- Email: [email protected]
Potential impact of former Zn ore extraction activities on dissolved uranium distribution in the Riou-Mort watershed (France)
Science of the Total Environment 382 (2007) 304 – 310 www.elsevier.com/locate/scitotenv
Potential impact of former Zn ore extraction activities on dissolved uranium distribution in the Riou-Mort watershed (France) Hanna-Kaisa Saari ⁎, Sabine Schmidt, Alexandra Coynel, Stéphanie Huguet, Jörg Schäfer, Gérard Blanc UMR CNRS 5805 “Environnements et Paléoenvironnements Océaniques”, Université Bordeaux 1, Avenue des Facultés, 33405 Talence Cedex, France Received 27 November 2006; received in revised form 17 April 2007; accepted 25 April 2007 Available online 4 June 2007
Abstract The industrial basin of Decazeville (Riou-Mort watershed, South-West France) is well known for its heavy metal pollution and its subsequent environmental effects on the Lot-Garonne River system. The source of this pollution is the Riou-Mort River, which drains smelting waste areas. A first survey after remediation works has revealed elevated dissolved uranium (UD) concentrations in the outlet of the Riou-Mort River. The objective of this research is to identify the origin of UD in the Riou-Mort watershed and to evaluate the impact of industrial activities on this element. Uranium was measured at 10 water sampling sites, located upstream and downstream the industrial basin, and in three smelting waste deposits. Uranium concentrations in the smelting waste deposits reach up to 14.4 mg kg− 1 and (234U/238U) activity ratios (AR) are near unity. Dissolved U concentrations in the Riou-Mort River and its main tributaries ranges over two orders of magnitude from 0.02 to 6.1 μg L− 1. The highest levels were measured in a site with no anthropogenic pollution, upstream from the industrial area. This observation suggests that UD is mainly linked to weathering; the elevated U concentrations originate from the naturally occurring radioactive materials (NORM) located in the Permian bedrock and no significant U pollution exists, at present, in the Riou-Mort watershed. This work demonstrates that spatial monitoring coupled with a long time-series are an essential prerequisite in assessment of spatiotemporal variations of UD, prior to a diagnostic of pollution in a small watershed. © 2007 Elsevier B.V. All rights reserved. Keywords: Watershed; Industrial impact; River; Uranium; Waste.
1. Introduction
⁎ Corresponding author. Tel.: +33 5 400 08880; fax: + 33 5 568 40848. E-mail address: [email protected] (H.-K. Saari). 0048-9697/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2007.04.030
A cadmium (Cd) contamination was discovered in the early 1980s in the Gironde Estuary (Fig. 1), with Cd concentrations measured in oysters, close to 100 μg g− 1 (dry 40 weight), about 10–50 times higher than those typically found in oysters from French non-industrial
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
305
Fig. 1. Location map of the Lot–Garonne River system and the Riou-Mort watershed shows the six main water sampling sites (white circles) and the four one-off sampling campaign sites (black circles). The locations of the Boisse Penchot and Temple sites are also shown in the map on the left.
areas, prohibiting mussel and oyster production within the estuary (Latouche, 1988; Claisse, 1989). The main source of this pollution is the Riou-Mort River in the upper part of the Lot-Garonne River system, which drains through the industrial basin of Decazeville (South-West France) (Fig. 1). In this basin opencast coal mining and coalfired Zn-ore treatment were practiced from 1842 to 1987, and coal-fired power production until 2001, conducting in a historic polymetallic pollution (Cd, Zn, Cu, Pb, Hg) (e.g. Latouche, 1992; Blanc et al., 1999; Audry et al., 2004; Schäfer et al., 2006). After 1987, remediation work was undertaken in the Riou-Mort watershed. The high-sulphide waste deposits rich in trace metals were collected and confined with clay-bearing mud and the acidic draining water systems were treating. Despite these treatments, the Lot River system continues to release heavy metals to the estuary (Schäfer et al., 2006). In the Joanis site, the outlet of the former mining and smelting areas (Fig. 1; site 4), with an 11-month time-series (November 2000–November 2001) of dissolved metals, total SO42− and alkalinity, Audry et al. (2005) have reported dissolved uranium (UD ) concentrations up to 1.1 μg L− 1. The presence of above UD concentration in this river was attributed to the alkaline reagents introduced as part of the remediation works in order to neutralise the acidified drainage waters. Uranium is a long-lived radionuclide, known for its chemical- and radio-toxicity. In surface continental waters and in sediments, U derives mainly from the presence of naturally occurring radioactive materials (NORM) in the earth's crust (average 1.7 mg kg− 1
(Wedepohl, 1995)). The U activities in rivers are primarily controlled by rates of weathering of the source rocks (e.g. Borole et al., 1982; Sarin et al., 1990; Chabaux et al., 2003). However human activities such as ore extraction and processing (e.g. Bolívar et al., 1995; Ketterer et al., 2000; Markich et al., 2002; Winde and van der Walt, 2004; Baborowski and Bozau, 2006; Mas et al., 2006), nuclear energy and weapon manufacturing (e.g. Jackson et al., 2005; Stubbs et al., 2006) and coal-fired power plants (e.g. Miljač and Križman, 1996; Papp et al., 2002) may lead to technologically enhanced naturally-occurring radioactive materials (TENORM) in the environment. To date, no studies have addressed in detail the variability of U within the Riou-Mort watershed and no studies have verified the risk for wider U pollution. According to Audry et al. (2005, 2006), in this small watershed uranium is mainly (98%) transported in the dissolved phase. Consequently, in this study we focus on UD to establish its origin and find out the possible environmental risks. The measurements of UD were carried out (2003–2004) on a daily basis on 4 sites, and once each month on a further 2 sites (Fig. 1), to investigate its spatial and temporal variations in relation to flow rates. With this long time-series and with supplementary one-off sampling campaign of 10 sites (including 6 monitoring programme sites and 4 additional sites (Fig. 1)) in 2005, the objective of this work is to determine whether the source of UD is; 1) as reported, the acid mine drainage (AMD) system treatments; 2) smelting waste deposits; or 3) NORM.
306
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
2. Materials and methods 2.1. Study area The Riou-Mort River is located in the South-western margin of the French Massif Central (Fig. 1). Before the confluence with the Lot River, one of the main effluents of the Gironde Estuary, the Riou-Mort River and its major tributaries, i.e. the Riou-Viou, Enne and Banel Rivers, flow through the industrial basin of Decazeville. The Riou-Mort River drains a watershed area of 165 km2 with a mean annual discharge of 2.2 m3 s− 1 (Table 1) (1968–2004; hydrological office of Toulouse, DIREN Midi-Pyrénées). The hydrology of the RiouMort River is characterised by a strong temporal variability of water flow with flash floods occurring in the space of 8–12 h. On the non-anthropogenically polluted upstream section of the Riou-Mort River (Fig. 1; site 1) the physical and chemical parameters pH, redox potential and electrical conductivity (EC) vary between 7.6–8.2, 100–300 mV and 30–500 μS cm− 1 respectively (Coynel, 2005). At the outlet of the Riou-Mort River (Fig. 1; site 4), the mean pH, redox potential, EC and SO42−concentration are 7.4, 223 mV, 817 μS cm− 1 and 384 mg L− 1 respectively (Audry et al., 2005). The enhanced ionic strength and dissolved SO42− content are attributed to the sulphide oxidation processes occurring in the smelting waste deposits. 2.2. Sampling strategy Four sampling sites on the Riou-Mort watershed were equipped with an automatic sampling system (Sigma 900P, Hach Lange) between June 2003 and September 2004 for the daily monitoring program (Fig. 1): (1) Firmi on the upstream Riou-Mort River before the industrial basin, which represents the nonanthropogenically polluted waters; (2) Moulin on the Riou-Viou before the confluence with the Enne and
Fig. 2. River flow diagram showing the smelting waste areas (SW) and the one-off sampling campaign sites (10), including the six monitoring programme sites (white circles) and the four additional sites (black circles). Dissolved U concentrations (μg L−1) and (234U/238U) activity ratios were measured once on each site.
