Assessment of Metals in Sediments in a Tributary of Guadalquivir River (Spain). Heavy Metal Partitioning and Relation Between the Water and Sediment System

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/ 10.1023/a:1005203012514”. ASSESSMENT OF METALS IN SEDIMENTS IN A TRIBUTARY OF GUADALQUIVIR RIVER (SPAIN). HEAVY METAL PARTITIONING AND RELATION BETWEEN THE WATER AND SEDIMENT SYSTEM A. ENGUIX GONZÁLEZ, M. TERNERO RODRíGUEZ, J. C. JIMÉNEZ SÁNCHEZ, A. J. FERNÁNDEZ ESPINOSA and F. J. BARRAGÁN DE LA ROSA∗ Department of Analytical Chemistry, University of Seville, Avda. Reina Mercedes s/n, E-41012 Seville, Spain Abstract. A physical-chemical characterisation of the sediments of the Guadaira River in Southwest Spain has been carried out, determining twenty-three parameters per sample. Non-treated urban sewage and industrial wastewaters from olive oil and table olives affect this river. The factor analysis of the data shows high organic matter contents with metallic pollution. Consequently, points downstream stand out because of very high organic and metallic pollution (Zn, Pb, Cr, Cu and Cd, in order of abundance, showing concentrations in sediments ranging from 50–7500 times those of waters). Heavy metals partitioning studies can be a good approach to show the real activity of these metals in the aquatic environment. We applied Tessier’s analytical sequential extraction scheme to sediments. Sediments would be in general mildly hazardous because we found low levels of easily interchangeable metals in the first fraction. Nevertheless, that is not the case in the river’s last stretch, precisely where sediments are most abundant, the heavy metal levels largest and dredgings are made periodically. Here, their disposal is of concern; since the actual use of these sediments as organic amendment for agricultural soils is unacceptable. Interaction between waters and sediments in the riverbed by means of correlation, factor and cluster analysis on waters and sediments data, was investigated. With this approach we concluded that in the final stretch there is a problem of contamination of both, sediments and waters. However, in the middle course only sediments are actually polluted. This corresponds to the previous research, when the situation of waters in this zone was even worse. Keywords: analysis of river sediments, cluster analysis, correlation analysis, factor analysis, Guadaira River, metal partitioning, water-sediment interaction 1. Introduction Pollution in river waters goes together with pollution of riverbeds, as a consequence of suspended solids sedimentation and fixation over these sediments of various pollutants. Thus, the organic or clay-like sludges that accumulate on the riverbed specially favour fixation to heavy metals, so sediments usually function as a trap, which removes some pollution from waters. In this way, sediment composition can indicate chemical forms that after dumping do not remain soluble in waters. On the other hand, sediments can also act as a contamination source under other conditions, such as the appearance of strong currents, dredgings, the presence of chelating agents, changes of pH, potential redox or salinity, etc. Then pollutants accumulated in sediments can return to waters in suspended or dissolved form. Therefore, they represent a potential risk for an aquatic environment and any water study ought to include an analysis of the sediment water comes into contact with. Many works about heavy metals in sediments (Förstner and Wittmann, 1983; Salomons and Förstner, 1984; Birch et al., 1996) refer to the overall metal concentration. These studies imply that all the existing forms of a metal in sediments have the same environmental effect and the same possibility of being extracted by waters, which is a long way from being true. Therefore, trace metals extraction studies gained in importance until 1990 (Samanidou and Fytyanos, 1987; Rauret et al., 1988; Pardo et al., 1990), on the basis of potential metals leaching, transport and bioavailability. Direct determination of specific sediment-trace metal associations is difficult, if not impossible, due to the great variety of solid phases that can bind trace metals, their amorphous character, and the low contents of metals involved. The partitioning of metals by a