application of a 3-d chemical fate prediction model (fate3d) to predict dioxin concentrations in the...

Estuarine, Coastal and Shelf Science 70 (2006) 621e632www.elsevier.com/locate/ecss

Application of a 3-D chemical fate prediction model (FATE3D)to predict dioxin concentrations in the Tokyo Bay

Norihiro Kobayashi a,*, Tomomi Eriguchi b, Kisaburo Nakata c,Shigeki Masunaga d, Fumio Horiguchi a, Junko Nakanishi a

a Research Center for Chemical Risk Management (CRM), National Institute of Advanced Industrial Science and Technology (AIST),

16-1 Onogawa, Tsukuba, Ibaraki 305-8569, Japanb Environmental Information Department, Chuden CTI Co., Ltd., 1-27-2 Meieki-Minami, Nakamura-ku, Nagoya 450-0053, Japan

c Faculty of Marine Science and Technology, Tokai University, 3-20-1 Orido, Shimizu, Shizuoka 424-8610, Japand Graduate School of Environment and Information Sciences, Yokohama National University, 79-7 Tokiwadai,

Hodogaya-ku, Yokohama, Kanagawa 240-8501, Japan

Received 1 June 2006; accepted 1 June 2006

Available online 26 July 2006

Abstract

A 3-D chemical fate prediction model (FATE3D) was applied to predict the dioxin concentrations in the seawater of Tokyo Bay, Japan. Thesimulations were carried out for a period of one year (from September 2002 to August 2003). Parameters such as meteorological data, flow fieldconditions, concentrations and sinking rates of organic particulate matter, initial and boundary conditions, and loading fluxes and physico-chemical properties of dioxins were used as the model inputs.

The simulation results compared favorably with the field measurements of dioxin concentrations in the bay for both the particulate and dis-solved phases, indicating the validity and predictive capability of the model. Furthermore, the differences in the seasonal cycles and distributionsbetween the particulate- and dissolved-phase dioxins in the bay were estimated from the simulation results.

However, the particulate-phase dioxin concentrations in the bottom layers (þ1 m from the bottom) were underestimated, probably becausethe resuspension process was not taken into account in the model. The improvement of the model’s predictive capability, including the resus-pension process, shall be the focus of our next study.� 2006 Elsevier Ltd. All rights reserved.

Keywords: dioxin; PCDD; PCDF; dioxin-like PCB; chemical fate; Tokyo Bay

1. Introduction

According to the Japanese Ministry of Health, Labour andWelfare (1999e2003), the consumption of fish and shellfish inJapan is the primary pathway for human exposure to dioxins;it accounts for approximately 70% of the total dioxinexposure. Thus, an investigation on the dioxin levels in anaquatic environment holds significant importance. However,

* Corresponding author.

E-mail address: [email protected] (N. Kobayashi).

0272-7714/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ecss.2006.06.013

monitoring data in terms of dioxin concentrations are verylimited, particularly in aquatic environments, such as rivers,lakes, and estuaries. The two reasons for this are e the dioxinanalysis involves a considerable amount of time, effort, andcost, and the dioxin concentrations in an aquatic environmentare usually extremely low (in pg L�1 or fg L�1), and conse-quently are very difficult to detect.

Furthermore, in order to estimate the human health and theecological risks posed by dioxins present in an aquatic envi-ronment, it is necessary to be aware of the average dioxinconcentration over a long period of time. However, this infor-mation is very difficult to obtain by field measurements alone.

mailto:[email protected]

http://www.elsevier.com/locate/ecss

622 N. Kobayashi et al. / Estuarine, Coastal and Shelf Science 70 (2006) 621e632

Therefore, mathematical models that are able to predict dioxinconcentrations in an aquatic environment will be very usefulfor human health and ecological risk assessments.

A 3-D chemical fate prediction model for an aquaticenvironment (FATE3D) has been developed by the ResearchCenter for Chemical Risk Management (CRM), National Insti-tute of Advanced Industrial Science and Technology (AIST),Japan. The FATE3D model can take into consideration thethree compartments of chemicals in the aquatic environment(particulate-phase chemicals, dissolved-phase chemicals, andchemicals in the sediment), and can simulate various transpor-tation processes (e.g., diffusion and sinking) that involve targetchemicals from their sources. In this study, the FATE3D modelwas applied to predict the dioxin concentrations in the seawa-ter of the Tokyo Bay, Japan.

