PII: Wat. Res. Vol. 34, No. 11, pp. 2941±2950, 2000 7 2000 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain S0043-1354(99)00370-X 0043-1354/00/$ - see front matter www.elsevier.com/locate/watres RADIOECOLOGICAL STUDY OF AN ESTUARINE SYSTEM LOCATED IN THE SOUTH OF SPAIN J. P. BOLIVAR1*, R. GARCIA-TENORIO2 and F. VACA1 1 Dept. Fõ sica Aplicada, E.P.S. La RaÂbida, Universidad Huelva, 21819-Palos de la Fra, Huelva, Spain and 2Dept. Fõ sica Aplicada, E.T.S. Arquitectura, Universidad de Sevilla, Avda Reina Mercedes s/n, 41012, Sevilla, Spain (First received 1 December 1998; accepted in revised form 1 November 1999) AbstractÐA big fertilizer industrial complex and a vast extension of phosphogypsum piles (12 km2), sited in the estuary formed by the Odiel and Tinto river mouths (south-west of Spain), are producing an unambiguous radioactive impact in their sourrounding aquatic environment, through radionuclides from the U-series. The levels and distribution of radionuclides in the di€erent water phases of this estuarine system (®ltered and particulate fractions) have been determined. The analyses of radionuclide concentrations and activity ratios have provided us with interesting information with which to evaluate the extension, degree and routes of the radioactive impact, as well with the knowledge of the di€erent pathways followed for the radioactive contamination which disturbs this natural system. The obtained results indicate that the main pathway of radioactive contamination of the estuary is through the dissolution in its waters of the radionuclides released by the industrial activities and their later ®xation on the particulate materials. On the other hand, the tidal activity plays an important role in the transport and homogeneization along the estuary of the radioactivity released from the fertilizer plants. Additionally, the results obtained for the di€erent studied natural radionuclides (U-isotopes, Thisotopes and 210Po) in the ®ltered and particulate fractions have provided very rich information about their di€erent environmental behaviour in this aquatic system. 7 2000 Published by Elsevier Science Ltd. All rights reserved Key wordsÐestuary, phosphogypsum, natural radioactivity, contamination pathways, alpha-spectrometry, dissolution/suspended matter INTRODUCTION The Odiel and Tinto rivers are located in the southwest of Spain, discharging their waters into the Atlantic Ocean. Their mouths conform an estuarine tidal system that sourrounds a large industrial chemical complex where several plants are placed devoted to the production of phosphoric acid and phosphate fertilizers (Fig. 1). In these factories, phosphoric acid is obtained through the chemical treatment of phosphate rock, mainly formed by phosphorite mineral, which contains high levels of radionuclides from the uranium series. Additionally, in this chemical process a by-product called phosphogypsum (PG) is formed, which is generally stored in the surroundings of the factories. Because this by-product accumulates a signi®cant fraction of the radioactivity contained originally in the mineral, it can be concluded that these industries could produce an important radioactive impact on their *Author to whom all correspondence should be addressed. Tel.: +34-959350651; fax: +34-959350311; e-mail: bolivar@uhu.es nearby environment (Bolivar, 1995; Rutherford et al., 1994; Bolivar et al., 1995a, 1996a). At the time of the sampling, these factories were processing annually over 2  106 metric tons of phosphate rock, and they generate about 3  106 tons of PG. The PG stacks cover about 12 km2 on the salt marshes of Tinto River and therefore they generate a high radioactive impact on this marsh-land (Bolõ var, 1995; Bolivar et al., 1995b). On the contrary, the salt-marshes of the Odiel river are not physically perturbed and have been recently declared a natural reserve by the regional goverment. In previous works we have determined the radioactivity distribution in the phosphoric acid production process of the factories sited in the estuary (Bolivar et al., 1993, 1995a, 1996a). We concluded in these studies that more than 90% of Po and Ra originally present in the phosphate rock remains in the phosphogypsum, while the percentage of U is much lower (<20%). It is also known that the behaviour of 210Pb in the production processs of phosphoric acid used in the factories is quite similar to the 226Ra (Guimond and Hardin, 1989). The radioactive impact generated by this indus- 2941 2942 J. P. Bolivar et al. trial complex in the sediments and waters of this estuary was widely studied in previous works (Martõ nez-Aguirre et al., 1994; Garcia-LeoÂn et al., 1997), but these studies were devoted to evaluating this impact in those zones of the rivers quite close to the possible sources of radioactive contamination: the fertilizers industries and the PG piles. In this work, we extend these studies to more remote zones in the rivers, electing for that two areas (Fig. 1): the common channel of the Odiel and Tinto river (called Ria de Huelva) and the Estero Domingo Rubio (a rivulet that ends in the Tinto river, opposite the PG piles). Through this study we can evaluate more properly the extension and degree of contamination produced by the factories, and