Modelling the spatio-temporal evolution of H-3 in the waters of the River Tagus

Modelling the spatio-temporal evolution of 3 H in the waters of the River Tagus A. Baeza a, *, E. Garcı ´a b , C. Miro ´ a , R. Peria ´n ˜ez c a Universidad de Extremadura, Facultad de Veterinaria, Dpto. de Fı´sica, Avda. Universidad sn, 10071 Ca´ceres, Spain b Universidad de Extremadura, Ingenierı´a Te´cnica Forestal, Dpto. de Fı´sica, 10600 Plasencia (Ca´ceres), Spain c Universidad de Sevilla, E.U. de Ingenierı´a Te´cnica Agrı´cola, Dpto. de Fı´sica Aplicada, 1, 41013 Sevilla, Spain Received 7 April 2005; received in revised form 2 October 2005; accepted 23 October 2005 Available online 15 December 2005 Abstract Measurements of tritium specific activity levels and of temperatures in waters of the Torrejo ´neTagus reservoir (Spain) showed that their radioactive characteristics were basically influenced by the radioactive liquid effluent from the Almaraz Nuclear Power Plant. This enters the Torrejo ´neTagus reservoir via the Arrocampo cooling reservoir, which exchanges water with the first. We studied the temporal and spatial (in two dimensions) evolution of the mentioned parameters for years 1997 and 1998. The tritium levels were found to be significantly correlated with temperature. Two numerical models were constructed for a quantitative study of the tritium levels along Torrejo ´n reservoir: a 1D model was used for the dispersion of tritium along the whole length of the reservoir, and a 2D depth-averaged model was used for a detailed study of the area where tritium is released into the reservoir. Both models solve the hydrodynamic equations, to obtain the currents induced by the exchanges of water between the reservoirs in the River Tagus and Arrocampo, and the advection/diffusion equation to calculate the dispersion of tritium. In general, the model results were in agreement with the experimental observations. Ó2005 Elsevier Ltd. All rights reserved. Keywords: Reservoir water; Tritium; Temperature; Numerical modelling; Dispersion * Corresponding author. Tel./fax: þ34 927 25 71 53. E-mail address: [email protected]s (A. Baeza). 0265-931X/$ - see front matter Ó2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvrad.2005.10.004 Journal of Environmental Radioactivity 86 (2006) 367e383 www.elsevier.com/locate/jenvrad 1. Introduction The quality of the planet waters notably influence their use, with the presence of radionuclides being one of the limiting factors. In this sense, the study of the tritium content of different waters has a threefold interest. Firstly, its universality, since tritium levels due to natural origins (Kaufman and Libby, 1954) may be raised by several orders of magnitude by an anthropogenic component (UNSCEAR, 1988; IAEA, 1990; Okada and Momoshima, 1993). Secondly, its radiotoxicity, which, while relatively low in comparison with that of other radionuclides, is 25 000 times greater when the 3 H is in the form of HTO instead of HT (ICRP, 1978). Thirdly, the practically indistinguishability for hydraulic systems of tritiated water from untritiated water (Thibodeaux, 1979), so that this radionuclide may be a good parameter for studying the dynamics of water masses (Fontes, 1983a,b; Baeza et al., 2002), and in particular for studying the spatial and temporal distribution of other substances contained in them, whether conservative or not. One of the Iberian Peninsula hydrological systems that is most affected by the presence of tritium is the River Tagus. The principal reason is that three nuclear power plants use its water for cooling (CSN, 1998). Its course is highly regulated by an almost completely coupled series of reservoirs, which are especially important in the river passage through the regions of Extremadura (Spain) and the Alentejo (Portugal), with five reservoirs in 200 km. The authors studied previously, for the years 1994e1996, the basic aspects of the tritium’s spatial and temporal evolution (Baeza et al., 1997), including various simple models that allowed us to understand certain facets of its dynamics, such as its mean transit speed (Baeza et al., 2001, 2004). One fact deduced was the importance