Original article
Simultaneous use of geochemical and geophysical methods to characterise abandoned landfills Jose´ F. Noguera Æ Lluı´s Rivero Æ Xavier Font Æ Andre´s Navarro Æ Francisco Martı´nez
Abstract Groundwater sampling and geophysical methods determined a serious contamination problem associated with refilled exploitation sites at the Cal Dimoni area, Llobregat delta, Barcelona, Spain. To characterise this process, hydrogeochemical analyses were performed and showed the following modifications to groundwater chemical composition: increasing pH values, changing redox conditions, significant increases in total organic carbon (TOC) and certain trace elements, and high groundwater conductivity values. Major ion content accumulations were found under the refilled area. In contrast, elements involved in the oxidation–reduction processes, such as iron, manganese and nitrates, clearly diminished. Electromagnetic prospecting methods were also performed and delineated the contamination plume extent. These methods also showed separate sources of contamination, one clearly related to the groundwater–refilled zone leachate interaction, another as a consequence of the manure–accumulation surface site. Geochemical and geophysical methods have shown similar results for
Received: 5 January 2001 / Accepted: 8 November 2001 Published online: 23 January 2002 ª Springer-Verlag 2002 J.F. Noguera (&) Æ F. Martı´nez Department of Geology, Petrology and Geochemistry Unit, Edifici C (s) Campus, Universitat Auto`noma de Barcelona, 08193 Bellaterra, Barcelona, Spain E-mail:
[email protected] Tel.: +34-93-5813092 Fax: +34-93-5811263 L. Rivero Æ X. Font Economical and Environmental Geology Group, Geochemistry, Petrology and Geological Prospecting Department, Zona Universita`ria de Pedralbes, Universitat de Barcelona 08071, Barcelona, Spain A. Navarro Department of Fluid Mechanics, Escola Te`cnica Superior d’Enginyers Industrials, Universitat Polite`cnica de Catalunya, c/ Colo´n 7, 08222 Terrassa, Barcelona, Spain
898
Environmental Geology (2002) 41:898–905
locating groundwater contamination sources, and for determining leachate generation mechanisms and flow paths. Keywords Electro-magnetic prospecting Æ Environmental assessment Æ Groundwater contamination
Study area The study area – Cal Dimoni – extends for about 5 km2 on the right margin of the Llobregat delta located close to Barcelona city, Spain (Fig. 1). In the early 1970s, a period of sand and gravel exploitation began that lasted up to the late 1980s. These exploitation sites were abandoned and refilled in a non-controlled way with a variety of waste materials including industrial, chemical and urban wastes, which created a major threat for groundwater and soils at such sites (Font et al. 1998). From a geological point of view, the exploited materials are mainly detritic sediments of Quaternary age discordantly overlying Pliocene materials. This association develops a typical cone-shaped delta similar to all Mediterranean rivers such as the Ebro and Rhone (Fig. 2). The stratigraphic column of the Llobregat delta is divided into four units (Marques 1984): • Blue marls, sands and shales with marine fauna (Pliocene). • Locally cemented sands and gravels interpreted as fluvial terraces of Pleistocene age. This unit constitutes the Llobregat delta lower aquifer, with a range in permeability values from 10 to 35 m day–1, and variable thickness. • Black clay siltstones with organic matter from a marine environment (Holocene to Present) can be considered as an aquitard, and confines the lower aquifer. • Clean sands with a local accumulation of gravel, and lacustrine shales with peat. This unit corresponds to the Llobregat delta upper shallow aquifer and covers the entire surface of the emerged delta; its sediments are of deltaic, fluvial and littoral origins. With variable
DOI 10.1007/s00254-001-0467-x
Original article
Fig. 1 Geological map of the study area (shaded)
Fig. 2 Stratigraphic section of Llobregat delta. Note the disposition of aquifer levels separated by an impermeable shale and silt layer (modified from MOPU 1966)
thickness, a mean saturated thickness of 6 m and a permeability value of 10 m day–1 this unit is the main source for sand and gravel exploitation, as well as the site of the non-controlled landfill.
