In a tributary watershed of the Kuji River of Japan, the hydrological components of runoff associated with a precipitation event were investigated using isotope tracers of hydrogen (2H) and oxygen (18O) in precipitations and stream water. The runoff was separated into "old water" (pre-existing in the ground before the precipitation event) and "new water"(from the precipitation). It was found that the discharge of several hazardous trace elements (Sb, Cu, Cr) was largely (24-54%) attributable to that of the new water in spite of its small contribution to the total water discharge. These investigations suggest that the new water may play an important role in the migration of atmospherically derived, hazardous trace elements to streams during precipitation events. The present findings will contribute to current necessity of assessment of a risk of long-term exposure to pollutants at low concentrations by providing information on their transport among different environmental media
Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 243 - ISOTOPE HYDROGRAPH SEPARATION FOR MODELING OF RUNOFF MECHANISMS OF ATMOSPHERICALLY DERIVED CHEMICAL AND RADIOACTIVE POLLUTANTS T. Matsunaga*, N. Yanase*, Y. Hanzawa*, K. Tsuduki* and H. Naganawa* *Department of Environmental Sciences, Tokai Research Establishment, Japan Atomic Energy Research Institute, Tokai-mura, Naka-gun, Ibaraki-ken, 319-1195, Japan ABSTRACT In a tributary watershed of the Kuji River of Japan, the hydrological components of runoff associated with a precipitation event were investigated using isotope tracers of hydrogen ( 2 H) and oxygen ( 18 O) in precipitations and stream water. The runoff was separated into "old water" (pre-existing in the ground before the precipitation event) and "new water"(from the precipitation). It was found that the discharge of several hazardous trace elements (Sb, Cu, Cr) was largely (24-54%) attributable to that of the new water in spite of its small contribution to the total water discharge. These investigations suggest that the new water may play an important role in the migration of atmospherically derived, hazardous trace elements to streams during precipitation events. The present findings will contribute to current necessity of assessment of a risk of long-term exposure to pollutants at low concentrations by providing information on their transport among different environmental media. KEYWORDS deuterium; isotope hydrograph; oxygen-18; pollutants; runoff INTRODUCTION Since the 1980s, non-point source contamination has been one of the central issues in water quality management, particularly with regard to the contamination of fresh water bodies with sources distributed over a catchment. Some of these sources can be considered to originate from the atmosphere. The widespread use of metals and their strict controls at point-sources (relevant firms of smelting, fabrication, etc.) have now raised concerns about spreading of atmospheric depositions of anthropogenic heavy metals at trace levels (NIES, 1991). While a number of related studies have been carried out, the process of inflow of contaminants to water bodies such as streams or impoundments has not been studied enough to allow quantification of its impact. An understanding of the discharge of trace elements will serve to predict migration of hazardous elements in the environment. Even radionuclides that may be released to the atmosphere in a nuclear accident can contribute to the concern of atmospherically derived, broadly spread contamination (Cooper et al., 2002). There are such instances, the Windscale and the Chernobyl accidents. Other than these instances, it is hardly possible to track those radionuclides in the actual environment, because an accidental release seldom occurs. Instead, mathematical modeling has been considered as a useful tool as for tracking those radionuclides (e.g. Monte et al., 2000). Thus, findings about the migration of stable trace elements will be also suggestive in providing a scientific basis for a migration model of those radionuclides by a similarity in their pathways including the atmosphere and the hydrosphere. We have discussed fluvial elemental loads in particulate and dissolved forms in relation with the discharge rate (Nagano et al., 2003). The present study focuses on the hydrological components of runoff after a precipitation event and on their relation to the load of dissolved elements. The Oda River and its watershed, located in a