Tokyo Bay (Fig. 1) is located southeast of the TokyoMetropolis, with its mouth opening to the Pacific Ocean inthe south. Its surface and catchment area are approximately1200 km2 and 7000 km2, respectively (Kaizuka et al., 1993).The seafloor gradient from north to south is generally fromshallow to deep, with an average depth of 15 m and a maxi-mum depth of approximately 50 m (Kaizuka et al., 1993).Six large rivers (Edogawa, Nakagawa, Arakawa, Sumidagawa,Tamagawa, and Tsurumi Rivers) flow into the Tokyo Bay. Theamount of water inflow from these six rivers to the bay ac-counts for more than 70% (7.4 � 109 m3 year�1) of the totalamount of the fresh water inflow (1.0 � 1010 m3 year�1)Kaizuka et al., 1993).

The Tokyo Bay basin is one of the most advanced industri-alized areas in Japan. The catchment area is densely inhabitedwith a wide variety of ongoing municipal, agricultural, and in-dustrial activities. All these activities are probable contributorsas emission sources of dioxins to the bay. Simultaneously,a considerable amount of fishery activity (more than2 � 105 t year�1) is going on in the bay (Shimizu, 1997).Thus, ecological risk assessments of dioxins in the TokyoBay seawater, in particular, need to be conducted.

In order to validate the simulation results, field measure-ments of dioxin concentrations in the Tokyo Bay seawater

and the sediment are essential. The dioxin concentrations inthe seawater of the Tokyo Bay were reported by Kobayashiet al. (2003a). That study measured the dioxin concentrationsin both the particulate and dissolved phases at three stations(Sta. A, B, and C) in the bay at three times (December 12,2002, March 24 and July 24, 2003). Water samples for eachlocation were collected from both the surface layers (�0.5 mfrom the surface) and the bottom layers (Sta. A: þ1 m fromthe bottom, Sta. B and C: þ10 m from the bottom). The loca-tions of these sampling stations are also shown in Fig. 1.

In this paper, the simulation results were validated by com-paring them with the concentrations reported by Kobayashiet al. (2003a). Further, the seasonal variations and distributionsof dioxin concentrations in the bay were discussed based onthe simulation results.

2. General description of the model

A schematic view of the FATE3D model is shown in Fig. 2.The following five processes were taken into consideration inFATE3D to predict dioxin concentrations in the seawater andthe sediment of the Tokyo Bay.

2.1. Loading fluxes of dioxins

The loading fluxes of dioxins from the rivers and air arebelieved to constitute a major portion of dioxin loadings thatenter the Tokyo Bay (Kobayashi et al., 2003a, 2004). Hence,both these loading fluxes were taken into account in theFATE3D model.

Furthermore, the loading fluxes of dioxins in both the par-ticulate and dissolved phases were considered in this study.

2.2. Partitioning between particulate- and dissolved-phase dioxins

Dioxins in the seawater of the Tokyo Bay exist in theparticulate or dissolved phase. Their partitioning between

Fig. 1. Location of the Tokyo Bay.

623N. Kobayashi et al. / Estuarine, Coastal and Shelf Science 70 (2006) 621e632

the particulate and dissolved phases was assumed to be deter-mined on the basis of the adsorption equilibrium.

Both organic and inorganic particulate matter are taken intoaccount in the FATE3D model. However, the dioxins were notassumed to be adsorbed by the inorganic particulate matter inthis study.

Furthermore, two types of organic particulate matter (phy-toplankton and detritus) are considered in this study, since or-ganic particulate matter mainly comprises phytoplankton anddetritus and their behaviors in the bay are somewhat differentbecause of the difference in their properties (e.g., sinking rate).

2.3. Horizontal and vertical transportation of dioxins

The particulate- and dissolved-phase dioxins existing in thebay are transported horizontally and vertically within or be-yond the bay by advection and diffusion. It was assumedthat vertical transportation of particulate-phase dioxins wasnot only affected by advection and diffusion but also by thesinking of organic particulate matter, and the sinking ratesof particulate-phase dioxins were determined by that of theorganic particulate matter.

2.4. Interaction between seawater and sediment

The FATE3D model takes into account the sinking processof particulate-phase dioxins with organic particulate matter tothe sediment. Other processes involving interactions betweenseawater and sediment, such as the resuspension of the sedi-ment and the adsorption (or desorption) between the dioxinsin the dissolved phase and those in the sediment are regardedas very important processes; however, these processes havenot been included in the present version of FATE3D.