• A silt and shale level (Fig. 4c) at 15 m depth, which is homogeneous and impermeable. • A sand and gravel level (Fig. 4b), with a total thickness of 2 m and high values of porosity and permeability. This is the exploited level of the gravel pits. • A refilled level (Fig. 4d). This is the former level of sand and gravel exploitation, and is now completely refilled with all kinds of refuse from industrial to urban wastes. Its approximate thickness is 8 m. Changes in phreatic level generate leachates that are responsible for groundwater pollution at this site.
On a local scale, two farming samples from wells (numbered 33 and 37), a swamp sample (40) and five sampling The stratigraphic column of sampling well number 32 wells (31, 32, 34, 35, 36) were used to characterise the presents, from bottom to top (Fig. 4): landfill zone (Fig. 3). These wells also allowed the subsoil • A gravel and sand level (Fig. 4b) between 5 and 2 m structure to be determined. depth that shows a level of shore sands with abundant The stratigraphic column of sampling well number 31 fossils, especially foraminifer shells. (Fig. 4) presents, from bottom to top:
Environmental Geology (2002) 41:898–905
899
Original article
Fig. 3 Location map of sampling wells (31, 32, 34, 35 and 36) and farming wells (33 and 37) close to the Cal Dimoni swamp (40). Shaded area shows swamp location, and striped area the gravel exploitation loci that was refilled. A–A¢ indicates the cross section position in Fig. 4. Arrows indicate the local groundwater flow
Fig. 4 Local structure reconstructed on the basis of sampling well data. a Sand and clay level; b gravel and sand level; c silt and clay level; d refill material; p.l. phreatic level
filled polyethylene sampling bottles were stored below 5 C until analysis was performed. Temperature (C), platinum electrode potential (mV), specific conductance (lS cm–1) and the pH of each sample were measured in the field. Each sample was analysed for the major cations (Na+, K+, Ca2+, Mg2+), major anions (Cl–, SO24 , HCO3 , NO3 , NO2 ), total organic content (TOC), trace elements [Fe (total), Mn, Al, Ba, Sr, Pb, Li, B, P) at the Scientific and Technical Survey of the Universitat de Barcelona. With these results, the possible extent of contamination caused by the interLocal flow direction was determined using groundwater depth values. A swamp at the centre of the study area was actions between groundwater and waste leachates could be formed as a consequence of sand and gravel exploitation, assessed. and now constitutes a protected area. The position of a Electromagnetic prospecting small landfilled zone and flow direction can be clearly Electromagnetic prospecting is a method of obtaining inidentified (Figs. 3 and 4). formation about subsoil structure by measuring the electromagnetic field propagation. The electromagnetic field is a combination of electric and magnetic fields that interact in a perpendicular way. The response to this propagation Methodology is referred to as apparent conductivity, and is defined as the electric current quantity that can propagate through a Systematic sampling of groundwater certain material. This type of conductivity is approxiSystematic sampling of groundwater was carried out to characterise and determine the evolution of the Llobregat mately the same as real conductivity when the material is homogeneous. Apparent conductivity is also defined as the delta upper shallow aquifer. To obtain the samples, an electric submersible pump with PVC tubing was used. The inverse of resistivity and its units in the International • A sand and clay level (Fig. 4a) with a maximum thickness of 2 m. This level forms delta plain deposits with clay and silt deposits (between 2 and 1 m depth) and vegetal soil (between 1 and 0 m depth). • No refilled level can be found at this sampling well because, in this study, it is considered as the base or local reference well because it has minor concentrations of dissolved ions and heavy metals.