central Japan, was investigated in the present study. Hydrogen ( 2 H) and oxygen ( 18 O) isotopes of precipitations and stream water samples are used as tracers of water to separate the runoff into i) that originating from the ground water and the sub-surface water existing in the ground before the precipitation event ("old water"), and ii) that from the storm precipitation ("new water"). This separation allows us to identify the component responsible for carrying trace Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 244 - elements to a stream. The analytical results of trace elements including hazardous metals in those water samples are then coupled with the isotope analysis. Finally, the importance of the component originating from the precipitation (new water) in a carrier of hazardous metals is considered. Our present findings will be included in our hydrological model to describe migration of accidentally released radionuclide and hazardous trace elements (Tsuduki et al., 2004). EXPERIMENT Site description The Oda River is a tributary of the Kuji River located in Fukushima Prefecture of Japan (Fig. 1). It is 10 km long and has a 40 km 2 watershed, 92 % of which is categorized as hilly forests (National Land Agency, 1980). Its plain area is limited to 2.8 km 2 (7 %) and the open surface of the stream is only 1.2 km 2 (3 %). Sampling and on-site monitoring River water. River water was collected at a site downstream of the Oda River (36 o 31' N, 140 o 26' E), 6 river-km to the confluence of the Oda tributary and the Kuji main stream. The collection was performed with the aid of an automated sampler at 2 h intervals. At a late stage of recession, this was adjusted to 4 h intervals. River water samples were collected in a 1 L polypropylene bottle and were transported in a cooler to our laboratory (60 km away from the sampling location) for filtration and analyses. Precipitation. Precipitations were collected at two locations at 600 - 650 m in altitude in a central part of the watershed. A supplemental sampling was done at a foot of the watershed (220 m). Each 1 mm of the first 8 mm after the start of the raining was collected separately. The remaining precipitations were collected as a whole. On-site monitoring. The water level was monitored continuously at the river water sampling location using a level gauge. This was calibrated to the water flow using manually observed data regarding the water level and flow rate at sectioned streams. The value of pH was also monitored. Analysis Isotopic composition of 2 H and 18 O. The compositions of the isotopes 2 H (D) and 18 O were analyzed to obtain δD (‰) and δ 18 O (‰), respectively, which are defined as the relative deviation of the parts per thousand ratios from that of a standard according to the following equation (Kendall and Caldwell, 1998): δ = {(R sample /R standard )-1}×1000, (1) where R refers to the ratio of 2 H/ 1 H, or 18 O/ 16 O in atomic concentrations. Oda River watershed 0 20 km P a c i f i c O c e a n Kuji River watershed N ■ Fig.1 Location of the Kuji River and the Oda River watersheds. The closed square denotes the location of our laboratory, where analytical works in the present study were conducted and the sampling of atmospheric depositions was conducted by Ueno and Amano (2003), see text. Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 245 - The standard employed in the present study was the standard mean ocean water (SMOW) distributed by the International Atomic Energy Agency (IAEA). The hydrogen isotopic composition was measured by reducing the water by passage over chromium at 840 o C, chiefly following Coleman et al. (1982), with a mass spectrometer (MAT252, Finnigan MAT GmbH). The composition of oxygen was analyzed according to a principle of CO 2 -H 2 O equilibration (Epstein and Mayeda, 1953; Horita and Kendall, 2004). Briefly, an aliquot of water sample (200 µl) was equilibrated with 3% CO 2 in hyperpure grade helium in a glass vessel (5 ml) at 40 o C for 12 hrs. Then, the oxygen composition of CO 2 was analyzed with a mass spectrometer (Isoprime-Multiflow System, VG Instruments) equipped with a gas chromatography column for separation of CO 2 . The standard deviation of our measurements was in the range of ±0.1 to ±0.3 ‰ and ±0.2 to ±0.4 ‰ for hydrogen and oxygen, respectively. Inorganic elements. After filtration, water samples were acidified with nitric acid for ultra-trace analysis (Wako Pure Chemical Industries Ltd.) to 0.3 