of Torrejo ´n reservoir in understanding the spatial and temporal evolution of tritium in this part of the River Tagus, since the water is used not only for hydroelectric power production and irrigation, but also to cool the Almaraz Nuclear Power Plant (henceforth, ANPP) through water exchanges with Arrocampo reservoir. In the present work, we study in detail the influence of this water exchange on the distribution of tritium in Torrejo ´n reservoir, using data for years 1997 and 1998. To this end, we collected monthly water samples, in which we determined their temperatures simultaneously with the collection. In this way we were able to analyze both the spatial distribution of tritium levels in two dimensions in Torrejo ´n reservoir, transversally and longitudinally, and their temporal evolution. The currents caused by the exchange of water between the different reservoirs in this part of the River Tagus make it difficult to give a detailed quantitative interpretation of the measured tritium levels, in a form allowing predictions of their future evolution. A prerequisite is to construct appropriate models for Torrejo ´n reservoir. We made a first approximation by using simple box models to simulate the temporal evolution of tritium levels in the Arrocampo (Baeza et al., 1997) and Torrejo ´n(Baeza et al., 2001) reservoirs, in particular studying the area where the exchange of water with the Arrocampo reservoir takes place. The present work describes the development of two new numerical models for the Torrejo ´n reservoir. The first is a one-dimen- sional model constructed to reproduce the large scale dispersion of tritium, covering the total length of the cited reservoir. This model cannot, however, describe the dispersion of tritium at a small scale, in particular in the area of the discharges from Arrocampo reservoir. A higher resolution two-dimensional depth-averaged model was therefore constructed for this area. The advance represented by these models is that they are physical models in which the instantaneous currents are obtained for each point of the Torrejo ´n reservoir and for each instant of time. These currents are then used to calculate the tritium dispersion. Hence, the tritium concentrations may be calculated for any position and time in this reservoir. 368 A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 2. Experimental 2.1. Study area The River Tagus has its source in the Sierra de Albarracı ´n (Spain), and its mouth in Lisbon (Portugal). Its total length is 1107 km (Fig. 1). The area of the Tagus basin is about 8.1 10 4 km 2 , of which 69% is in Spain. The ANPP, with two 930 MWe PWR-type units is located on the river right bank, in the region of Extremadura and 180 km westesouthwest of Madrid. The River Tagus receives discharges from the ANPP (Baeza et al., 1991) and from other two nuclear power plants located below 300 km upstream eJose ´Cabrera and Trillo eas well as several smaller contributions from a significant number of laboratories and hospitals (CSN, 1998). The flow regime of the River Tagus is controlled through a chain of reservoirs. One of them, in the region of Extremadura (Fig. 1), is, in cubic hectometres (1 hm 3 ¼10 6 m 3 ), the 166 hm 3 capacity Torrejo ´n reservoir, which is the object of the detailed study of the present work. It is 42 km long, with a mean width of 130 m and a mean depth of 20 m. Immediately upstream, the 500 hm 3 capacity Valdecan ˜as reservoir is located. The Alca ´ntara reservoir, with a normal volume of 1600 hm 3 , is situated downstream from Torrejo ´n. The water of Torrejo ´n reservoir is periodically partially exchanged with that of Arrocampo, since the latter is used as the ANPP’s Fig. 1. Map of the River Tagus with the locations of the sampling points. The location of the dam of each reservoir is indicated by <. 