2.5. Degradation of dioxins

The degradation processes of dioxins in the particulate, dis-solved, and sediment phases are taken into account in theFATE3D model. However, on weighing the half-lives of dioxincongeners in environmental media, it was concluded that the

Fig. 2. Schematic view of the FATE3D model.

degradation of dioxins in the three phases could be omittedfrom the calculations (described later).

3. Formulations of the model

Formulations of the temporal changes in terms of concen-trations of the particulate- and dissolved-phase dioxins in theseawater are expressed in Eqs. (1) and (2), respectively. Theformulation of the dioxin concentrations in the sediment isexpressed in Eq. (3).

Temporal changes in concentrations of particulate-phasedioxins:

vCPðjÞvt¼�u

vCPðjÞvx� v

vCPðjÞvy� ðw�wsðjÞÞ

vCPðjÞvz

þ v

vx

�KX

vCPðjÞvx

�þ v

vy

�KY

vCPðjÞvy

�

þ v

vz

�KZ

vCPðjÞvz

�� lPðjÞCPðjÞ

�KðjÞðCPðjÞ �CSðjÞKdðjÞCWÞ þQPðjÞ ð1Þ

Temporal changes in concentrations of dissolved-phasedioxins:

vCW

vt¼�u

vCW

vx� v

vCW

vy�w

vCW

vzþ v

vx

�KX

vCW

vx

�

þ v

vy

�KY

vCW

vy

�þ v

vz

�KZ

vCW

vz

�

�lWCW�X2

j¼1

KðjÞðCSðjÞKdðjÞCW�CPðjÞÞ þQW ð2Þ

where

CP ( j ): concentration of particulate-phase dioxins bound toparticulate matter j ( j ¼ 1: phytoplankton, j ¼ 2: detritus);CW: concentration of dissolved-phase dioxins;u, v: horizontal components of the flow velocity;w: vertical component of the flow velocity;ws ( j ): sinking rate of particulate matter j;KX, KY: coefficients of the horizontal eddy diffusivity;KZ: coefficient of the vertical eddy diffusivity;lP ( j ): degradation rate of particulate-phase dioxins boundto particulate matter j;lW: degradation rate of dissolved-phase dioxins;K ( j ): adsorption rate of dissolved-phase dioxins to partic-ulate matter j;Kd ( j ): partition coefficient between dissolved-phase di-oxins and particulate-phase dioxins bound to particulatematter j;CS ( j ): concentration of particulate matter j;QP ( j ): loading flux of particulate-phase dioxins fromrivers, which are bound to particulate matter j;QW: loading flux of dissolved-phase dioxins from rivers andair.


Concentrations of dioxins in the sediment

CB ¼X2

j¼1

MPðjÞ=X3

j¼1

MSðjÞ ð3Þ

where

CB: concentration of dioxins in the sediments;MP ( j ): amount of particulate-phase dioxins bound to par-ticulate matter j deposited on the sediment;MS ( j ): amount of particulate matter j deposited on thesediment.

4. Model inputs

Running the FATE3D model requires various input datasuch as meteorological data, flow field conditions, concentra-tions and sinking rates of organic particulate matter, initial andboundary conditions, and loading fluxes and physico-chemicalproperties of dioxins.

The values of these parameters that were used in this studyare described as follows.

4.1. Calculation period and time step

The calculation period of the simulation extended to oneyear from September 1, 2002 to August 31, 2003. The timestep of the calculation was 200 s. The calculation output wasstored at 3600 s (1 h) intervals.

4.2. Target area

The entire Tokyo Bay was selected as the target area forthis study. For the model computations, the bay was dividedinto 1 km � 1 km horizontal meshes e 71 in the north-southdirection and 45 in the east-west direction (Fig. 3). In addition,the water column was divided into 10 layers (Table 1).

4.3. Target chemicals

The dioxin-related compounds in Japan include polychlori-nated dibenzo-p-dioxins (PCDDs) and polychlorinated diben-zofurans (PCDFs) as well as the dioxin-like polychlorinatedbiphenyls (dioxin-like PCBs). Typically, these three com-pounds are simultaneously evaluated in terms of toxicity sincethey are considered to be one toxic agent for the purpose ofassigning a measure of toxicity. Thus, all of these wereselected as target chemicals for this study.