gain knowledge about the behaviour of several natural radionuclides in estuarine systems. The actual situation of the Ria de Huelva was conformed 30 years ago when an arti®cial dike was built (Fig. 1) for the construction of a new harbour, the water exchange rate of the rivers with the Atlantic Ocean being changed as a consequence. On the other hand, the water ¯ow that comes from the rivers is much smaller than the water volume exchanged with the ocean, a fact that can be deduced from the uniform values of pH (in the region of 6.5±7) and from the salinity determined in the waters of the estuary, which are quite similar to those found in coastal seawaters of this geographical area. On the contrary, the pH values in the waters of both rivers, upstream of the estuary, are very low (between 2 and 3) due to the in¯uence of some mining activities developed historically within their margins. Additionally, the study of the distribution of the di€erent radionuclides between the ®ltered and particulate phases, can provide interesting information about their behaviour in this aquatic system. It can also allow the rati®cation of the theory about the main contamination pathways of the estuary. MATERIALS AND METHODS The main criterion in the sampling strategy was to select a set of points distributed uniformly along the zone in study, which was the estuarine area more remote from the fertilizer industrial complex and the PG piles. A total of 13 sampling points were selected, as illustrated in Fig. 1. At each sampling location, two water samples were collected: one corresponding to high tide (code {H}) and the other one to low tide (code {L}). Eight sites (OT1±OT8) correspond to the Ria de Huelva, while only ®ve sites (DR1±DR5) were elected from the Estero Domingo Rubio rivulet. In both cases the numbering increases in the direction of the water ¯ow. About 20 L of water per sample was collected. Immediately after the collection, they were ®ltered using polycarbonate ®lters of 14.2 cm diameter and 0.2 mm pore size. After ®ltration, about 100 ml of concentrated HNO3 was added to each ®ltered fraction to establish its pH at around 1±2 and consequently to avoid the adsorption of the actinides on the container walls and the growth of micro-organisms. The code {W} was assigned to the ®ltered fraction and {F} to the particulate fraction. Alpha emitting radionuclides (210Po, U and Th isotopes) were measured in the di€erent phases by alpha-particle spectrometry using ion-implanted silicon detectors. The suspended matter determinations were done placing the ®lters in an ultrasonic bath containing hydrochloric acid for Fig. 1. Map of the Odiel and Tinto river mouths. The location of Huelva Town, fertilizer plants, phosphogypsum piles and the 13 sampling points are specially marked. Radioecological study of an estuarine system 30 min. Then, the ®lters were cleaned with nitric acid and the liberated particulate matter was dissolved applying an acid atmospheric digestion with aqua regia (3HCl:1HNO3). The dry residue was redissolved in 8 M nitric acid. Afterwards, the isolement of U, Th and Po was carried out through a sequential solvent extraction technique, intially developed by Holm and Fukai (1977) and slightly modi®ed for us. In the case of water samples, 1 L of the ®ltered water was taken for 210Po and U-isotopes analyses, while 5 L were necessary for the Th-isotopes studies. The Fe3+ carrier was added and actinides and Po were co-precipitated with Fe(OH)3 by adding drop-wise NH4OH until pH=8± 9. Then, the Fe(OH)3 precipitate, containing U, Th and Po, was isolated by ®ltration, re-dissolved in 10 ml 8 M HNO3 and the TBP method previously described for the suspended matter was applied. 210 Po was self-deposited onto silver discs from a 2 M HCl solution (El-Daoushy et al., 1991), while the ®nal solutions of U and Th were electroplated onto stainless steel discs by the method of Hallstadius (Hallstadius, 1984). For the total recovery determination, and at the beginning of the radiochemical process, known amounts of 208 Po,232U or 229Th were added into each sample as isotopic tracers. RESULTS AND DISCUSSION Table 1 shows the physical properties measured in all the water samples. It can be seen that the average temperature of the Ria de Huelva water at low tide (18.4 2 0.28C) is slightly higher than that obtained at high tide (17.6 2 0.18C). This fact indicates that the temperature of the oceanic water entering in the estuary was a little bit lower in comparison with the fresh-water coming from the rivers. In relation to the pH, a smooth increase of this 2943 value towards the Atlantic Ocean is observed in Ria de Huelva and in both tidal cycles. This trend is probably generated by the mixing of the oceanic waters (pH 0 8) with the river waters (pH 0 2.5), which is very low due to the mining activities developed in their basins since the 19th century, and the in¯uence of the industrial complex e‚uents. Although the river ¯ow water is very small in relation to the marine ¯ow, a di€erence in the order of magnitude of 5±6 in the H+ concentration can change the pH of the estuarine water a little bit, mainly formed by sea water. In Table 1, it is observed that pH 1 8 (typical of sea water) is found at high tide in the samples