369A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 secondary cooling circuit. The purpose of the exchange is to maintain Arrocampo’s volume practically constant at 35 hm 3 (ANPP, 1998) and its water temperature sufficiently low to fulfil its cooling function. The lower part of Fig. 2 shows the temporal evolution of the flows between the Arrocampo and Torrejo ´n reservoirs. The strong 12-month period seasonality of these exchanges, as has been described in previous studies (Baeza et al., 2001), is observed. The upper part of the Fig. 2 shows the monthly discharges of water between the Tagus reservoirs present in Fig. 1, i.e., from Valdecan ˜as to Torrejo ´n, and from the latter to Alca ´ntara. It can be seen that the management of the discharges between these three reservoirs is perfectly coupled, which, as was indicated above, is a common characteristic to all the River Tagus reservoirs. On the other hand, this management is radically different from that between Torrejo ´n and Arrocampo, since the observed seasonality in this case is conditioned by the availability of water in the Tagus River, which in turn is related to the intensity and frequency of rainfalls in the area. 2.2. Sampling Two sampling programs eone monthly and the other intensive ewere carried out in 1997 and 1998 in the above described study area. The intensive program of sampling was designed to determine the spatial distribution of tritium levels in two dimensions in Torrejo ´n reservoir: longitudinally, for 17 sampling points distributed along the length of the reservoir, and transversally, in the centre and on the two banks of the reservoir for each of those 17 points. The monthly program was to allow us to analyze the temporal evolution on the basis of monthly DISCHARGES 0 500 1000 1500 1 3 5 7 9 11 13 15 17 19 21 23 1997-1998 1357 9 11 13 15 17 19 21 23 1997-1998 (hm 3 ) (hm 3 ) Valdecañas Torrejón 0 25 50 75 100 Arrocampo to Torrejón Torrejón to Arrocampo Fig. 2. Upper part: monthly discharges of water from Valdecan ˜as to Torrejo ´n, continuous line, and from Torrejo ´nto Alca ´ntara, dotted line. Lower part: monthly exchange of water from Arrocampo reservoir to Torrejo ´n reservoir, continuous line, and from Torrejo ´n to Arrocampo reservoir, dotted line. 370 A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 records at three of the aforementioned sampling points in Torrejo ´n reservoir and at another point in Arrocampo reservoir. Water temperature was simultaneously recorded with each sample collection. 2.2.1. Monthly sampling Every month, surface water samples were collected at the three sampling points of the Torrejo ´n reservoir identified as 1, 7, and 8 in Fig. 1, and at one sampling point of the Arrocampo reservoir identified as A in the figure. As was noted above, Arrocampo is the cooling reservoir for the ANPP. There is a constant forced circulation of water at a rate of 88 m 3 /s around the separation dyke constructed in the centre of the reservoir. As a result, its water has perfectly homogeneous physicochemical properties (Baeza et al., 1991), so that point A is entirely representative of the characteristics of water in the reservoir. The locations of points 1, 7, and 8 in Torrejo ´n reservoir are the following. Point 1 is some 10 m from the Torrejo ´n dam, on the left bank of the Tagus. Point 7 is downstream the Arrocampo dam and close to the outlet from that reservoir into Torrejo ´n, on the right bank of the Tagus. Finally, point 8 is some 400 m upstream of point 7, on the left bank of the river and clearly in the opposite direction to the natural flow of the water. These last two points are logically those that will be most intensely affected by the mixing of waters between the Torrejo ´n and Arrocampo reservoirs, which, as we shall see below, have very different radiological and thermal characteristics. 2.2.2. Intensive sampling Intensive sampling was carried out on two specific dates: 1997-08-01 and 1998-02-12. As can be seen in the lower part of Fig. 2, the former date corresponds to the time of year when the exchange of water between Torrejo ´n and Arrocampo reservoirs is maximum, while the latter is representative of the time of year when the volumes of exchanged water are relatively low. On both dates, surface water samples were collected in a number of longitudinal and transversal transects of the Torrejo ´n reservoir. Fig. 1 shows the location of the points in the three longitudinal transects, covering the 42 km length of the reservoir. The separation of these points was kept as homogeneous as possible, except for a zone of the reservoir that is especially protected because of its wildlife. The number of sampling points on the 1998-02-12 date was increased in light of the results of the analysis of the 1997-08-01 samples. 