PCDDs and PCDFs have 75 and 135 congeners in mono-through octa-chlorinated homologues, respectively, and thedioxin-like PCBs have 12 congeners in tetra- through hepta-chlorinated homologues.

Since the physico-chemical properties of dioxins (e.g., parti-tion coefficient) differ widely within each dioxin congener butare similar in the same homologue, calculations were performed

for each homologue. The loading fluxes and concentrations weresummed for each homologue and their average physico-chemicalproperties for each homologue were used as the model inputs.Calculations for mono- through tri-chlorinated homologues ofPCDDs and PCDFs were omitted since the six homologues arenot considered to possess any toxicity. Thus, the concentrationsof the tetra- through octa-chlorinated PCDDs (TeCDD, PeCDD,HxCDD, HpCDD, OCDD), tetra- through octa-chlorinatedPCDFs (TeCDF, PeCDF, HxCDF, HpCDF, OCDF), and tetra-through hepta-chlorinated dioxin-like PCBs (TeCB, PeCB,HxCB, HpCB) were calculated in this study.

4.4. Loading fluxes of dioxins from rivers

The loading fluxes of the particulate- and dissolved-phasedioxins from the six rivers (Edogawa, Nakagawa, Arakawa,

Fig. 3. Horizontal mesh division of the target area.

Table 1

Layer locations of the water columns

Layer Location

[m]

Layer Location

[m]

1st layer surface to �2 6th layer �10 to �13

2nd layer �2 to �4 7th layer �13 to �16

3rd layer �4 to �6 8th layer �16 to �19

4th layer �6 to �8 9th layer �19 to �22

5th layer �8 to �10 10th layer �22 to bottom


Sumidagawa, Tamagawa, and Tsurumi Rivers) that flow intothe Tokyo Bay were considered as sources of dioxins to thebay. The loading fluxes from the rivers were provided foreach river and each homologue, and vary on a daily basisin response to the river water discharge in this study. Thelocations of the mouths of these six rivers are shown inFig. 3. These loading fluxes were obtained from the studyconducted by Kobayashi et al. (2003a, 2004). In that paper,the dioxin concentrations in the six rivers were measured,and the daily loading fluxes of dioxins from these riverswere estimated as functions of river water discharges. Theloading fluxes of the particulate- and dissolved-phase di-oxins from the Edogawa River used in this study are shownin Fig. 4.

4.5. Loading fluxes of dioxins from air

The loading fluxes of dioxins from air were provided foreach mesh and each homologue in this study. It was assumedthat the loading fluxes of dioxins from air were constantthroughout the calculation period and that the dioxins existingin the air were deposited on the sea surface of the bay in theform of the dissolved-phase dioxins. These loading fluxeswere obtained from the study conducted by Kobayashi et al.(2003b, 2004). In that paper, the dioxin concentrations ofatmospheric depositions around the Tokyo Bay were mea-sured, and the annual average loading fluxes of dioxins fromair were estimated. The fluxes for each homologue used inthis study are shown in Table 2.

Fig. 4. Input data for the FATE3D model.


Table 2

Loading fluxes of dioxins from air

Homologue Flux

[ng-TEQ km�2 day�1]

Homologue Flux


Homologue Flux


TeCDD 4.06 � 102 TeCDF 5.25 � 102 TeCB 9.74

PeCDD 2.57 � 103 PeCDF 3.97 � 103 PeCB 9.57 � 102

HxCDD 1.20 � 103 HxCDF 3.75 � 103 HxCB 51.0

HpCDD 4.33 � 102 HpCDF 3.73 � 102 HpCB 8.40 � 10�1

OCDD 17.2 OCDF 3.04

4.6. Meteorological data

Meteorological parameters required for running theFATE3D system, such as temperature, wind conditions,cloud amount, relative humidity, solar radiation, tide level,and river water discharge were obtained from the JapanMeteorological Agency (2002e2003) and Japanese Ministryof Land, Infrastructure and Transport (2002e2003). Theseparameters were provided for each mesh and each day inthis study.

4.7. Flow field conditions

The FATE3D model requires various flow field conditionssuch as current velocity, water temperature, salinity, and coef-ficients of vertical and horizontal eddy diffusivity as inputdata. The values for these parameters are required for eachmesh and each day in the model. However, it is very difficultto acquire all these parameters using the field measurementsalone.