nearest to the Atlantic Ocean, ratifying the previous hypothesis. On the contrary, no special trends were observed in the pH values along the Domingo Rubio rivulet (around seven in all the samples). Finally, and concerning the values of the electrical conductivity, only to notice its constancy along the Ria de Huelva with values typical of surface sea waters, while this magnitude in the Domingo Rubio waters increases towards its mouth, showing this fact the mixing with waters from the estuary. In fact, it can be observed that the value of the electrical conductivity in sample 1 is typical of freshwater, while the value in sample 5 (collected in the mouth of the rivulet) is quite similar to the obtained ones in the Ria de Huelva channel. Radioactivity in the ®ltered fraction Ria de Huelva channel. The concentrations of Uisotopes, Th-isotopes and 210Po determined in the ®ltered waters of the Ria de Huelva at high and Table 1. Some physico-chemical properties of the water samples collected in the analysed zones of the Odiel and Tinto river estuary. {RH}=Ria de Huelva channel, {DR}=Domingo Rubio rivulet Sample High tide (RH) 1 2 3 4 5 6 7 8 Average Low tide (RH) 1 2 3 4 5 6 7 8 Average Global average (RH) DR 1 2 3 4 5 T (8C) pH Suspended matter (mg/l) Conductivity (mS/cm) 17.7 17.7 17.8 17.7 17.7 17.6 17.6 17.1 17.620.1 6.7 6.7 7.0 7.4 7.7 7.8 8.2 8.1 7.420.2 42.3 37.6 39.5 54.0 27.3 37.1 33.1 36.3 38.422.7 38.3 38.4 39.1 39.0 39.1 39.2 38.6 38.2 38.820.1 18.0 18.0 18.1 18.3 18.1 20.1 18.3 18.2 18.420.2 18.020.1 6.3 6.9 6.7 6.6 6.9 6.9 7.0 7.3 6.820.1 7.120.1 21.3 22.4 18.2 36.2 64.6 48.5 50.3 28.6 36.325.9 37.323.1 37.3 35.3 38.0 38.2 38.9 39.0 38.4 39.7 37.120.5 38.423.0 17.2 20.5 17.6 17.5 18.0 7.2 6.7 7.2 7.6 7.2 148 13.2 25.1 27.1 33.8 0.47 10.9 20.9 25.5 33.0 2944 J. P. Bolivar et al. Table 2. Activity concentrations (Bq/kg) measured in the ®ltered water fractions. {RH}=Ria de Huelva channel, {DR}=Domingo Rubio rivulet. N.M.=not measured; N.D.=not detected 238 U Sample High tide (RH) HW1 HW2 HW3 HW4 HW5 HW6 HW7 HW8 Average Low tide (RH) LW1 LW2 LW3 LW4 LW5 LW6 LW7 LW8 Average Global average DR W1 W2 W3 W4 W5 235 U 234 U 210 Po 232 Th 230 Th 228 Th 28.622.3 24.221.9 30.823.8 36.622.9 40.823.2 43.424.0 43.723.8 45.823.2 37.722.8 1.0320.3 0.920.2 N.M. 1.420.4 1.120.3 2.320.6 1.420.5 1.720.3 1.4020.17 28.922.3 26.022.0 41.824.7 49.523.7 43.523.4 48.224.3 51.924.4 45.423.2 41.923.4 10.921.0 5.220.6 2.1420.30 47.023.1 4.2020.77 5.3320.40 3.6320.26 2.0920.17 10.125.3 N.M. N.D. 0.1920.08 0.7920.15 0.6020.25 0.5920.08 0.5320.07 0.3420.06 0.5120.09 N.M. 2.4920.30 2.5220.32 1.9120.25 1.8920.47 3.0220.21 2.1920.17 2.1120.18 2.3020.15 N.M. 1.8320.25 1.3820.23 1.5320.22 2.2920.53 1.9420.15 1.7120.14 1.5620.15 1.7520.12 63.824.8 14.521.9 38.923.0 33.122.8 34.522.8 35.022.8 85.326.0 36.822.9 42.727.7 39.725.9 2.820.6 0.720.4 N.M. 1.820.4 1.220.3 1.720.4 3.220.6 1.620.4 1.8620.32 1.6320.19 61.624.7 15.522.0 43.623.3 37.223.1 41.823.3 39.023.1 147210 45.523.4 53.9214 47.927.7 22.721.6 1.2420.30 23.021.6 5.6320.58 2.5620.43 3.0720.45 11.121.1 3.1720.43 9.123.2 9.623.0 1.5420.29 N.M. 0.9020.13 0.3720.07 0.2520.06 0.1020.04 N.M. N.M. 0.6320.26 0.5620.12 5.6820.67 N.M. 2.0120.21 1.2020.13 1.9120.18 1.0120.12 N.M. N.M. 2.3620.85 2.3320.36 1.3020.26 N.M. 1.6920.19 2.2620.19 2.4220.22 2.0020.18 N.M. N.M. 1.9320.20 1.8320.11 9.820.7 8.120.6 21.921.3 29.822.6 30.822.5 0.6220.11 0.6620.11 1.0220.14 1.520.4 1.220.3 10.420.7 8.220.6 23.821.4 29.222.5 30.122.5 1.1820.13 2.4720.21 5.9120.48 5.5120.54 11.620.9 1.2720.22 1.1020.30 0.9520.16 2.1020.40 N.M. 2.1820.30 1.0220.28 2.3620.24 1.9620.39 N.M. 1.9820.28 5.420.79 2.3920.27 2.0920.40 N.M. low tide are shown in Table 2. From the obtained results, it is clear than the 234U and 238U activity concentrations are very uniform in solution during high tide, while there are two activity peaks during the low tide at sampling points 1 and 7 (Fig. 2), indicating the existence of some contamination ``bags'' along the channel with its origin probably in the contamination associated with the fertilizer activities. These ``bags'' are more easily detectable Fig. 2. 238 at low tide due to the smaller amount of clean sea water entering the estuary. The average values of the 238U concentrations in low and high tides are 43 2 7 mBq/l and 38 2 3 mBq/l respectively, in the range of the concentrations found in uncontaminated surface sea waters, 40 mBq/l (Ku et al., 1977; Borole et al., 1977). This fact indicates that except in some speci®c locations, the possible contamination of U- U activity concentrations in solution for the Ria de Huelva at high and low tide. Radioecological study of an estuarine system isotopes in the dissolved form is masked for the high mobility of these radionuclides and consequently for their relatively high values in clean sea waters. Concerning the 210Po concentrations, they are generally lower than the U-isotopes along the channel but generally higher than the normal values found in unperturbed estuarine waters, 2 mBq/l (Iyengar et al., 1981), indicating the possible 210Po impact of the fertilizer activities in this compartment. The average 210Po activity concentration along the channel, shown in Table 2, is 1023 mBq/ l with activity peaks up to 4723 mBq/l, and with a