900
Environmental Geology (2002) 41:898–905
Original article
System are mho m–1 or siemens m–1. If these values are very low, the units can be millisiemens m–1 (mS m–1). The rocks’ electrical conductivity is a function of porosity, the conductivity of the saturation liquid, texture and size of the pores, degree of saturation (fraction of saturated pores), temperature and the presence of intermediate to high cationic exchange capacity clays. The method of determining apparent electric conductivity consists of a transmission coil fed with alternating current and a receiver coil located at a certain distance from the transmission coil. The current at the transmission coil generates a primary magnetic field that propagates freely on the subsurface; in the presence of a conductive material, the magnetic component generates parasitic currents. These current types also generate a secondary magnetic field that can be detected by the receiver. The receiver additionally detects the primary magnetic field and transforms the combination value directly to apparent electrical conductivity in mS m–1. This information can be used to determine the geometry, size and extension of conducting bodies. This is a conservative method, and does not require any physical interaction with the measured terrain. As a consequence, electromagnetic prospecting is faster than other non-induced geophysical methods. Another important Table 1 Conductivity measurement depths as a function of frequency. HD Horizontal coil disposition; VD vertical coil disposition
Coil separation (m)
advantage is that the facility of transportation and measurements can provide information of a practically constant depth (3, 6, 7, 5 and 15 m.) as a function of coil spacing and orientation (Table 1). Ground conductivity measurements were carried out by electromagnetic methods using Geonics EM-31 and EM-34 instruments. Both devices use two portable coplanar coils. In this study, EM-31 was used with the basic aim of developing an intensive prospecting campaign specifically for the landfilled zone. EM-34 has been used on a regional basis (210,000 m2) to characterise leachate behaviour, and includes the EM-31 area.
Results Hydrogeochemical measurements A wide range of chemical compositions for groundwater samples were found at the Cal Dimoni area (Table 2). These samples show a neutral–alkaline trend (pH range 7.0–8.0), even in the refilled zone. The redox potential presents negative values or reducing conditions (from –40 to –90 mV) with the exception of sample numbers 35, 36 and 40, which had oxidising values.
Frequency (Hz)
4 (EM-31) 10 (EM-34) 20 (EM-34) 40 (EM-34)
Investigation depth
6,400 1,600 400
(HD)
(VD)
3 7.5 15 30
6 15 30 60
Table 2 Analytical results. Samples from the Cal Dimoni area. Sample 32 represents the local unaltered groundwater composition Sample T Eh pH Conductivity CO3 HCO3 SO4 Cl NO3 NO2 NH4 Na K Ca Mg TOC Fe Mn Al Ba Sr Pb Li B P
(C) (mV) (lS/cm) ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppb ppb ppb ppb ppb ppb ppb ppb
31
32
33
34
35
36
37
40
19 –82 7.2 3,300
19 –40 7.1 2,510
15.6 -90 7.1 5,200
18 –76 7.3 5,100
18 3 7.5 4,300
17.4 0.01 7.0 3,960
551 1,117 407 0.23 0.1 37 300 90 422 84 30 0.2 40 71 53 7,320 85 160 1,682 103
659 40 363 8.72 1.42 0.1 213 51 263 51 4 2.5 1113 67 372 3,609 9 13 294 249
676 701 841 1.47 2.04 1.4 519 44 240 160 14 12.3 433 7 193 4,845 52 51 689 1,423
1,410 375 921 0.1 2.97 25.9 801 84 134 141 74 0.2 159 24 208 4,416 49 230 7,177 608
18 18 7.3 2,730 25 553 253 288 15.3 0.1 0.1 194 60 268 97 21 0.6 606 19 71 3,329 22 48 289 1,747
621 1,352 653 0.03
729 605 884 1.2 1.94 0.1 655 21 159 163 19 0.5 255 41 92 2,859 40 60 885 96
25 207 8.0 5,330 33 446 509 1,009 0.37 0.1 0.1 634 127 112 165 11 0.02 20 73 157 2,775 8 62 793 177
1.2 475 14 329 213 14 0.2 136 22 55 5,883 17 81 866 44
Environmental Geology (2002) 41:898–905
901
Original article
Fig. 5 Conductivity value distribution map. Marked areas indicate swamp and landfilled zone. Dashed lines delineate EM-31 and EM-34 electromagnetic prospecting locations
a double triangular system that, on a major ion content basis, separates groundwater samples into different categories. A variety ranging from a calcic–sodic bicarbonate pole (sample 32) to a sulphate (sample 31) or chloride pole (sample 37) can be noted. Samples 34 and 40, located close to the refilled zone, are plotted at the chloride and sulphate poles. On the cations triangle, the general tendency is towards the calcic pole. The major tendency for anions is towards the sulphate and chloride poles. As a result of the different geochemical classification methods used, the main major ion distribution was determined. In Fig. 7, this major anion distribution shows a tendency for chloride composition of groundwater towards the north. This distribution reflects the local flow regime induced by over-pumping at farm wells 33 and 37. Classification of groundwater samples It also shows the theoretical spread of the contaminant in the study area In order to establish a classification of groundwater