Banel Rivers; (3) Usine, on the Riou-Viou before the confluence with the Riou-Mort, close to the former factories and (4) Joanis on the outlet of the Riou-Mort downstream of the mining and smelting area. In addition two sampling sites on the Lot River were sampled monthly (with an interval of 24–30 days): (5) Boisse Penchot, before the confluence between the Lot and Riou-Mort Rivers and (6) Temple, downstream of the Lot River. In April 2005, a one-off sampling campaign was undertaken to detail the UD concentrations of the main rivers of the Riou-Mort watershed. In addition to the aforementioned sites (Firmi, Moulin, Usine, Joanis, Boisse Penchot and Temple), four water sampling sites
Table 1 Data synthesis of the four main sampling sites of the Riou-Mort watershed and of the Temple site in the Lot River: river catchments, surface areas, mean annual discharges and dissolved U concentrations (μg L− 1; mean, minimum and maximum) Site
Moulin Usine Firmi Joanis Temple
Basin
Riou-Viou Riou-Viou + (Enne + Banel) Upstream Riou-Mort Downstream Riou-Mort Downstream Lot
A
Qa
Dissolved uranium
km2
m3 s− 1
μg L−1
68 97 21 165 11 254
0.88 1.12 0.33 2.2 180
Mean
Minimum
Maximum
0.2 0.7 3.8 1.0 0.4
0.02 0.1 0.2 0.1 0.2
0.4 1.4 6.1 2.6 0.7
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
were chosen, three on the Enne and Banel Rivers and one on the Riou-Mort River just before the confluence with the Riou-Viou River (Figs. 1 and 2). During the same campaign, representative smelting waste samples
307
Table 2 Uranium concentrations (mg kg− 1) and (234U/238U) and (226Ra/238U) activity ratios in three smelting waste deposits (Fig. 2) SW sample
U mg kg− 1
234
226
(1) (2) (3a) (3b)
6.8 ± 0.3 8.3 ± 0.3 13.4 ± 0.6 14.4 ± 0.6
1.04 ± 0.06 1.00 ± 0.04 0.99 ± 0.05 1.03 ± 0.05
1.16 ± 0.08 1.00 ± 0.05 1.03 ± 0.09 1.02 ± 0.03
U/238U
Ra/238U
Two different types of sample 3 (a and b) were analysed in order to determine the heterogeneity of these materials.
were collected from three deposit sites within the industrial basin of Decazeville (Fig. 2). The smelting waste deposits consist of high-sulphide residues, produced by Zn-ore treatment and coal-fired power production. 2.3. Analytical methods
Fig. 3. Spatial-temporal variations of dissolved U concentrations (μg L− 1) in the Riou-Viou (A), Riou-Mort (B) and Lot Rivers (D). Arrow on graph B underlines dry period at Firmi. Flow rates (m3 s− 1) at Joanis are shown in C. The grey stripes underline the variation of dissolved U concentrations during the high and low flow rate periods.
The water samples, of the monitoring program (June 2003–September 2004), were stocked in thoroughly decontaminated, acid pre-cleaned polypropylene bottles and were filtered afterward in the laboratory (0.2 μm Sartorius® polycarbonate filters). Dissolved U concentrations were measured by ICP-MS (Elan 5000, Perkin– Elmer) under standard conditions (Audry et al., 2004). The applied analytical methods were continuously quality checked by analysis of international certified reference materials (Thames water, SLRS4, CRM 320, NCS). Accuracy was within 10% of the certified values and the analytical error better than 5% (r.s.d.) for concentrations ten times higher than detection limits. The additional water samples (10 L) collected in the one-off campaign 2005 were filtered (Millipore membrane: diameter 142 mm, 0.45 μm pore size) immediately after sampling. The 0.45 μm pore size filters were used to facility the filtration of large volumes of water. In the Riou-Mort watershed, 90% of UD has been identified in colloids smaller than 0.02 μm (Coynel et al., in press). Thus the filters can be compared in terms of dissolved and “truly dissolved” species. Uranium was subsequently purified by anionic exchanges and measured by αcounting as described in Schmidt and Reyss (2000). The α-counting was used for these samples to determine (234U/238 U) activity ratio (AR). The mean analytical errors for UD concentrations and for (234U/238U) AR are ± 8% and ± 13% respectively. About 300 mg of smelting waste of each sample, previously oven dried at 60 °C, were spiked with yield monitors (232 U), before being digested in mixtures of HCl, HNO3, HClO4 and HF (in volume ratio of 1:2:2:2 respectively). Uranium in smelting waste samples was
308
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
purified and measured by α-counting as described above. The dried smelting waste samples were additionally measured using a low background-high efficiency well type γ-counter (Schmidt et al., 2001), to verify secular equilibrium within the 238 U decay series.