Therefore, these flow field conditions were obtained fromthe simulation results of the hydrodynamic model, COSMOS.FATE3D is programmed to run by directly using the simula-tion results as input data, which was obtained from COSMOS.The details of COSMOS have been described by Taguchi andNakata (1998), Taguchi et al. (1999a, b), and Horiguchi et al.(2001); therefore, a description of the model has been omittedfrom this paper.

The time series of water temperature and salinity at thesurface layer of Sta. A, which were calculated by COSMOS,are shown in Fig. 4. In this figure, the field measurements ofthe water temperature and salinity obtained from this studyare also plotted to compare them with the calculated values.It is evident from this figure that both the water temperatureand salinity calculated using COSMOS agreed fairly well

with the field measurements. Thus, it is reasonable to usethe simulation results obtained from COSMOS as input datafor FATE3D.

4.8. Organic particulate matter concentrations

The organic particulate matter (phytoplankton, detritus)concentrations were obtained from the simulation results ofthe coastal ecosystem model, EUTROP. FATE3D is pro-grammed to run by directly using the simulation resultsobtained from EUTROP as input data. A description ofEUTROP has been omitted from this paper since the detailsof the model have been described by Taguchi and Nakata(1998) and Taguchi et al. (1999b).

The time series of the concentrations of phytoplankton anddetritus at the surface layer of Sta. A, which were calculatedby EUTROP, are shown in Fig. 4. In this figure, the field mea-surements of the phytoplankton concentrations obtained fromthis study are also plotted to compare them with the calculatedconcentrations. It is evident from this figure that the phyto-plankton concentrations obtained from EUTROP could befavorably compared with the field measurements. Thus, it isreasonable to use these simulation results obtained fromEUTROP as input data for FATE3D.

4.9. Dioxin concentrations at the initial time ofcalculation

Dioxin concentrations at the initial time of calculation wereobtained from the study conducted by Kobayashi et al.(2003a). The averages of dioxin concentrations from the fieldmeasurements in the Tokyo Bay reported in that paper wereused in the calculations. The concentrations of the particulate-and dissolved-phase dioxins for each homologue used in thisstudy are shown in Tables 3 and 4, respectively.

Table 3

Particulate-phase dioxin concentrations at the initial time of calculation

Homologue Concentration

[fg-TEQ L�1]


[fg-TEQ L�1]


[fg-TEQ L�1]

TeCDD 1.29 TeCDF 8.75 � 10�1 TeCB 2.28 � 10�2

PeCDD 3.08 PeCDF 3.05 PeCB 1.27

HxCDD 1.55 HxCDF 3.73 HxCB 2.07 � 10�1

HpCDD 1.05 HpCDF 5.29 � 10�1 HpCB 1.87 � 10�3

OCDD 1.25 � 10�1 OCDF 8.68 � 10�3


Table 4

Dissolved-phase dioxin concentrations at the initial time of calculation


[fg-TEQ L�1]


[fg-TEQ L�1]


[fg-TEQ L�1]

TeCDD 2.60 � 10�1 TeCDF 1.14 TeCB 5.05 � 10�2


HxCDD 4.06 � 10�1 HxCDF 1.30 HxCB 1.42 � 10�1

HpCDD 1.76 � 10�1 HpCDF 1.17 � 10�1 HpCB 8.23 � 10�4

OCDD 1.52 � 10�2 OCDF 1.30 � 10�3

4.10. Dioxin concentrations at the open boundary

Since the reported dioxin concentrations in coastal seawateraround the Tokyo Bay are extremely low (Yamashita et al.,1998), the dioxin concentrations at the open boundary (seeFig. 3) were assumed to be 0 [fg-TEQ L�1] for all the dioxinhomologues, for both the particulate and dissolved phases.

4.11. Sinking rates of organic particulate matter

The sinking rates of organic particulate matter (phytoplank-ton and detritus) were obtained from the study conducted byTaguchi and Nakata (1998). Values corresponding to theseparameters that were used in this study are shown in Table 5.

4.12. Adsorption and desorption rates of dioxins

The adsorption and desorption rates of dioxins from seawa-ter to organic particulate matter (phytoplankton or detritus)were obtained from the study conducted by Hirai et al.(1992). It was assumed that the rates of desorption and adsorp-tion were identical. In this study, the rates of desorption andadsorption were assumed to be identical. Values correspondingto these parameters that were used in this study are shown inTable 6.