sam- plume generated by the landfill focus. ples, they were plotted on a Piper diagram (Fig. 6). This is Calcic–sodic bicarbonate samples Samples 32 and 35 represent the original chemical composition in the Cal Dimoni area because they are located outside the refilled zone and out of the leachate influence area. As contaminated site leachates interact with them, these waters show a degradation process. Their degree of mineralization is considered to be medium, and conductivity values are 2,500 lS cm–1. The ionic content or degree of mineralization for the samples varied from high to very high values (conductivity range: 2,500–4,000 lS/cm); sample numbers 33, 34 and 40 were the exception with conductivity values that were much higher. Figure 5 shows the distribution of conductivity values and also delineates the location of electromagnetic prospecting for EM-31 and EM-34. Maximum values are coincident with the refilled zone. A general trend to the north can also be noted, i.e. where extraction wells 33 and 37 are located. The original chemical composition of the water samples from the Cal Dimoni area is determined by sample 32; this can be classified as calcic or calcic–-sodic bicarbonate water (Table 2).
Fig. 6 The Cal Dimoni groundwater samples on the Piper diagram
902
Environmental Geology (2002) 41:898–905
Fig. 7 General distribution map of the samples’ major chemical composition. The Cal Dimoni area
Original article
Calcic sulphate samples Samples 31 and 36 have a high ion content; conductivity values range from 3,300 to 3,950 lS/cm and represent the original calcic bicarbonate samples contaminated by leachate interaction or by their own circulation through refilled zones. This contamination process is shown by the progressive increment of the sulphate content. Sodic chloride samples Samples 33, 34, 37 and 40 have a sodic chloride composition in the Cal Dimoni area. Their ion composition is very high and determines conductivity values higher than 4,000 lS cm–1. These samples are wholly affected by leachate interaction. Their contamination can be determined by a high degree of mineralization, reducing conditions and alkaline pH values (Table 2). A common chemical composition, especially for sodium and chloride, can also be noted in the nitrate and sulphate content, which can be interpreted as a landfill contamination process. Another important way to determine and characterise the contamination process by landfill leachate–groundwater interaction is to analyse trace elements and their possible implication (Table 2). Groundwater values for iron, manganese, strontium, lead, lithium, boron and phosphorous show significant variation from the original chemical values and refilled zone. In the early 1980s, the sand and gravel exploitations generated holes that were refilled with a variety of refuse, from industrial to chemical and even urban waste. In spite of the time that has lapsed, these refilled zones are still generating leachates because of variations in the piezometric level. Leachate–groundwater interaction results in a contamination process; this can be seen in Table 3 where the chemical composition for sampling well 33 varies
Table 3 Groundwater chemical composition before pumping (sample 33), and after 2 h pumping (sample 33¢). Samples were collected in a different sampling campaign from that reported in Table 2
significantly after intensive pumping of approximately 2 h. These variations include increments on most of the analysed chemical parameters, particularly for conductivity and pH values. Redox conditions also change from oxidising to reducing. Total organic carbon (TOC) shows a significant increase and as do certain trace elements such as aluminium, barium, phosphorus and boron; all of these also determine the changing conditions. Special mention should be made of elements involving redox reactions, such as iron, manganese and nitrate, because their concentration clearly diminishes. These results are coincident with other similar studies on landfills such as those of Pinto (Madrid; Dorronsoro et al. 1996) and Nabarniz in the Basque country (Morales et al. 1994). Abandoned landfills and associated problems have been intensely studied, principally because the generated leachates can contaminate groundwater and soils. MacFarlane and others, in 1983, produced a now classic work on this type of study. More specific work was developed on heavy metal thermodynamics and mathematical model validation at Borden site, Canada (Cherry et al. 1996). Electromagnetic study EM-31 prospecting at the Cal Dimoni area was carried out over almost 7,000 m2 (Figs. 5 and 8a) and directly on the abandoned landfill area. Another source of potential contamination is a current manure storage site located at the NW part of the study area. Electromagnetic measurements were carried out at 5-m intervals (340 total measurements). Figure 8b shows the conductivity map at 3 m depth and clearly marks an increase towards the NW sector (manure accumulation site), with conductivity maximum values of 170 mS m–1.