2.3. Analytical methods At each sampling point, 1.5 L of water was collected in polyethylene bottles, and temperature was measured with a precision of 0.1 C. After transport to the laboratory, the samples were filtered through 0.45 mm pore filters prior to their preparation for 3 H assay. For the tritium assay, a 50 mL aliquot was taken and subjected to a process of simple distillation to eliminate colouration, organic matter, and salts that might interfere with the analysis. The distilled sample was then mixed with a scintillation solution of Optiphase Hisafe 3 from LKB-Wallac at a sample/scintillator ratio of 8/12 (Scho ¨nhoffer and Henrich, 1987)in20mL teflon-coated polyethylene vials. Tritium content was determined using a Quantulus-1220, LKB-Wallac, liquid-scintillation alphaebeta spectrometer. The quenching calibration was performed with tritium standards quenched to increasing degrees with CCl 4 . The background in the region of the tritium b-emission was of the order of 1.5 cpm. The systematically used counting time was 1000 min (200 min 5 measurement cycles). The resulting Lower Detection 371A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 Limit at a 95% confidence for tritium was 0.5 Bq/L and the figure of merit, (efficiency) 2 /background, was 320 (Baeza et al., 1999). 3. Results and discussion 3.1. Temporal evolution (1997e1998) Fig. 3 shows the temporal evolution of 3 H levels and the temperature at the surface water for the sampling points identified in Fig. 1 as 1, 7, and 8, belonging to Torrejo ´n reservoir, and point A in Arrocampo reservoir. Table 1 lists the mean values and ranges recorded at these sampling points for the cited variables etritium and temperature. The following results stand out. The highest activity levels were recorded at point A, as was to be expected since this is the cooling reservoir for the ANPP. Activity levels and their temporal evolution are consistent with the findings of previous studies (Baeza et al., 2001). The observed temporal evolution is a consequence of the management of these waters, whose regime of exchange with those of the Torrejo ´n reservoir (see lower part of Fig. 2) is conditioned by the maximum temperature that the water can reach to be used to cool the ANPP. Thus, point A not only had the highest temperatures of all the study area, but it also had the narrowest range of temperatures as well. This is because in the cold months of the year, when it is unnecessary to exchange water in order to satisfy the thermal requirements, its temperature falls to less than that of the Torrejo ´n reservoir waters. The reiterated use of the Arrocampo water as coolant during winter is also the reason why 3 H levels at point A were higher in these months. In particular, there was a significant negative correlation between tritium levels and temperature at point A (see the last row of Table 1). For the three points sampled monthly in the Torrejo ´n reservoir epoints 1, 7, and 8 eboth tritium levels and temperature decreased from point 7 to 8, and from the last to point 1. This is readily understandable given the spatial location of these three points (see Fig. 1) relative to the 0 50 100 150 200 250 300 0 10 20 30 40 Sampling point A Sampling point 7 0 100 200 300 400 500 1357911131517192123 1997-1998 1357911131517192123 1997-1998 1 3 5 7 9 11 13 15 17 19 21 23 1997-1998 1 3 5 7 9 11 13 15 17 19 21 23 1997-1998 3H (Bq·L-1) 3H (Bq·L-1) 3H (Bq·L-1) 3H (Bq·L-1) 0 10 20 30 T (ºC) T (ºC) T (ºC) T (ºC) H-3 T H-3 T H-3 T H-3 T 600 40 Sampling point 8 0 50 100 150 200 250 300 Sampling point 1 0 20 40 60 80 0 10 20 30 40 0 10 20 30 40 Fig. 3. Specific activity of tritium and temperature measured during 1997e1998 at sampling points A, 1, 7, and 8 of the study area. 372 A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 local source of tritium and heat into Torrejo ´n