sequential extraction procedure is an alternative. Although the distribution of a metal does not reflect necessarily the association with discrete sediment phases, it can define the availability of some chemical entities by the analytical scheme itself. In this simple approach, the metals are analysed ingroups after a chemical scheme. There are several studies about sample integrity after sampling and during analysis. In order to guarantee accurate representation of the concomitant groups, it is necessary to take some experimental precautions (Thompson et al., 1980; Saeki et al., 1993). In this work we use the Tessier scheme considered as the most general (Tessier et al., 1979) with recommended precautions for sample preservation and manipulations procedures to prevent exposure to atmospheric oxygen. The sediments of the Guadaira River are studied in this work for the first time, implementing a former study method for waters (Enguix et al ., 1993). In the present paper, we establish the physical-chemical properties of the sediments and the availability to the environment of their trace metals. The interest of this work is to present the impact of very specific human activities on the sediment quality of a river in Southwest Spain in two clear recognisable areas. The first area is the upper and middle basin, with natural lands, agriculture and many olive industries, and the second is the lower basin, under the influence of important industrialised towns. An aim is also to know the interaction between waters and sediments by applying multivariate statistical analysis to waters and sediments data jointly. This approach demonstrated value for interpretation of pollution sources. This study also interfaces with policymakers involved at present with a Restoration plan in Guadaira’s basin and specifically with the assessment for the potential environmental problems in the use of sediments for soil amendment. Figure 1. Map of the area with sampling points, human activities, effluents and discharge types. 13 RELATION BETWEEN WATER AND SEDIMENT SYSTEMS TABLE I General characterisitics of surface sediment samples taken from Guadaira river Sampling pH ECa Carbonate OCb Grain size distribution site (mS cm−1) (%) (%) clay (%) silt (%) sand (%) 1 7.9 1.40 26.3 1.63 36.3 11.7 52.1 2 8.0 2.07 47.1 1.65 61.4 21.2 17.7 3 7.7 2.01 24.4 1.62 34.6 4.7 60.9 4 7.8 1.75 48.8 1.58 39.4 10.1 50.0 5 7.7 2.13 42.0 1.76 63.2 11.0 26.3 6 7.7 1.57 27.8 0.43 34.3 4.3 61.5 7 7.5 2.56 26.0 2.02 55.8 38.3 10.4 8 7.0 2.66 19.7 13.10 77.7 17.8 5.2 9 7.0 2.17 27.5 16.60 67.5 22.0 10.5 10 7.6 1.83 30.5 0.34 60.9 12.8 25.2 a Electrical conductivity. b Organic carbon. 2. Materials and Methods Ten sampling sites of sediment in the main river course were chosen (Figure 1). Water samples from the same points were also collected. In a previous study (Enguix et al., 1993) using a 30 point network, waters were analysed for two years to gain a better idea about the environmental situation in the basin. A single field collection of sediment and water samples was carried out in February 1993. All the parameters considered for sediments and the main of waters (among the 30 analysed samples) appear in Tables I and II. Sampling and analysis of sediments were done using standard methods (ASASSSA, 1982). Surface sediment grab samples were taken (0–10 cm). After sampling, samples were quickly packed in a propylene flask (about 1.5–2 kg in dry weight), hermetically closed and frozen as soon as they reached the laboratory. Subsamples were de-frozen, dried at ambient temperature and a fraction <2 mm was separated by sieving to determine granulommetry (Day, 1965) and total metal contents. Sample digestions for total metals determinations were performed using the method of Bello et al. (1994). Referring to the partitioning of trace metals in sediments, we followed the sequential scheme of Tessier (Tessier et al., 1979). Five fractions are obtained. These fractions related to metal leaching from sediments by successive extractants simulating stronger chemical conditions. These fractions can be operationally defined and obtained as indicated below: TABLE II Intervals values and average values of the parameters analyzed in sediments and waters of the Guadaira river Parameter Sediments Waters Interval Average