which feed the river, flow over a greater period of time in the aquifers and contain more dissolved elements. On the other hand, during the period with high flow rates, groundwater inputs are diluted with rainwater poor in U, revealing lower UD concentrations in the river water. 3.2. Dilution of dissolved U within the watershed
3. Results and discussion 3.1. Spatiotemporal variations of dissolved U The monitoring program (June 2003–September 2004) has permitted the development of a long-term record of UD, for a large range of flow rates within the Riou-Mort watershed (Fig. 3 and Table 1). The UD concentrations in the Riou-Mort watershed appear to behave differently from the Lot River. In the Lot River (Boisse Penchot and Temple) the UD concentrations are almost constant (Fig. 3D). On the opposite, within the Riou-Mort watershed, UD presents a high variability (Fig. 3A and B). The four sampling sites in the Riou-Mort watershed can be separated into two groups. The first group consists of Moulin and Usine on the Riou-Viou River. Both sites present comparably low (0.02–1.4 μg L− 1) and little fluctuation levels of UD (Fig. 3A and Table 1). A peculiar pattern is the anti-correlated variations of UD concentrations. At Moulin, highest concentrations were measured from November until June. In contrast, at the same time the lowest concentrations were measured at Usine site. The confluence of the Banel, Enne and RiouViou Rivers, which all drain different lithologies, and display variation of flow rates, could provide an explanation of this feature. The second group consists of Firmi and Joanis sites on the Riou-Mort River. The UD vary considerably in phase from 0.1 to 6.1 μg L− 1 (Fig. 3B). These variations depend on the flow rate. The maximum UD levels in Firmi and Joanis sites were measured in November 2003, when the flow rates varied between 0.22–0.26 m3 s− 1. On the contrary, during the high flow period (Q up to 107 m3 s− 1) the UD concentrations decreased to the minimum levels. The small tributaries with low discharge, like the RiouMort River (Table 1), are more susceptible to register changes in weathering or dilution by rain for example. The Firmi station on the Riou-Mort is fed by the groundwater inputs and the river is occasionally dry, as in July–November 2003 (Fig. 3B). During this dry period the UD concentrations at Joanis are relatively stable and close to values reported for Usine, which demonstrates the importance of the inputs from Firmi. During the period with low flow rates, the groundwaters
The long time-series seems to indicate a dilution effect of UD concentrations within the watershed. The one-off sampling campaign in 2005, including all the rivers of this system (Enne, Banel) was undertaken to test this hypothesis (Fig. 2). Flow rates during this campaign were near the mean annual discharges and thus the UD concentrations represent the values of the average hydrological conditions. The samples from one-off campaign were measured by α-counting to determinate the AR of 238U and its greatgranddaughter 234U, which is a powerful tool in the comprehension of weathering processes and U river transport (Chabaux et al., 2003, and references therein). (234U/238U) AR varies due to the lithological nature of the formations constituting the watersheds (e.g. Sarin et al., 1990; Plater et al., 1992). (234U/238U) AR is in secular equilibrium in rocks older than a few million years. However, the natural waters have typically (234U/238U) AR significantly greater than 1.0, due to the physical process of alpha-recoil during the energetic decay of 238U (Osmond and Ivanovich, 1992). Although the low precision of our (234U/238U) AR precludes further interpretation, there are noticeable differences of (234U/238U) AR throughout the watershed (Fig. 2). These differences could be explained by the different lithologies by dilution with mixing processes within the basin. The maximum UD concentrations, 1.6–1.7 μg L− 1, are observed on the upstream of the Riou-Mort River before the confluence with the Riou-Viou River. These concentrations are almost five times the concentrations measured in the Riou-Viou and Banel Rivers (i.e. 0.3 μg L− 1). The Enne River differs from these two latter rivers with slightly enhanced values (i.e. 0.9 μg L− 1), which may account for the anti-correlative variation of UD in Moulin and Usine (Fig. 3A). The UD concentrations in all four rivers are diluted with mixing processes after the confluences (in agreement with measured flow rates). For example, the Riou-Mort River waters (2.37 m3 s− 1) are mixed with the Riou-Viou waters (1.35 m3 s− 1) decreasing the UD concentrations in the Riou-Mort River by a factor of ∼ 2. The results of the specific experiment confirm the dilution effect of UD concentration within the watershed, already evidenced by the monitoring program.
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
3.3. U in smelting waste The U concentrations in the smelting waste deposits range between 6.8 and 14.4 mg kg− 1 (Table 2). In mining district of Mansfeld (Germany), the measured U concentrations in the dumps are as high as 40–150 mg kg− 1 (Baborowski and Bozau, 2006), which have lead to U contamination. Typical U concentrations in waste of coal power production are 1.6–3.2 mg kg− 1 (UNESCEAR, 2000). Thus, the measured U concentrations in the smelting waste deposits in Decazeville are slightly elevated, but neither the Zn ore extraction activities nor the coal power production, have not resulted in environmentally hazardous TENORM. Activity ratios (234U/238U) and (226Ra/238U) are near unity for all the smelting waste samples, indicating equilibrium within the U-decay series (Table 2). The equilibrium appears in materials which are not disturbed by chemical and physical processes. Therefore there is no evidence of significant weathering effects of the smelting waste deposits. 3.4. Source of U Both sampling programs show that the source of the UD is located upstream of the Riou-Mort River and is not related to the occurrence of smelting wastes. The groundwaters on the upstream of the Riou-Mort River, before combining with the river waters, are flowing in Permian bedrock, which comprises the highly weathered Permian hematite-rich conglomerates, sandstones and mudstones. The limit of these Permian layers is located at Firmi. The upstream sections of the Enne and Banel Rivers flow on the carboniferous and metamorphic formations. Thus, the Permian bedrock is probably the source for the U concentrations. In the southern part of the Massif Central, France, the distribution of U ores in the Rouergue area is related to the unconformity of the Permian–Carboniferous sedimentary terrains (Lévêque et al., 1988). The radioactive anomalies in Permian layers have been found also, for example, in the River Saale area in Germany (Baborowski and Bozau, 2006). 3.5. Statement about the potential environmental risks In the Riou-Mort basin, the monitoring programme sites Firmi, Usine and Joanis have the mean U D concentrations (Table 1) higher than the global average river water UD concentration, 0.3 μg L− 1 (Palmer and Edmond, 1993; Windom et al., 2000; Schmidt, 2004). However, this global value hides a wide variability of UD concentrations in natural waters (from 0.01 to 100 μg L− 1 (Osmond and Ivanovich, 1992)). In addition highly
309