value in sample HW8 (high tide and therefore a sample with the highest marine in¯uence) of 2.1 2 0.2 mBq/l, which could be considered the background value for 210Po of this geographical area. This value was also obtained in another water sample collected in open sea, located 40 km from this estuary. On the other hand, the 232Th concentrations are two orders of magnitude smaller (average value 0.56 2 0.12 mBq/l) than those found for U-isotopes, indicating its tendency to be associated to the particulate matter (Koczy et al., 1957). The 232Th concentrations are also smaller than those of its daughter 228Th (average value of 1.8 2 0.1 mBq/l). This last fact can be explained by the higher solubility of Ra than Th in sea waters, and because the 228 Th concentrations in solution are mainly governed by its progenitor 228Ra due to its major halflife. Clear enhancements can also be observed in the concentrations of 230Th in relation to232Th (about ®ve times higher), which could be generated 2945 by the wastes from the fertilizer plants released into the estuary. Some of the activity ratios in the ®ltered waters along the channel, together with their average values, are compiled in Table 3 for high and low tidal cycles. The average 234U/238U activity ratio was 1.17 2 0.05, similar to the one found in unperturbed sea waters (Moore and Sackett, 1964; Borole et al., 1982), and re¯ecting an undetectable input of U from the factories to this area. On the contrary, the 230Th/232Th activity ratios are clearly higher than one, ranging from 2.4 to 9, and with an average value equal to 5.421.0, which shows the arti®cial input of 230Th because in unperturbed waters this ratio is around the unity. The average 230Th/234U activity ratio was 0.0532 0.006, very far from the secular equilibrium, showing the extremely low mobility of thorium in sea water in relation to uranium. However, this activity ratio is clearly higher than the value found in uncontaminated sea waters (R10ÿ2) (Scott, 1982) which con®rms the detectability of some anthropogenic inputs in solution of 230Th in these estuarine waters. Obviously the average 232Th/234U activity ratio is also very low (0.012 2 0.002) due to the strong anity of Th with the particulate matter and the high mobility of U. On the other hand, the average 210Po/234U activity ratio in the ®ltered waters along the estuary (0.2120.06), higher than the value of 0.05 found in unperturbed sea waters (Bacon and Elzerman, 1980), indicates, as was stated before commenting on the absolute 210Po concentrations, a detectable enrichment of 210Po in relation to U in these waters. Table 3. More signi®cant activity ratios determined in solution. N.M.=not measured Sample High tide (RH) HW1 HW2 HW3 HW4 HW5 HW6 HW7 HW8 Average Low tide (RH) LW1 LW2 LW3 LW4 LW5 LW6 LW7 LW8 Average Global average DR W1 W2 W3 W4 W5 230 Th/ 232 Th 234 U/238U 228 Th/232Th 232 Th/234U (10ÿ3) 230 Th/234U (10ÿ2) 210 Po/234U (10ÿ1) 210 Po/230Th N.M. N.M. 1325 2.420.5 3.221.4 5.220.7 4.120.6 6.121.2 5.721.4 1.0120.07 1.0720.08 1.3620.17 1.3520.09 1.0720.07 1.1120.09 1.1920.09 0.9920.05 1.1420.05 N.M. N.M. 723 1.920.4 3.821.7 3.320.5 3.220.4 4.520.9 3.920.8 N.M. N.M. 4.521.9 16.023.2 13.821.0 12.221.9 10.221.5 7.521.4 10.721.7 N.M. 9.621.3 6.020.9 3.920.6 4.321.1 6.320.7 4.220.5 4.620.5 5.620.7 3.820.4 2.020.3 0.5120.09 9.520.9 0.9620.18 1.1120.12 0.7020.07 0.4620.05 2.421.1 N.M. 2.120.3 0.8520.15 24.623 2.2220.68 1.7620.17 1.6620.17 0.9920.11 4.923.3 3.720.7 N.M. 2.220.3 3.320.6 7.521.8 923 N.M. N.M. 5.121.3 5.421.0 0.9720.06 1.0620.16 1.1220.07 1.1220.09 1.2120.08 1.1220.08 1.7220.07 1.2420.08 1.2020.08 1.1720.05 0.8420.22 N.M. 1.920.3 6.121.2 9.522.3 1826 N.M. N.M. 7.323.0 5.421.5 2525 N.M. 2023 9.922.2 6.021.5 2.621.0 N.M. N.M. 12.724.2 11.622.0 9.221.3 N.M. 4.620.5 3.220.4 4.620.5 2.620.4 N.M. N.M. 4.821.1 5.320.6 3.720.4 0.8020.20 5.320.5 1.5220.19 0.6220.10 0.7920.12 0.7620.09 0.7020.12 1.820.5 2.120.6 4.020.3 N.M. 1121 4.720.6 1.420.2 3.020.5 N.M. N.M. 4.821.6 4.922.0 1.720.3 0.920.3 2.520.4 0.920.2 N.M. 1.0720.06 1.0120.06 1.0920.04 0.9820.08 0.9820.07 1.620.3 4.921.4 2.520.4 0.9720.24 N.M. 122221 134230 4027 74215 N.M. 2123 1223 9.921.1 6.721.4 N.M. 1.1320.12 3.020.3 2.520.2 1.920.2 3.920.4 0.5420.11 2.420.7 2.520.3 2.820.6 N.M. 2946 J. P. Bolivar et al. This activity ratio reaches the highest values also in the samples nearest to the contamination source. Then we can conclude that in general there are no signi®cant di€erences between the average concentrations and activity ratios at high and low tide, considering the experimental uncertainties, and from these data we can establish the following order in activity concentrations (®ltered fraction) along the channel: [U] > [210Po] > [230Th] > [228Th] > [232Th]. Domingo Rubio rivulet. Table 2 (code {DR}) also shows the U-isotopes, Th-isotopes and 210Po concentrations in solution. In Fig. 3 a good correlation between the electrical conductivity and U concentrations can be seen. This fact, together with the decrease of the U-isotopes and 210Po concentrations with the distance from its mouth, re¯ects the entrance of waters from the Tinto river channel enriched in U-series radionuclides. In relation to the activity ratios, we can indicate that in general they present some variations in comparison with the obtained ones in the Ria de Huelva channel. The 234U:238U activity ratio is close to 1, which is lower than the typical value of 1.14 measured in uncontaminated surface waters. In addition, the 230Th/232Th activity ratios (Table 3) are lower than in the Ria de Huelva, with values for 232Th and 230Th concentrations higher and similar to those found in the Ria de Huelva channel respectively, even in the sample W1 not a€ected by the tidal cycles (fresh water sample). Finally, it is possible to observe the decrease of the 230Th/234U toward the rivulet mouth, showing again that the Fig. 3. 