Myrand, B., Sevigny, J.M., Guderley, H., 1998. Bioenergetic and geneticparameters in relation to susceptibility of blue mussels, Mytilus edulis (L) tosummer mortality. Journal of Experimental Marine Biology and Ecology 221,27e58.

Wong, W.H., Levinton, J.S., 2004. Culture of the blue musselMytilus edulis (Linnaeus,1758) fed both phytoplankton and zooplankton: a microcosm experiment.Aquaculture Research 35 (10), 965e969.

Workman, W.G.J., 2000. Procedure for the Analysis of Tritium on the QuantulusLiquid Scintillation Counter, ETB-ERM-601, Rev. 1.0.

Workman, W.J.G., Brown, R.M., 1992 May. Methods for the measurement of tritiumin environmental materials.

Yankovich, T.L., Kim, S.B., 2005. EMRAS Tritium/C14 Working Group: the dynamicPerch Lake mussel transplantation scenario description (revision 1). AECLEnvironmental Technologies Branch Memorandum No. ETB-05-99 (Revision 1),2005 July 27.

Yankovich, T.L., Killey, R.W.D., Klukas, M.H., Cornett, R.J.J., Zach, R., LaFontaine, C.R.,O’Donnell, B.C., Eve, T.L., Chaput, T.J., Benz, M.L., Haas, M.K., 2000. The impor-tance of environmental monitoring data in Environmental Risk Assessment: anecosystem approach. In: Proceedings from the Canadian Nuclear SocietyMeeting, Toronto, Ontario, 20 pp.

Yankovich, T.L., Koarashi, J., Kim, S.B., Davis, P.A., 2008. International study on thevalidation of models for the environmental transfer of tritium and carbon-14.Applied Radiation and Isotopes 66, 1721e1729.

Yankovich, T.L., Kim, S.B., Baumgärtner, F., Galeriu, D., Melintescu, A., Miyamoto, K.,Saito, M., Siclet, F., Davis, P., December 2006. Intermodel comparison of tritiumConcentrations in freshwater Barnes mussels (Elliptio complanata) following anabrupt increase in ambient tritium exposure conditions. AECL Technical ReportNo. 153-121240-440-007.

Zotin, A.A., Ozernyuk, N.D., 2004. Age-related changes in oxygen consumption inthe edible mussel Mytilus edulis from the White Sea. Biology Bulletin of theRussian Academy of Sciences 31 (5), 465e468.