value Interval Average value pH 7.0–8.0 7.6 7.2–7.8 7.5 Conductivity 1401–2664 2017 (µS cm − 1) 2090–5720 3088 (µS cm − 1) Dampness at 105 ◦ C 0.9–4.5 2.5 (%) – – Losses at 500 ◦ C 2.1–20.0 6.8 (%) – – Losses at 800 ◦ C 11.6–24.4 16.9 (%) – – Carbonates 19.7–48.8 32.0 (%) – – Bicarbonates – – 154–344 248 (mg L − 1) Sulphates 0.08–1.92 0.48 (%) 53–579 171 (mg L − 1 ) COD (chemical – – 7–392 127 (mg L − 1 ) oxygen demand) BOD5 (biological – – 2–362 82 (mg L − 1) oxygen demand) Organic carbon 0.34-16.60 4.07 (%) – – Ammonium 2.2–218.7 42.7 (mg kg − 1) 0.1–19.0 8.2 (mg L − 1) Nitrites 0.00-19.22 4.31 (mg kg − 1 ) 0.01–0.73 0.14 (mg L − 1 ) Nitrates 0.01–2.06 0.34 (mg kg − 1 ) 0.1–17.1 5.6 (mg L − 1 ) Total Kjeldahl N 715–6548 2060 (mg kg − 1 ) – – Total phosphorous 653–5757 1861 (mg kg − 1 ) 0.1–9.4 5.0 (mg L − 1 ) Iron 11.14–33.71 20.21 (g kg − 1 ) 0.06–0.40 0.21 (mg L − 1 ) Aluminium 101.3–50.4 72.3 (g kg − 1 ) 0.05–1.44 0.14 (mg L − 1 ) Manganese 27.4–754.2 336.0 (mg kg − 1 ) 0.109–0.276 0.156 (mg L − 1 ) Zinc 43.1-1033.0 233.6 (mg kg − 1 ) 0.02–0.19 0.04 (mg L − 1 ) Lead 15.8–388.7 68.9 (mg kg − 1 ) 0.036–0.0205 0.0081 (mg L − 1 ) Chromium 17.0-188.8 64.1 (mg kg − 1 ) 0.0–0.05 0.02 (mg L − 1 ) Copper 13.9-142.6 43.2 (mg kg − 1 ) 0.01–0.02 0.01 (mg L − 1 ) Nickel 18.1–46.7 32.1 (mg kg − 1 ) 0.001–0.021 0.007 (mg L − 1 ) Cadmium 2.0–9.1 3.7 (mg kg − 1 ) 0.00 0.0006 (mg L − 1 ) Arsenic 0.7–5.2 1.9 (mg kg − 1 ) 0.0001–0.0033 0.0019 (mg L − 1 ) Mercury 0.0–0.059 0.014 (mg kg − 1 ) 0.0005–0.0020 0.00 (mg L − 1 ) – Interchangeable metal (FI) (treating sediment with 50 mL of Mg2Cl 1 M and agitation during 60 min at room temperature). – Metal linked to carbonates (FII) (treating the beforehand residue with 25 mL of acetic acid 0.11 M with agitation during 60 min at room temperature). – Metal linked to iron and manganese oxide (FIII)(extract the residue agitating with 25 mL of hydroxylammonium chloride 0.1 M in acetic acid 25% (v/v) during 2 hr at 96 ◦C). – Metal bound to organic matter and sulphides (FIV) (adding 25 mL of HNO3 0.02 M and 15 mL of 30% (v/v) H2O2, heating at 85 ◦C under agitation for 2 hr). – Residual metal (FV) (treat with 10 mL of concentrated HF, 15 mL of concentrated HNO3 and 1 mL of concentrated HClO4 in a teflon vessel. Heat for 5 hr at 140 ◦C on a stove). Tessier’s scheme was applied directly on duplicates of defrozen samples (wet subsamples of about 1 g on dry basis) taking into account the precautions proposed by Rapin et al. (1986). Thus, the samples were manipulated throughout under oxygen-free conditions to suppress the anoxic character of sediments. The extracting solutions were previously deaerated by N2 gas bubbling, shakings and centrifugations as fractions I to IV were made under N2 atmosphere. The centrifugal tubes were closed with a rubber stopper with small connecting tubes to withdraw the air in the upper part of the tube with N2 and to remove through them the supernatants. Immediately after this, the residue from the previous step is added through the connections with the next extracting solution. Metals concentrations in the whole sample and the lechates were determined by hydride generation (As) and flame atomic absorption spectrometry, when concentrations were close to detection limits Graphite Furnace atomic absorption was used (PERKIN ELMER 2380 spectrometer and HGA programmer-300). The analysis procedures have been validated by estuarine sediment certified sample (CRM 277). Conventional and multivariate statistical techniques for data treatment were used, such as correlation between parameters, Factor Analysis (Principal Components Analysis followed by varimax rotation) and Cluster Analysis (Single linkage of nearest neighbour) for both variables and cases. Water data (Enguix, 1995) and sediment data (present work) were submitted to the aforementioned statistical techniques in the waters-sediments interaction study. We considered the average results of seven samplings of river water (taken from 1990–1993) and one sample of sediment (taken in February 1993). The water average values were considered more representative because they are less sensitive to actual pluviometry. The statistical studies were performed by means of a set of computer programs (CCSS-Statsoft version 3.1, Statsoft Tulsa, Oklahoma). 