polluted groundwaters and river waters can have UD concentration as high as 240–540 μg L− 1 (Winde and van der Walt, 2004) and 56 μg L− 1 (Markich et al., 2002) respectively. For drinking water, WHO guidelines (WHO, 2007) propose a limit of 15–30 μg L− 1. The maximum UD concentrations in the Riou-Mort River do not exceed this limit. The mean UD concentration measured in the Lot River after the confluence with the Riou-Mort River is twice lower than at Joanis site in the Riou-Mort. Due to the ratio of flow rates, the Lot waters dilute the UD input from the Riou-Mort to mean riverine levels (Table 1). 4. Conclusion The purpose of this work was to elucidate the origin of the high levels of UD in the Riou-Mort watershed and in particular to clarify the potential effects of former industrial activities. The strategy was to privilege spatial monitoring coupled with a long time-series of UD at key sites across the basin, to investigate the spatial and temporal variations within the watershed. The main conclusions are: 1) The highest UD concentration (6.1 μg L− 1) was measured at the non-anthropogenically polluted Firmi site on the upstream section of the Riou-Mort River. This confirms that the acid mine drainage system treatments have not affected the Riou-Mort River waters by raising the UD concentrations at Joanis, on the downstream section of the Riou-Mort River. The lower and more stable UD concentrations in the Riou-Viou River show that neither this tributary nor smelting waste deposits are significant sources of UD. The UD concentrations at Joanis originate from Firmi inputs, where the relatively high UD concentrations are due to leaching of NORM in Permian bedrock. Contrary to expectations, the Zn ore extraction activities have not brought out any significant U-pollution in the Riou-Mort watershed. 2) Extensive industrial activities throughout the past centuries or decades have led to the formation of numerous large, heavily contaminated sites throughout the world. This example clearly illustrates that one prerequisite to suitable remediation is the monitoring of the appropriate contaminants. In polluted environments, a good knowledge of the natural background is essential. Human activities do not always provide the explanation, especially for ubiquitous elements such as U. This study reveal that more expansive and rigorous temporal sampling is necessary for rivers. Limited sampling may not capture the spatiotemporal variations in U concentrations, or of other elements, and may therefore, be a cause of error in any conclusions concerning the pollution diagnostic. Special care is
310
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
required in studies of small watershed of limited surface area (10–1000 km2) and flow discharges. Mean U concentrations are mainly driven by the weathering, but temporal changes are expected due to local factors such as meteorological conditions, groundwater discharge, and geological parameters. Acknowledgements This work was supported by the Agence de l'Eau Adour-Garonne, the GIS-Ecobag, the INSU — ECODYN and ACI JC ARTTE programs. We thank E. Maneux, C. Bossy, G. Lavaux and J.P. Lissalde for their contribution to field and laboratory work. The suggestions kindly provided by William Fletcher and Liam Goodes greatly improved the spelling of this manuscript. References Audry S, Blanc G, Schäfer J. Cadmium transport in the Lot–Garonne River system (France) — temporal variability and a model for flux estimation. Sci Total Environ 2004;319:197–213. Audry S, Blanc G, Schäfer J. The Impact of sulphide oxidation on dissolved metal (Cd, Zn, Cu, Cr, Co, Ni, U) inputs into the Lot– Garonne fluvial system (France). Appl Geochem 2005;20:919–31. Audry S, Blanc G, Schäfer J. Solid state partitioning of trace metals in suspended particulate matter from a river system affected by smelting-waste drainage. Sci Total Environ 2006;363:216–36. Baborowski M, Bozau E. Impact of former mining activities on the uranium distribution in the River Saale (Germany). Appl Geochem 2006;21:1073–82. Blanc G, Lapaquellerie Y, Maillet N, Anschutz P. A cadmium budget for the Lot–Garonne fluvial system (France). Hydrobiol 1999;410:331–41. Bolívar JP, García-Tenorio R, García-León M. Enhancement of natural radioactivity in soils and salt-marshes surrounding a non-nuclear industrial complex. Sci Total Environ 1995;173–174:125–36. Borole DV, Krishnaswami S, Somayajulu BLK. Uranium isotopes in rivers, estuaries and adjacent coastal sediments of western India: their weathering, transport and oceanic budget. Geochim Cosmochim Acta 1982;46:125–37. Chabaux F, Riotte J, Dequincey O. U–Th–Ra fractionation during weathering and river transport. Rev Mineral Geochem 2003;52: 533–76. Claisse D. Chemical contamination of the French coasts. The results of a ten year mussel watch. Mar Pollut Bull 1989;20:523–8. Coynel A. Erosion mécanique des sols et transferts géochimiques dans le basin Adour-Garonne. Ph.D Thesis, Univ. Bordeaux1. 2005. Coynel A, Schäfer J, Dabrin A, Girardot N, Blanc G. Groundwater contributions to metal transport in a small river affected by mining and smelting waste. Water Res, in press. Jackson BP, Ranville JF, Bertsch PM, Sowder AG. Characterization of colloidal and humic-bound Ni and U in the “dissolved” fraction of contaminated sediment extracts. Environ Sci Technol 2005;39: 2478–85. Ketterer ME, Wetzel WC, Layman RR, Matisoff G, Bonniwell EC. Isotopic studies of sources of uranium in sediments of the Ashtabula River, Ohio, U.S.A. Environ Sci Technol 2000;34: 966–72.
Latouche C. Cadmium pollution in the Gironde estuary. Bulletin de l'Institut Géologique du Bassin d'Aquitaine. Bordeaux 1988;44: 15–21. Latouche C. La pollution par le cadmium des huîtres sauvages de l'Estuaire de la Gironde. Origine mécanismes responsables de la fixation du Cadmium. Ichtyophysiol Acta 1992;15:139–52. Lévêque MH, Lancelot JR, George E. The Bertholène uranium deposit — mineralogical characteristics and U–Pb dating of the primary U mineralization and its subsequent remobilization: consequences upon the evolution of the U deposits of the Massif Central France. Chem Geol 1988;69:147–63. Markich SJ, Jeffree RA, Burke PT. Freshwater bivalve shells as archival indicators of metal pollution from a copper–uranium mine in Tropical Northern Australia. Environ Sci Technol 2002;36:821–32. Mas JL, San Miguel EG, Bolívar JP, Vaca F, Pérez-Moreno JP. An assay on the effect of preliminary restoration tasks applied to a large TENORM wastes disposal in the south-west of Spain. Sci Total Environ 2006;364:55–66. Miljač L, Križman M. Radioactive contamination of surface waters from a fly-ash depository at Velenje (Slovenia). Environ Int 1996;22:339–45. Osmond JK, Ivanovich M. Uranium-series mobilization and surface hydrology. In: Ivanovich M, Harmon RS, editors. Uranium Series Disequilibrium 2nd ed.; 1992. p. 259–89. Palmer MR, Edmond JM. Uranium in river water. Geochim Cosmochim Acta 1993;57:4947–55. Papp Z, Dezsö Z, Daróczy S. Significant radioactive contamination of soil around a coal-fired power plant. J Environ Radioact 2002;59:191–205. Plater AJ, Ivanovich M, Dugdale RE. Uranium series disequilibrium in river sediments and waters: the significance of anomalous activity ratios. Appl Geochem 1992;7:101–10. Sarin MM, Krishnaswami S, Somayajulu BLK, Moore WS. Chemistry of uranium thorium and radium isotopes in the Ganga– Brahmaputra river system: weathering processes and fluxes to the Bay of Bengal. Geochim Cosmochim Acta 1990;54:1387–96. Schäfer J, Blanc G, Audry S, Cossa D, Bossy C. Mercury in the Lot– Garonne River system (France): sources fluxes and anthropogenic component. Appl Geochem 2006;21:515–27. Schmidt S, Reyss JL. Improvement in radiochemical and γ-counting procedures for the determination of 234Th in seawater. Radiochim Acta 2000;88:459–63. Schmidt S, de Stigter HC, van Weering TCE. Enhanced short-term sediment deposition within the Nazare Canyon North-East Atlantic. Mar Geol 2001;173:55–67. Schmidt S. Investigation of dissolved uranium content in the watershed of the Seine River (France). J Environ Radioact 2004;78:1–10. Stubbs JE, Elbert DC, Veblen DR, Zhu C. Electron microbeam investigation of uranium-contaminated soils from Oak Ridge TN USA. Environ Sci Technol 2006;40:2108–13. UNESCEAR. Report of the United Nations Scientific Committee on the Effects of Atomic Radiation. Sources and effects of ionizing radiation; 2000. Annex B. Wedepohl KH. The composition of the continental crust. Geochim Cosmochim Acta 1995;59:1217–32. Who guidelines, Mars 2007: http://www.who.int/water_sanitation_ health/dwq/chemicals/uranium290605.pdf. Winde F, van der Walt IJ. The significance of groundwater–stream interactions and fluctuating stream chemistry on waterborne uranium contamination of streams—a case study from a gold mining site in South Africa. J Hydrol 2004;287:178–394. Windom H, Smith R, Niencheski F, Alexander C. Uranium in rivers and estuaries of globally diverse smaller watershed. Mar Chem 2000;68:307–21.