N HNO 3 . Inorganic elements other than Si were analyzed with an inductively coupled plasma mass spectrometer (ICP-MS) (HP-4500, Yokogawa Analytical Systems). The elements were classified into several groups by concentration level as appropriate for the analysis. Mixed standards (XSTC-1, -7 , -8 and -13, SPEX) were used for calibration. The concentration of Si was determined by spectrophotometry using ammonium molybdate. RESULTS AND DISCUSSION Runoff and isotopic composition This paper deals with an observation for a precipitation event during Oct. 22-23 of 2003. The amount of precipitation was 30 mm in total. Figure 2 shows the precipitation record of the studied event (Fig. 2a) and the record of the river flow rate (solid line, Fig. 2b) of the Oda River at the water sampling location. Fig. 3 Variation in the δD of precipitations (a) , the river water (c), and the accumulated rainfall (b). -40 -35 -30 -25 -20 -15 0 5 10 15 20 25 30 δD (‰) (a) accumulated precipitation (mm) 0 10 20 30 Oct. 21 Oct. 22 Oct. 23 Oct. 24 accumulated precipitation (mm) (b) -52 -50 -48 -46 -44 -42 -40 -38 δD (‰) 0 24 48 72 96 (c) Elapsed time (hrs) since 0000 Oct. 21 0 2 4 6 8 10 Oct. 21 Oct. 22 Oct. 23 Oct. 24 Hourly precipitation (mm) (a) 0 1 2 3 Flow rate (m 3 /s) (b) total old water new water 0.0 0.1 0.2 0.3 0.4 0.5 0 24487296 Fraction of new water, f 1 (t) (c) Elapsed time (hrs) since 0000 Oct. 21 Fig. 2 Record of precipitation (a) and river water flow rate (b) (total, solid line) with evaluated fraction of new water (c) and separated flow rates (b) (dashed and dotted lines). Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 246 - In response to the major part of the raining (0600 to 1400 of Oct. 22, 22 mm in total), the flow rate increased to its peak value of 2.2 m 3 /s (1400 of Oct. 22). It then decreased gradually with two small peaks that corresponded to light rains. Four days after the major raining ceased, the flow rate maintained a certain value (0.8 m 3 /s, 1400 of Oct. 26) that was higher than the lowest value (0.5 m 3 /s) before the precipitation event. Variations in the δD of precipitations and the river water are depicted in Fig. 3a and Fig. 3c, respectively. The isotopic records in Fig. 3 are of the highest location (650 m). The records at this location were considered to be representative in the central part of the watershed and are discussed hereafter. This is because records of the δD of precipitations at another location in the same part differed only by 0.5 ‰ from the former in terms of amount-weighed mean. The δD of precipitations decreased from -17.2 ‰ (the first 1 mm) to -38.3 ‰ (6-7 mm), and then returned to -27.6 ‰ in the last 14 mm (1000 - 1400 of Oct. 22, Figs. 3a. and 3b). Its amount-weighed mean was -28.1 ‰. On the other hand, the δD of river water was -47.9 ‰ at the start (0600, Oct. 22) and reached a maximum value of -44.1 ‰ at 1600 of Oct. 22. It then returned to -48.2 ‰ 12 hrs later, at which time it most resembled the earlier low value (-47.9 ‰) before the precipitation event. Interestingly, it continued to decrease with time, reaching -50.1 ‰ at 1200 of Oct. 24. The oxygen isotopic composition δ 18 O of the precipitations varied from -4.9 to -8.7 ‰ in separate samples. Its amount-weighed mean was -6.0 ‰ (Fig. 4). Machida and Kondoh (2001) showed that the relationship between δD and δ 18 O in shallow ground water and river water in Japan statistically is given by the equation: -60 -50 -40 -30 -20 -10 0 -10 -8 -6 -4 -2 0 δD (‰) δ 18 O (‰) river water (during the main part of the flooding) river water (after the main part of the flooding) □ ■ △ ▲ river water (pre-event) precipitations precipitations (weighted mean) river water (dry season, mean) Fig. 4 Variation in the δ 18 O of precipitations at a highest sampling location (650m) and the river water with a relation to those in the δD. The dashed lines correspond to probable relationships reported for Japanese shallow ground waters and river waters (see text). Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 247 - δD = 6.72 δ 18 O + 3.94 ± 11.88 (3σ). (2) The relationship of the studied precipitations was reasonably found between the two expected lines (dashed lines in Fig. 4). The δ 18 O of river water exhibited less variations than the δD. The δ 18 O was -7.6 ‰ before the start of raining (0600 of Oct. 22). It then decreased slightly to its lowest value -8.0 ‰, at 0200 Oct. 23, and then increased to -7.4 ‰ (1200 of Oct. 24). This ambiguity may come from the intrinsic few variation in the δ 18 O relative to that in the δD as indicated by Eq.