3. Area Description Guadaira River is one of the last tributaries of the Guadalquivir River from its east (Figure 1). Its length is 130 km and the extension of its drainage basin is about 1360 km2 with a population density of 150 inh km−2. On its banks, Sevilla, Dos Hermanas and Alcalá de Guadaira are the largest populated towns and the most industrialised areas in the basin. There are many medium size factories on their outskirts, which imply potential sources of pollutants: cement, building materials, glass, pottery, painting, chemicals and especially food and drink processing industries. Along its middle and upper basin are another six main towns with an economy based mainly on agriculture and farming. Here, the river has been traditionally polluted with alpechines (olive-oil mill effluents), and other effluents from a very specific industry in the Guadaira basin: that is a pickled olives factory (including its canning). Nevertheless, the pollution caused by sewage waters from all towns in the basin is even worse than industrial pollutants (Enguix, 1995). The Guadaira River system is mainly made up of little streams originated by superficial run-off, that generally only run in the rainy season. In the upper (sampling points Nos. 1–3) and middle (Nos. 4–6) basin the Guadaira River has noticeable ecological and landscape values despite the present pollution, remarkable by a gallery forest along it. In the lower basin (Nos. 7–10) the river is channelized until it empties into Guadalquivir estuary (the marine flood enters the Guadaira River’s last stretch). This 22 km long channel must be dredged completely every ten years due to the large amount of sediments that form an artificial swampy area across which waters flow very slowly. The sediment disposal is of great concern, because agricultural destination is prefered. In summer or periods of persistent draught, it can be said that the Guadaira River is made up almost entirely of sewage waters (river flow of 2 m3 s−1). It can be described in several spots as a real outdoor sewer with foams. Stinking water problems are common in the area. In fact, the painful situation of the Guadaira River has often been published in the press. 4. Results and Discussion The inclusion of all considered analytical results is not considered necessary. Table I shows the most general characteristics of all sediment samples. Nevertheless, the most relevant individual results of sediments (referred to dry weight of sediment) will be commented. Average values and intervals of studied parameters are shown in Table II to give a global idea of the Guadaira’s waters (Enguix et al., 1993) and sediments. 4.1. G ENERAL CHARACTERISTICS OF SEDIMENTS Samples 1, 3, 4 and 6 have sandy/clay-like textures, and the others are predominantly clay-like, especially 7–10. Samples 8 and 9 stand out by extremely rich organic matter (Table I). The conductivity of sediments is, in general, high and pH is close to 7.7 except for samples 8 and 9 (Table I). Figure 3. Spatial evolution of the factors in the factor analysis of the sediments. 24 A. ENGUIX GONZA ´ LEZ ET AL. have been discarded, because that would add unnecessary complexity to the data set and to the interpretation. 7.1. C ORRELATION BETWEEN SEDIMENT AND WATER PARAMETERS According to the correlation matrix, it stands out that ignition losses at 800 ◦C and carbonates in sediments are negatively correlated with the following parameters of waters: COD, BOD5, ammonium, phosphates, oils and fats, and others; that is, waters with urban sewage pollution. This is because the points, whose sediments have higher levels of carbonates, correspond to situations where the organic pollution of their waters has been already degraded by oxidation. Likewise, carbonates in sediments correlate well positively with some natural components of the waters: chlorides, bicarbonates, sodium and lithium. Ammonium in sediments is correlated with water parameters such as colour, COD, BOD5, oxidability, ammonium, phosphates, sulphates and certain metals, and negatively with pH, dissolved oxygen, redox potential and nitrates. In general, total phosphorus and ammonium in sediments behave in the same way. This points out the permanence in sediments of faecal contamination in river stretches where the stream is steadily