4.13. Partition coefficients of dioxins

The organic carbon-water partition coefficients (Koc) wereused as the partition coefficients between the particulate-and dissolved-phase dioxins. The partition coefficients be-tween the particulate- and dissolved-phase PCDDs, PCDFs,and dioxin-like PCBs for each homologue were obtainedfrom the study conducted by Mackay et al. (1992) and Hansenet al. (1999). The averages of the reported values for eachhomologue that were used in this study are shown in Table 7.

Table 5

Sinking rates of organic particulate matter

Species Rate

[cm s�1]

Phytoplankton 2.0 � 10�4

Detritus 5.0 � 10�4

4.14. Degradation rates of dioxins

Dioxins are known as the persistent organic pollutants(POPs) because they are resistant to decay in the environment.Mackay et al. (1992) reported that the half-lives of dioxin con-geners in environmental media extend up to several years.

On the other hand, it is reported that the estimated waterresidence time in the Tokyo Bay is 0.8e3.5 months (average1.6 months), depending on flow field conditions in the bay(Kaizuka et al., 1993).

On comparing these values, it is considered that the degra-dation of dioxins in the Tokyo Bay can be neglected when pre-dicting the dioxin concentrations in the Tokyo Bay seawater,particularly in case of short-term simulations (e.g., oneyear). Therefore, degradation rates of dioxins were assumedto be 0 [s�1] for all the dioxin homologues in this study.

5. Results and discussion

The dioxin concentrations in the particulate, dissolved, andsediment phases for each homologue were obtained from thecalculations results obtained by FATE3D. Firstly, the model’spredictive capability was estimated by comparing the concen-trations calculated by the model with the measured concentra-tions. Then, the factors affecting the dioxin concentrations inthe Tokyo Bay were estimated based on the simulation results.Lastly, the distributions of the dioxin concentrations in the baywere discussed based on the simulation results.

In this paper, the simulated concentrations in the sedimentsare not discussed, since the amount of dioxins accumulated inthe sediments during the calculated period (one year) was verysmall and therefore difficult to compare against the fieldmeasurements.

5.1. Time series of dioxin concentrations in the bay

Figs. 5 and 6 show the time series of the total dioxin con-centrations (sum of concentrations of the 14 dioxin homo-logues) in the particulate and dissolved phases, respectively,

Table 6

Adsorption and desorption rates of dioxins

Species Rate

[sec�1]

Phytoplankton 2.0 � 10�5

Detritus 2.0 � 10�5


Table 7

Partition coefficients of dioxins (Koc)

Homologue log Koc Homologue log Koc Homologue log Koc

TeCDD 6.41 TeCDF 5.71 TeCB 5.61


HxCDD 7.41 HxCDF 6.61 HxCB 6.12

HpCDD 7.61 HpCDF 7.01 HpCB 6.37

OCDD 7.81 OCDF 7.61

as predicted by FATE3D, at locations for which the measuredconcentrations are available. The measured concentrationsreported by Kobayashi et al. (2003a) were also plotted in thesefigures for the purpose of comparing them with the predictedconcentrations. Generally, the concentrations predicted by themodel were within a reasonable range, which was two to threetimes higher or lower than the field measurements. This not

Fig. 5. Time series of the total concentrations of the particulate-phase dioxins in the Tokyo Bay predicted by FATE3D.


Fig. 6. Time series of the total concentrations of the dissolved-phase dioxins in the Tokyo Bay predicted by FATE3D.

only clarifies the model’s predictive capability, but alsoindicates that the model inputs, such as loading fluxes andphysico-chemical properties of dioxins, have moderate values.

However, there are visible differences between the pre-dicted and measured concentrations at the bottom layer(þ1 m from the bottom) of Sta. A. On March 24, 2003, thepredicted concentrations of the particulate-phase dioxinswere underestimated by approximately ten times in

comparison with the measured concentrations, whereas thepredicted concentrations of the dissolve-phase dioxins werein good agreement with the measured concentrations. Thiscould be due to the fact that the resuspension process of thesediment in the seawater was not considered in the model. Itwas found that the measured concentrations of the particu-late-phase dioxins near the bottom sometimes became veryhigh due to the influence of resuspension.


On the basis of these results, it can be stated that theFATE3D model is very useful in predicting the dioxin concen-trations in the seawater of the Tokyo Bay, except near thebottom. For the next step that involves improving the model’spredictive capability, inclusion of the resuspension process inthe formulae of the model is required.