reservoir, i.e., the discharge from the Arrocampo reservoir. Local temporary increases in activity were directly correlated with increases in the exchange of water between the Arrocampo and Torrejo ´n reservoirs (see lower part of Fig. 2), and therefore had a statistically significant positive correlation with rises in Torrejo ´n reservoir water temperature (see the last row of Table 1). These temperature increases of course have a natural component during the hot months, reinforced by the cited exchanges with the hotter water in Arrocampo reservoir. The coupled regulated transit of water in this zone of the River Tagus (see upper part of Fig. 2) is ultimately responsible for the transport of tritium discharged into the Torrejo ´n reservoir to waters below Torrejo ´n dam. This transit occurs preferentially during winter, so that tritium is spread out along the mentioned reservoir during the summer months. 3.2. Spatial evolution To check the veracity of the foregoing statement, we performed an intensive study on the spatial distribution of tritium and temperature on two dates that were clearly identifiable with the previously typified characteristics for the months of winter and of summer. Results are presented in Table 2. The table lists, firstly, the distance of the 17 points selected along the length of the Torrejo ´n reservoir (see Fig. 1) from the outlet of the discharges from the Arrocampo reservoir (point 7). Secondly, for each of the two sampling dates and each sampling point, it lists the 3 H level and, in parentheses, water temperature in the centre and near the banks. It can be observed that the results were radically different in the conditions identified as summer (sampling date 1997-08-01) and winter (sampling date 1998-02-19). The summer samples showed a good transversal mixing in both tritium levels and temperatures at each sampling point. Indeed, the maximum relative transversal variation was 13.8% for 3 H and 5.2% for temperature, both corresponding to point 17. Likewise, the overall spatial homogeneity was also very significant. In particular, omitting the activities and temperatures measured at point 17 which is 12.4 km away in the opposite direction to normal water flow in the reservoir, we can consider that the rest of the reservoir has a well defined tritium level and temperature in the conditions that were identified as summer. In particular, for that sampling date, the values were 56.6 3.2 (SD) Bq/L and 29.2 1.5 (SD) C, respectively. The two parameters were closely correlated, with a correlation coefficient of 0.880 at a significance level quantified by the value of the parameter p¼510 4 . Table 1 Mean values and ranges of tritium and temperature recorded during years 1997 and 1998 in waters of Arrocampo reservoir (point A), and Torrejo ´n reservoir (points 1, 7, and 8) Sampling point A 1 7 8 3 HSD (Bq/L) 270 144 22 19 101 77 57 63 Range 3 H (Bq/L) 59e496 1.4e6.0 2.4e243 1.2e199 TSD (C) 28 5185218207 Range T(C) 19e36.2 9e26.5 10e30.5 10e28.5 Correlation 3 HeT(significance level) 0.774 (<0.0001) 0.676 (0.0003) 0.484 (0.0165) 0.711 (<0.0001) Details of the correlation between these two parameters, at the four sampling points that were selected for the study of the temporal evolution, are also given. 373A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 Table 2 Details of the two intensive sampling campaigns, identifying the distance (m) of each sampling point from the outlet of water discharged from Arrocampo reservoir into Torrejo ´n reservoir Notation Distance (m) Sampling 1997-08-01 Sampling 1998-02-19 Left side Centrum Right side Left side Centrum Right side 1þ28 600 53 3 (26.0) 54 3 (26.5) 56 3 (26.0) 1.5 1.5 (11.5) 1.5 1.5 (11.5) 1.5 1.5 (11.5) 2þ13 100 58 3 (27.0) 58 3 (27.0) 61 4 (27.0) 17.5 2.0 (10.5) 17.7 2.0 (10.5) 17.8 2.0 (10.5) 3þ2600 56 3 (28.7) 55 3 (29.0) 56 3 (29.0) 34 3 (11.0) 16.6 1.9 (10.0) 10.7 1.7 (11.0) 4þ1600 57 3 (29.0) 58 3 (29.5) 57 3 (29.0) 44 3 (11.5) 48 3 (12.0) 7.6 1.7 (11.0) 5þ600 57 3 (29.5) 58 3 (29.5) 60 4 (30.0) 32 3 (12.0) 87 5 (13.0) 58 4 (12.0) 6þ300 53 3 (30.2) 57 3 (30.5) 56 3 (30.0) 88 5 (13.5) 111 6 (14.0) 81 5 (13.0) 70 594 (31.0) 58 3 (31.5) 55 