238 U and 210 inputs of dissolved U from the estuarine waters are dominant. Radioactivity in particulate matter Ria de Huelva channel. The activity concentrations in the particulate matter fraction in the Ria de Huelva channel are shown in Table 4, while their activity ratios are shown in Table 5. In general, we can observe that the activity concentrations of U-isotopes are higher than in unperturbed estuaries. Martin et al. (1978) estimated a world-wide average concentration in particulate matter of 3 ppm, about 35 Bq/kg of 238U, with some ¯uctuations depending on the geographical areas. The higher concentrations are particularly found at the ®rst three sampling points (for both high and low tide) decreasing until point 4, and are, from this location, quite uniform until reaching the Atlantic Ocean. It is interesting to note that it is even possible to ®nd relatively high levels of U-isotopes at high-tide in the zone close to the Atlantic Ocean (about 100 Bq/kg of 238U at location 8). The 210Po concentrations again have values higher than in typical estuaries, being in similarity to U the most contaminated samples are located nearest the fertilizer industries and the PG piles. Speci®cally, it is observed that point 1 at low tide, is the most contaminated sample for U-isotopes and that 210Po, occurs the same in the dissolved fraction. Consequently the maximum U and 210Po concentrations observed in a determined zone of the channel can be associated with the movement down it of some of the e‚uents released directly or indirectly from the fertilizer plants. These contami- Po activity concentrations, and electrical conductivity in the ®ltered water fractions from Domingo Rubio rivulet water samples. Radioecological study of an estuarine system 2947 Table 4. Activity concentrations (Bq/kg) measured in the particulate matter fraction. {RH}=Ria de Huelva channel, {DR}=Domingo Rubio rivulet. N.M.=not measured; N.D.=not detected 238 Sample High tide (RH) HF1 HF2 HF3 HF4 HF5 HF6 HF7 HF8 Average Low tide (RH) LF1 LF2 LF3 LF4 LF5 LF6 LF7 LF8 Average Global average DR F1 F2 F3 F4 F5 235 U 234 U 210 U 232 Po 230 Th Th 339234 635257 388231 123211 112220 100214 122216 144212 245268 N.M. 25.624.8 18.823.0 6.021.4 1224 3.321.9 6.422.9 4.421.1 10.922.9 345234 714263 424234 127212 81216 98214 129217 176214 262278 14129 270224 198222 12628 134210 10228 208230 161213 167219 9.222.1 9.922.0 16.323.3 12.022.1 20.026.7 15.022.3 39.028.6 29.224.4 18.823.7 49.625.9 63.626.2 126212 69.626.2 115218 69.526.0 146221 99210 92212 1165283 131210 337232 55.624.6 45.824.1 70211 123212 24.522.4 2442136 245273 43.724.7 N.M. N.M. N.M. N.M. N.M. 5.621.4 N.M. ± ± 1021273 131210 330231 53.324.4 41.423.8 70211 138214 28.122.7 2272118 244264 450237 93.827.5 308220 64.525.1 68.326.5 70.725.9 10626 48.923.9 151252 159227 N.M. N.D. 19.224.2 3.521.2 15.824.0 1023 13.224.1 8.722.4 11.722.1 15.822.4 N.M. 41.729.1 124213 24.523.5 33.326.0 2629 58210 26.724.5 48213 72211 28.724.1 473239 10029 113211 274220 1.920.8 2525 4.221.2 5.721.7 13.021.8 34.724.7 508242 114210 121212 285221 102210 249227 117210 9829 102211 31.022.7 30.326.7 11.723.3 11.223.0 7.821.5 34.723.0 53.829.3 110213 53.726.3 52.724.7 nation ``bags'' (PeriaÂnÄez, 1995) generated during both high and low tides (minimum water speed) travel across the principal channels of the estuary due to the tidal activity and they will give activity peaks in certain zones. However it is also interesting to point out that the zone where the highest values for U and Po in particulate matter are observed, represent the zone where the Tinto and Odiel waters mix, where some increase in the adsorption and pre- cipitation of the dissolved radioactivity in the water onto the particles in suspension could be produced. Concerning the Th-isotopes (Table 4), it can be indicated that the 232Th concentrations are in the range found in uncontaminated estuarine waters (Martin et al., 1978; Moore, 1967) and present normal average values for both high and low tidal cycles (1924 Bq/kg and 1223 Bq/kg, respectively). On the contrary, enhancements in relation to 232Th Table 5. More signi®cant activity ratios determined in the particulate matter fraction. N.M.