Next, the important factors that lead to variations in the di-oxin concentrations in the Tokyo Bay were estimated based onthe simulation results. With regard to the particulate-phase di-oxins, it was estimated that the dioxin concentrations werestrongly affected by the loading fluxes of dioxins from therivers. The particulate-phase dioxin concentrations shown inFig. 5 are highly correlated with the loading fluxes of theparticulate-phase dioxins from the rivers shown in Fig. 4.

On the other hand, seasonal cycles were observed in thedissolved-phase dioxin concentrations. At every station, thedissolved-phase dioxin concentrations were found to be highduring winter (from December to February) and low duringsummer (from June to August). The dissolved-phase dioxinconcentrations are considered to be affected by the seasonalcycle of the organic particulate matter (phytoplankton and de-tritus) concentrations in the bay. As shown in Fig. 4, the con-centrations of phytoplankton and detritus were the highestduring summer and the lowest during winter. Thus, a transferof dioxins from the dissolved phase to the particulate phasewill occur mostly during summer, as determined by the parti-tion coefficient, and the dioxins will move in the opposite di-rection mostly during winter.

Fig. 7. Budgets of the dioxins in the Tokyo Bay during summer and winter cal-

culated from the simulation results (expressed in g-TEQ).

Fig. 7 shows the budgets of the dioxins in the seawater ofthe Tokyo Bay during summer and winter, as calculatedfrom the simulation results. It was shown from this figurethat the loading fluxes of dioxins were high during summerand consequently the amount of particulate-phase dioxinswas large during summer. However, the amount of dis-solved-phase dioxins was large during winter.

In summary, it was determined that the loading flux of di-oxins from the rivers was the most important factor affectingthe particulate-phase dioxin concentrations and that the sea-sonal cycle of the concentrations of organic particulate matterwas the most important factor affecting the dissolved-phasedioxin concentrations in the bay.

5.2. Distributions of dioxin concentrations in the bay

Since the dioxin concentrations predicted by FATE3D werereasonable when compared with the field measurements inboth the particulate and dissolved phases, the distributions ofdioxin concentrations in the Tokyo Bay were discussed basedon the simulation results.

Fig. 8 shows the distributions of the total dioxin concentra-tions in the particulate and dissolved phases in the Tokyo Bay,as predicted by FATE3D. The concentrations expressed inFig. 8 are the one-year averages in the surface layers.

It was predicted that the particulate- and dissolved-phasedioxin concentrations in the bay ranged from approximately5 to 150 fg-TEQ L�1 and from 2 to 25 fg-TEQ L�1, respec-tively. It was shown that the dioxin concentrations were thehighest at the mouths of the four large rivers (Edogawa, Naka-gawa, Arakawa, Sumidagawa rivers), and the lowest at theopen boundary in both the particulate and dissolved phases.

Further, it was shown that the concentrations of the particulate-phase dioxins account for the major portion of the sum of the par-ticulate- and dissolved-phase dioxin concentrations; however, thepercentages of the particulate-phase dioxins varied depending onthe location. From the simulation results, the particulate-phase di-oxin concentrations near the mouths of these four rivers are ap-proximately six times higher than those of the dissolved-phasedioxins, whereas they were two to three times higher near theopen boundary. These variations are considered to indicate thedifferences in the transporting behavior between the particulate-and dissolved-phase dioxins. The particulate-phase dioxinsfrom the rivers were immediately diluted and deposited on thesediment after they flowed into the bay, whereas the dissolved-phase dioxins transported from rivers and air were only dilutedby seawater, revealing almost no marked variation comparedwith the particulate-phase dioxins in the bay.

6. Conclusion

Using a 3-D chemical fate prediction model (FATE3D), thetime series and distributions of dioxin concentrations in theseawater of the entire Tokyo Bay were predicted. These sim-ulation results could be favorably compared with the fieldmeasurements of dioxin concentrations in the bay for both


Fig. 8. Distributions of the total dioxin concentrations in the particulate and dissolved phases in the surface layers of the Tokyo Bay predicted by FATE3D

(one-year averages).

the particulate and dissolved phases, indicating the validityand predictive capability of the model.

Furthermore, the differences in the seasonal cycles and dis-tributions between the particulate- and dissolved-phase dioxinsin the bay were estimated from the simulation results. Thus, itcan be stated that the FATE3D model is very useful in predict-ing the dioxin concentrations in the seawater of the Tokyo Bay,and the results obtained from this model can be applied tohuman health and ecological risk assessments of dioxins.