(2), and also from relatively large uncertainty (±0.2 to ±0.4 ‰) in the analysis of the δ 18 O with our instrument. The isotopic composition of river water samples collected in a dry season (Feb. 2004) indicated that the composition of the river waters in the studied precipitation event (Oct. 2003) were affected by seasonal variations to some extent. Hydrograph separation A hydrograph separation was carried out based on data of the flow rate and the isotopic composition. The isotopic composition of hydrogen (δD) was used in the separation because its value was more significant than that of oxygen in our measurement, considering of their respective analytical errors. A concept of two-component mixing was employed, in which it was assumed that runoff is composed of a mix of two components: one originating from the ground water and the sub-surface water existing before the precipitation event ("old water" or pre-event water), and one from the precipitation event ("new water" or event water) (Buttle, 1994; Kendall et al., 1995). The δD of river water is expressed by the following equation: C r (t) = f 1 (t) C 1 (t) + (1- f 1 (t)) C 2 (t) (3) where C r (t), C 1 (t), and C 2 (t) are the isotopic compositions of hydrogen (δD) of the river water, new water, and old water, respectively, and f 1 (t) represents the fraction of new water in the runoff at time t. Further, C 1 (t) and C 2 (t) are assumed to be constant. In the present case, C 1 was set to the amount-weighed mean δD of the precipitation samples (-28.1 ‰), and C 2 was set to the δD of the river water just before the precipitation event (-47.9 ‰). For the main part of the runoff that is described later, it was considered that these two (precipitations and pre-event river water) could be end members of mixing (see Fig. 4). Although there are many arguments for the general applicability of this two component concept (e.g. Harris and McDonnell, 1995), it was employed in the present study only in order to grasp a first insight into the relationship between the discharge of new water and the migration of hazardous elements, which has not been sufficiently analyzed or reported previously. As a result of the calculation, and by setting C 1 and C 2 as mentioned above, the fraction of new water evaluated for the main part of the runoff (0000 of Oct. 22 to 1600 of Oct. 23, 40 h in total) was found to be as shown in Fig. 2c. The fraction reached its maximum, 0.21 (21 %) at 1600 of Oct. 22, while the peak intensity of the precipitation event occurred 6 hrs earlier and the peak of the flow rate occurred 2 hrs earlier. Figure 2b shows the hydrograph of the two components. This separation suggests that the increase of water discharge in the precipitation event was mainly caused by old water, and that the temporal contribution of new water was limited to about 20 % at most. This feature has been commonly found in rural areas, but not in urban areas. It has been reported that the statistical mean of the fraction of new water is 0.23 (i.e. 0.77 is the fraction of peak discharge consisting of old water) for 32 forested areas in precipitation events (Buttle, 1994). It seems that the Oda River watershed also is one of these much-studied forested areas, at least from this viewpoint. Discharge of major and trace inorganic elements In the present study, 68 major and trace inorganic elements in river water were analyzed. This large number made it possible to consider the discharge of inorganic elements as a function of their origins, including anthropogenic, detrital, and other natural sources. Figure 5 shows the temporal variations in concentrations of several selected elements in river water. Sodium, which must have come from the weathering of base rocks (detrital origin) and also from sea salt fallout, showed a slight decrease at the time of peak flow rate, possibly due to dilution by the precipitations. The variation in concentrations of potassium has not been explained. Chlorine, which is often Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 248 - ubiquitously distributed in the sub-surface and ground water, exhibited no change in response to this increased runoff, as well as magnesium. In contrast, concentrations of copper and antimony exhibited clear increase in response to increase in the flow rate. To determine the relative importance of the separated water components as