Potential impact of former Zn ore extraction activities on dissolved uranium distribution in the Riou-Mort watershed (France) Hanna-Kaisa Saari ⁎, Sabine Schmidt, Alexandra Coynel, Stéphanie Huguet, Jörg Schäfer, Gérard Blanc UMR CNRS 5805 “Environnements et Paléoenvironnements Océaniques”, Université Bordeaux 1, Avenue des Facultés, 33405 Talence Cedex, France Received 27 November 2006; received in revised form 17 April 2007; accepted 25 April 2007 Available online 4 June 2007
Abstract The industrial basin of Decazeville (Riou-Mort watershed, South-West France) is well known for its heavy metal pollution and its subsequent environmental effects on the Lot-Garonne River system. The source of this pollution is the Riou-Mort River, which drains smelting waste areas. A first survey after remediation works has revealed elevated dissolved uranium (UD) concentrations in the outlet of the Riou-Mort River. The objective of this research is to identify the origin of UD in the Riou-Mort watershed and to evaluate the impact of industrial activities on this element. Uranium was measured at 10 water sampling sites, located upstream and downstream the industrial basin, and in three smelting waste deposits. Uranium concentrations in the smelting waste deposits reach up to 14.4 mg kg− 1 and (234U/238U) activity ratios (AR) are near unity. Dissolved U concentrations in the Riou-Mort River and its main tributaries ranges over two orders of magnitude from 0.02 to 6.1 μg L− 1. The highest levels were measured in a site with no anthropogenic pollution, upstream from the industrial area. This observation suggests that UD is mainly linked to weathering; the elevated U concentrations originate from the naturally occurring radioactive materials (NORM) located in the Permian bedrock and no significant U pollution exists, at present, in the Riou-Mort watershed. This work demonstrates that spatial monitoring coupled with a long time-series are an essential prerequisite in assessment of spatiotemporal variations of UD, prior to a diagnostic of pollution in a small watershed. © 2007 Elsevier B.V. All rights reserved. Keywords: Watershed; Industrial impact; River; Uranium; Waste.
1. Introduction
⁎ Corresponding author. Tel.: +33 5 400 08880; fax: + 33 5 568 40848. E-mail address: [email protected] (H.-K. Saari). 0048-9697/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2007.04.030
A cadmium (Cd) contamination was discovered in the early 1980s in the Gironde Estuary (Fig. 1), with Cd concentrations measured in oysters, close to 100 μg g− 1 (dry 40 weight), about 10–50 times higher than those typically found in oysters from French non-industrial
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
305
Fig. 1. Location map of the Lot–Garonne River system and the Riou-Mort watershed shows the six main water sampling sites (white circles) and the four one-off sampling campaign sites (black circles). The locations of the Boisse Penchot and Temple sites are also shown in the map on the left.
areas, prohibiting mussel and oyster production within the estuary (Latouche, 1988; Claisse, 1989). The main source of this pollution is the Riou-Mort River in the upper part of the Lot-Garonne River system, which drains through the industrial basin of Decazeville (South-West France) (Fig. 1). In this basin opencast coal mining and coalfired Zn-ore treatment were practiced from 1842 to 1987, and coal-fired power production until 2001, conducting in a historic polymetallic pollution (Cd, Zn, Cu, Pb, Hg) (e.g. Latouche, 1992; Blanc et al., 1999; Audry et al., 2004; Schäfer et al., 2006). After 1987, remediation work was undertaken in the Riou-Mort watershed. The high-sulphide waste deposits rich in trace metals were collected and confined with clay-bearing mud and the acidic draining water systems were treating. Despite these treatments, the Lot River system continues to release heavy metals to the estuary (Schäfer et al., 2006). In the Joanis site, the outlet of the former mining and smelting areas (Fig. 1; site 4), with an 11-month time-series (November 2000–November 2001) of dissolved metals, total SO42− and alkalinity, Audry et al. (2005) have reported dissolved uranium (UD ) concentrations up to 1.1 μg L− 1. The presence of above UD concentration in this river was attributed to the alkaline reagents introduced as part of the remediation works in order to neutralise the acidified drainage waters. Uranium is a long-lived radionuclide, known for its chemical- and radio-toxicity. In surface continental waters and in sediments, U derives mainly from the presence of naturally occurring radioactive materials (NORM) in the earth's crust (average 1.7 mg kg− 1
(Wedepohl, 1995)). The U activities in rivers are primarily controlled by rates of weathering of the source rocks (e.g. Borole et al., 1982; Sarin et al., 1990; Chabaux et al., 2003). However human activities such as ore extraction and processing (e.g. Bolívar et al., 1995; Ketterer et al., 2000; Markich et al., 2002; Winde and van der Walt, 2004; Baborowski and Bozau, 2006; Mas et al., 2006), nuclear energy and weapon manufacturing (e.g. Jackson et al., 2005; Stubbs et al., 2006) and coal-fired power plants (e.g. Miljač and Križman, 1996; Papp et al., 2002) may lead to technologically enhanced naturally-occurring radioactive materials (TENORM) in the environment. To date, no studies have addressed in detail the variability of U within the Riou-Mort watershed and no studies have verified the risk for wider U pollution. According to Audry et al. (2005, 2006), in this small watershed uranium is mainly (98%) transported in the dissolved phase. Consequently, in this study we focus on UD to establish its origin and find out the possible environmental risks. The measurements of UD were carried out (2003–2004) on a daily basis on 4 sites, and once each month on a further 2 sites (Fig. 1), to investigate its spatial and temporal variations in relation to flow rates. With this long time-series and with supplementary one-off sampling campaign of 10 sites (including 6 monitoring programme sites and 4 additional sites (Fig. 1)) in 2005, the objective of this work is to determine whether the source of UD is; 1) as reported, the acid mine drainage (AMD) system treatments; 2) smelting waste deposits; or 3) NORM.
306
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
2. Materials and methods 2.1. Study area The Riou-Mort River is located in the South-western margin of the French Massif Central (Fig. 1). Before the confluence with the Lot River, one of the main effluents of the Gironde Estuary, the Riou-Mort River and its major tributaries, i.e. the Riou-Viou, Enne and Banel Rivers, flow through the industrial basin of Decazeville. The Riou-Mort River drains a watershed area of 165 km2 with a mean annual discharge of 2.2 m3 s− 1 (Table 1) (1968–2004; hydrological office of Toulouse, DIREN Midi-Pyrénées). The hydrology of the RiouMort River is characterised by a strong temporal variability of water flow with flash floods occurring in the space of 8–12 h. On the non-anthropogenically polluted upstream section of the Riou-Mort River (Fig. 1; site 1) the physical and chemical parameters pH, redox potential and electrical conductivity (EC) vary between 7.6–8.2, 100–300 mV and 30–500 μS cm− 1 respectively (Coynel, 2005). At the outlet of the Riou-Mort River (Fig. 1; site 4), the mean pH, redox potential, EC and SO42−concentration are 7.4, 223 mV, 817 μS cm− 1 and 384 mg L− 1 respectively (Audry et al., 2005). The enhanced ionic strength and dissolved SO42− content are attributed to the sulphide oxidation processes occurring in the smelting waste deposits. 2.2. Sampling strategy Four sampling sites on the Riou-Mort watershed were equipped with an automatic sampling system (Sigma 900P, Hach Lange) between June 2003 and September 2004 for the daily monitoring program (Fig. 1): (1) Firmi on the upstream Riou-Mort River before the industrial basin, which represents the nonanthropogenically polluted waters; (2) Moulin on the Riou-Viou before the confluence with the Enne and
Fig. 2. River flow diagram showing the smelting waste areas (SW) and the one-off sampling campaign sites (10), including the six monitoring programme sites (white circles) and the four additional sites (black circles). Dissolved U concentrations (μg L−1) and (234U/238U) activity ratios were measured once on each site.