3 (32.0) 3.7 1.6 (10.0) 229 11 (18.5) 371 18 (24.0) 8400 58 3 (30.5) 54 3 (30.5) 55 3 (30.5) 1.6 1.5 (10.0) 2.5 1.5 (10.0) 1.5 1.5 (9.5) 92400 54 3 (29.5) 56 3 (29.5) 53 3 (29.5) 3.9 1.6 (10.5) 5.0 1.7 (9.5) 4.4 1.6 (10.0) 10 2900 52 3 (28.0) 51 3 (29.0) 53 3 (29.0) 3.7 1.5 (11.0) 3.8 1.5 (10.0) 1.5 1.5 (9.5) 11 3900 No data No data No data 3.3 1.6 (10.0) 3.0 1.5 (10.0) 2.0 1.5 (10.0) 12 4900 60 4 (30.3) 67 4 (30.2) 61 4 (29.0) 2.2 1.5 (9.5) 2.5 1.5 (9.5) 2.4 1.5 (9.5) 13 6400 No data No data No data 2.9 1.5 (9.5) 3.8 1.5 (9.5) 2.8 1.6 (9.5) 14 7900 No data No data No data 2.2 1.5 (9.5) 2.2 1.5 (9.5) 2.2 1.5 (9.5) 15 8900 No data No data No data 2.2 2.0 (10.5) 1.5 1.5 (9.5) 2.5 1.5 (9.5) 16 11 400 No data No data No data 1.5 1.5 (9.5) 1.5 1.5 (9.5) 2.6 1.6 (9.5) 17 12 400 2.2 1.3 (19.0) 2.0 1.3 (19.0) 2.3 1.3 (20.0) 1.5 1.5 (9.5) 1.5 1.5 (9.5) 2.8 1.6 (9.5) The sign þor indicates whether it is in the same direction or the opposed one, respectively, as the normal advance of the water in the Torrejo ´n reservoir. The table lists, for each of the two sampling dates, the 3 H levels (Bq/L) and, in parentheses, the temperature (C) for the surface water close to the two banks and in the centre of the reservoir. 374 A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 With respect to the spatial distribution in the Torrejo ´n reservoir determined on the sampling date typical of winter conditions, i.e., when there was a small flow of relatively warm water from the Arrocampo reservoir containing relatively high levels of tritium, as well as a large water flow in the direction of the natural current in this tract of the River Tagus, the following facts may be observed. Firstly, the radioactive and thermal contributions from the Arrocampo discharges were only notable in a limited portion of the Torrejo ´n reservoir. This zone was in the direction of the natural advance of the water, also favoured by the direction of the Arrocampo discharge canal on the right bank of the River Tagus at point 7. One observes in Table 2that it was precisely at that point where the highest temperature and 3 H activity were detected. Both the transverse and the longitudinal homogeneities of these parameters in the Torrejo ´n reservoir water have disappeared under the operating conditions that were identified as being typical of winter. Indeed, during these winter months, the impact of the Arrocampo discharges was only partially detectable downstream from the outlet point into the Torrejo ´n reservoir. It could not even be detected at the head of the reservoir, as indeed was the case on the summer date of our intensive sampling. Nonetheless, the two parameters e 3 H activity and temperature ewere again perfectly correlated, reflecting their common origin in the discharges from Arrocampo into the Torrejo ´n reservoir. The correlation coefficient was 0.974, with a significance level quantified by the value of the parameter p¼10 4 . 4. The modelling study 4.1. 1D model The use of a longitudinal model to study the dispersion of tritium in the whole Torrejo ´n reservoir is justified by its narrowness and shallowness relative to its length (see the Section 2.1). Also, discharges from Arrocampo are designed to achieve a rapid mixing of the pollutants both in depth and transversally. Thus, we are dealing, essentially, with a one-dimensional dispersion problem. Indeed, 1D models have been found to be adequate tools for studying the dispersion of pollutants in channeled regions and rivers (Abril and Abdel-Aal, 2000; Abril et al., 2000; Prandle, 1974). The hydrodynamic equations are solved to obtain the currents along the reservoir for the time period of interest. In the present case, the advection/diffusion dispersion equation for tritium is solved simultaneously. The 1D hydrodynamic equations are (Prandle, 1974) dSvz vtþvðAuÞ vxdxQ¼0 vu vtþuvu vxþgvz vxþCd ujuj hþz¼0 where Sis the horizontal surface area, zthe height of the water surface above the mean level, u the water velocity along the xdirection, Athe cross-section area of the reservoir, gthe acceleration due to gravity, Qthe external flow of water entering/leaving the reservoir at each point, and