=not measured Sample High tide (RH) HF1 HF2 HF3 HF4 HF5 HF6 HF7 HF8 Average Low tide (RH) LF1 LF2 LF3 LF4 LF5 LF6 LF7 LF8 Average Global average DR F1 F2 F3 F4 F5 230 Th/232Th 234 U/238U 230 Th/234U 232 Th/234U 210 Po/234U 210 Po/230Th 5.421.3 6.421.4 7.721.6 5.821.1 5.822.1 4.620.8 3.720.8 3.420.5 5.320.5 1.0220.07 1.1220.05 1.0920.05 1.0320.07 0.7220.16 0.9820.15 1.0520.15 1.2220.07 1.0320.05 0.1420.02 0.08920.012 0.2920.11 0.5520.07 1.4220.36 0.7120.12 1.1320.20 0.5620.07 0.6120.16 0.02620.007 0.01420.003 0.03820.015 0.09520.021 0.2420.10 0.1520.03 0.3020.07 0.1620.03 0.1320.04 0.4120.05 0.3820.05 0.4720.06 0.9920.11 1.6520.35 1.0420.15 1.6120.30 0.9120.10 0.9320.18 2.820.3 4.220.5 1.5720.23 1.8120.20 1.1620.20 1.4720.17 1.4220.24 1.6320.21 2.0020.36 N.M. N.M. 6.421.5 7.022.6 2.120.6 2.521.4 4.421.6 3.120.9 4.320.8 4.920.5 0.8820.02 1.0020.06 0.9820.07 0.9620.06 0.9020.07 1.0020.16 1.1320.08 1.1420.11 1.0020.03 1.0120.03 N.M. 0.3220.07 0.3820.05 0.4620.07 0.8020.17 0.3720.14 0.4220.09 0.9520.17 0.5320.09 0.5720.10 N.M. N.M. 0.05920.015 0.06620.026 0.3820.13 0.1520.10 0.09620.034 0.3120.10 0.1820.05 0.1520.03 0.4420.05 0.7220.08 0.9320.11 1.2120.12 1.6520.22 1.0120.18 0.7720.09 1.7820.20 1.0620.16 1.0020.12 N.M. 2.220.5 2.520.3 2.620.4 2.120.4 2.720.8 1.820.3 1.820.3 2.2420.14 2.1220.20 1.1220.09 1.8020.05 9.422.7 4.821.4 6.721.3 1.2120.07 1.0720.06 1.1520.09 1.0720.10 1.0420.04 1.0020.15 0.1120.02 0.9620.14 0.4420.07 0.1820.02 0.8920.14 0.06120.02 0.1020.03 0.0920.02 0.02720.005 2.9420.40 0.4920.07 1.0320.12 0.8120.11 0.3620.04 2.920.4 4.620.09 1.0620.15 1.8220.29 1.9320.27 2948 J. P. Bolivar et al. are found for 230Th, but lower than those found for U-isotopes and 210Po, being relatively uniform in its concentration along the channel. In fact, and ordering the studied radionuclides for their average concentrations in the suspended matter, we can indicate: [U] > [210Po] > [230Th] > [232Th]. Also, and of particular interest, is that for 230Th there can be observed some small di€erence between the average concentrations measured at high (92212 Bq/kg) and low tide (48213 Bq/kg). The arti®cial origin of most of the U-series radionuclides enhancements found in the particulate fraction of the waters along the Ria de Huelva channel can be also deduced from the values obtained for several activity ratios. For example, the average 234 U/238U activity ratio is 1.01 2 0.03, very close to 1 which is the value found in the phosphate rocks used in the fertilizer plants (Guimond and Hardin, 1989; Bolivar et al., 1996a) while the average 230 Th/232Th activity ratio (5.1 2 0.6) indicates the existence of enrichment from uranium series radionuclides. However, complementary information can be obtained by analysing the 230Th/232Th activity ratio (similar chemical behaviour), which in suspended matter (4.9 2 0.5) is similar to both average values in sediments (5.920.3) (Bolõ var et al., 1995a,b) and in solution (5.4 2 1.0), and indicating that the contamination pathways of the sediments and suspended matter are very similar. The global average 210Po/234U ratio is 1.0020.12 while in the fresh PG (and in the PG piles) this ratio is about or higher than 3, illustrating that the main pathway of contamination is through the dissolution (and, as a consequence, its content in radionuclides) of PG released by the fertilizer plants into the estuary and their later ®xation onto the suspended material. This fact is rati®ed through the other activity ratios shown in Table 5, which are very di€erent than the ratios measured in the PG (Bolõ var et al., 1996a). Domingo Rubio rivulet. With regard to the suspended matter, it can also be observed in Table 2 that in the Domingo Rubio rivulet the activity concentrations for several radionuclides from the Useries are higher than those of the Th-series, in common with the Ria de Huelva channel data. Only the sample F1 has a normal content of natural radioactivity, with concentrations similar to those found in suspended matter of unperturbed estuaries (which range from 10 to 50 Bq/kg for both series), except for 210Po where its concentration (102 2 12 Bq/kg) can be considered similar to the one determined in other samples of this rivulet. Additionally, observing the activity ratios (Table 5), it is again found that the levels of 230Th are enhanced in relation to 232Th, specially in the samples nearest to its mouth. This isotopic activity ratio ranges from 5 to 9 in samples F3 to F5. This clear enrichment in the U-isotopes can be deduced from the low values obtained for the 232Th: 234U activity ratios. The high content of 210Po in location 1 could be due to the high proportion in this sample of surface materials from its surrounding soils, since this sampling was done after a rainy day which introduced a large amount of eroded soil into this rivulet. The sample was collected in a place not a€ected by the tidal activities as it was demonstrated through its electrical conductivity typical of freshwaters and the other radioactive concentration values. Then the particulate fraction of this sample can be considered of terrestrial origin and it can contain a considerable amount of atmospheric (or unsupported) 210Pb (210Po) (El-Daoushy, 1987). Partition coecients in Ria de Huelva Particulate matter is an important medium of transport for many elements. For this reason is very important to determine the distribution of elements between the dissolved and particulate phases of the waters, which can be expressed by a partition coecient, K. This coecient is de®ned in our case as the ratio between the concentration of a radionuclide in particulate matter and its concentration in solution, and gives information about the adsorption degree of a