carriers of trace elements to a stream, we evaluated the mass load of those elements associated with each component. This load was evaluated as a cumulative value during the main part of the runoff of interest, for which f 1 (t) was evaluated (0000 of Oct. 22 to 1600 of Oct. 23, 40 hrs in total). First, the total cumulative load (ML total ) in gram was calculated for each element as follows: ML total = Σ i C i × Q i × T (4) , where C i (g/m 3 ) represents concentration of the element in river water at time i, Q i (m 3 /s) is river water flow rate at time i, and T is time interval of observation (s). The cumulative load attributed to the old water, ML old was calculated in a similar manner by assuming that C i was constant at its pre-storm value. Then, ML new , the cumulative load attributed to the new water, was determined as the difference between ML total and ML old : ML new = ML total - ML old (5) The resultant load is listed in Table 1 for selected elements. The fraction of ML new in ML total , which is the new water contribution, is summarized in Table 2 as well as other elements listed in the order of the contribution. In Table 2, the new water contribution is limited for the major detrital elements such as Ca and Mg. This indicates reasonability of this method of load separation in this case. It was found that Al, Ti, Y, U, and Th have similarly Fig.5 Temporal variations in concentration of several elements in river water during the freshet period. Elapsed time is since 0000 of Oct. 21, 2004. 0 400 800 1200 0 1 2 3 Cu (ng/l) Flow rate (m 3 /s) Oct. 21 Oct. 22 Oct. 23 Oct. 24 0 24 48 72 96 Elapsed time (hrs) (c) Cu 0 20 40 60 0 1 2 3 Sb (ng/l) Flow rate (m 3 /s) (d) Sb Oct. 21 Oct. 22 Oct. 23 Oct. 24 0 24 48 72 96 Elapsed time (hrs) 0 4 8 12 0 1 2 3 Cl - (mg/l) Flow rate (m 3 /s) 0 24 48 72 96 Oct. 21 Oct. 22 Oct. 23 Oct. 24 Elapsed time (hrs) (b) Cl 0 2 4 6 0 1 2 3 Major cations (mg/l) Na K Mg (a) Na, K, Mg Flow rate (m 3 /s) Oct. 21 Oct. 22 Oct. 23 Oct. 24 0 24 48 72 96 Elapsed time (hrs) Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 249 - high contributions (30-60 %). Generally, aluminum hydroxide forms colloids in neutral river water, as in the present circumstances (pH 7.1-7.3). It also tends to be associated with other types of colloids including Ti, Y, U, and Th in freshwater (e.g. Pokrovsky and Schott, 2002). Indeed, in this case their concentrations in river water are strongly correlated each other. Taking into account these reported colloidal size distributions and the observed correlation, the present result implies that a discharge of these elements and a formation of colloids including these elements are closely related to the discharge of new water. Table 1 Cumulative mass load of selected elements in fluvial discharge of the studied precipitation event. Element Cumulative mass load ( g/km 2 ) * ML total ML old ML new Na 26 x10 3 25.7 x10 3 0.3 x10 3 Si 41 x10 3 41 x10 3 0 ( < 2 x10 3 ) Cr 0.4 0.34 0.06 Cu 3.2 1.5 1.7 Sb 0.19 0.12 0.07 runoff (mm) 5.6 5.3 0.3 * The load is normalized for the area of the watershed ( 40 km 2 ). Table 2 Evaluated new water contribution for selected elements in stream water in the studied precipitation event. New water contribution* (%) Trace elements Major elements 0-10 Li, B Na, Mg, Ca, Si 10-20 V, Cr, Fe, Ni, Rb, Cs K 20-30 Zn, As, Pb 30-60 Al, Ti, Cu, Y, Sb, U, Th * A fraction of ML new in ML total . Table 3 Ratio of relative abundance of inorganic elements to aluminum in atmospheric depositions in the Tokai village, Japan. Ratio* Trace elements Major elements 1-2 Sc, Ti, V, Fe Mn 2-5 Co, Rb, Ta, Mg 5 < Cr(24) # , Br(16), Sb(54) Na(27) *[X deposit /Al deposit ]/[X soil /Al soil ],where X represents the element of interest (see text). # Figure in parenthesis denotes the ratio of that element. The most notable feature in Table 2 is that some trace, hazardous elements were increased in the new water contribution, namely Sb (36 %) and Cu (52 %). Ueno and Amano (2003) reported on recent atmospheric depositions of a number of major and trace elements in the Tokai village, which is located at the foot of the Kuji River watershed (Fig. 1). Note that the village faces the Pacific Ocean, and the deposition is influenced by the Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 250 - ocean to some extent. Based on their data, we estimated the anthropogenic influence on the depositions by comparing their content with that of aluminum. Aluminum was chosen as a