However, the particulate-phase dioxin concentrations in thebottom layers were sometimes underestimated because theprocess of resuspension was not taken into account in themodel. As described in a previous section, since most of theparticulate-phase dioxins deposit on the sediment of the TokyoBay, improving the model’s predictive capability for thebottom layers is important. Thus, including the resuspensionprocess to the formulae of the model shall be the focus ofour next study. Following that study, the ecological risk assess-ment of dioxins under the concentrations predicted by themodel will be carried out.

References

Hansen, G., Paya-Perez, A.B., Rahman, M., Larsen, B.R., 1999. QSARs for

Kow and Koc of PCB congeners: a critical examination of data, assumptions

and statistical approaches. Chemosphere 13, 2209e2228.

Hirai, M., Ishikawa, K., Nakata, K., Sayama, M., Horiguchi, F.,

Urushigawa, Y., Yonezawa, Y., Masunaga, S., Fukui, M., 1992. Behaviors

of synthetic compounds in coastal environments. Reports on Pollution

Study vol. 44, 1e17 (in Japanese).

Horiguchi, F., Yamamoto, J., Nakata, K., 2001. A numerical simulation of sea-

sonal cycle f temperature, salinity and velocity fields in Tokyo Bay. Marine

Pollution Bulletin 43, 145e153.

Japan Meteorological Agency, 2002e2003. Yearly Reports from Japan Mete-

orological Agency (in Japanese).

Japanese Ministry of Health, Labour and Welfare, 1999e2003. Results of the

Total Diet Studies in FY 1998e2002 (in Japanese).

Japanese Ministry of Land, Infrastructure and Transport, 2002e2003. Chrono-

logical Table of Daily River Water Discharge (in Japanese).

Kaizuka, S., Kikuchi, T., Matsuda, I., Miyamura, T., Unoki, S., Ishiwatari, R.,

1993. Topography, Geology and Hydrology of the Tokyo Bay. Tsukiji

Shokan Publishing, Tokyo, 211 pp (in Japanese).

Kobayashi, N., Masunaga, S., Kameda, Y., Hanai, Y., Nakanishi, J., 2003a.

Transportation of PCDD/Fs and dioxin-like PCBs by rivers into the Tokyo

Bay, Japan. Organohalogen Compounds 61, 345e348.

Kobayashi, N., Masunaga, S., Nakanishi, J., 2003b. Source identification and

behavior of PCDD/Fs and dioxin-like PCBs in Japanese River Water. Journal

of Japan Society on Water Environment 26, 655e662 (in Japanese).

Kobayashi, N., Masunaga, S., Nakanishi, J., 2004. Transportation and sources

of polychlorinated dibenzo-p-dioxins, dibenzofurans and dioxin-like PCBs

in rivers flowing into the Tokyo Bay. Journal of Japan Society on water

Environment 27, 465e472 (in Japanese).

Mackay, D., Shiu, W.Y., Ma, K.C., 1992. Illustrated Handbook of Physical-

Chemical Properties and Environmental Fate for Organic Chemicals,

vol. 2. Lewis Publishers, 608 pp.

Shimizu, M., 1997. Aquatic organisms. In: Numata, M., Furota, T. (Eds.),

Ecology of Tokyo Bay. Tsukihi Shokan Publishing, Tokyo, pp. 143e155

(in Japanese).


Taguchi, K., Nakata, K., 1998. Analysis of water quality in Lake Hamana us-

ing a coupled physical and biochemical model. Journal of Marine Systems

16, 107e132.

Taguchi, K., Nakata, K., Ichikawa, T., 1999a. A 3-D simulation of long term

variability in the flow field and T-S structure in the Ise-Mikawa Bay

estuary. Journal of Advanced Marine Science and Technology Society 5,

37e48.

Taguchi, K., Nakata, K., Ichikawa, T., 1999b. A 3-D simulation of lower

trophic ecosystem in the Ise-Mikawa Bay estuary using a coupled physical

and biochemical model. Journal of Advanced Marine Science and Technol-

ogy Society 5, 49e62.

Yamashita, N., Imagawa, T., Miyazaki, A., 1998. Dynamics of PCDDs and

PCDFs in coastal sea water around Tokyo Bay in Japan. Organohalogen

Compounds 39, 389e392.

application of a 3-d chemical fate prediction model (fate3d) to predict dioxin concentrations in the...

Documents