Sample no.
33
33¢
Temperature Eh pH Conductivity HCO3 SO4 Cl NO3 TOC Na K Ca Mg Fe Mn Al Ba Sr Zn Cu P Li B Ni Se Co
16.6 C +36 mV 7.36 3,090 lS/cm 684 ppm 552 ppm 633 ppm 0.41 ppm 16.06 ppm 568 ppm 48 ppm 223 ppm 137 ppm 12.5 ppm 568 ppb 115 ppb 211 ppb 12 ppb 22 ppb 10 ppb 1,583 ppb 51 ppb 819 ppb 58 ppb 50 ppb 26 ppb
16.8 C –240 mV 7.56 3,950 lS/cm 811 ppm 751 ppm 705 ppm 0.28 ppm 43.1 ppm 815 ppm 49 ppm 231 ppm 149 ppm 4.6 ppm 325 ppb 1,235 ppb 295 ppb 13 ppb 5 ppb 12 ppb 15,380 ppb 41 ppb 1,170 ppb 34 ppb 50 ppb 10 ppb
Environmental Geology (2002) 41:898–905
903
Original article
At 60 m depth, conductivity values show a particular distribution of high conductivity values in the SE and a diminishing trend to the north. These are lower aquifer values and the distribution pattern can be interpreted as a seawater intrusion and salinization process generated by It seems that a regional gradient from SE to NW is present. over-pumping of an exploitation well that is located partly In order to determine its effect, total conductivity values outside the study area (SE). were adjusted to a first degree polynomial equation using the least squares method. Results are shown in Fig. 8c. The area influenced by manure storage is clearly marked, and was determined by subtracting the polynomial adjustment Conclusions to total conductivity (Fig. 8d). On a regional scale, electromagnetic prospecting method- Almost 20 years have passed since the early 1980s saw the ology can be described as follows. The measurement instart of the refilling process in the Cal Dimoni area. In terval was approximately 2,500 m2 (i.e. 50 · 50 m grid) spite of this, it is still possible today to verify the leachate– over a total area of 210,000 m2. EM-34 was used at each groundwater interaction and the contamination process point to give the following different depth conductivity on the Llobregat delta upper shallow aquifer. This convalues of 7.5, 15, 30 and 60 m, and it generated a threetamination can be either through leachate–groundwater dimensional (3D) evolution of leachate generation (Fig. 9). mixing or through piezometric level variation. The maps of 7.5- and 15-m depth in Fig. 9 show the general Modification of groundwater chemical composition is tendency of the study area to have high conductivity values mainly expressed in increasing pH values and changing on the surroundings of sampling wells 31 and 34, i.e. a redox conditions: from oxidising to reducing after a few significant presence of leachate generation in the central years. Total organic content (TOC) and concentrations of refilled area. This conductivity trend also shows an exten- certain trace elements, namely aluminium, barium, phossion towards the N–NW sector. This trend does not show phorous and boron, increase significantly. the same behaviour for regional hydraulic gradient and is In contrast, elements involved in oxidation–reduction clearly oriented in a NE–SW direction. As a consequence, it processes (iron, manganese and nitrates) show an is easy to deduce leachate movement influenced by inten- important decrease in concentration. sive pumping at the location of farming well 33. The electromagnetic study showed an increase in the At 30 m depth, conductivity values can be interpreted as conductivity process as a consequence of waste leachate characteristic of