radionuclide to the solid fraction. From our previously shown data, we have calculated these partition coecients for several radionuclides in the Ria de Huelva channel at high and low tide, by using the formula Kˆ As l=kg† Ad where As is the activity concentration in the particulate matter in Bq/kg and Ad is the activity concenTable 6. Partition coecients (l/kg) in the water samples collected at the Ria de Huelva Sample High tide (RH) HF1 HF2 HF3 HF4 HF5 HF6 HF7 HF8 Average Low tide (RH) LF1 LF2 LF3 LF4 LF5 LF6 LF7 LF8 Average Global average 238 U (103) 210 Po (104) 230 Th (104) 232 Th (104) 11.821.2 1.2920.08 ± ± 26.223.1 5.1520.45 2.5520.24 ± 12.621.8 9.221.0 5.020.5 8.621.7 3.3620.40 0.26720.017 3.6420.32 1.5220.26 2.7520.53 3.1920.23 6.121.0 3.321.0 2.3020.38 1.9120.15 2.3020.20 2.5420.38 2.7920.44 5.720.8 6.720.9 7.321.5 3.1420.34 7.720.6 4.720.5 8.621.2 8.123.0 4.321.1 4.420.6 5.321.3 18.321.9 9.021.4 8.721.1 1.6820.19 1.3320.16 2.0020.35 1.4420.17 0.6720.08 5.422.2 6.821.8 1.9820.16 7.620.4 1.3420.09 1.1520.08 2.6420.25 2.3020.19 0.9520.05 1.5420.12 2.420.7 3.421.1 ± ± 6.120.6 2.0420.29 1.7420.33 2.620.9 ± ± 3.121.0 4.020.6 ± ± 2.120.4 1.020.3 6.321.5 1023 ± ± 4.822.0 5.121.1 Radioecological study of an estuarine system tration in solution expressed in Bq/l. The obtained K values are compiled in Table 6. The so de®ned partition coecient coincides under the equilibrium conditions between the two phases with the well known equilibrium distribution coecient (Kd). The degree of equilibrium conditions for the di€erent radionuclides is not known for the analysed samples, and in fact will be evaluated here using the values obtained for the partition coecients together with previously shown data and literature information about Kd values in estuarine systems. The K values obtained for uranium do not present signi®cant di€erences between both high and low tide, with an average value of (6.8 2 1.8) 103 l/ kg, and ranging along the Ria de Huelva from 1.3  103 to 1.8  104 l/kg. These K values are in good agreement with the Kd values found in the literature that ranges from 0.2  103 to 5  103 l/kg, with an average of 3  103 l/kg (IAEA, 1985; Eisma, 1993). This agreement can indicate a good degree of equilibrium for uranium in the waters of the estuary. Concerning the obtained partition coecients for 210 Po, it can also be commented that again there are no signi®cant di€erences between both low and high tide, ®nding a mean value of (3.4 2 1.0) 104 l/ kg, which is ®ve times higher than the partition coecients determined for Uranium. This fact shows the bigger tendency of the 210Po to be associated to the particulate material in comparison with U. The Kd recomended in the literature (IAEA, 1985) for Po is 1  104 l/kg, but this value must only be considered as an order of magnitude since it presents big variations in unperturbed systems. The partition coecients for 230Th and 232Th are very similar for the majority of the samples. It is observed in Table 6 that they are very uniform and the di€erences between the average values for 230Th and 232Th are not evident considering the experimental uncertainties: (4.0 2 0.6)  104 l/kg for 230 Th and (5.121.1)  104 l/kg for 232Th. However, these K values are clearly smaller than the Kd found in the literature (IAEA, 1985), being outside the delimited range associated with unperturbed systems: 1  105 to 1 107 l/kg, but in the same order of magnitude as the lower limit. These low values found for thorium K could be related to several facts previously commented on as the low concentrations of Th-isotopes in the suspended matter or the existence of non-equilibrium between solid-solution phases due to the residence time of the suspended particles in the estuary and the slow rate of transfer for Th from the water onto the particles. CONCLUSIONS The results obtained in this paper have shown the signi®cant radioactive impact produced by radionuclides from uranium series in the estuary formed by the Tinto and Odiel rivers. In this sense, activity 2949 concentrations of several uranium series radionuclides were found in the waters around one order of magnitude higher than those of uncontaminated. These enhancements were observed either in solution or attached onto the suspended material. It was demonstrated that this radioactive contamination comes from the wastes produced by a phosphate fertilizer complex located close to this estuary. The distribution of radionuclides in the di€erent phases, analysed by this aquatic system as well as the study of their several activity ratios, have allowed us to con®rm the arti®cial origin of the radioactive contamination (fertilizer plants), to deduce that the main pathway of contamination is through the solution in the estuarine waters of the radionuclides released by the industrial activities with their later deposition on the sediment material, and to obtain information about the di€erent behaviour of the isotopes analysed in this environmental system. The relative uniformity found in the concentrations of U and daughters along the estuary, even in quite remote locations from the contamination sources, and the uniformity in some activity ratios, indicates that the tidal activity plays an essential role in the dispersion of the radioactive