Banel Rivers; (3) Usine, on the Riou-Viou before the confluence with the Riou-Mort, close to the former factories and (4) Joanis on the outlet of the Riou-Mort downstream of the mining and smelting area. In addition two sampling sites on the Lot River were sampled monthly (with an interval of 24–30 days): (5) Boisse Penchot, before the confluence between the Lot and Riou-Mort Rivers and (6) Temple, downstream of the Lot River. In April 2005, a one-off sampling campaign was undertaken to detail the UD concentrations of the main rivers of the Riou-Mort watershed. In addition to the aforementioned sites (Firmi, Moulin, Usine, Joanis, Boisse Penchot and Temple), four water sampling sites
Table 1 Data synthesis of the four main sampling sites of the Riou-Mort watershed and of the Temple site in the Lot River: river catchments, surface areas, mean annual discharges and dissolved U concentrations (μg L− 1; mean, minimum and maximum) Site
Moulin Usine Firmi Joanis Temple
Basin
Riou-Viou Riou-Viou + (Enne + Banel) Upstream Riou-Mort Downstream Riou-Mort Downstream Lot
A
Qa
Dissolved uranium
km2
m3 s− 1
μg L−1
68 97 21 165 11 254
0.88 1.12 0.33 2.2 180
Mean
Minimum
Maximum
0.2 0.7 3.8 1.0 0.4
0.02 0.1 0.2 0.1 0.2
0.4 1.4 6.1 2.6 0.7
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
were chosen, three on the Enne and Banel Rivers and one on the Riou-Mort River just before the confluence with the Riou-Viou River (Figs. 1 and 2). During the same campaign, representative smelting waste samples
307
Table 2 Uranium concentrations (mg kg− 1) and (234U/238U) and (226Ra/238U) activity ratios in three smelting waste deposits (Fig. 2) SW sample
U mg kg− 1
234
226
(1) (2) (3a) (3b)
6.8 ± 0.3 8.3 ± 0.3 13.4 ± 0.6 14.4 ± 0.6
1.04 ± 0.06 1.00 ± 0.04 0.99 ± 0.05 1.03 ± 0.05
1.16 ± 0.08 1.00 ± 0.05 1.03 ± 0.09 1.02 ± 0.03
U/238U
Ra/238U
Two different types of sample 3 (a and b) were analysed in order to determine the heterogeneity of these materials.
were collected from three deposit sites within the industrial basin of Decazeville (Fig. 2). The smelting waste deposits consist of high-sulphide residues, produced by Zn-ore treatment and coal-fired power production. 2.3. Analytical methods
Fig. 3. Spatial-temporal variations of dissolved U concentrations (μg L− 1) in the Riou-Viou (A), Riou-Mort (B) and Lot Rivers (D). Arrow on graph B underlines dry period at Firmi. Flow rates (m3 s− 1) at Joanis are shown in C. The grey stripes underline the variation of dissolved U concentrations during the high and low flow rate periods.
The water samples, of the monitoring program (June 2003–September 2004), were stocked in thoroughly decontaminated, acid pre-cleaned polypropylene bottles and were filtered afterward in the laboratory (0.2 μm Sartorius® polycarbonate filters). Dissolved U concentrations were measured by ICP-MS (Elan 5000, Perkin– Elmer) under standard conditions (Audry et al., 2004). The applied analytical methods were continuously quality checked by analysis of international certified reference materials (Thames water, SLRS4, CRM 320, NCS). Accuracy was within 10% of the certified values and the analytical error better than 5% (r.s.d.) for concentrations ten times higher than detection limits. The additional water samples (10 L) collected in the one-off campaign 2005 were filtered (Millipore membrane: diameter 142 mm, 0.45 μm pore size) immediately after sampling. The 0.45 μm pore size filters were used to facility the filtration of large volumes of water. In the Riou-Mort watershed, 90% of UD has been identified in colloids smaller than 0.02 μm (Coynel et al., in press). Thus the filters can be compared in terms of dissolved and “truly dissolved” species. Uranium was subsequently purified by anionic exchanges and measured by αcounting as described in Schmidt and Reyss (2000). The α-counting was used for these samples to determine (234U/238 U) activity ratio (AR). The mean analytical errors for UD concentrations and for (234U/238U) AR are ± 8% and ± 13% respectively. About 300 mg of smelting waste of each sample, previously oven dried at 60 °C, were spiked with yield monitors (232 U), before being digested in mixtures of HCl, HNO3, HClO4 and HF (in volume ratio of 1:2:2:2 respectively). Uranium in smelting waste samples was
308
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
purified and measured by α-counting as described above. The dried smelting waste samples were additionally measured using a low background-high efficiency well type γ-counter (Schmidt et al., 2001), to verify secular equilibrium within the 238 U decay series.