the lower aquifer. As the geological col- generation and the manure–accumulation surface site. umn describes, under a silt and clay barrier there must be This manure accumulation clearly constitutes a source of ‘clean’ groundwater. But, as the measurements indicate, an contamination. From geophysical values, this contaminaincreasing degree of permeability over the theoretical tion gives positive values for residual conductivity. impermeable level can be deduced, and leachates are able On a regional scale, conductivity increases towards the to reach deeper levels. NW part of the study area. This trend is against the Fig. 8a–d Electromagnetic prospecting results (EM-31). a Study area location; b total conductivity map; c polynomial adjustment; d residual conductivity map
904
Environmental Geology (2002) 41:898–905
Original article
regional flow and can be explained by over-pumping from farmer wells 33 and 37. These are the only irrigation wells on the entire area. Geochemical and geophysical methods have shown similar results for locating groundwater contamination sources and for determining leachate generation mechanisms and flow paths. The principal aim of this study was the simultaneous use of these methods and the optimisation of resources for prospecting abandoned refilled zones, and to assess the possible contamination processes. Acknowledgements The authors wish to thank David Owen’s suggestions on the English used and Mr Miquel Mestres i Palou who provided us with his farm products and groundwater samples at the Cal Dimoni area (farming wells 33 and 37).
References
Fig. 9 Electromagnetic prospecting map using Geonics EM-34 at 7.5-, 15-, 30- and 60-m depths
Cherry JA, Barker JF, Feenstra S, Gillham RW, Mackay DM, Smyth DJ (1996) The Borden site for groundwater contamination experiments: 1978–1995. In: Kobus H, Barczewski B, Koschitzky HP (eds) Groundwater and subsurface remediation. Springer, Berlin Heidelberg New York, pp 101–127 Dorronsoro JL, Carreras N, Sanchez D, Quejido A, Sanchez M, Herraez I, Fernandez ME (1996) Estudio de la influencia del vertedero de residuos so´ lidos urbanos de Pinto (Comunidad de Madrid) en su entorno. Residuos VI(31):36–40 Font X, Navarro A, Rivero L, Casas A, Noguera JF, Martinez F (1998) Assessment of non-controlled land-fillings by geochemical and geophysical methods. Llobregat delta. IV Meeting of the Environment and Engineering Geophysical Society (European Section), pp 135–138 MacFarlane DS, Cherry JA, Gillham RW, Sudicky EA (1983) Migration of contaminants in groundwater at a landfill: a case study. 1. Groundwater flow and plume delineation. J Hydrol 63:1–29 Marques MA (1984) Les formacions quaterna`ries del delta del Llobregat. Institut d’Estudis Catalans, Barcelona MOPU (1966) Estudio de los recursos hidra´ ulicos totales de las cuencas de los rı´os Beso´ s y Bajo Llobregat, vol 4. Comisarı´a de Aguas del Pirineo Oriental y Servicio Geolo´ gico de Obras Pu´ blicas Morales T, Go´ mez MA, Antigu¨ edad I, Zubiaga R, Barcenilla A, Bonilla A (1994) Reconocimiento del impacto en aguas subterra´ neas de dos vertederos clausurados en la reserva de la Biosfera de la comarca de Urdabai (Paı´s Vasco). In: Rebollo LF (ed) Ana´ lisis y Evolucio´ n de la Contaminacio´ n de las Aguas Subterra´ neas, T.II. Madrid, pp 309–324
Environmental Geology (2002) 41:898–905
905