reference because of its naturally high content in soil. In principle, the ratio of their content of the depositions to that of the surface soil can reflect the enrichment of the element through the atmosphere. In the case of no increase, the ratio will be equal to 1. Table 3 shows the obtained ratios in the present study. Copper was not analyzed by Ueno and Amano due to methodological limitations. While the elements Na, Cl, and Br have high ratios because they are mostly contained in incident sea salt spray, the elements Sb (ratio 54) and Cr (24) also have high ratios and may be of anthropogenic origin. The industrial use of Sb and Cr is constantly increasing in Japan, along with the use of Cu. Especially, the domestic use of Sb in Japan markedly increased from 1800 tons (1980) to 11000 tons (2000) (Metal Mining Agency of Japan, 2001). Thus, atmospheric depositions of Sb, Cu, and Cr probably do have major anthropogenic origins. Implication in studies of hazardous trace elements Other than several serious episodes caused by localized sources, concentrations of heavy metals in Japanese aquatic environments have been generally far lower than legal standards and guideline values (e.g. Ministry of Environment, 2005). However, Nriagu (1988) posed the question that the risk caused by "trace levels" has been underestimated. This might be especially true for substances that have not been used until now. For such substances, few studies have been carried on their environmental behavior and associated toxicity. Antimony, which is known as a carcinogen (Eisler, 2000), is one of these substances. Although its toxic properties have not been well understood (Urano, 2001; Fiella et al., 2002), it is used as a flame retardant additive in large amounts (Fiella et al., 2002). The frequency of the detection of antimony over 0.002 mg l -1 in Japanese water bodies has recently reached 20 to 30 %, while the guideline value for drinking water is 0.015 mg l -1 . Increasing is the risk of long-term, low-dose exposure to toxicants which are broadly distributed in the environment and whose toxicities are not well known. As Arizono (1999) pointed out, assessing such type of risk has become very necessary. The present investigation will contribute to knowledge of transport of heavy metals of non-point sources in different environmental media. The above assessment requires such knowledge. CONCLUSIONS In the present study, runoff from a precipitation event was successfully separated into two components which were designated old water and new water, based on their hydrogen isotopic compositions. The sources of these components correspond respectively to the ground and sub-surface water existing before the precipitation event, and the precipitation. This separation approach made it possible to attribute the fluvial discharge of major and trace elements associated with the precipitattion event to the respective components. It was found that higher levels of Sb, Cu and Cr occurred in new water, along with Al and its related elements. An analysis of the reported atmospheric depositions near the studied watershed and the statistics regarding metal use suggests that there has been an anthropogenic enrichment of Sb and Cr, at least, in their atmospheric depositions. These investigations suggest that new water may play an important role in the migration of atmospherically-derived, hazardous trace elements to streams during rain storms. The present findings will contribute to current necessity of assessment of a risk of long-term exposure to heavy metals at low concentrations. Acknowledgements This research has been supported by the River Environment Fund of the Foundation of River & Watershed Environment Management. Necessary allowances related to the field work by the authorities of the Fukushima Prefecture and the Ministry of Land, Infrastructure and Transport are appreciated. The authors thank Dr. Mariko Atarashi-Andoh for her aid in mass spectrometry. We also thank Mr. Takashi Ueno and Mr. Morio Takada for their support in laboratory analysis and field sampling. Critical comments for this work from Dr. Masahiro Kumata are acknowledged. Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 251 - REFERENCES Arizono, K. (2000). Heavy metal toxicity in water environment, J. J. Soc. Water Environ., Vol.22, 341-345. Buttle, J. M. (1994). Isotope hydrograph separation and rapid delivery of pre-event water from drainage basins. Progress in Physical Geography, Vol.18, No.1, 16-41. Coleman, M. L., Shepherd, T. J., Durham, J. J., Rouse, J. E. and Moore, G. R. (1982). Reduction of water with zinc for hydrogen isotope analysis. Anal. Chem., Vol.54, 993-995. Cooper, J. R., Randle, K. and Ranjeets, S. S. (eds.) (2002). Radioactive Releases in the Environment -Impact and Assessment-. John Wiley & Sons, Chichester. Eisler, R. (2000). Chemical Risk Assessment - Health hazards to human, plants, and animals -, Volume 1 (Metals), Lewis Publishers, New York. Epstein, S. and Mayeda, T. (1953). Variation of 18 O content of waters from natural sources. Geochim. Cosmochim. Acta, Vol.4, 213-24. Fiella, M., Belzile, N. and Chen, Yu-Wei (2002). Antimony in the environment: a review focused on natural waters I. Occurrence, Earth-Science Reviews, Vol. 57, 125-176. Harris, D. M., McDonnel,l J. J. (1995). Hydrograph separation using continuous open system isotope mixing. Wa t er Res. Res., Vol.31, No.1, 157-171. Horita, J. and Kendall, C. (2004). Stable isotope analysis of water and aqueous solutions by conventional dual-inlet mass spectrometry. In: Handbook of Stable Isotope Analytical Techniques (Volume 1), P. A. de Groot (ed.), Elsevier Science B. V., Amsterdam, pp. 1-37. Kendall, C., Sklash, M. G. and Bulle D. (1995). Isotope tracers of water and solute sources in catchments. In: Solute Modelling in Catchment Systems, S. T. Trudgill (ed.), John Wiley & Sons, Chichester, pp. 261-303. Kendall, C. and Caldwell, E. A. (1998). Fundamentals of isotope geochemistry. In: Isotope Tracers in Catchment Hydrology, C. Kendall and J. J. MacDonnell (eds.), Elsevier Science B. V., Amsterdam, pp. 51-86. Machida, I. and Kondoh, A. (2001). Stable isotope ratios of natural water in Japan. J. Japn. Soc. Hydrol. Water Resource, Vol.16, No.5, 556-569. Metal Mining Agency of Japan (ed.) (2001). The Mining Industry of Japan. Metal Mining Agency of Japan, Tokyo. Ministry of Environment (Japan) (2005). Results of the Water Quality Survey of Public Water Areas, http://www.env.go.jp/en/news/2004/1216a.html. Monte, L., Van der Steen, J., Bergstroem, U., Gallego, D. E., Hakanson L. M. and Brittain J. (2000). The project MOIRA: a model-based computerised system for management support to identify optimal remedial strategies for restoring radionuclide contaminated aquatic ecosystems and drainage areas: Final report, RT/AMB/2001/13, Ente per le Nuove tecnologie l'Energia e l'Ambiente, Rome. Nagano, T., Yanase, N., Tsuduki, K. and Nagao, S. (2003). Particulate and dissolved elemental loads on the Kuji River related to discharge rate. Environ. Int., Vol.28, 649-658. National Institute for Environmental Sciences (NIES) (1991). Study on behavior and characteristics of hazardous compounds in soil and ground water environment, Report of Special Research from the National Institute for Environmental Sciences, SR-6-'91, NIES, Tsukuba, Japan. (In Japanese) National Land Agency (ed.) (1980). A survey data on the major river basins in the Ibaraki Prefecture, the Tochigi Prefecture and the Tokyo Municipal. National Land Agency of Japan, Tokyo. Nriagu, J. O. (1988). A silent epidemic of environmental metal poisoning?. Environmental Pollution, Vol. 50, 139-161. Pokrovsky, O. S. and Schott, J. (2002). Iron-colloids/organic matter associated transport of major and trace elements in small boreal rivers and their estuaries (NW Russia). Chemical Geology, Vol.190, 141-179. Tsuduki, K. (2004). Development of a mathematical model to describe the transport of artificial radionuclides and trace elements in a river watershed. In: Consideration for Modeling Studies of Migration of Accidentally Released Radionuclides in a River Watershed, JAERI-Review 2003-039, Japan Atomic Energy Research Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 252 - Institute, Tokai-mura, Ibaraki-ken, pp. 49-55. Ueno, T. and Amano, H. (2003). Deposition of Radionuclides and Stable Elements in Tokai-mura, JAERI-Data/Code 2003-004, Japan Atomic Energy Research Institute, Tokai-mura, Ibaraki-ken. Urano, K. (2001). Toxicity ranking and chemical properties of chemical substances listed in PRTR-MSDS systems, Kagaku Kohgyo Nippou Sha (publisher), Tokyo. (in Japanese) . Na 26 x10 3 25.7 x10 3 0 .3 x10 3 Si 41 x10 3 41 x10 3 0 ( < 2 x10 3 ) Cr 0.4 0 .34 0.06 Cu 3. 2 1.5 1.7 Sb 0.19 0.12 0.07 runoff (mm) 5.6 5 .3 0 .3 * The. 1 2 3 Cu (ng/l) Flow rate (m 3 /s) Oct. 21 Oct. 22 Oct. 23 Oct. 24 0 24 48 72 96 Elapsed time (hrs) (c) Cu 0 20 40 60 0 1 2 3 Sb (ng/l) Flow rate (m 3 /s)