polluted. An analogue conclusion is valid also with respect to metals, because the anoxic stretches ease also their retention in sediments with levels up to ten thousand times greater than in waters. These correlations let us check that a high pollution of some sediments corresponds parallel to a highly polluted stream, clearly with regard to organic matter and sometimes to trace metals. Cd and Pb in waters are well correlated with Cr, Cu and Zn in sediments. This reveals, that although all of them have same sources, Cd and Pb are more leachable from sediments and in fact are present mostly in fraction FII and FIII, while Zn is preferably found in FIII and FIV, and finally Cr and Cu have affinity for fractions FIV and FV (according to the earlier partitioning study in sediments). 7.2. F ACTOR AND CLUSTER ANALYSIS APPLIED TO SEDIMENTS AND WATERS RESULTS Factor analysis applied to the ten sampling sites renders five factors, which explain 86.8% of the total variance (Table IV study of waters-sediments). Maps with variable positions for all different pairs of factors have been evaluated, considering the sediments and/or waters parameters picked up by each of them. Factor F1WS is the only one related to sediment contamination, by means of organic and metallic components, since it comprises only sediment parameters. Factor F2WS shows waters hardness associated largely with gypsum coming from the southwest of Guadaira’s basin, corresponding also to absence of urban or industrial pollution. Factor F3WS represents the stream pollution from urban sewages and Factor F5WS the salinity related now to the predominant natural sodic chlorided chemical facies of river waters. Finally, Factor F4WS, which gathers Figure 4. Spatial evolution of the factors in the waters-sediments joint study. iron, aluminium and nickel in waters, stands out in point 10, maybe because of elemental resolubilization into the stream from the associated sediments under the saline flood influence this site. In this study position maps of the observations for all possible factor pairs have been represented, although not showed here. Factor variations (scores) along the river are showed in Figure 4. Thus, the sediment pollution in points 8 and 9 are confirmed (F1WS high scores). However it might be emphasised that point No. 8 does not presently show water contamination (F3WS low scores). Point Nos. 1 and 3 have similar characteristics, because of significant hardness in the waters (F2WS high scores), and subsequent lack of pollution in sediments and waters (F1WS low scores). The high pollution level that waters reach at points 3 and 9 (high F3WS values) is consistent with their respective situation after the two most degraded tributaries of Guadaira River. Other sites present some uniform characteristics. Figure 5. Dendrogram of sampling points obtained from a Cluster analysis on water and sediment data. The dendrogram, corresponding to the cluster analysis of the cases applied to the waters and sediments, reinforces the factor analysis conclusions (Figure 5). Points 8 and 9, and points 1 and 2 make well-defined clusters because of their respective aforementioned similarities and the remaining points constitute another cluster, except point No. 3 which has a slightly different character because of its water pollution. 8. Conclusions The methodology applied in this study has proven to be adequate for sediment quality and aquatic pollution assesement of a river in South West Spain. It is at present subjected to a Corrective Plan. Sediments for points located in the upper basin show salinity (sulphates and chlorides) due to lithological causes. On the other hand, the most polluted sediments are in the river’s final stretch, showing high values for many parameters owing to the progressive river degradation. Heavy metal average levels reveal that sediments according to USEPA disposal criteria (Salomons and Förstner, 1984) are moderately polluted for Cd, Cr, Ni and Cu. The tolerable limits are surpassed for Pb and Zn. River sediments in the final stretch (Nos. 8, 9, 10) have to be dredged periodically and they are used nearby as organic amendment for agricultural soils. Our study reveals that this practice is environmentally unacceptable. The partitioning study of heavy metals reveals