which feed the river, flow over a greater period of time in the aquifers and contain more dissolved elements. On the other hand, during the period with high flow rates, groundwater inputs are diluted with rainwater poor in U, revealing lower UD concentrations in the river water. 3.2. Dilution of dissolved U within the watershed
3. Results and discussion 3.1. Spatiotemporal variations of dissolved U The monitoring program (June 2003–September 2004) has permitted the development of a long-term record of UD, for a large range of flow rates within the Riou-Mort watershed (Fig. 3 and Table 1). The UD concentrations in the Riou-Mort watershed appear to behave differently from the Lot River. In the Lot River (Boisse Penchot and Temple) the UD concentrations are almost constant (Fig. 3D). On the opposite, within the Riou-Mort watershed, UD presents a high variability (Fig. 3A and B). The four sampling sites in the Riou-Mort watershed can be separated into two groups. The first group consists of Moulin and Usine on the Riou-Viou River. Both sites present comparably low (0.02–1.4 μg L− 1) and little fluctuation levels of UD (Fig. 3A and Table 1). A peculiar pattern is the anti-correlated variations of UD concentrations. At Moulin, highest concentrations were measured from November until June. In contrast, at the same time the lowest concentrations were measured at Usine site. The confluence of the Banel, Enne and RiouViou Rivers, which all drain different lithologies, and display variation of flow rates, could provide an explanation of this feature. The second group consists of Firmi and Joanis sites on the Riou-Mort River. The UD vary considerably in phase from 0.1 to 6.1 μg L− 1 (Fig. 3B). These variations depend on the flow rate. The maximum UD levels in Firmi and Joanis sites were measured in November 2003, when the flow rates varied between 0.22–0.26 m3 s− 1. On the contrary, during the high flow period (Q up to 107 m3 s− 1) the UD concentrations decreased to the minimum levels. The small tributaries with low discharge, like the RiouMort River (Table 1), are more susceptible to register changes in weathering or dilution by rain for example. The Firmi station on the Riou-Mort is fed by the groundwater inputs and the river is occasionally dry, as in July–November 2003 (Fig. 3B). During this dry period the UD concentrations at Joanis are relatively stable and close to values reported for Usine, which demonstrates the importance of the inputs from Firmi. During the period with low flow rates, the groundwaters
The long time-series seems to indicate a dilution effect of UD concentrations within the watershed. The one-off sampling campaign in 2005, including all the rivers of this system (Enne, Banel) was undertaken to test this hypothesis (Fig. 2). Flow rates during this campaign were near the mean annual discharges and thus the UD concentrations represent the values of the average hydrological conditions. The samples from one-off campaign were measured by α-counting to determinate the AR of 238U and its greatgranddaughter 234U, which is a powerful tool in the comprehension of weathering processes and U river transport (Chabaux et al., 2003, and references therein). (234U/238U) AR varies due to the lithological nature of the formations constituting the watersheds (e.g. Sarin et al., 1990; Plater et al., 1992). (234U/238U) AR is in secular equilibrium in rocks older than a few million years. However, the natural waters have typically (234U/238U) AR significantly greater than 1.0, due to the physical process of alpha-recoil during the energetic decay of 238U (Osmond and Ivanovich, 1992). Although the low precision of our (234U/238U) AR precludes further interpretation, there are noticeable differences of (234U/238U) AR throughout the watershed (Fig. 2). These differences could be explained by the different lithologies by dilution with mixing processes within the basin. The maximum UD concentrations, 1.6–1.7 μg L− 1, are observed on the upstream of the Riou-Mort River before the confluence with the Riou-Viou River. These concentrations are almost five times the concentrations measured in the Riou-Viou and Banel Rivers (i.e. 0.3 μg L− 1). The Enne River differs from these two latter rivers with slightly enhanced values (i.e. 0.9 μg L− 1), which may account for the anti-correlative variation of UD in Moulin and Usine (Fig. 3A). The UD concentrations in all four rivers are diluted with mixing processes after the confluences (in agreement with measured flow rates). For example, the Riou-Mort River waters (2.37 m3 s− 1) are mixed with the Riou-Viou waters (1.35 m3 s− 1) decreasing the UD concentrations in the Riou-Mort River by a factor of ∼ 2. The results of the specific experiment confirm the dilution effect of UD concentration within the watershed, already evidenced by the monitoring program.
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
3.3. U in smelting waste The U concentrations in the smelting waste deposits range between 6.8 and 14.4 mg kg− 1 (Table 2). In mining district of Mansfeld (Germany), the measured U concentrations in the dumps are as high as 40–150 mg kg− 1 (Baborowski and Bozau, 2006), which have lead to U contamination. Typical U concentrations in waste of coal power production are 1.6–3.2 mg kg− 1 (UNESCEAR, 2000). Thus, the measured U concentrations in the smelting waste deposits in Decazeville are slightly elevated, but neither the Zn ore extraction activities nor the coal power production, have not resulted in environmentally hazardous TENORM. Activity ratios (234U/238U) and (226Ra/238U) are near unity for all the smelting waste samples, indicating equilibrium within the U-decay series (Table 2). The equilibrium appears in materials which are not disturbed by chemical and physical processes. Therefore there is no evidence of significant weathering effects of the smelting waste deposits. 3.4. Source of U Both sampling programs show that the source of the UD is located upstream of the Riou-Mort River and is not related to the occurrence of smelting wastes. The groundwaters on the upstream of the Riou-Mort River, before combining with the river waters, are flowing in Permian bedrock, which comprises the highly weathered Permian hematite-rich conglomerates, sandstones and mudstones. The limit of these Permian layers is located at Firmi. The upstream sections of the Enne and Banel Rivers flow on the carboniferous and metamorphic formations. Thus, the Permian bedrock is probably the source for the U concentrations. In the southern part of the Massif Central, France, the distribution of U ores in the Rouergue area is related to the unconformity of the Permian–Carboniferous sedimentary terrains (Lévêque et al., 1988). The radioactive anomalies in Permian layers have been found also, for example, in the River Saale area in Germany (Baborowski and Bozau, 2006). 3.5. Statement about the potential environmental risks In the Riou-Mort basin, the monitoring programme sites Firmi, Usine and Joanis have the mean U D concentrations (Table 1) higher than the global average river water UD concentration, 0.3 μg L− 1 (Palmer and Edmond, 1993; Windom et al., 2000; Schmidt, 2004). However, this global value hides a wide variability of UD concentrations in natural waters (from 0.01 to 100 μg L− 1 (Osmond and Ivanovich, 1992)). In addition highly
309