contamination. This conclusion has also been con®rmed through the analysis of the water samples collected from a small rivulet that ends in the estuary, and that it is also a€ected by inputs from the estuary coming from it due to the tidal activities. AcknowledgementsÐThis work has been partially supported by ENRESA, Junta de Andalucia and the CICYT projet: 1FD 97-0900-C02-02. REFERENCES Bacon M. P. and Elzerman A. W. (1980) Enrichment of 210 Pb and 210Po in the sea-surface microlayer. Nature 284, 332±334. Bolivar J. P. (1995) Aplicaciones de las espectrometrõ as alfa y gamma al estudio del impacto radiactivo producido por industrias no nucleares, Ph. D. Thesis, University of Sevilla, Spain (In Spanish). Bolivar J. P., Martõ nez-Aguirre A. and Garcõ a-Tenorio R. (1993) Radioactivity involved in the production of fertilizers: radioecological impact. In Proc. Chemical Industry and Environment, Vol. III, eds N. Pellicer and M. Rigola, pp. 311±320. University of Gerona, Gerona, Spain. Bolivar J. P., Garcõ a-Tenorio R. and Garcõ a-LeoÂn M. (1995a) Fluxes and distribution of natural radionuclides in the production and use of fertilizers. Appl. Radiat. Isotopes 46, 717±718. Bolivar J. P., Garcõ a-Tenorio R. and Garcõ a-LeoÂn M. (1995b) Enhancement of natural radioactivity in soils and salt-marshes surrounding a non nuclear industrial complex. Sci. Total Envir. 173/174, 125±136. Bolivar J. P., Garcõ a-Tenorio R. and Garcõ a-LeoÂn M. (1996a) On the fractionation of natural radioactivity in the production of phosphoric acid by the wet acid method. J. Radioanal. Nucl. Chem. Lett. 214, 77±88. Borole D. V., Krishnaswami S. and Somoyajulu B. L. 2950 J. P. Bolivar et al. (1977) Investigations on dissolved uranium, silicon and particulate trace elements in estuaries. Estuar. Coast. Mar. Sci. 5, 743±754. Borole D. V., Krishnaswami S. and Somoyajulu B. L. (1982) Uranium isotopes in rivers, estuaries and adjacent coastal sediments of western India: their weathering, transport and oceanic budget. Geochim. Cosmochim. Acta. 46, 125±137. El-Daoushy F. (1987) The Pb-210 global cycle: dating and tracing applications. In Proceedings of the 1st International Summer School of Low-Level Measurements and their Applications to Environmental Radioactivity, La RaÂbida, Huelva, 28 September to 9 October, eds M. Garcõ a-LeoÂn and G. Madurga, pp. 222±273. El-Daoushy F., Olsson K. and Garcõ a-Tenorio R. (1991) Accuracy parameters in alpha spectrometry for lead-210 studies. Hydrobiologia 214, 43±52. Eisma D. (1993) Suspended Matter in the Aquatic Environment. Springer-Verlag, Berlin. Guimond R. J. and Hardin J. M. (1989) Radioactivity released from phosphate-containing fertilizers and from gypsum. Rad. Phys. Chem. 34, 309±315. Hallstadius L. (1984) A method for electrodeposition of actinides. Nucl. Instrum. Meth. 223, 226±238. Garcõ a-LeoÂn M., Abril J. M., Bolõ var J. P., Martõ nezAguirre A., PeriaÂnÄez R. and Garcõ a-Tenorio R. (1997) Radioactive environmental impact during the phosphoric acid production. In Proceedings of the International Symposium on Radiological Problems with Natural Radioactivity in the Non-Nuclear Industry, 8±10 September, Amsterdam, Netherlands. Holm E. and Fukai R. (1977) Method for multi-element alpha-spectrometry of actinides and its applications to environmental radioactivity studies. Talanta 24, 659± 664. IAEA (1985) Sediment Kds and Concentration Factors for Radionuclides in the Marine Environment. Technical Reports Series No 247. Iyengar M. A., Kannan V., Ganapathy S. and Kakmath P. R. (1981) 210Po in the coastal waters at Kalpakkam. In Proceedings of the 2nd Symposium on Natural Radiation Environment, 19±23 de Enero, Bombay, India. Koczy F. F., Picciotto E., Poulaert G. and Wilgain S. (1957) Mesure des isotopes du Thorium dans l'eau de mer. Geochim. Cosmochim. Acta. 11, 103±110. Ku T. L., Knauss K. G. and Mathieu G. G. (1977) Uranium in open oceans: concentration and isotopic composition. Deep Sea Res. 24, 1005±1017. Martin J. M., Nijampurkar V. N. and Salvadori F. (1978) Uranium and thorium isotopes behaviour in estuarine systems. In Biogeochemistry of Estuarine Sediments, ed. E. D. Goldberg, pp. 111±127. UNESCO, Paris. Martõ nez-Aguirre A., Garcõ a-LeoÂn M. and Ivanovich H. (1994) Identi®cation and e€ects of anthropogenic emissions of U and Th on the composition of sediments in a river/estuarine system in southern Spain. J. Environ. Radioact. 23, 231±248. Moore W. S. and Sackett W. M. (1964) Uranium and thorium series disequilibrium in sea waters. J. Geophys. Res. 79, 5401±5415. Moore W. S. (1967) Amazon and Mississipi river concentrations of uranium, thorium and radium isotopes. Earth Planet. Sci. Lett. 2, 231±240. PeriaÂnÄez R. (1995) Un modelo matemaÂtico para la simulacioÂn de la dispersioÂn de radionuÂclidos no conservativos en un sistema estuario: aplicacioÂn a la Rõ a de Huelva, PhD., Dept. Fõ sica AtoÂmica, Molecular y Nuclear, University of Sevilla. (In Spanish). Rutherford P. M., Dudas M. J. and Samek R. A. (1994) Environmental impacts of phosphogypsum. Sci. Total Envir. 149, 1±38. Scott M. R. (1982) The chemistry of U- and Th-series nuclides in rivers. In Uranium Series Disequilibrium: Applications to Environmental Problems, eds M. Ivanovich and R. S. Harmon. Clarendon Press, Oxford.