that contents in the first fractions, which are the most hazardous from the environmental point of view, are imperceptible in most of the samples. Lead and cadmium, which present the lowest values in the residual fraction, would represent the highest danger of leaching. The statistical multivariate analysis applied to the sediments analytical data has led to recognition of points 8 and 9 as the most polluted by organic contamination and secondarily by heavy metals; mainly Zn, Pb, Cr, Cu and Cd, in this order. The presence of a great amount of phosphorus and ammonium in these sediments proclaims their source in sewage waters, but to a lesser extent the influence of industrial wastewaters containing metals can be observed. The influence of the polluted stream over sediments has been confirmed through a multivariate analysis applied now to both analytical data from sediments and from their contacting waters. This has let us classify the sampling points depending on its analytical characteristics. Points 1 and 2 show a high salinity in waters related to their respective natural sulphated and chlorided facies. They have also a remarkable absence of pollution in waters and sediments. Other points stand out for water pollution (point Nos. 3 and 9), because they directly receive the influence of very polluted streams. Points 7, 8 and 9 show sediment pollution (the water anaerobiosis increases in this order). Point 7 and even more 8 reflect pollution above all in sediment, possibly due to an old contamination before installation of treatment plants in this area, as well as vegetation contributing to settle the suspended matter in the stream. Thus, waters are not so polluted as sediments are. Sediment No. 9 shows the worst situation and reveals that the river flows there permanently with an excessive faecal and, in small measure, industrial contamination. Finally in point No. 10, since it is located in a zone affected by Guadalquivir’s estuarine flood, the sediment is frequently removed and does not show the characteristics of the former points 8 and 9. Acknowledgements We thank the ‘Consejería de Educación y Ciencia de la Junta de Andalucía’ for supporting this study with the help of a scholarship of ‘Formación del Personal Investigador’. We are grateful to ‘Confederación Hidrográfica del Guadalquivir’ and implied city councils for their facilities. References ASA-SSSA: 1982, Methods of Soil Analysis part 2-Chemical and Microbiological Properties, 2nd Edition. ASA-SSSA, Madison, Wisconsin. Bello, M. A., Callejón, M., Jiménez, J. C., Pablos, F. and Ternero, M.: 1994, Toxicological Environmental Chemistry 44, 203. Bejarano Bravo, M. and Madrid, L.: 1995, Environmental Thechnology 17, 427. Birch, L., Hanselmann, K. W. and Bachofen, R.: 1996, Water Research 30, 679. Campos, J. M., Usero, J. and Gracia, I.: 1990, ‘Contaminación por metales en los sedimentos del rio Tinto’, Tecnología del Agua No. 2◦ monog., pp. 49–56. Day, P. R.: 1965, ‘Particle Fractionation and Particle-size Analysis’, Methods of Soil Analysis, Part 1, American Society of Agronomy, pp. 545–567. Enguix González, A., Ternero Rodríguez, M., Jiménez Sánchez, J. C. and Barragán de la Rosa, F. J.: 1993, Tecnología del Agua 114, 65. Enguix González, A.: 1995, Ph. D. Thesis. Facultad de Químicas, Univ. Sevilla. Förstner, U. and Wittmann, G. T. W.: 1983, Metal Pollution in the Aquatic Environment, Springer- Verlag, Berlín. Pardo, R., Barrado, E., Pérez, L. and Vega, M.: 1990, Water Research 24, 373. Pinto, J., Galindo, M. D., Martínez, M. P. and García-Vargas, M.: 1996, Tecnología del Agua 148, 89. Rapin, F., Tessier, A., Campbell, P. G. C. and Carignan, R.: 1986, Environmental Science and Technology 20, 836. Rauret, G., Rubio, R., López-Sánchez, J. F. and Cassasas, E.: 1988, Water Research 22, 449. Saeki, K., Okazadi, H. and Matsumoto, S.: 1993, Water Research 27, 11243. Salomons, W. and Förstner, U.: 1984, Metals in the Hydrocycle, Springer-Verlag, Berlin. Samanidou, V. and Fytyanos, K.: 1987, Science of Total Environment 67, 279. Tessier, A., Campbell, P. G. C. and Bisso, M.: 1979, Sequential extraction procedure for the speciation of particulate trace metals. Analytical Chemistry 51, 844. Thomson, E. A., Luoma, S. N., Cain, D. J. and Johansson, C.: 1980, Water, Air, and Soil Pollut. 14, 215.