polluted groundwaters and river waters can have UD concentration as high as 240–540 μg L− 1 (Winde and van der Walt, 2004) and 56 μg L− 1 (Markich et al., 2002) respectively. For drinking water, WHO guidelines (WHO, 2007) propose a limit of 15–30 μg L− 1. The maximum UD concentrations in the Riou-Mort River do not exceed this limit. The mean UD concentration measured in the Lot River after the confluence with the Riou-Mort River is twice lower than at Joanis site in the Riou-Mort. Due to the ratio of flow rates, the Lot waters dilute the UD input from the Riou-Mort to mean riverine levels (Table 1). 4. Conclusion The purpose of this work was to elucidate the origin of the high levels of UD in the Riou-Mort watershed and in particular to clarify the potential effects of former industrial activities. The strategy was to privilege spatial monitoring coupled with a long time-series of UD at key sites across the basin, to investigate the spatial and temporal variations within the watershed. The main conclusions are: 1) The highest UD concentration (6.1 μg L− 1) was measured at the non-anthropogenically polluted Firmi site on the upstream section of the Riou-Mort River. This confirms that the acid mine drainage system treatments have not affected the Riou-Mort River waters by raising the UD concentrations at Joanis, on the downstream section of the Riou-Mort River. The lower and more stable UD concentrations in the Riou-Viou River show that neither this tributary nor smelting waste deposits are significant sources of UD. The UD concentrations at Joanis originate from Firmi inputs, where the relatively high UD concentrations are due to leaching of NORM in Permian bedrock. Contrary to expectations, the Zn ore extraction activities have not brought out any significant U-pollution in the Riou-Mort watershed. 2) Extensive industrial activities throughout the past centuries or decades have led to the formation of numerous large, heavily contaminated sites throughout the world. This example clearly illustrates that one prerequisite to suitable remediation is the monitoring of the appropriate contaminants. In polluted environments, a good knowledge of the natural background is essential. Human activities do not always provide the explanation, especially for ubiquitous elements such as U. This study reveal that more expansive and rigorous temporal sampling is necessary for rivers. Limited sampling may not capture the spatiotemporal variations in U concentrations, or of other elements, and may therefore, be a cause of error in any conclusions concerning the pollution diagnostic. Special care is
310
H.-K. Saari et al. / Science of the Total Environment 382 (2007) 304–310
required in studies of small watershed of limited surface area (10–1000 km2) and flow discharges. Mean U concentrations are mainly driven by the weathering, but temporal changes are expected due to local factors such as meteorological conditions, groundwater discharge, and geological parameters. Acknowledgements This work was supported by the Agence de l'Eau Adour-Garonne, the GIS-Ecobag, the INSU — ECODYN and ACI JC ARTTE programs. We thank E. Maneux, C. Bossy, G. Lavaux and J.P. Lissalde for their contribution to field and laboratory work. The suggestions kindly provided by William Fletcher and Liam Goodes greatly improved the spelling of this manuscript. References Audry S, Blanc G, Schäfer J. Cadmium transport in the Lot–Garonne River system (France) — temporal variability and a model for flux estimation. Sci Total Environ 2004;319:197–213. Audry S, Blanc G, Schäfer J. The Impact of sulphide oxidation on dissolved metal (Cd, Zn, Cu, Cr, Co, Ni, U) inputs into the Lot– Garonne fluvial system (France). Appl Geochem 2005;20:919–31. Audry S, Blanc G, Schäfer J. Solid state partitioning of trace metals in suspended particulate matter from a river system affected by smelting-waste drainage. Sci Total Environ 2006;363:216–36. Baborowski M, Bozau E. Impact of former mining activities on the uranium distribution in the River Saale (Germany). Appl Geochem 2006;21:1073–82. Blanc G, Lapaquellerie Y, Maillet N, Anschutz P. A cadmium budget for the Lot–Garonne fluvial system (France). Hydrobiol 1999;410:331–41. Bolívar JP, García-Tenorio R, García-León M. Enhancement of natural radioactivity in soils and salt-marshes surrounding a non-nuclear industrial complex. Sci Total Environ 1995;173–174:125–36. Borole DV, Krishnaswami S, Somayajulu BLK. Uranium isotopes in rivers, estuaries and adjacent coastal sediments of western India: their weathering, transport and oceanic budget. Geochim Cosmochim Acta 1982;46:125–37. Chabaux F, Riotte J, Dequincey O. U–Th–Ra fractionation during weathering and river transport. Rev Mineral Geochem 2003;52: 533–76. Claisse D. Chemical contamination of the French coasts. The results of a ten year mussel watch. Mar Pollut Bull 1989;20:523–8. Coynel A. Erosion mécanique des sols et transferts géochimiques dans le basin Adour-Garonne. Ph.D Thesis, Univ. Bordeaux1. 2005. Coynel A, Schäfer J, Dabrin A, Girardot N, Blanc G. Groundwater contributions to metal transport in a small river affected by mining and smelting waste. Water Res, in press. Jackson BP, Ranville JF, Bertsch PM, Sowder AG. Characterization of colloidal and humic-bound Ni and U in the “dissolved” fraction of contaminated sediment extracts. Environ Sci Technol 2005;39: 2478–85. Ketterer ME, Wetzel WC, Layman RR, Matisoff G, Bonniwell EC. Isotopic studies of sources of uranium in sediments of the Ashtabula River, Ohio, U.S.A. Environ Sci Technol 2000;34: 966–72.
Latouche C. Cadmium pollution in the Gironde estuary. Bulletin de l'Institut Géologique du Bassin d'Aquitaine. Bordeaux 1988;44: 15–21. Latouche C. La pollution par le cadmium des huîtres sauvages de l'Estuaire de la Gironde. Origine mécanismes responsables de la fixation du Cadmium. Ichtyophysiol Acta 1992;15:139–52. Lévêque MH, Lancelot JR, George E. The Bertholène uranium deposit — mineralogical characteristics and U–Pb dating of the primary U mineralization and its subsequent remobilization: consequences upon the evolution of the U deposits of the Massif Central France. Chem Geol 1988;69:147–63. Markich SJ, Jeffree RA, Burke PT. Freshwater bivalve shells as archival indicators of metal pollution from a copper–uranium mine in Tropical Northern Australia. Environ Sci Technol 2002;36:821–32. Mas JL, San Miguel EG, Bolívar JP, Vaca F, Pérez-Moreno JP. An assay on the effect of preliminary restoration tasks applied to a large TENORM wastes disposal in the south-west of Spain. Sci Total Environ 2006;364:55–66. Miljač L, Križman M. Radioactive contamination of surface waters from a fly-ash depository at Velenje (Slovenia). Environ Int 1996;22:339–45. Osmond JK, Ivanovich M. Uranium-series mobilization and surface hydrology. In: Ivanovich M, Harmon RS, editors. Uranium Series Disequilibrium 2nd ed.; 1992. p. 259–89. Palmer MR, Edmond JM. Uranium in river water. Geochim Cosmochim Acta 1993;57:4947–55. Papp Z, Dezsö Z, Daróczy S. Significant radioactive contamination of soil around a coal-fired power plant. J Environ Radioact 2002;59:191–205. Plater AJ, Ivanovich M, Dugdale RE. Uranium series disequilibrium in river sediments and waters: the significance of anomalous activity ratios. Appl Geochem 1992;7:101–10. Sarin MM, Krishnaswami S, Somayajulu BLK, Moore WS. Chemistry of uranium thorium and radium isotopes in the Ganga– Brahmaputra river system: weathering processes and fluxes to the Bay of Bengal. Geochim Cosmochim Acta 1990;54:1387–96. Schäfer J, Blanc G, Audry S, Cossa D, Bossy C. Mercury in the Lot– Garonne River system (France): sources fluxes and anthropogenic component. Appl Geochem 2006;21:515–27. Schmidt S, Reyss JL. Improvement in radiochemical and γ-counting procedures for the determination of 234Th in seawater. Radiochim Acta 2000;88:459–63. Schmidt S, de Stigter HC, van Weering TCE. Enhanced short-term sediment deposition within the Nazare Canyon North-East Atlantic. Mar Geol 2001;173:55–67. Schmidt S. Investigation of dissolved uranium content in the watershed of the Seine River (France). J Environ Radioact 2004;78:1–10. Stubbs JE, Elbert DC, Veblen DR, Zhu C. Electron microbeam investigation of uranium-contaminated soils from Oak Ridge TN USA. Environ Sci Technol 2006;40:2108–13. UNESCEAR. Report of the United Nations Scientific Committee on the Effects of Atomic Radiation. Sources and effects of ionizing radiation; 2000. Annex B. Wedepohl KH. The composition of the continental crust. Geochim Cosmochim Acta 1995;59:1217–32. Who guidelines, Mars 2007: http://www.who.int/water_sanitation_ health/dwq/chemicals/uranium290605.pdf. Winde F, van der Walt IJ. The significance of groundwater–stream interactions and fluctuating stream chemistry on waterborne uranium contamination of streams—a case study from a gold mining site in South Africa. J Hydrol 2004;287:178–394. Windom H, Smith R, Niencheski F, Alexander C. Uranium in rivers and estuaries of globally diverse smaller watershed. Mar Chem 2000;68:307–21.