ha depth parameter defined as h¼A/bwhere bis the width of the reservoir. The bed friction coefficient is defined as Cd¼g C2 h 375A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 evolution of the levels of tritium in the reservoirs of Arrocampo and Torrejo ´n hence depends strongly on the management of their waters. The temporal evolution of the water temperature is also conditioned by the natural evolution of the environmental temperature. The two magnitudes consequently presented seasonal variations with an annual period, which were six months out of phase in the Arrocampo reservoir, and perfectly synchronized at the three sampling points that were systematically analyzed in the Torrejo ´n reservoir. For the study of the spatial evolution of tritium in the Torrejo ´n reservoir, we carried out intensive sampling at numerous points of the reservoir on two dates clearly typical of winter and of summer. In winter, due to the minimal exchange of water with the Arrocampo reservoir and to the large water flow in the natural direction of the River Tagus waters, the Torrejo ´n reservoir was found to behave in a form similar to a river with a quasi-natural flow regime. Consequently, the relatively insignificant inputs of tritium and heat from the Arrocampo reservoir fundamentally affected the area close to and downstream from the discharge point (point 7 in our study), with major transverse and longitudinal gradients in both temperature and tritium activity. On the contrary, in summer, the Torrejo ´n reservoir was found to behave in a form similar to a lake, since there were minimal flows of water from Valdecan ˜as upstream, and into Alca ´ntara downstream. The exchange of water with the Arrocampo reservoir, where the tritium levels and the temperature are higher, was maximal at that time, however. In these conditions, both the 3 H and the heat diffused practically over the whole Torrejo ´n reservoir, with the consequent longitudinal and transversal homogenization of these two magnitudes. A 1D model of the Torrejo ´n reservoir was developed to study the behaviour of tritium. In general, the model yielded a temporal evolution of tritium specific activities along the reservoir, and tritium profiles along the length of the reservoir for the two intensive sampling campaigns, that were in agreement with the measurements. It was necessary, however, to use a larger value for the diffusion coefficient during the summer months. A 2D model of the release area was developed to study the behaviour of tritium near the ANPP discharge point. It was found that the different dispersion patterns measured in February and August (transverse gradients and mixing of tritium, respectively) were reproduced with a single formulation of the diffusion coefficient implemented over a 2D domain. The different behaviour in winter and summer arises from differing regimes of water circulation in those two seasons. Acknowledgements The authors are grateful to the Consejerı ´a de Educacio ´n, Ciencia y Tecnologı ´a (project 2PR02A022) and to the Agricultura y Medio Ambiente for their financial support. Also the authors thank the companies IBERDROLA and Almaraz Nuclear Power Plant for providing information about the transferred flows of water between the reservoirs. References ANPP, 1998. Nuclear Power Plant of Almaraz, Private communication. Abril, J.M., Abdel-Aal, M.M., 2000. Marine radioactivity studies in the Suez Canal: hydrodynamics and transit times. Estuar. Coast. Shelf Sci. 50, 489e502. Abril, J.M., Abdel-Aal, M.M., Al-Gamal, S.A., Abdel-Hay, F.A., Zahar, H.M., 2000. Marine radioactivity studies in the Suez Canal: experiments and a modelling study of dispersion. Estuar. Coast. Shelf Sci. 50, 503e514. Baeza, A., del Rı ´o, L.M., Miro ´, C., Paniagua, J.M., 1991. Radiological impact of the Almaraz Nuclear Power Plant (Spain) during 1986 to 1989 on the surrounding environment. J. Radioanal. Nucl. Chem. 152, 175e188. 382 A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383 Baeza, A., del Rı ´o, L.M., Garcı ´a, E., Jime ´nez, A., Miro ´, C., Paniagua, J.M., Rufo, M., 1997. Temporal evolution of the 3 H levels in the surface waters around the Almaraz Nuclear Power Plant. J. Radioanal. Nucl. Chem. 219, 25e31. Baeza, A., Garcı ´a, E., Miro ´, C., 1999. A procedure for the determination of very low activity levels of tritium in water samples. J. Radioanal. Nucl. Chem. 241 (1), 93e100. Baeza, A., Brogueira, A.M., Carreiro, M.C.V., Garcı ´a, E., Gil, J.M., Miro ´, C., Sequeira, M.M., Teixeira, M.M.R., 2001. Spatial and temporal evolution of the levels of tritium in the Tagus River in its passage through Ca ´ceres (Spain) and the Alentejo (Portugal). Water Res. 35 (3), 705e714. Baeza, A., Dı ´az, M., Garcı ´a, E., Miro ´, C., 2002. Influence of interbasin transfers between the Alca ´ntara and Guadiloba reservoirs on the radiological quality of the drinking water of the city of Ca ´ceres (Spain). J. Radioanal. Nucl. Chem. 252 (3), 441e449. Baeza, A., Garcı ´a, E., Miro ´, C., Rodrı ´guez, A., Sequeira, M.M., 2004. Behaviour of tritium in the waters of the River Tagus. In: Seventh International Conference on Tritium Science and Technology. Com. Num. F15-P2, Baden-Baden, Germany. Breton, M., Salomon, J.C., 1995. A 2D long term advection dispersion model for the Channel and the southern North Sea. Part A: validation through comparison with artificial radionuclides. J. Mar. Syst. 6, 495e513. CSN. Consejo de Seguridad Nuclear, 1998. Informe al Congreso de los Diputados y al Senado. CSN/I, C.S.N, Madrid, Espan ˜a. Fontes, J.Ch., 1983a. Dating of groundwater. In: Guidebook on Nuclear Techniques in Hydrology. International Atomic Energy Agency, Technical Report Series, vol. 91. International Atomic Energy Agency, Vienna, pp. 285e317. Fontes, J.Ch., 1983b. Groundwater in fractures rocks. In: Guidebook on Nuclear Techniques in Hydrology. International Atomic Energy Agency, Technical Report Series, vol. 91. International Atomic Energy Agency, Vienna, pp. 337e350. Garcı ´a, E., 2000. Utilizacio ´n del tritio para modelar la evolucio ´n radiactiva del rı ´o Tajo a su paso por Extremadura. PhD thesis (in Spanish). Universidad de Extremadura, Espan ˜a. IAEA. International Atomic Energy Agency, 1990. Status Report of Controlled Thermonuclear Fusion: Executive Summary and General Overview. Prepares by the International Fusion Research Council. IAEA, Vienna Report number: STI/PUB/872. Ed. IAEA, Vienna, Austria. ICRP. International Council on Radiological Protection, 1978. Annals of the International Commission on Radiological Protection. Part I: limits of intake of radionuclides by workers. ICRP Publication 30. Pergamon Press, Oxford, U.K. Kaufman, S., Libby, W.F., 1954. The natural distribution of tritium. Phys. Rev. 93, 1337e1344. Okada, S., Momoshima, N., 1993. Overview of tritium: characteristics, sources and problems. Health Phys. 65, 595e609. Okubo, A., 1971. Oceanic diffusion diagrams. Deep Sea Res. 18, 789e802. Peria ´n ˜ez, R., 2000. Modelling the tidal dispersion of 137 Cs and 239,240 Pu in the English Channel. J. Environ. Radioact. 49, 259e277. Peria ´n ˜ez, R., 2002. The enhancement of 226 Ra in a tidal estuary due to the operation of fertilizer factories and redissolution from sediments: experimental results and a modelling study. Estuar. Coast. Shelf Sci. 54, 809e819. Peria ´n ˜ez, R., 2005. Modelling the Dispersion of Radionuclides in the Marine Environment. Springer-Verlag, Berlin. Prandle, D., 1974. A Numerical Model of the Southern North Sea and River Thames. Institute of Oceanographic Sciences, Bidston, U.K. Report number 4. Schonfeld, W., 1995. Numerical simulation of the dispersion of artificial radionuclides in the English Channel and the North Sea. J. Mar. Syst. 6, 529e544. Scho ¨nhoffer, F., Henrich, E., 1987. Recent progress and application of low level liquid scintillation counting. J. Radioanal. Nucl. Chem. 2, 317e322. Streeter, V.L., Wylie, E.B., 1988. Fluid Mechanics. McGraw-Hill, Mexico. Thibodeaux, L.J., 1979. Chemodynamics Environmental Movement of Chemical in Air, Water and Soil. John Wiley, New York, USA. UNSCEAR. United Nations Scientific Committee on the Effects of Atomic Radiation, 1988. Sources, Effects and Risks of Ionizing Radiation. United Nations, New York, USA. 383A. Baeza et al. / J. Environ. Radioactivity 86 (2006) 367e383